Introduction to Behavioral Neuroscience

Brain cells are eukaryotic cells

The first cell is thought to have arisen on earth at least 3.8 billion years ago. Today we classify cellular life into three domains: Bacteria, Archaea and Eukarya. All cells on this planet contain a barrier to the outside world called a cell membrane and genetic material (DNA). However, unlike prokaryotic cells, which are members of the Bacteria and Archaea domains, eukaryotic cells belong to the Eukarya domain which has as its major cellular feature membrane-bound organelles. Membrane-bound organelles allow for functional compartmentalization and specialization. DNA is segregated in the nucleus where it is copied (via a process called replication) and/or converted into mRNA (messenger RNA). Once converted (transcribed) into mRNA, this material exits the nucleus. It is outside of the nucleus that the translation of mRNA to protein occurs in eukaryotic cells The Central Dogma.

Brain cells, neurons and glia, also adhere to the characteristics specified above and are fundamentally just eukaryotic cells. However, neurons have additional features that make them uniquely specialized for communication, a feature which defines nervous systems. Before we get into the specifics of neurons and glia, let us establish a few fundamentals of eukaryotic cells.

Introduction to the central dogma: cell identity is determined by gene expression

The human body is thought to consist of trillions of cells. Because a single fertilized egg gives rise to all the cells in the adult body, every cell within an organism contains basically the same unique DNA genome. Yet, the human body is not made of identical cells. Depending on the organ system, cells differ in their shape, size, function, and modes of interaction. What makes a heart or liver cell different from a brain cell is not the DNA contained in its nucleus, which is identical in both. Rather, the difference is due to the selective process of transcription of DNA into RNA. Which parts of the DNA are read, and by extension which parts of the DNA are excluded, is the first step in making proteins. The chosen regions of DNA are transcribed into mRNA and translated into specific proteins.

One way to envision the genome is as a comprehensive cookbook containing all the written instructions of life and found in every cell. However, only certain recipes (DNA parts called genes) are chosen to ultimately be cooked (made into proteins) in different kitchens (different cells). Cell identity therefore arises from cell-specific gene choices. This is called gene regulation and is controlled by both intrinsic (within cells) and extrinsic factors (the environment that the cell matures in). For example, a neuron may require a specific channel protein that allows electrical communication while a pancreatic cell requires the expression of an insulin gene to regulate glucose levels in our body. Furthermore, different types of neurons found in different brain regions require different proteins to produce unique chemical messages known as neurotransmitters (e.g. dopamine versus acetylcholine). Proteins do all the work in our cells. This information flow of DNA to RNA to protein is called the central dogma. The central dogma is universal and is a feature of all living cells on our planet: DNA is transcribed into RNA which can be translated into proteins. The process of translation is facilitated by ribosomes which are large macromolecular complexes found floating inside cells or attached to organelles like the endoplasmic reticulum (see below).

Organelles: carry out functions of cells

Brain cells, like all eukaryotic cells contain a complement of organelles necessary for the proper cell function. This includes a nucleus (houses the DNA), endoplasmic reticulum, Golgi apparatus, lysosomes, mitochondria (singular is mitochondrion) and a few others (Typical eukaryotic cell). Cells store genetic material in their nucleus, transport materials across the cell membrane and within the cell, produce energy in the form of ATP, breakdown and build macromolecules, and communicate and interact with other cells. Organelles serve in these multiple functional roles. Of notable importance are the mitochondria, which facilitate most of the processes of cellular respiration that produce ATP. While the human brain only takes up about 2% of body mass, it uses about ~25% of daily energy production. Neurons can have up to 2 million mitochondria each and thus have the capability to produce a large amount of ATP. This high ATP demand is needed for the basic function of neurons and the communication between neurons (see Introduction).

Neurons

The human brain is made up of ~86 billion neurons. Neurons are highly specialized for communication that occurs via electrical and chemical signals. With some exceptions, neurons have special structural features that branch from the cell body (soma), called axons and dendrites (Basic neuronal structure). Functionally, neurons receive and integrate signals, and send information in both electrical and chemical forms. Neurons communicate with their partners at regions called synapses.

Neurons come in many shapes and sizes. When asked to describe a neuron, most students would draw a multipolar neuron containing one axon and several dendrites branching from the cell body (for example a motor neuron). This is the classic depiction found in the popular press and many biology texts. However, the nervous system contains other neuronal varieties, including bipolar neurons (two processes jetting off from the cell body; one dendrite and one axon; olfactory (smell) neuron being an example) and unipolar neurons (a single process from the cell body; dendrite and axon are continuous; often used for sending sensory information) (Neuron heterogeneity). Neuronal shape and structure can inform function. For example, bipolar neurons are uniquely suited for providing a pathway between input and outputs in the visual system.

Neuroscientists have used many lenses for classifying neurons, grouping them by size, shape, or the connections they make (e.g. whether those connections have predominantly inhibitory or excitatory effects). For example, afferent neurons (e.g., sensory neurons) carry messages towards the central nervous system (CNS). Efferent neurons (e.g., motor neurons) carry messages away from the CNS. Interneurons allow communication within the nervous system and are located between sensory and motor neurons. Classification of neurons remains difficult, though, in part because they are so dynamic. Huge efforts are underway to catalog all the different types of neurons in the brain

Whatever their morphology or function, neurons have all the typical eukaryotic organelles, all of which are found in the cell body, called the soma (Parts of a neuron). Below, we will discuss in more detail the unique cellular features of neurons, using a common multipolar neuron depiction to help visualize them.

Dendrites

Dendrites are branches/processes that extend from the cell bodies (or soma) of most neurons. The word dendrite comes from Greek and means tree. We often talk about neurons having tree-like branching, arborization (also a tree analogy). Functionally, dendrites are specialized for receiving information at a synapse and transferring that information toward the cell body. Dendritic branching can be extensive or can be more minimal depending on the type of neuron. In some animal nervous systems, dendrites are studded with bumps called dendritic spines, which provide specialized compartmentalization for synapses and, among other functions, seem to be important in learning and memory (Dendritic spines).

Axons

A neuron can have thousands of dendrites, but only one axon. An axon is a branch that like dendrites, emerges from the soma or cell body of the neuron. At the point of connection with the soma, the axon region is called the axon hillock (Parts of a neuron). Functionally, axons are specialized for sending information. Some cells, such as those in the retina, lack axons or have axons that are very hard to identify. An axon emerging from the soma can branch out at its terminal end to communicate with multiple cell partners. These branches are called axon collaterals and at the ends of these collaterals are axon terminals. Axons can vary in length depending on their location. For example, axons extending from the spinal cord to your toes are very long in comparison to axons found within the brain. Axons can be surrounded by a fatty substance called myelin, which helps insulate axons and allows more efficient/faster sending of signals over long distances in the body. Multiple sclerosis (MS) is an autoimmune disease where an individual’s immune system attacks myelin, causing neuronal communication slow-down and disruption. This leads to severe motor problems in the affected individual.

Not all axons in our nervous system are myelinated, especially those that travel shorter distances. Interestingly, there are many types of animals including some vertebrates that do not have myelin at all. Other biological adaptations, such as larger diameters of axons, are used in these cases to create efficient communication.

The Synapse

The synapse is the place where one neuron communicates with its partner. Neurons do not touch other neurons/cells directly, but rather interact via the space between them called a synaptic cleft. Neurons can form synapses not only with neurons, but also glands (neuroglandular junctions) and muscles (neuromuscular junctions). (Types of synapses).

In the case of neuronal synapses, axons can synapse with dendrites (axo-dendritic) (Types of synapses), with other axons (axo-axonic) or with the soma (axo-somatic). The number of synapses that one neuron can participate in varies from tens to thousands. In chemical synapses, the presynaptic side of the synapse sends chemical signals (neurotransmitters) from one neuron to the postsynaptic side (membrane of the receiving cell). Most synapses are chemical synapses, but another type of synapse exists: the electrical synapse. It allows communication via channel proteins that physically connect adjacent cells and no neurotransmitter is utilized (see Introduction).

Glia

The name glia comes from the Greek word for ‘nerve glue’. Historically, these cells were first identified as connective cells that “glued” neurons together. However, over many years of research, glia have emerged as key players of the nervous system with a variety of sophisticated functions and not just nerve glue. Glia are non-neuronal cells in the nervous system that make up ~30-60% of the total brain mass. They are found in both the central and peripheral nervous systems. In many parts of the nervous system, they outnumber neurons. Glia are not thought to generate electrical signals, but function to support neuronal signaling (regulate the spread of signals in the brain). They can function to guide developing neurons to their proper destination. They can provide nutrients and other chemicals to the neurons that they surround. There is a rich and evolving literature around the interesting roles that glia play in the nervous system, but we will limit ourselves to a general overview. Below we will discuss the major types of glia: astrocytes, microglia, oligodendrocytes, and Schwann cells (Types of glia and example functions).

Astrocytes

Astrocytes have a characteristic star shape and represent the most abundant fraction of glial cell type in the adult human brain. Astrocytes have many functions, such as the maintenance of the blood-brain barrier. The blood-brain barrier (BBB) is a structure that filters chemicals and pathogens entering the brain. For example, the larger and less water soluble a molecule is, the less likely it is to cross this barrier (Astrocytic functions). In addition, these glial cells provide neurons with metabolic and structural support and help modulate neuronal function in some cases. One very important role for astrocytes involves recycling neurotransmitters by reuptake back into neurons (Astrocytic functions). Dysfunction of astrocytes has been associated with pathological conditions including epilepsy, brain tumors and neurodegenerative diseases like Alzheimer’s disease.

Microglia

Microglia serve an immunological role in the nervous system and act like phagocytic cells. Phagocytes are cells that ingest foreign particles like bacteria, for example (see Introduction). Microglia are named for being very small. They destroy invaders that get through the blood-brain barrier and then clear out damaged neurons and unused synapses and thus play an important role in synaptic pruning and elimination of extra synapses during brain development. Improper synaptic pruning may be involved in neuropsychiatric disorders such as autism and schizophrenia.

Oligodendrocytes

Myelin is the insulating axon wrapping made up of lipid and protein, and oligodendrocytes are the myelinating cells of the central nervous system. Each oligodendrocyte provides myelin for several neurons. The axon of one neuron can be myelinated by multiple oligodendrocytes. As you will see in later chapters, myelin wrapped around axons is necessary for rapid communication in the nervous system (see Introduction).

Schwann Cells

Schwann cells are the main glial cells of the peripheral nervous system and serve similar functions to oligodendrocytes. Unlike oligodendrocytes, one Schwann cell provides myelin for the axon of a single neuron (see Types of glia and example functions).

History of Neuroscience: Visualizing Brain Cells: from Cajal to Brainbow

Throughout this section, we have discussed the cellular components of the nervous system and the unique specializations that allow neurons to communicate with each other. In the late 1880s, a debate brewed in the field of neuroscience. Without sophisticated tools and relying on simply observing the human brain, it was almost impossible to ascertain its components. Camillo Golgi, a prominent scientist of his time and the inventor of the Golgi stain (see below), championed a brain model where the building material of the brain was a single, dense, intricate, and continuous network akin to a material like gauze. This was called the reticular theory. On the other side of the debate was another Spanish scientist, Ramon y Cajal. Based on his findings, Cajal championed the Neuron Doctrine—the idea that the brain was composed of individual, unconnected cells. Ultimately, Cajal’s theory proved to be correct. Our ability to visualize the brain has advanced and deepened our understanding of Neuroscience. Beginning with the Golgi stain, this section will outline some common approaches to visualizing brain cells (Examples of neuronal stains).

Chemical Stains: Focus on the Golgi Stain

The driver of scientific progress is often fueled by technical discoveries (a new instrument, a new technique) that allow scientists to visualize, analyze or probe. The Golgi stain (silver nitrate) was first discovered by Camillo Golgi as a way to visualize neurons in postmortem tissue. Its improvement and application to brain tissue by Ramon y Cajal in the late 1880s opened the door to our ability to see brain cells as individual entities for the first time. A unique feature of this technique is that, if applied to a forest of neurons, only a few random cells will become stained and stand out when visualized under a microscope. The reason for this selectivity is still unknown to this day. The neurons that do get stained become visible in a very complete way, allowing visualization of cell bodies/soma, dendrites, and axons (Examples of neuronal stains).

Cajal’s work revolutionized the field of neuroscience. For the first time, it was clear that neurons were separated from other individual cells and that they were morphologically distinct in different brain regions. He used Golgi’s own methods to refute Golgi’s belief that brain cells were a continuous cellular web. Through his careful work, Cajal gained many additional insights, including the fact that neurons were polarized (receiving information on their dendrites and cell bodies and sending information via axons). The Golgi stain continues to be a valuable tool for modern day neuroscientists. In addition to the Golgi stain, we have a number of chemical stains that can be used to visualize neural tissue, including but not limited to Nissl and H&E (Hematoxylin and Eosin) staining for cell bodies and nuclei and Luxol fast blue for myelin.

Immunofluorescence

All the chemical stains mentioned above allow scientists to visualize parts of a neuron such as a nucleus, cell body or processes. However, as we learned in the first section of this chapter, the types of genes expressed within a cell dictate its function. If a scientist is interested in determining the distribution of a specific protein (like an enzyme that produces a certain type of neurotransmitter) or the presence of myelin proteins, a more specific method of visualization is necessary. Immunofluorescence is a technique that uses antibodies (designed against a specific protein of interest) equipped with fluorescent tags (which come in many different colors) to localize and visualize a protein of interest using a fluorescent microscope (Examples of neuronal stains) (see Methods: Immunohistochemistry). There are well established antibodies that can be used as neuronal cell markers that allow scientists to distinguish neurons from glia, visualize synaptic connections, measure protein expression levels in different neurons, etc. We have the technology to generate antibodies to almost any protein that we are interested in studying, giving scientists a large toolkit to answer basic questions in neuroscience.

GFP Labeling/Brainbow

Following its discovery in the 1960s, green fluorescent protein (GFP) has been used to label cells, organelles and specific proteins. Because we know the exact DNA sequence that codes for the green fluorescent protein and because all living cells on this planet use the same genetic code (see central dogma above), we can insert the GFP DNA directly or attach it to any sequence that we may want to express in a cell. Once that DNA is taken up and expressed, the cell will glow green wherever the protein is found in the cell. The GFP DNA can be introduced into cells grown in culture or brain slices using specialized delivery tools like viruses. The GFP DNA can also be genetically engineered into gametes (reproductive cell of an animal or plant) prior to fertilization, which leads to GFP transgenic mice allowing visualization of specific regions in the brain throughout the lifetime of an animal. Beyond just green, genetic manipulations of the GFP DNA have been developed to introduce the genetic machinery to randomly mix colors (such as green, cyan and yellow fluorescent proteins) in individual neurons thereby creating a palette of different color combinations and shades, allowing visualization of whole neuronal circuits. This process produces a Brainbow, where each individual neuron stands out due to a different shade of color (Examples of neuronal stains).

Neuroscience across species: neural nets, ganglia and centralized brains

All multicellular organisms with the exception of sponges have neurons and nervous systems, albeit varying in structure and complexity. Large differences are observed between animals with or without a backbone (vertebrates vs. invertebrates). Two general principles in nervous system organization across the animal kingdom include: neural nets and ganglia (Diversity of the nervous system organization).

Cnidaria (includes hydra, corals, jellyfish, and anemones) have neural nets. These are the simplest—and, considered evolutionarily, the oldest—form of nervous system organization. We can describe them as mesh-like systems of separate, yet interconnected neurons (not clustered into nerves or centralized). Nerve nets can be diffuse or can be connected to each other via long nerve cords, as in the case of flatworms. Nerves are defined as bundles of fibers (axons) that transmit information.

Nerve cords can be arranged in a ladder-like conformation and connected to ganglia. Ganglia are structures that contain collections of neuronal cell bodies and are found in animals more complex than Cnidaria, such as flatworms. Many invertebrates actually have more than one ganglia, each with specific functions, such as control of different muscles or, in the case of earthworms, different body segments. Often there is a cerebral ganglion which serves a more coordinative role. In contrast, vertebrates have a spinal cord encased in vertebral bones. The spinal cord runs along their back and most of the nervous system that is important for complex behavior is centralized and concentrated in a brain located in the head.

Animal symmetry

Humans and many other animals, including other vertebrates and some invertebrates, have bilateral body plans meaning that the left and right sides of the body are mirror images of each other. Other animals like jellyfish and sea stars are arranged centrally around a gastrointestinal core. This is called radial symmetry. While bilateral organisms often have one nerve cord and centralized ganglia or a brain that controls both sides of the body, radially symmetrical organisms lack a central brain and have radial nerves (sea star) or nerve nets (hydra) (Diversity of the nervous system organization and Bilateral vs radial symmetry).

Centralization and cephalization

Nervous systems of animals with bilateral symmetry exhibit centralization and cephalization. Centralization refers to a nervous system organization where neurons are consolidated into specific areas of integration rather than just being randomly arranged throughout the body. Cephalization refers to the concentration of the nervous system at the anterior part of the body or the head. Flatworms, considered one of the simplest organisms with bilateral symmetry, exhibit both centralization and cephalization, having ganglia at the anterior end of their body (Diversity of the nervous system organization). All vertebrate nervous systems contain a central nervous system (CNS) (brain and spinal cord) and a peripheral nervous system (PNS) (peripheral motor and sensory nerves).

The basic organization and structure of the vertebrate nervous system

The human nervous system shares the same basic plan found in all vertebrates and we humans are not unique in our brain organization. The human brain is a vertebrate brain and basically a large primate brain. Our nervous system can be divided according to anatomical structures and/or functional regions. This section will introduce the CNS and PNS and highlight some important features of the two systems (Vertebrate nervous system organization).

CNS and PNS

The human nervous system, like other vertebrate nervous systems, is organized into the central nervous system (CNS), which is composed of the brain and spinal cord, and peripheral nervous system (PNS), which includes all nerves and ganglia outside of the spinal cord and brain. Groups of neurons are segregated into nuclei in the CNS and ganglia in the PNS. Axon bundles are called nerves in the PNS and tracts in the CNS (Nervous system basics). Ascending tracts take information to the brain and descending tracts away from the brain. Vertebrate nervous systems are highly centralized with higher neural functions like memory, learning, perception and movement carried out within the brain. The spinal cord sends sensory information to the brain and conveys motor commands from the brain to the entire body. In addition, there are circuits found in the spinal cord that perform local processing and allow for simple behaviors such as reflexes to occur without the direct involvement of the brain. The spinal cord also contains central pattern generator circuitry that allows for rhythmic behavior such as walking (see Introduction). The PNS is responsible for carrying messages between the CNS and all body parts, including muscles, organs, and the body periphery in general. A more detailed discussion of the CNS and PNS can be found below.

Gray and white matter

The nervous system is made up of gray and white matter (Nervous system basics). This terminology refers to gross histological distinctions. Gray matter, named for its pinkish-gray color, has a high concentration of neuronal cell bodies, dendrites and axons that are not myelinated. White matter is composed of axon fibers covered in myelin, giving it its whitish color. Gray matter is found in the outermost layer of the brain and centrally within the spinal cord. White matter is in the deeper tissues of the brain and surrounds the gray matter in the spinal cord.

Meninges

The CNS is covered in bone. A skull encases the brain while the spinal cord is housed in a collection of 33 small bones called vertebrae. Between the bones and the nervous tissue of the CNS, there are layers of cushiony tissue called the meninges. The three layers of the meninges are dura mater, arachnoid mater and pia mater. The dura mater layer is found directly against the skull and vertebrae, and is the thickest layer of the three meninges (The meninges).

Ventricles

Ventricles are interconnected cavities or open spaces located in the brain and spinal cord that contain cerebrospinal fluid (CSF). They serve two main functions: 1) to cushion the brain and 2) to allow exchange of materials between the brain and blood vessels. CSF is produced by the choroid plexus, a tissue rich in blood vessels found in the lining of the ventricles. It is a clear liquid, essentially filtered blood. Spinal taps are medical procedures to remove a small amount of CSF to test for medical conditions or infections in the brain. The ventricular system is composed of 4 ventricles and an aqueduct (The cerebrospinal fluid (CSF)). Think of the ventricles as being similar to the great lakes: independent bodies of water but connected by rivers (aqueduct). The 2 large lateral ventricles in each cerebral hemisphere (left and right) connect to a third ventricle (in the diencephalon) which opens into the cerebral aqueduct (in the midbrain) and a fourth ventricle (in the hindbrain). The fourth ventricle extends to the central canal which is filled with CSF and runs across the spinal cord. CSF fills the ventricular system and circulates over the brain and spinal cord. It flows through the subarachnoid space and exits through the arachnoid villi into venous sinuses, a group of blood channels found throughout the brain. Eventually CSF is reabsorbed into the circulatory system. CSF is also found surrounding the brain below the arachnoid membrane called the subarachnoid space (The cerebrospinal fluid (CSF)).

Vascular system of the brain

The human brain requires a disproportionate amount of energy in the form of ATP to function properly. ATP production requires oxygen, which is supplied by a constantly circulating blood supply. As a result, the brain has a complex system of blood vessels. Three main arteries (anterior, middle and posterior cerebral) are the main suppliers of blood to the cerebral hemispheres. The anterior and middle arteries arise from the carotid arteries on the left and right sides of the neck. The vertebral arteries run through the spinal column in the neck to provide blood to the brain and spinal cord, and give rise to the basilar artery. There are many vessels and capillaries that branch off from the main arteries delivering nutrients and removing waste (Brain vasculature). The endothelial cells that form the walls of brain capillaries are extremely well connected to each other and form the blood-brain barrier (BBB), which we discussed in the previous section. This serves as a layer of protection not found in other organs that prevents the flow of some substances like toxins or drugs and also protects against infections.

Neural circuits

Neurons do not function independently. Rather, neurons are organized into simple or complex functional arrangements called neural circuits. A neural circuit represents a group of neurons connected by synapses that process specific types of information. Circuits range from very simple to exceptionally complex. One of the big challenges in modern neuroscience is to understand the complex ways in which circuits are built and function. Large scientific efforts are focusing on the enormous task of building detailed maps of brain circuitry. In this section, we will explore the basic building blocks and functionality of a simple circuit and introduce you to the research focused on more complex circuitry.

Basic building blocks of neural circuits

While neural circuits vary in function, anatomic arrangements and complexity, we can establish some basic principles. As introduced in Building a Nervous System, afferent (sensory) neurons carry information about the inside of the body and external environment to the brain or spinal cord while efferent (motor) neurons carry information from other neurons away from the brain/spinal cord or specific circuit. Interneurons serve as a bridge between other neurons and are found in the CNS. They are local circuit neurons that make intra-circuit connections and can act as regulators or adjustors to the circuit (Afferent, efferent and interneurons).

Simple human circuit example

A reflex arc is a simple neuronal circuit that controls a reflex beginning with a sensory neuron at a receptor (for example, pain receptors on a finger) and ending with a motor neuron at an effector (like a muscle). Reflex arcs are important for the generation of fast reactions to external stimuli such as accidentally placing your finger into an open flame or on a hot surface. Reflex actions are essential to the survival of organisms. To prevent bodily damage, a movement or response that is almost instantaneous and involuntary is required.

In a reflex arc, the sensory neurons do not pass information directly to the brain but rather information travels to the spinal cord, which, in turn, causes direct activation of spinal motor neurons. These neurons can then activate muscles that allow your hand to move away from the hot surface. Of course, information about such an event must eventually get to the brain to allow learning and future avoidance. Another example of a simple human circuit is the knee-jerk reflex (Simple circuit example), which allows for involuntary movement of the leg in response to a sharp tap to the tendon below the knee cap. The knee-jerk reflex is an example of a reflex pathway that is both excitatory and inhibitory (once the extensor muscle on the thigh contracts, the flexor muscle on the back of the thigh must relax). Stimulation of stretch receptors via the tap leads to afferent sensory neuron(s) activation, which in turn synapses either onto interneurons or directly onto efferent motor neurons that cause contraction of the extensor muscles. The interneuron serves to inhibit the motor neurons that connect to flexor muscles which keeps the flexor muscle relaxed. This combined action leads to leg extension (Simple circuit example). This reflex prevents knee collapse in the event of a sudden leg bend.

Mapping complex brain circuits

A major goal in neuroscience is the mapping of complex neural circuits in the human brain. Because the human brain is exceptionally complex with its 86 billion neurons, each communicating with hundreds or even thousands of other neurons, this is an extremely difficult endeavor. Scientists hope to one day complete the task of mapping the human brain connectome (a comprehensive map of neuronal connections in the brain akin to a wiring diagram). In the meantime, neuroscientists have been using simple animals like worms and flies to develop wiring maps that can provide clues to how brains work in general. Presently, two comprehensive wiring diagrams have been published: the roundworm (C. elegans) nervous system which has a very small number of neurons (302 neurons and 7000 connections) and most recently the entire brain of a fruit fly larva (D. melanogaster), made up of 3016 neurons and 548,000 connections.

There are several large, international, ongoing human mapping projects. These include the Human Connectome Project started in 2009 whose aim was to build a healthy human brain connectome and develop connectomes related to human disease. Building onto the Human Connectome Project, the BRAIN Initiative was started in 2013 with the aim of producing a dynamic picture of the brain which would map individual brain cells and the circuits that they take part in both time and space. The goal with such large projects and initiatives is to provide a deeper understanding of brain function and as a result develop new treatments for brain disorders. In 2021, the BRAIN initiative unveiled a detailed atlas of a small brain region, the mammalian primary motor cortex, derived from studies in mice, monkeys and humans.

Anatomical orientation and planes of orientation

Specific terminology is applied to describing anatomical structures and orientation in the nervous system. Dorsal is an anatomical direction that refers to both the back of the spinal cord and to the top of the brain. The opposite of dorsal is ventral which means toward the bottom of the brain or the front of the spinal cord. Superior and inferior refer to above or below another structure. The bend in the human neuraxis causes a dissociation between the dorsal/ventral and superior/inferior directions. Anterior means toward the face and head, and posterior means toward the back or toward the feet. Rostral is toward the brain and or the top of the spinal cord while the opposite, caudal, is towards the back of the brain. Medial means toward the middle or center while lateral means left or right of another structure (toward the outside) (Anatomical orientation). Ipsilateral means the same side and contralateral means the opposite side. For example, the right motor cortex controls the left side of the body and the left motor cortex controls the right side of the body (contralateral).

There are three planes for sectioning the brain for anatomical analysis: coronal sectioning (cutting the brain front to back); sagittal sectioning (cutting the brain left to right) and horizontal or axial sectioning (cutting the brain top to bottom) (Planes of orientation).

Developmental perspective: vertebrate nervous system development

Neurodevelopment is a complex and multi-step process that requires a tremendous amount of coordination. Turning on the right genes at the right time is crucial and any missed steps or abnormal development can have catastrophic consequences. The embryonic development of the vertebrate nervous system begins with the outer tissue of the embryo called the ectoderm. At about 3 weeks in development, the ectoderm begins to thicken and form a flat structure called the neural plate. During the 3rd and 4th week of development, the neural plate begins to invaginate into a groove that eventually closes to form a neural tube. In the initial stages of formation, the neural tube is made up of one layer of cells (epithelial), which serve as the progenitors (embryonic neural stem cells) of glia and neurons. The neural tube elongates to form the spinal cord while the anterior region enlarges and expands to form the brain. Three swellings or bulges eventually become apparent, which become the prosencephalon (forebrain), mesencephalon (midbrain) and rhombencephalon (hindbrain). As development progresses, the prosencephalon subdivides into the telencephalon and the diencephalon. The mesencephalon does not further subdivide. The rhombencephalon becomes the myelencephalon and the metencephalon. The spinal cord forms from the end of the neural tube next to the myelencephalon. By 11 weeks of deveopment, the shape of the human brain is similar to a brain at birth. In the sections that follow, we will discuss the adult structures that derive from the five embryonic regions above (Development of human nervous system). More details on development can be found later in this text (see Introduction).

Telencephalon and diencephalon

The telencephalon becomes the cerebral hemispheres, which makes up the largest part of the brain (~85% of the total brain mass). The cerebral hemispheres include the cerebral cortex, the outermost layer of the cerebrum. The word cortex comes from Latin for ‘bark’ and represents an extremely thin layer that if flattened is the size of a regular pillowcase. Below the cortex are deeper, subcortical structures that include the hippocampus, amygdala, and basal ganglia. The lateral ventricles arise from the telencephalon. The posterior part of the forebrain is the diencephalon. This eventually becomes the thalamus, hypothalamus, and posterior pituitary gland. Additionally, the diencephalic cavity becomes the third ventricle (see Organization of the Nervous System for ventricles).

Mesencephalon, metencephalon and myelencephalon

The mesencephalon (midbrain) forms into the tectum and tegmentum. These eventually form the roof and floor of the cerebral aqueduct which arises from the mesencephalic cavity. Behind the mesencephalon are the hindbrain divisions: 1) metencephalon, which becomes the cerebellum and the pons, and 2) myelencephalon, which becomes the medulla. The mesencephalon and myelencephalon develop from the rhombencephalon whose cavity becomes the fourth ventricle (see Organization of the Nervous System).

Spinal cord

The spinal cord forms from the neural tube and retains this tube-like shape once fully developed. It extends from the medulla oblongata in the brainstem down the body as a solid structure that ends in the middle of the lower back and continues as a spray of fibers called cauda equina. The cauda equina is part of the PNS rather than the CNS and it sends and receives information to and from the pelvic organs and lower limbs.

The spinal cord is enclosed by three membrane layers (the meninges) and is encased in the vertebral column made up of bony vertebrae separated by intervertebral discs. Each of these vertebrae has an opening that allows 31 spinal nerves to pass on each side. The spinal cord is divided into 5 major regions: cervical, thoracic, lumbar, sacral and coccyx. Each spinal nerve is named based on which of the 4 major regions of the spinal cord it is connected to. The area of the skin innervated by the spinal nerve is the nerve’s dermatome. These are named according to the 5 major regions above (for example T1-T12 is thoracic) (Spinal nerve organization). The central canal (which is continuous with the 4th ventricle, see Organization of the Nervous System) runs through the spinal cord and is filled with cerebrospinal fluid. It provides cushioning and a transportation system to the spinal cord.

Gray and white matter in the spinal cord

When looking at a cross section of the spinal cord, the center contains gray matter while the white matter is found towards the outside. The gray matter contains glia, interneurons, and motor neuron cell bodies, and it is butterfly-shaped. In keeping with this analogy, the wing areas are called horns. Dorsal and ventral horns are found on each side of the cord (Spinal cord organization).

The white matter consists of myelinated nerve tracts (axons) that travel up and down the spinal cord (Sensory and motor tracts). Along the spinal cord, the 31 pairs of spinal nerves emerge at regular intervals laterally serving each side of the body (Spinal nerve organization). There are two distinct nerve bundles, roots that branch off. The dorsal roots contain sensory neurons that project from the body to the spinal cord whose cell bodies are organized in a ganglion called the dorsal root ganglion which is adjacent to the spinal cord. The ventral roots house motor neurons that leave the spinal cord and project to muscles (motor) (Spinal cord organization).

Ascending and descending pathways

Information that travels to the brain is divided along compartmentalized pathways via tracts. Ascending and descending tracts make up the white matter. Sensory information travels via ascending pathways from sensory receptors to processing centers in the brain. These sensory tracts carry specific information, for example pain and temperature or pressure and crude touch or fine touch, vibration, and proprioception. The sensory tracts are made of 1st, 2nd and 3rd order neurons connected to each other. The 1st order neurons receive the sensory information at the periphery. These pass information to the 2nd order neurons in the spinal cord. Eventually 3rd order neurons pick up information from the 2nd order neurons in the thalamus and carry it to the cortex (see The Brain: Structure and Function). Motor information moves via descending pathways that carry commands from specialized CNS centers to skeletal muscles. Upper motor neurons originate in the cerebral cortex and travel down to the spinal cord, while the lower motor neurons begin in the spinal cord and send information to the periphery. Motor tracts can be specialized in regulation of voluntary movement or balance and posture (Sensory and motor tracts).

Cerebrum and cerebral cortex

The telencephalon becomes the cerebrum, the largest part of the brain, divided into two halves, or hemispheres, and makes up about three fourths of the total brain volume (Development of human nervous system and Basic cortical features). In the human nervous system, the left hemisphere controls the right half of the body, and the right hemisphere controls the left side of the body. The right and left cerebral hemispheres are connected by a thick fiber bundle (collection of axons) called the corpus callosum. This allows information to pass from one side to the other side of the brain. The cerebrum includes the cerebral cortex and cerebral nuclei (clusters of neurons). In many animals, the cortex is smooth. However, in larger brains like ours, the cortex is wrinkled and thin (on average about 2.5 mm). It is folded, to create an increased surface area, thus maximizing the amount of cortex in the brain. This folding leads to the formation of ridges or bumps called gyri (gyrus is the singular) and grooves called sulci (singular is sulcus). The deeper grooves of the brain are called fissures (Basic cortical features). The two hemispheres are divided by the longitudinal fissure. The two lateral ventricles (see The Brain: Structure and Function) are located in the cerebrum.

The cerebral cortex is packed with neurons, about 16 billion of the total 86 billion neurons in the human brain, and is composed mostly of gray matter. In humans, the cerebral cortex houses most higher brain functions and complex cognition. This includes reasoning, language, consciousness, perception, emotion, personality, decision-making, and memory. Neurons in the cortex are typically arranged in six layers. Each layer contains unique neuron types and architecture. The cortex is divided into specific functional areas: sensory, association and motor areas. The sensory areas receive sensory information from touch, sight, hearing, smell and taste. The association areas help integrate sensory information to give meaning to the incoming information. Finally, the motor areas initiate and regulate voluntary movement.

Lobes of the cerebral cortex

Different areas in the cerebral cortex are associated with specific functions. Each hemisphere’s cerebral cortex is divided into four lobes: frontal, temporal, parietal, and occipital (Lobes of the cortex). Each of these four lobes is involved in a number of functions. This section highlights the major functions but is not necessarily comprehensive.

The largest of the four lobes is the frontal lobe. It also contains the olfactory bulb, which is key for our sense of smell. Towards the posterior end of the frontal lobe are the premotor and motor regions that help plan and carry out movement throughout the body. The primary motor cortex sends axons through the brain and into the spinal cord via a descending tract (corticospinal tract). Once these axons reach the brainstem (the stalk of the brain that joins with the spinal cord, see The Central Nervous System: CNS), they crossover so that movement on the left side of the body is controlled by the right side of the brain and vice versa. The prefrontal cortex of the frontal lobe is crucial for personality; higher level-cognition and decision making, problem solving, attention and also some memory functions.

The frontal and parietal lobes are separated by a central sulcus (Basic cortical features). The parietal lobe is important for somatic sensation integration. The somatosensory cortex, which is found in the anterior part of the parietal lobe, receives information about touch on all areas of our body. In addition, the parietal lobe has a role in our ability to know the location of our body is space, also known as proprioception.

The occipital lobe is located posterior and inferior to the parietal lobe and primarily functions in vision processing. It contains a primary visual cortex and other areas necessary for visual processing of color, movement and patterns.

The temporal lobe is the second largest in size following the frontal lobe. It is located underneath the frontal and parietal lobes and is key for language, memory, hearing and face perception. It is divided from the other lobes by a lateral sulcus called the Sylvian fissure (Basic cortical features).

Hemispheres

In addition to the cortex which covers the cerebral hemispheres, the inner core of the cerebral hemispheres is composed of white matter. The cerebral hemispheres are largely redundant in function with a few small exceptions. For example, in most humans the left hemisphere controls language and speech while the right interprets spatial, visual information and face recognition (see Introduction). The left and right hemispheres are divided by a deep longitudinal fissure and are connected by axonal tracts called cerebral commissures. Tracts that cross the midline of the brain are called commissures. As mentioned above, the largest of these is the corpus callosum, which connects the two hemispheres and allows information to be passed between sides (Lobes of the cortex). If the corpus callosum is damaged or purposely severed (rare approach to treat epilepsy), it leads to a condition called split-brain syndrome. This is characterized by neurological abnormalities related to communication between the two sides of the brain. For example, a split-brain patient may not be able to read a specific word when it is presented to the right hemisphere but can identify the drawing that goes with that word.

Limbic System and Basal Ganglia

Underneath the cerebral cortex are a group of specialized brain areas called subcortical structures or nuclei. Two major subcortical structure systems are the limbic system and the basal ganglia, which also derive from the telencephalon.

Limbic System

The limbic system is a collection of structures that border the brainstem and are located under the cerebral cortex (Limbic system). The word limbic is derived from the Latin “limbus,” which means border. The limbic system is an anatomical concept and continues to be somewhat controversial when it comes to brain anatomy. Interestingly, neuroscientists can’t fully agree on a set of criteria for regions that count as part of the limbic system. Not even anatomy is set in stone.

Traditionally, the limbic system is defined as a connected set of structures that is involved in emotion, aggression, appetites and sexual behavior. While still up for debate, it is generally agreed that the limbic system includes the hippocampus, amygdala, and the cingulate gyrus. The cingulate gyrus is part of the cortex and sits right above the corpus callosum and is very important for emotional and cognitive functions. The amygdala is especially important for fear emotions and fear memory. The hippocampus is associated with memory and specifically the ability to form long term memories. Additional regional structures of the limbic system may include the fornix (connects hippocampus to hypothalamus), septum (often linked to emotion) and olfactory bulb (important for our sense of smell).

Basal Ganglia

Basal ganglia (nuclei) are a group of structures that play an important role in controlled and coordinated movements (Basal ganglia). This includes initiating and facilitating movement while also inhibiting unwanted movement. In addition, some regions of the basal ganglia are also important in motivation, reward and addiction. The basal ganglia system can be subdivided into the caudate nucleus and putamen (collectively called the striatum), globus pallidus and subthalamic nucleus. The striatum receives commands from the cortex and passes them along to the other regions of the basal ganglia.

Eventually basal ganglia information goes to the thalamus which relays information back to the cortex. Damage to the basal ganglia causes movement dysfunctions and is associated with neurodegenerative diseases like Parkinson’s and Huntington’s diseases.

Diencephalon

The diencephalon is made up of the thalamus and hypothalamus, located inferior to the cerebrum and next to the third ventricle. The thalamus provides a relay station for sensory information coming in from the spinal cord and PNS, heading to the cortex. The hypothalamus is important for homeostasis control of basic functions (Diencephalon).

Thalamus

The thalamus is an oval structure, made of two symmetrical parts, one in each hemisphere (Diencephalon). It is the sensory relay station and gateway to and from the cerebral cortex. All sensory information goes through the thalamus with the exception of olfaction (smell). The thalamus is composed of many nuclei that receive sensory information from the body and sorts, processes, and directs the information to the appropriate cortical sensory areas. As an example of one thalamic nucleus, the lateral geniculate nucleus (LGN) receives visual information from the retina via the optic nerve (see cranial nerves) and sends it to the primary visual cortex found in the occipital lobe (see Introduction). The thalamus not only sends information to the cortex but also receives information from the cortex for control and regulation (often associated with motor commands).

Hypothalamus

The hypothalamus is located ventral to the thalamus (Diencephalon). Like the thalamus, it contains several nuclei. It is responsible for regulation of hormones and maintenance of homeostatic mechanisms such as those that keep our bodies at the right temperature and having the right levels of sugar. The hypothalamus regulates the release of hormones from the pituitary gland and thus plays a direct role in connecting the nervous system to the endocrine (hormonal) system of the body. Overall, the hypothalamus is important for many vital functions including: thermoregulation, sex, thirst, hunger, sleep. It is a key regulator of the autonomic nervous system (see The Peripheral Nervous System: PNS below).

Brainstem

The brainstem connects the cerebrum to the spinal cord and also to the cerebellum. It is composed of the midbrain and the hindbrain. The hindbrain is further divided into the pons and medulla oblongata. The fourth ventricle is found at the caudal region of the pons and the dorsal region of the medulla. The brainstem is crucial for vital functions which are discussed below. Stretching from the midbrain to the medulla and into the spinal cord is a network of nuclei and neurons called the reticular formation. These diffusely distributed neurons function in several physiological states such as arousal, consciousness, sleep-wake cycles, heart and respiratory control, and even some pathways involved in pain (Brainstem).

Midbrain

The midbrain has its embryonic origins in the mesencephalon. It is located anterior to and above the pons, and has both sensory and motor components (Brainstem). The midbrain is involved in auditory and visual sensory processing. The top of the midbrain is the tectum (from the Latin word for roof) and the bottom is the tegmentum (from the Latin word for floor) with the cerebral aqueduct passing right through its center. Other regions in the midbrain include the substantia nigra and the ventral tegmental area (VTA). The substantia nigra (one on each side of the brainstem) are considered to be part of the basal ganglia circuitry (The Peripheral Nervous System: PNS) and are named from the Latin words for ‘black substance’ due to the pigment that is found naturally in neurons in this region. The substantia nigra is rich in dopamine neurons and extremely important for motor control. Damage to the substantia nigra (death of dopamine neurons in this region) is associated with Parkinson’s disease, a neurodegenerative movement disorder. Adjacent to the substantia nigra is the red nucleus which also plays a role in motor coordination. The VTA is found adjacent to the substantia nigra and also houses dopamine neurons. This region is part of the reward circuitry in the brain.

Pons

The pons and cerebellum originate from the embryonic metencephalon. The pons is superior to the medulla oblongata (Brainstem). It contains many tracts and nuclei and provides a connection between the cerebellum and the medulla. The word pons is derived from the Latin word for ‘bridge’ and is so named because of the transverse fibers that appeared as a bridge between the cerebellar hemispheres. Due to its many nuclei and tracts, the pons is associated with many different functions. For example, the pons is important for the regulation of sleep and basic bodily functions such as breathing. The pons is associated with 4 of the 12 pairs of cranial nerves and is important for facial expressions, chewing and control of eye movement.

Medulla oblongata

The medulla oblongata derives from the myelencephalon. The medulla is possibly the most important part of our brain as it is essential in the control of vital body functions like breathing, heart rate and blood pressure Brainstem. Additional functions include swallowing, vomiting, and digestion. The medulla contains a number of nuclei in addition to ascending and descending tracts of the spinal cord. Like the pons, it is also associated with cranial nerves (4 of them). These are key for gag reflexes and muscle control that allows for the turning of the head, speaking and swallowing. Of note is the vagus nerve (10th cranial nerve), which is one of the most important nerves in our body. It is a main component of the parasympathetic system and provides sensory and motor circuitry connecting the brain to the neck, lungs, heart and gut. The vagus nerve provides an important back and forth connection between the CNS and the enteric nervous system (the brain-gut axis). It can relay signals from the microbiota (collection of bacteria) in our gut to our brain. Some research suggests that disruption of the microbiota-gut-brain axis may play a role in depression and other psychiatric disorders. Interestingly, the vagus nerve has been studied as a target for treating these types of disorders. Stimulation of the vagus nerve is an approved treatment for hard-to-treat depression.

Cerebellum

The cerebellum (aka little brain) is a structure that originates in the metencephalon and functions in motor coordination and some forms of simple learning. Like the cerebrum, the cerebellum is also divided into two hemispheres Cerebellum. The cerebellum is divided via a midline connector called the vermis (the Latin word for worm). While the cortex gives motor commands, the cerebellum coordinates smooth/timely movement and fine-tunes/adjusts movement based on incoming sensory information such as proprioceptive information. It is necessary for balance and fine motor control, as well as learning and memory of motor tasks, such as learning to ride a skateboard or learning to dance. This motor coordination can be inhibited by alcohol intoxication. Lack of coordination due to intoxication can be assessed by standard tests (finger to nose tests, walking in a straight line, etc.). Alcohol misuse or overuse can cause cerebellar dysfunction and excessive amounts of alcohol consumption actually leads to permanent changes/degeneration in the cerebellum. Furthermore, alcohol exposure during development leads to cerebellar defects such as those observed in children born with FAS (fetal alcohol symptom). The cerebellum has a specific arrangement of three cell layers. While the human brain contains about 86 billion neurons, roughly 80% of these are found in the cerebellum. Note that the cerebellum only makes up 10% of brain volume. Interestingly, case studies of individuals lacking a cerebellum indicate that for the most part, these individuals function relatively well, often not even discovering their lack of a cerebellum until adulthood. (Cerebellum)

Brain lateralization

Hemispheric lateralization refers to different brain functions and/or behaviors being controlled by one hemisphere rather than the other. Most of the evidence for this type of specialization is related to language control. For example, there are two left hemisphere specific regions that are important for language: a region called Broca’s area associated with speech production and speech articulation and Wernicke’s area associated with language comprehension. Other examples of lateralization include handedness, the preference of the right or left hand for different tasks. Beyond that, the differences that have been found in associating different behaviors to one or the other hemispheres are generally small.

Visualizing the human brain

A number of imaging techniques have been developed to aid neuroscientists and medical professionals to visualize the human brain at the level of structure and function. Structural techniques such as PET, CT and MRI scans result in images of anatomical features of the brain while functional techniques such as functional MRI (fMRI) produce images that represent neural activity (see Methods: fMRI). Functional imaging techniques can be used in experiments where a subject performs a specific behavior or cognitive task, allowing scientists to correlate brain anatomy with function. Brain imaging techniques are important experimental tools, allowing us to delve deeper into the inner workings of the brain. Imaging can also be used for diagnosing and studying brain disorders. However, it is important to remember that the brain is exceptionally complicated and that these techniques are not perfect in measurement or definitive in conclusion. Each has its own advantages and limitations. For example, PET and CT scans are sensitive and prone to motion artifacts and fMRI measurements are an indirect measurement of brain activity (increased blood flow and oxygenation) (Imaging the live human nervous system).

Afferent and efferent divisions

The PNS is made up of afferent (sensory) and efferent (motor) components. The afferent division brings information from the periphery (receptors found within skin, joints, muscle, visceral organs) to the spinal cord and brain via sensory nerves. The information brought in by afferent neurons is decoded and processed by the CNS. The efferent component, on the other hand, utilizes motor neurons that transmit information from the CNS to the periphery. This includes our muscle tissue (effectors) causing muscle contraction and thus voluntary body movements. In addition, the efferent division also controls smooth muscle in visceral organs, cardiac muscle and glands. It can function with or without consciousness (voluntary or involuntary).

Cranial and spinal nerves

A substantial portion of the PNS is made up by cranial and spinal nerves that transmit efferent and afferent information. There are 12 pairs of cranial nerves that emerge from the back of the brain, and 31 pairs of spinal nerves that emerge from the spinal cord (Spinal and cranial nerves).

The cranial nerves pass through the cranium (skull) via holes called foramina and canals. Most of the cranial nerves control the head and neck including the jaw, throat, tongue, neck and face. They also carry visual (optic), smell (olfactory) and taste (gustatory) information, and participate in hearing and balance (vestibulocochlear). Of the 12 cranial nerves, 3 are exclusively sensory (afferent), 5 are exclusively motor (efferent) and the remaining 4 cranial nerves are both motor and sensory. Nerve number X (10) is called the vagus nerve (named for the word ‘wandering’). It wanders/extends very far and innervates the heart and intestines, receiving and sending information regarding internal organs. The vagus nerve has been actively researched and doctors can stimulate it with electricity to treat a number of disorders including difficult to treat cases of epilepsy and depression.

The spinal nerves arise from the spinal cord as nerve roots (ventral and dorsal). They jet out through the openings of the vertebral column thus serving both sides of the body. Like the cranial nerves, the spinal nerves carry both sensory information and motor information to and from the spinal cord. The spinal nerves are named based on the spinal column region from which they exit. There are 8 neck (cervical), 12 trunk (thoracic), 5 lower back (lumbar), 5 pelvic (sacral) and 1 bottom (coccyx) spinal nerve (Spinal nerve organization).

Ganglia

The nerve cell bodies that are found in the PNS are organized into structures called ganglia which can be seen as enlargements along the peripheral nerves. The word ganglia originates from the Latin word for ‘swelling’. Functionally, PNS ganglia serve as relay points for information being received and transmitted. They fall into two major categories: sensory ganglia and autonomic ganglia. The sensory ganglia (of spinal nerves and some cranial nerves) contain thousands of cell bodies of neurons that extend their processes to the periphery. These are generally unipolar neurons (see Building a Nervous System above) that receive input from the periphery and send that information via axons to the spinal cord and the brain. The other category of PNS ganglia are the autonomic ganglia, which contain the cell bodies of autonomic nervous system neurons (ANS) and can be found in both the sympathetic and parasympathetic divisions of the ANS (Nervous system divisions).

The somatic and autonomic divisions of the PNS

The PNS can be divided into the somatic nervous system (SNS) and the autonomic nervous system (ANS). Both systems have motor and sensory components and both systems utilize the neurotransmitter acetylcholine. The ANS also uses norepinephrine. The main function of the SNS is the control of voluntary muscles and thus conscious actions such as walking. It consists of sensory and motor nerves. It receives somatosensory information from our skin, muscles, bones and joints, and targets skeletal muscles via motor innervation (Somatic vs autonomic nervous system). The cell bodies of these motor neurons are found in the gray matter of the spinal cord.

The ANS controls generally involuntary (autonomic, meaning self-governing) processes such as our breathing and heartbeat. It primarily controls the internal organs of the body, cardiac and smooth muscles, and glands (Sympathetic and Parasympathetic systems). We often hear the ANS referred to as a visceral system. The reason for this is because it communicates with visceral organs. These are organs in the chest, abdomen, and pelvis. Visceral sensory neurons monitor changes in temperature and stretch of visceral organs allowing the CNS to control and regulate their action via visceral motor outputs. We can further divide the ANS into three distinct parts: sympathetic, parasympathetic, and enteric systems. The sympathetic and parasympathetic systems have generally antagonistic or opposite functions in regulating organ functions. The same organs are innervated by both the parasympathetic and sympathetic systems causing opposite effects. These two systems differ in function, anatomical organization, and types of neurotransmitters that they use. We will briefly discuss each of these systems below (Sympathetic and Parasympathetic systems).

Sympathetic nervous system

The sympathetic system underlies the fight or flight response. This is a threat/emergency response mechanism which is not unique to only humans. The sympathetic nervous system is responsible for a lot of body functions related to hyperarousal such as increases in heart rate and blood pressure, inhibition of salivation/digestion and pupil dilation (Sympathetic and Parasympathetic systems). At the same time, non-necessary functions like digestion are put on hold while the body gets more blood flow to muscles. The sympathetic nervous system neurons arise from the thoracic and lumbar regions of the spinal cord. Fibers extend from the spinal region to its target organ such as the heart, liver, intestines, stomach, bladder, glands, etc.

Parasympathetic nervous system

The parasympathetic system acts in exact opposition to the sympathetic system and helps the body relax and digest. The parasympathetic system is responsible for slowing the heartbeat, stimulating salivation, digestion, urination, etc. (Sympathetic and Parasympathetic systems). Like the sympathetic nervous system, it also targets organs like the heart, intestines, bladder, etc. but causes opposite effects. The neurons of the parasympathetic nervous system arise from the brain (cranial nerves) and the sacral spinal cord region.

The enteric nervous system

The enteric nervous system is a net-like system of neurons that orchestrates gastrointestinal function via innervation of the digestive tract. It has more neurons than the spinal cord and is exceptionally complex. For example, it uses 30 different neurotransmitters. It is not entirely controlled by the CNS but is connected to the CNS via a communication network called the brain-gut axis. The enteric nervous system has both sensory and motor properties and controls the digestive system, smooth muscle, and glands (Enteric nervous system).

Communication between neurons happens chemically, at synapses

The chemical signaling that occurs between neurons happens primarily at chemical synapses (Chemical communication by neurons), specialized communication structures where a broadcasting neuron and a receiving neuron draw very close to one another, separated only by a small pocket of extracellular space called a synaptic cleft. One neuron releases transmitter into the synaptic cleft—we call this the pre-synaptic neuron and say that it contains the pre-synaptic terminal. The presynaptic terminal stores neurotransmitter in small membrane-bound bubbles called vesicles (Step 1 in Chemical communication by neurons). In response to an action potential within the neuron, these vesicles fuse with the pre-synaptic membrane and dump their neurotransmitter into the synaptic cleft (Step 2 in Chemical communication by neurons). The other neuron is studded with specialized receptors that recognize the transmitter and respond to it—we call this the post-synaptic neuron and say that it contains the post-synaptic terminal. There are many different neurotransmitters and many different transmitter receptors. Despite this complexity, each chemical message received is translated by receptors into one of just three responses in the post-synaptic neuron (Step 3 in Chemical communication by neurons):

• An Excitatory Post-Synaptic Potential (EPSP), a brief electrical change in the post-synaptic neuron that excites the neuron, pushing it towards firing an action potential.
• An Inhibitory Post-Synaptic Potential (IPSP), a brief electrical change in the post-synaptic neuron that inhibits the neuron, pushing it away from firing an action potential.
• Neuromodulation, a change in intracellular signaling in the post-synaptic neuron that modulates that neuron, changing its patterns of growth, connectivity, or signaling.

Almost as soon as it is released, neurotransmitter is broken down and recycled, inactivating the receptors on the post-synaptic membrane (Step 4 in Chemical communication by neurons). This ensures your synapses don’t clog up with accumulated neurotransmitter over time. More importantly, it tunes neural communication to the here and now, making their messages to one another extremely transient (short-lived). Each EPSP and IPSP lasts only a few milliseconds (bottom of Chemical communication by neurons). The one exception to this rule is modulation, which is usually short-lived, but which can also become long-lasting; more details on modulation are in Introduction.

The chemical communication that occurs between neurons is a specialization of systems that appeared very early in the history of life. Even bacteria can secrete chemicals and have receptors that stick through their membranes to detect and respond to chemicals in their environment. Many of the neurotransmitters and transmitter receptors used in your nervous system have long evolutionary histories, so it is not uncommon to find similar chemicals and proteins both in other forms of life and in other parts of your body. For example, histamine is an important neurotransmitter, but is also produced in white blood cells as part of an immune response to injury. This is part of the reason why substances in the natural world can influence your nervous system (caffeine!) and why drugs developed to treat brain disorders often have unwanted side effects in other parts of the body.

Although chemical communication between neurons has ancient roots, it has become highly specialized in neurons. One key specialization is that chemical communication is precisely targeted. Neurons release neurotransmitter almost exclusively at synapses. The tiny volume of the synaptic cleft ensures the post-synaptic neuron will get the message, and that other neurons (for the most part) will not. The pre-synaptic terminal also has specialized protein machinery to maintain a steady supply of neurotransmitter for release, and the post-synaptic terminal is loaded with receptors as well as other proteins that anchor the receptors and help fine-tune their responses to synaptic signals.

Although most communication between neurons occurs via chemical synapses, neurons can also signal to each other through electrical synapses (Electrical synapses); At an electrical synapse, neurons express a specialized protein, called connexon, which forms a protein bridge called a gap junction between the neurons. This enables electrical signals to pass directly from one neuron to another. In addition, small molecules can pass through an electrical synapse, so they actually allow for both electrical and chemical communication.

If you find communication between neurons fascinating, you’re not alone and you’re also in luck. When you get to Introduction, you’ll read more about chemical communication between neurons.

Information spreads within a neuron electrically

Neurons are specialized not only for communicating with each other, but also for generating internal electrical signals (Action potential spread and neurotransmitter release). Like almost all cells, neurons maintain an overall negative charge called a resting potential. In addition, neurons are electrically excitable, generating action potentials that sweep through a neuron at speeds up to 60 meters per second (134 miles per hour; Todnem et al., 1989). Each action potential is translated back into chemical messages released to partner neurons.

How does a neuron decide when to “fire” an action potential? Based, in part, on the constant barrage of chemical messages it is receiving on the post-synaptic side of its synaptic contacts. The EPSPs and IPSPs produced by these messages push a neuron above or below its threshold for generating an action potential (Step 1 in Action potential spread and neurotransmitter release). When the balance of excitation and inhibition being received is below a neuron’s threshold, the neuron does not fire an action potential or release transmitter to its partner neurons. When the balance of excitation and inhibition rises above a neuron’s threshold, it fires an action potential (Step 2 in Action potential spread and neurotransmitter release). The action potential spreads from the cell body along the length of the axon and all of its branches (Step 3 in Action potential spread and neurotransmitter release). As it spreads, the action potential is translated back into chemical messages, triggering the release of neurotransmitter from the pre-synaptic side of all the neuron’s synaptic contacts (Step 4 in Action potential spread and neurotransmitter release). A neuron firing an action potential is said to be activated or excited, with each action potential producing a burst of chemical signals to its synaptic partners.

The more excitation a neuron receives, the more frequently it fires action potentials. But this doesn’t make the action potential itself taller, stronger, or faster. Action potentials are ‘all-or-nothing’—each action potential a neuron generates is basically just like every other. There can be, however, considerable diversity between neurons. For example, some neurons fire action potentials that spread very quickly; others fire action potentials that spread more slowly. In addition, neurons are highly dynamic. Experience, disease, and maturation can change a neuron’s threshold, which synaptic partnerships it maintains, and more. For example, modulatory signaling in your nervous system is constantly fine-tuning action potential thresholds, lowering them to produce more activity during times of concentration and raising them to produce less activity during rest and sleep (more on this in Introduction).

Excitation and inhibition from partner neurons are not the only factors that determine when a neuron fires an action potential. Some neurons fire action potentials in response to changes in the outside world. We call these sensory neurons. In addition, most neurons in your nervous system generate action potentials spontaneously, meaning that they generate action potentials from time to time even without excitatory messages from partner neurons. For example, motor neurons, the neuron which release transmitter onto muscles, are often spontaneously active. This regular activity in your motor neurons maintains muscle tone (see Introduction). In spontaneously active neurons, EPSPs and IPSPs from partner neurons serve to speed up (when excited) or slow down (when inhibited) the rate at which action potentials are fired. In Internal and external triggers for action potentials you can see an example of a sensory neuron (left) responding to the outside world and a spontaneously-active motor neuron (right) that changes its rate of firing to control muscle tension.

The electrical signaling system in neurons is relatively new in the history of life, emerging at the dawn of the Animal kingdom at least 500 million years ago (though with important precursors in other life forms; Anctil, 2015). Electrical signaling enables neurons to coordinate information throughout your body with speed and precision, something that would be difficult to do with chemical communication alone. For example, you have sensory neurons in your toe (Simple reflex circuit). The axons of these neurons ascend in your sciatic nerve to the base of your spine (depending on how tall you are, that can be a distance of up to 1 meter!). Pinch your big toe and you will trigger action potentials in your toe sensory neurons that spread to the spine within a couple of milliseconds, releasing transmitter onto partner neurons in the spine. Some of these partner neurons will send signals back down to the muscles of your leg and foot to cause muscle contractions to jerk your foot away from the pain. Other neurons will generate action potentials that spread rapidly to the brain, hopefully triggering you to question your life choices:

Why would you pinch your big toe? Just because your neuroscience textbook told you to? These circuits, showing neural circuit from sensory neuron to spinal interneuron to motor neuron, are diagrammed at the top of Simple reflex circuit.

Neurons don’t just transfer information; they also process information. Each neuron synthesizes the excitation and inhibition received from thousands of synaptic partners, instantly tallying up these different influences to help determine when an action potential will be generated. In addition, the threshold for an action potential means that neurons filter information. Inputs that push a neuron above threshold produce action potentials, which then cause neurotransmitter release to partner neurons. Inputs that do not reach threshold are essentially ‘ignored’ or filtered out. For example, recall again the toe-pinching experiment shown in Simple reflex circuit. The interneuron in that circuit filters out gentle stimuli, ensuring only painful stimuli produce a withdrawal. If your toe is only lightly touched (bottom left), there won’t be very much sensory neuron activity, so the interneuron will receive some excitation but not quite enough to get to threshold. In that case, the interneuron does not fire an action potential, and the motor neuron never gets the message: the light touch is ignored. A strong pinch, though (Ow!), produces enough sensory activity to drive the interneuron to threshold (bottom right), and it passes along the message to the motor neuron, producing a behavior to get away from the painful stimulus. Thresholds give neural circuits the ability to ignore some events and respond to others, a key ability for using the body’s energy wisely.

In the swim circuit of Tritonia diomedea a few neurons produce rhythmic behavior important for survival

Tritonia is a species of slimy mollusk that glides along the ocean floor off the west coast of the United States and Canada. The mortal enemy of a Tritonia is the Pacific sea star, a voracious predator that loves to dine on Tritonia. To avoid this fate, a Tritonia “swims” away if it feels the touch of a sea star. Well, actually the Tritonia kind of thrashes about, arching its back, then relaxing, then arching again, in a rhythmic motion that helps it move up into the ocean current. When the Tritonia stops swimming, it sinks back down to the ocean floor, hopefully far away from the hungry sea star. Check out a video of a Tritonia escaping a sea star here:

Researchers have found that Tritonia swim away from danger with the help of a relatively simple neural circuit (Willows and Hoyle, 1969; Getting, 1983), consisting of sensory neurons, motor neurons, a partner neuron, and an inhibitory neuron.

The sensory neurons in the skin are tuned to specific chemicals on the tentacles of the Pacific sea star. When these are detected, the sensory neurons become very excited, firing a long-lasting barrage of action potentials (Step 2 in The swim circuit in Tritonia diomedea; each hash line represents an action potential). The sensory neurons release excitatory transmitter onto a set of 3 motor neurons which control the muscles of the back. When these motor neurons reach threshold they fire action potentials, contracting the back muscles to produce the arching movement that forms the first half of the ‘swim’ rhythm (a behavior that is cyclical or periodic).

This is not all the motor neurons do. They also release excitatory transmitter onto a partner neuron, which in turn excites an inhibitory neuron (Step 3 in The swim circuit in Tritonia diomedea). This inhibitory neuron provides feedback to the circuit, inhibiting the motor neurons. When this happens, the inhibition overwhelms the excitatory input from the sensory neurons, and the motor neurons are pushed below threshold. With the motor neurons inactive, the back muscles relax and the Tritonia straightens out. At this point, it has arched back and then relaxed forward, completing a ‘swim’ rhythm.

Can you predict what will happen next? Stop and think about it for a moment. If you predicted that the whole cycle would begin again, you’re spot on: with the motor neurons inactive, the partner neuron and, in turn, the inhibitory neuron are no longer being excited, so they stop firing action potentials and the motor neuron is released from inhibition. That means the long-lasting activity in the sensory neurons is again able to activate the motor neurons, producing another arching of the back (Step 4), followed by another round of inhibition that again relaxes the animal forward (Step 5 in The swim circuit in Tritonia diomedea). The whole cycle repeats over and over again until the sensory neurons stop firing, which usually takes 10 to 30 seconds (far longer than the brief touch of the predator). Thus, just a few neurons in the Tritonia nervous system transform an outside event (a sea star touch) into a complex and long-lasting rhythm that ‘swims’ it away from danger.

The Tritonia swim network is not only rhythmic; it is also dynamic, changing based on the Tritonia’s experience. For example, if a Tritonia is injured, the sensory neurons become hyper-excitable, firing at a lower threshold and for longer times. This shifts the Tritonia to be more likely to swim and to swim more vigorously, hopefully protecting it from further injury (but also using precious energy). The swim network can also shift in the other direction. If a Tritonia is gently touched over and over again, the sensory neurons produce smaller and smaller EPSPs, and the network begins to ignore this gentle touch, learning that it does not pose a real danger (Brown, 1998; Hoppe, 1998). The dynamic nature of the swim network helps the Tritonia strategically allocate its energy based on its experiences—to swim away from real danger while also saving energy by ignoring events that are innocuous (meaning harmless or non-threatening).

Even though the Tritonia swim circuit is simple, it can still malfunction. If the connection to the inhibitory neuron is weakened, there will not be enough inhibition to pause activity in the circuit, and instead the circuit will produce constant activity that could lock up the back muscles, freezing the Tritonia in an arched position that will make it into an easy dinner. Too much inhibition is also problematic. If the inhibitory neuron releases too much neurotransmitter, the circuit would pause for too long, letting the animal drop back down to the ocean floor before it has managed to get away from the sea star. Generating the swim rhythm that keeps a Tritonia safe requires just the right balance between excitation and inhibition to keep the circuit moderately but not excessively activated (Katz and Frost, 1997; Calin-Jageman et al., 2007).

Studying neural circuits reveals important principles about how nervous systems generate behavior

You probably never thought about the swimming behaviors of slimy mollusks before, but hopefully you found it exciting (pun intended!) to learn how just a few neurons can produce a life-saving behavior. One of the key goals of neuroscience is to uncover the secrets of more complex neural networks, such as the ones operating in your nervous system to produce language, emotions, and thought. There is still a lot we do not know, but even from the simple swim circuit of Tritonia we can glean a few key insights (Five principles of neural networks):

Neural networks feature parallel processing, meaning that information spreads along multiple pathways. In the Tritonia swim network, the motor neurons form synapses onto the muscles and onto a partner neuron that activates inhibition, sending messages along two distinct pathways at the same time.

Feedback, where a neuron influences the inputs it will later receive, makes even seemingly simple networks capable of producing complex patterns of activity.

Neural networks are often rhythmic or cyclical, exhibiting repeating patterns of activity and inactivity. This is reflected in the fact that many of our behaviors are also rhythmic: walking, sleep/wake cycles, breathing, and more. Introduction digs into detail on these fascinating cycles of neural activity.

Neural circuits can malfunction. They operate best at moderate levels of activity. They can easily be overwhelmed with excitation (which shows up in behavior as seizures or muscle spasticity) or inhibition (which shows up in behavior as torpor or muscle flaccidity). To work well, networks need both inhibition and excitation, and in the right balance.

Neural networks are highly efficient. Tritonia can swim away from danger with a network of less than a dozen neurons. Even with the large numbers of neurons in the mammalian nervous system, the efficiency of operation is incredible. Your entire brain uses about 20 watts of electrical power (Balasubramanian, 2021). For comparison, a modern Xbox or Playstation draws up to 160 watts of power. It is true that your brain uses a large fraction of your daily energy budget (about 500 kilocalories per day, or 25% of a typical 2,000 kilocalorie daily energy budget; Herculano-Houzel, 2012), but nervous systems are still remarkably efficient relative to the electronics around us.

What makes neural networks even more incredible is that they are self-assembled, following genetic and environmental signals to create and maintain the functioning of the network. How, exactly, this happens is still deeply mysterious. Introduction explains some of what we’ve learned about how neural circuits assemble.

Fundamentals of electricity: Charge, current, conductance, and potential

Electricity is the movement of charge. Charge is a fundamental property of sub-atomic particles: electrons have a negative charge; protons have a positive charge. For reasons even physicists don’t completely understand, charges exert force on one another: opposites attract and likes repel (Electrochemical force).

If a molecule has equal numbers of protons and electrons, then its net charge is 0 and the molecule is electrically neutral. When a molecule has an imbalance of protons and electrons, though, it is said to carry a charge, meaning that it has a net positive or negative charge that will exert electrostatic force on other charged particles. We call a charged particle an ion. In physics, charge is given the symbol $Q$ and measured in coulombs.

Forces induce movement. Because of the electrostatic forces between them, ions move towards their opposites and away from their likes, if they can. The movement of ions is an electrical current. It is this movement of charge that we call electricity. Anywhere we see work being done by electricity there must be a current flowing: a light bulb glows as electrical current flows through its light-emitting diode, your laptop responds to your commands as electrical current surges through its transistors. In physics, the symbol for current is $I$. We measure electrical currents in amperes (often shortened to amps), which tells us about the rate at which charge is flowing (number of Coulombs of charge per second).

When an electrical current flows, the path it takes can impede the movement of the ions or it can allow an easy passage. We call this conductance. Physicists use the symbol $G$ for conductance, which is measured in siemens. For example, water has a fairly high conductance, meaning that charged particles in water can rapidly respond to electrostatic forces, moving to disperse from their likes and to unite with their opposites. Cell membranes, on the other hand, have very low conductance, and act as insulator. (If you’ve had a physics class that discussed resistance rather than conductance, don’t panic: these are the same idea but expressed from different perspectives. Conductance is a measure of how easy it is to move charge; resistance is a measure of how difficult it is to move charge. You can convert back and forth between conductance and resistance: $G=\frac{1}{R}$; $R=\frac{1}{G}$.)

By acting as an insulator, cell membranes can hold electrostatic forces “in check”, holding ions in place even if electrostatic forces are pushing to move them. We call this pent-up energy electrical potential; it is a push or pressure just waiting to generate a current should a conductive pathway become available. In physics, the symbol for electrical potential is $V$, and it is measured in volts. You’ll often hear electrical potential referred to as voltage. Technically, this mixes up the unit (volt) with the concept (electrical potential), but the term has become so popular that we can accept voltage as a synonym for electrical potential.

Current, conductance, and potential are inter-related: $I=G\cdot V$ (Ohm’s law).

This relationship is known as Ohm’s law. This means the electrical current that flows in a system is determined by the product of the conductance (how easy it is for charge to move) and the potential (how much pressure there is for charge to move). You can use algebra to re-arrange Ohm’s law to solve for different electrical concepts (for example, $V=I/G$ is also Ohm’s law); this is because in all of its forms Ohm’s law expresses the idea that potential, conductance, and current are inter-related in such a way that knowing any two of these values instantly tells you the third.

Electrical signaling in neurons involves changes in potential and can be measured with a voltmeter

Electrical signaling in neurons involves changes in current, conductance, and potential. The easiest of these to measure in neurons, though, is potential. Because of this, scientists first characterized and named the electrical signals in neurons in terms of electrical potential: the resting potential, excitatory and inhibitory post-synaptic potentials, and the action potential.

We can measure a neuron’s membrane potential by connecting it to a voltmeter, a device for measuring electrical potential (Measuring changes in neuronal membrane potential with a voltmeter). All that is needed is a voltmeter and two wires, or electrodes. For neurons, this usually means placing one wire, called the recording electrode, inside the neuron, and a second wire, called a reference electrode, outside the neuron. The voltmeter measures the electrical potential between these two points, indicating the difference in charge across the neuron’s membrane, its membrane potential. For neurons, this is usually reported in millivolts (mV). A negative membrane potential (as a neuron has at rest) indicates the inside of the neuron has a net negative charge relative to the outside of the neuron. A positive membrane potential (as a neuron has at the peak of an action potential) indicates the neuron has a net positive charge relative to the outside. A membrane potential of 0 volts would indicate no imbalance of charge: the neuron is electrically neutral relative to the reference electrode.

In some ways, neurophysiology, the recording of electrical signals from neurons, is no different than when a mechanic checks the charge on your car’s battery. There are some huge technical difficulties, though, including the challenge of trying to place the tip of an electrode inside a microscopic neuron (!). The methods section in this book explains some of the specialized equipment that is used to make neurophysiology possible (see Introduction).

If you get your car battery checked at an auto shop, you will observe a stable potential (unless something has gone very wrong). In contrast, membrane potential in neurons is highly dynamic: it is negative at rest (Part A in Measuring changes in neuronal membrane potential with a voltmeter), becomes slightly more positive during an EPSP, becomes slightly more negative during an IPSP, and flips to positive and back to negative during action potentials. Because of this, we graph membrane potential as a function of time. When a neuron’s membrane potential increases, it means a current has flowed that either brought positive charge into the neuron or that removed negative charge (Part B in Measuring changes in neuronal membrane potential with a voltmeter). When a neuron’s membrane potential decreases, a current has flowed that either brought negative charge into the neuron or removed positive charge (Part C in Measuring changes in neuronal membrane potential with a voltmeter). When a neuron has a stable membrane potential, there is no net current, maintaining the neuron’s current charge (Part D in Measuring changes in neuronal membrane potential with a voltmeter). The currents that flow in neurons are typically quite small and are usually reported in nanoAmps (nA).

Another complexity of neural electricity is that membrane potential varies over the length of a neuron’s dendrites and axons. EPSPs and IPSPs alter membrane potential right at the synapse. Action potentials travel down the axon, with one segment flipping to a positive potential and then the next and then the next. Thus, we have to keep track of not only how membrane potential changes over time but also where membrane potential is changing.

Voltage-sensitive dyes enable us to measure potential along an entire neuron at once, so we can see the locations of different signals and how they spread (this figure). In contrast, a voltmeter can only measure membrane potential at the exact location of the recording electrode. Because of this, the first recordings of action potentials captured only the up and down of electrical potential at a single point, showing up as a “spike” in the membrane potential recording. The spread of an action potential had to be deduced from systematically moving the electrode along the length of the axon. When you see an action potential “spike” from a voltmeter, try to keep in mind that it is just a limited view of what is actually an electrical wave that spreads through the axon.

Electrical currents in neurons are the movement of four electrolytes dissolved in your cellular fluids: Na+, K+, Cl–, and Ca2+

What are the moving charges that generate electrical currents in neurons? They are trace minerals dissolved in your intracellular and extracellular fluids (Key electrolytes for neuronal signaling). Nearly all electrical signaling in neurons is due to the movement of just 4 minerals: sodium (chemical symbol Na⁺), potassium (chemical symbol K⁺), calcium (Ca²⁺) and chloride (Cl^–). Notice the positive and negative symbols next to the chemical symbols. That is because each is not only a mineral but also an electrolyte, something that dissolves into water as an ion, a charge-carrying molecule. Because electrolytes dissolved in water carry a charge, their movement is an electrical current. It is the movement of Na⁺, K⁺, Cl^–, and Ca²⁺ into and out of neurons that we observe as electrical signals in neurons.

In the next section, we’ll examine the pressures that drive this movement: electrostatic force and diffusion. Before we move on, you might be wondering: where do electrolytes come from? The K⁺, Na⁺, Cl^–, and Ca²⁺ dissolved in your cellular fluids come from your diet. For example, bananas are loaded with K⁺. You lose electrolytes in your urine and in your sweat, so you need daily intake from your diet to keep the proper balance of electrolytes in your body. That’s why sports drinks are always promising to “replenish your electrolytes”—they are selling you sugar water supplemented with a few milligrams of these basic minerals (though some leave out the calcium). Electrolytes are essential not only for generating neural electricity, but for nearly all cellular processes, including transcription and translation. This reflects the ancient origins of life in the Earth’s oceans: the Na⁺, K⁺, Cl^–, and Ca²⁺ dissolved in your cellular fluids are the same trace minerals found in salt water and are ubiquitous (ever-present) in all cellular forms of life. At the dawn of the animal kingdom, new mechanisms evolved to harness electrolytes for a new function: neural signaling.

Electrolytes are moved by electrostatic force and diffusion but their movement is blocked by the cell membrane

What moves the molecules of electrolytes dissolved in your cellular fluids? Two things: electrostatic force and diffusion (Diffusion).

We’ve already discussed electrostatic force. Every molecule of Na⁺, K⁺, Ca²⁺, and Cl^– dissolved in your cellular fluids carries a charge. Because of this, each is subject to electrostatic forces that push them towards their opposites and away from their likes.

Because they are dissolved in water, your electrolytes also diffuse, spreading out, as much as possible, towards equal concentration. The rules of diffusion are simple: when possible, solutes will always move down a concentration gradient, spreading from a region of high concentration to a region of low concentration, movement that would eventually produce equal concentrations. The steeper the concentration gradient (the more different the concentrations between regions) the stronger the push of diffusion.

Diffusion and electrostatic force sum and can produce strong pressure on ions to move. Water has a high conductance, so diffusion and electrostatic force easily move ions around inside your cellular fluids. The cell membrane, on the other hand, has a low conductance; ions cannot move directly through the cell membrane. Thus, a molecule of K⁺ inside a neuron cannot directly leave the neuron, even if electrostatic force and diffusion are pushing it to do so. You could say that the cell membrane is the Gandalf of neural signaling (“You shall not pass!”), but being that corny could lose you friends.

If the membrane blocks electrolyte movement, how do they travel into and out of neurons, generating the currents that carry signals in neurons? All ion traffic occurs via transmembrane proteins—amino acid machines that stick through the membrane. In the next sections we’ll discuss two families of proteins that control the movement of electrolytes: ion pumps and ion channels.

Ion pumps create concentration-gradient batteries, concentrating K+ inside neurons and Na+, Ca2+, and Cl– outside

For your cell phone to work, it needs a battery, something that can generate a steady electrical potential to push electrical currents through the circuits in your phone. Neurons also have batteries. They use ion pumps to create concentration gradients of the electrolytes found in your cellular fluids (Ion pumps).

All the cells in your body express ion pumps, specialized proteins that stick through the membrane. Pumps use the energy-molecule ATP to actively transport ions across the cell membrane, working against diffusion to concentrate ions inside or outside of the cell. These concentration gradients are chemical batteries, giving each ion a distinct “push” from diffusion. The bottom of Ion pumps shows several of the most prominent ion pumps that help maintain the concentration gradients for the 4 major electrolytes. For example, one type of pump pushes Ca²⁺ out of the cell. Once pushed outside the cell, Ca²⁺ cannot directly re-enter (because the cell membrane acts like Gandalf). Thus, pumps cause calcium to become much more concentrated in the extracellular fluid than in the intracellular fluid. The specific concentration gradient can vary in different tissues, but in your brain most neurons have about 10,000x the concentration of Ca²⁺ outside the cell membrane compared to inside (Erecińska and Silver, 1994). This represents a strong push of diffusion for Ca²⁺ to enter cells, a positive current.

Like Ca²⁺, Cl^– and Na⁺ are also pushed out by ion pumps. In neurons, Cl^– is usually 20x more concentrated outside the cell membrane compared to inside, and Na⁺ typically about 7x more concentrated. Thus, both Na⁺ and Cl^– have a pressure of diffusion to enter cells.

K⁺ stands out as the only major electrolyte pumped into cells. In neurons K⁺ is usually ~25 times more concentrated in the intracellular fluid than in the extracellular fluid. This means that there is a constant pressure of diffusion for K⁺ to leave cells, a negative current. this figure summarizes the concentration gradients pumps create for each of the electrolytes that play a key role in neural signaling.

Notice an interesting balancing act in the actions of the pumps. They import K⁺ into the neuron, filling the neuron with positive charge. This is mostly offset, however, by the fact that the pumps expel Na⁺ and Ca²⁺ from the neuron, removing those positive charges. In fact, if we tally up all the Na⁺ and Ca²⁺ and Cl^– outside the neuron against all the K⁺ inside we end up very close to an even balance sheet in terms of both concentration and charge (though not quite even, as we’ll discuss in the section on the resting potential). Thus, pumps, on the whole, do not directly charge the membrane, and they do not upset the important balance of net electrolyte concentrations that all cells must maintain.

What pumps do achieve is the build-up and maintenance of strong concentration pressures that the cell membrane can then hold in place: Na⁺ and Ca²⁺ are just raring to charge the neuron to a positive potential, K⁺ and Cl^– stand ready to charge the neuron to a negative potential. Note that K⁺ and Cl^– are both on “team negative” despite having different charges; this is because K⁺ has a pressure to leave but Cl^– has a concentration gradient to enter.

Ion channels are selective but passive conductors; some ion channels are gated

Look inside a laptop and you’ll see a tangle of wires, pathways of high conductance that allow electrical currents to flow. What are the ‘wires’ that carry electrical currents into and out of a neuron? It’s not the cell membrane, which has a low conductance. Instead, neurons (and all other cells) express ion channels (Ion channels). Ion channels are proteins that stick through the membrane. Each has a water-filled central pore through which specific ions can flow with high conductance. Ion channels allow currents of electrolytes to flow into and out of neurons.

Unlike pumps, ion channels are passive. Any ion that fits through an ion channel is “free” to move through it in either direction. The channel does not add any “push” or use energy to direct the flow of ions. Instead, ion movement through channels is determined purely by the pressures operating on that ion (diffusion and electrostatic force). this figure gives a summary comparison between pumps and channels.

Although ion channels are passive they are nevertheless selective. Some channels are conductive only to Na⁺; they are called Na⁺ channels. Other channels are only conductive to K⁺; they’re called K⁺ channels. As you can probably already guess, there are also Ca²⁺ channels and Cl^– channels.

Ion channels shows 3 common types of channels in the bottom panel. Some ion channels, called leak channels, are simply a pore with a selectivity filter—they provide a constant selective conductance. Other ion channels are gated, meaning that they can switch from open (a conformation with a high conductance) to closed (a conformation with a low conductance).

Gated channels don’t just open and close willy-nilly —they have sensors, specialized sections of their protein structure that determine when they open and when they close. For example, ligand-gated ion channels have sensors that bind to specific neurotransmitters, opening the channel only when the right transmitter is present. Voltage-gated ion channels have sensors that detect electrical events in a neuron, opening the channel only when the neuron reaches a specific membrane potential. This is just the tip of the iceberg. Other chapters will introduce you to channels that respond to many different events from the outside world (light, physical pressure, chemicals, and more). For this chapter, we’ll focus in on 3 families of ion channels that are critical for neural signaling: 1) leak K⁺ channels that produce the resting potential, 2) ligand-gated channels that produce EPSPs and IPSPs, and 3) voltage gated Na⁺ and K⁺ channels that produce the action potential.

Ionic Currents and Equilibrium Potentials

Pumps create concentration gradients that act like chemical batteries, storing up pressures of diffusion for Na⁺, Ca²⁺, and Cl^– to enter neurons and K⁺ to leave. What happens when an ion channel is open, providing a pathway to release that pressure? We get a current, the flow of ions into or out of a neuron. This current charges the neuron, changing its membrane potential.

Let’s think about this in detail (this figure).

	At rest, Na+ is high outside the neuron. Both electrostatic and diffusive forces pull Na+ in but it cannot flow with no channels open.
	Na+ channels open. Electrostatic and diffusive forces pull Na+ in the cell rapidly. This increases the membrane potential, weakening the elctrostatic attraction for Na+.
	As the neuron's potential becomes more positive, electrostatic repulsion starts to push Na+ out of the cell. At +52 mV, diffusion pushing Na+ in the cell is equally opposed by electrostatic repulsion pushing Na+ out.

Imagine a neuron at rest (top of this figure). Its membrane potential is negative, around -70mV, meaning that it has an excess of unpaired negative charges in the intracellular fluid. Pumps have concentrated Na⁺ outside the neuron, but it cannot re-enter directly through the membrane. What happens if Na⁺ channels now open? Na⁺ will come flowing in: it is pulled in both by the pressure of diffusion (going from the high concentration outside the neuron to the lower concentration inside) and by electrostatic force (positive Na⁺ is attracted to the negative charge in the neuron). Each molecule of Na⁺ that enters, though, offsets some of the neuron’s negative charge, so the membrane potential begins to rise (middle of this figure). If Na⁺ channels stay open, the neuron will eventually become neutral (its membrane potential will rise to 0mV); this occurs when enough Na⁺ has entered to fully offset all the negative charges that had given the neuron a negative resting potential. At this point, there is no longer any electrostatic pull for Na⁺ to enter, but there is still a strong pressure of diffusion, so Na⁺ will continue to enter, charging the neuron up to a positive potential. Will this continue forever? No: as the neuron becomes more positive it becomes increasingly repulsive to Na⁺ (likes repel), and eventually this electrostatic force will be able to stand up to the pressure of diffusion, preventing any further increase in Na⁺ (bottom of this figure). We call this the equilibrium potential. An ion’s equilibrium potential depends (mostly) on the concentration difference created by the pumps: the bigger the concentration difference the more charged the neuron has to become before electrostatic force can stand up to it. For Na⁺ in a typical neuron, the equilibrium potential is about +52mV. That is: when Na⁺ channels open in a resting neuron, Na⁺ will rush in and keep rushing in until the neuron is charged up to +52mV, but at that point it will stop further charging the neuron even if the Na⁺ channels remain open.

When Na⁺ enters the neuron through a Na⁺ channel, is all the hard work of the pumps undone? To some extent, yes, but not by much. Amazingly, only a few thousand Na⁺ molecules have to enter a neuron to charge it from rest to +52mV, a small fraction of all the Na⁺ the pumps have pushed out of the neuron. This is because the neuron is small and the membrane is very thin: each ion inside a neuron has little space to spread out and is held tantalizingly close to opposite charges on the other side of the membrane. So, when Na⁺ channels open, Na⁺ rushes in to charge the neuron to the Na⁺ equilibrium potential (around +52mV) but this “spends” relatively little of the concentration gradient pumps have worked so hard to create.

What happens when there is a K⁺ channel open? The pressure of diffusion is for K⁺ to leave a neuron, and this departure is a negative current that builds a negative charge for the neuron. As the neuron becomes more and more negative, though, there is an increasing electrostatic attraction for K⁺ to come back in (opposites attract), and eventually this electrostatic charge can stand up to the pressure of diffusion for K⁺ to leave. In a typical neuron, the equilibrium potential for K⁺ is around -87mV, meaning that when K⁺ channels are open K⁺ brings the neuron’s charge to a very negative potential. Again, though, only a small fraction of all the K⁺ in the neuron has to leave to charge a neuron to this point.

this figure lists the equilibrium potential for all 4 of the important electrolytes in neural signaling. These numbers are not there to memorize but to help you build some context and understanding. Focus especially on the last column: Na⁺ and Ca²⁺ generate currents that pull a neuron up to positive potentials while Cl^– and K⁺ generate currents that pull a neuron down towards negative potentials. Keeping those general trends in mind is key for understanding how resting, post-synaptic, and action potentials are generated.

The resting potential is produced by leak channels and concentration gradients

The resting potential is an overall negative charge maintained by a neuron, typically around -65mV (though the exact value varies from neuron to neuron and can change). Neurons are sometimes said to “defend” their resting potential, meaning that after any stimulation they have a strong tendency to return quickly to rest. Neurons are not alone in having a resting potential—most of the cells in your body also maintain an overall weak negative charge.

The resting membrane potential is primarily due to the concentration-gradient for K⁺ (more inside than out) and special ion channels called leak K⁺ channels. As their name suggests, leak K⁺ channels are selective for K⁺ and always open (Leak channels).

Not only do leak K⁺ channels allow K⁺ to produce a negative charge on a neuron, they allow the neuron to keep returning to that negative charge. Leak channels defend the resting potential shows how this works.

For a neuron at rest (step 1), the K⁺ leak channels have allowed a few molecules of K⁺ to leave the neuron, leaving the inside of a neuron with a negative charge. Imagine we now make the neuron more positive by using an electrode to inject some positive charges (step 2). This upsets the balancing act between diffusion and electrostatic force for K⁺; the electrical pull into the neuron is now weaker (because the neuron is not as negative) so the push of diffusion to eject K⁺ temporarily “wins”, expelling a few more molecules of K⁺. Thus, the leak channels allow K⁺ to work against the positive current we are injecting and limit the degree we can charge the neuron (step 3). When we stop injecting current (step 4), K⁺ continues to depart until it pulls the neuron right back down to its normal resting potential. This is what we observe as the “defense” of the resting potential: the leak K⁺ channels provide a constant conductance that always allows K⁺ to pull the neuron’s membrane potential back towards the negative equilibrium potential for K⁺.

If leak K⁺ channels allow K⁺ to work against changes to a neuron’s membrane potential, how can there still be EPSPs, IPSPs, and action potentials? Shouldn’t K⁺ flow through leak channels to instantly cancel out any change in a neuron’s charge? The answer is that leak K⁺ channels are only weakly expressed in neurons; there are relatively few of them embedded in each neuron’s membrane. This low conductance keeps the flow of K⁺ fairly slow, so it does work against changes in membrane potential, but not instantly.Interestingly, neurons can change their excitability by changing how many leak K⁺ channels they have: the more leak channels expressed the more rapidly a neuron returns to rest and the harder it is to get it to threshold; the less leak channels a neuron has the easier it is for the neuron to reach threshold and fire action potentials.

Post-synaptic potentials are produced by ligand-gated channels

At the beginning of the chapter we found that neurons don’t just rest; they chatter away, firing action potentials that release chemical messages that excite (EPSP) or inhibit (IPSP) their partners (see Neural Communication). Each EPSP and IPSP is a small, temporary change in the membrane potential that is produced because neurotransmitter has been released by a synaptic partner. EPSPs and IPSPs are like “Yes” and “No” votes that determine if a neuron will get to threshold and fire an action potential. We’ll get to the action potential in the next section. For now we consider: how do chemical messages from a partner neuron get translated into the electrical “votes” we observe as EPSPs and IPSPs?

EPSPs and IPSPs are produced by ligand-gated ion channels (Ligand-gated channels).

These are transmembrane proteins with a central pore. Each has an external binding site that recognizes a specific neurotransmitter. When a molecule of neurotransmitter fits into the binding site the channel is pulled open; when no transmitter is present, the channel stays closed. The post-synaptic side of each synapse is studded with hundreds of ligand-gated ion channels, so each burst of transmitter release from the pre-synaptic neuron produces a temporary but substantial opening of channels on the post-synaptic membrane, translating the chemical transmitter signal into a sharp change in conductance that allows current to flow.

There is tremendous diversity in ligand-gated ion channels, with the human genome containing genes encoding several hundred ligand-gated ion channels (Viscardi et al., 2021). Part of this diversity is due to differences in binding sites, with different receptors specialized for detecting different neurotransmitters. Introduction will discuss the different classes of neurotransmitters and their receptors in more detail.

What happens when a ligand-gated ion channel opens? That depends on its selectivity. If the channel is selective for Cl^–, the channel produces IPSPs (top of EPSPs and IPSPs).

That is because in a neuron at rest there is pressure for Cl^– to enter the neuron (see Principles of Bioelectricity). The higher outside concentration of Cl^– enables diffusion to push Cl^– through the open ligand-gated channels, producing a negative current that decreases the neuron’s membrane potential, moving it further away from the threshold for generating an action potential.

If the channel is selective for Na⁺ or Ca²⁺, the channel produces EPSPs (bottom of EPSPs and IPSPs). That is because both Na⁺ and Ca²⁺ carry a positive charge and are more concentrated outside the neuron than in, giving them a strong pressure to enter a neuron at rest. Therefore, an increase in conductance for these ions provides an opportunity for diffusion to produce a positive current that moves the neuron’s membrane potential towards threshold.

As you’ll have noticed, there are multiple steps for an EPSP or IPSP to be produced: transmitter has to be expelled from the pre-synaptic neuron, diffuse across the synaptic cleft, and then bind to a ligand-gated channel (Chemical communication by neurons). Only then does the post-synaptic neuron begin to get the influx of Na⁺, Ca²⁺, or Cl^– that creates an EPSP or IPSP. This multi-step process creates a “synaptic delay”, a gap of up to 5 ms between the partner neuron firing the action potential and the post-synaptic neuron exhibiting the EPSP or IPSP. This time penalty for converting from electrical message (action potential) to chemical message (neurotransmitter) and back again (EPSP or IPSP) is one of the key factors limiting how quickly we can respond to outside events.

Once an EPSP or IPSP finally gets started, it very rapidly ends, usually within just 1ms. The transient (short-lived) nature of EPSPs and IPSPs is due, in part, to the fact that the neurotransmitter released by partner neurons is rapidly inactivated and recycled. As neurotransmitter is cleared from the synapse, the ligand-gated channels close. A second limiting factor are the K⁺ leak channels: almost as quickly as Na⁺ or Cl^– can enter to produce an EPSP or IPSP, leak channels allow K⁺ to undo that work and pull the neuron back towards the neuron’s normal resting potential. Thus, each message from a partner is processed but quickly cleared; this keeps neurons tuned to the here and now (though neuromodulation and neuroplasticity enables long-term changes in the nervous system). Each synaptic message leaves the neuron with a few ions “out of place” (e.g. some extra Cl^– inside with each IPSP). Pumps work continuously to put things back to normal. The constant need for pumps to reset ion concentrations is why they consume a large portion (about 28%) of your brain’s energy budget (Lennie, 2003).

Another notable feature of EPSPs and IPSPs is that they are usually quite small relative to a neuron’s threshold. For example, in a human cortical neuron, a single EPSP is estimated to increase the membrane potential at the soma by only 0.3 mV, whereas the threshold for generating an action potential usually requires about a 30mV increase in potential (Eyal et al., 2018).

If each partner message is so small, how does a neuron ever reach threshold? The answer is that EPSPs and IPSPs summate over space and over time. Temporal and spatial summation shows how summation works.

At the top of the figure we see a neuron receive an EPSP from a partner neuron (E1) that is subthreshold, meaning it is too small to take it to threshold. This partner neuron can try again later (second arrow labeled E1), but the result is the same: poor E1 is being ignored by its synaptic partner! Things change, though, if E1 fires action potentials more rapidly (bottom left of Temporal and spatial summation), delivering the second EPSP before the first one has fully faded. In this case, a second round of Na⁺ or Ca²⁺ can enter before the postsynaptic neuron has returned to rest from the first EPSP, and this overlap brings the neuron much closer to threshold. We call this temporal summation because it adds up EPSPs from the same partner that occur close together in time. Neurons can only fire so fast, though. A far more effective way to get a partner neuron to threshold is to team up with its other inputs. As shown in the middle part of Temporal and spatial summation, if a neuron receives EPSPs from two different partners (E1 and E2) at the same time this creates a bigger EPSP than either could produce on its own. We call this spatial summation because it sums inputs from multiple locations across a neuron’s dendritic tree. Of course, neurons aren’t just tallying the “yes” votes of EPSPs: IPSPs also factor in, summing with ongoing EPSPs to pull the neuron away from threshold (bottom right of Temporal and spatial summation). Synaptic summation is just the simple addition and subtraction of charges, and yet it is what underlies the incredible ability of neurons to synthesize and integrate information.

The key point for synaptic summation is the axon hillock (also called the initial segment). The hillock is the very first part of the axon where action potentials are initiated, so this is where the sum of current EPSPs and IPSPs can tip a neuron past threshold. It’s worth reflecting again that most EPSPs and IPSPs make a very small impact on the axon hillock relative to its threshold: it is estimated that it would typically require the summation of about 100 EPSPs to trigger an action potential in a resting human cortical neuron (Eyal et al., 2018). That sounds like a lot, but keep in mind that a typical cortical neuron has about 7,000 synaptic contacts! Thus, receiving 100 excitatory messages at about the same time is a fairly common occurrence, though not so common that your cortical neurons are relentlessly activated.

The action potential is produced by sequential opening of voltage-gated Na+ and K+ channels.

The action potential is a rapid up-and-down of electrical potential that spreads through a neuron to trigger transmitter release. this figure diagrams each step of the action potential.

We start with a neuron at its typical negative resting potential (step 1) that then receives excitation strong enough to bring the neuron to threshold (step 2). Once at threshold, the action potential is initiated, starting in the cell body or right where an axon branches off from the cell body, a point called the initial segment or axon hillock (these terms are interchangeable). We observe the action potential as a rapid and dramatic rise and fall of electrical potential (step 3). During the rising phase (step 3), the neuron’s membrane potential flips from negative all the way up to a positive potential of around 40mV (though this varies from neuron to neuron). In a typical cortical neuron this upward climb in potential takes only about 0.5ms. This is followed immediately by the falling phase, when the neuron’s membrane potential descends just as quickly, falling back down to a negative potential and becoming even more negative than during the typical resting potential, often reaching about -85mV. The descent “below” rest is called the undershoot (step 4), and it takes a typical cortical neuron about 4-5 ms to gradually return to the typical resting potential (step 5). Critically, this same rise, fall, and undershoot then occurs a bit further down the axon, then even further down, and so on. As the action potential propagates down the axon, it triggers transmitter release at each presynaptic membrane it passes.

The action potential is due to the operation of two specialized ion channels that are relatively unique to neurons: inactivating voltage-gated Na⁺ channels and non-inactivating voltage-gated K⁺ channels.

Inactivating voltage-gated Na⁺ ion channels are highly expressed in neurons, and are often inserted in the cell membrane specifically within the initial-segment and axon. Scientists classify these channels as voltage-gated because they have a sensor that holds them closed when a neuron is near its negative resting potential, but that pulls them open when a neuron’s membrane potential rises to threshold. A better name, though, would be excitation-sensing because it is excitation (EPSPs) from partner neurons that makes a neuron’s membrane potential positive enough to begin opening these channels. This gating of the voltage-gated Na⁺ channels is what we observe as the threshold for an action potential.

Action potential propagation shows how the inactivating voltage-gated Na⁺ channels produce the rising phase and propagation of the action potential.

When threshold is reached, these channels open (step 1), producing a sharp increase in conductance for Na⁺: there are now many pathways available for diffusion to push Na⁺ into a neuron (recall that pumps have concentrated Na⁺ outside the neuron, so the pressure of diffusion is for Na⁺ to enter, see Principles of Bioelectricity). The result is a rapid influx of Na⁺ that quickly pulls the neuron’s membrane potential all the way up to a positive membrane potential. This is what we observe as the rising phase of an action potential.

The rising phase of the action potential propagates down the axons because the opening of one voltage-gated Na⁺ channel produces a chain reaction (step 2 in Action potential propagation). The first voltage-gated Na⁺ channel to open allows Na⁺ to enter, causing an increase in membrane potential—the very thing that can trigger the opening of more voltage-gated Na⁺ channels! This positive-feedback loop, where each Na⁺ channel that opens can help trigger additional openings causes the action potential to propagate down the axons, triggering transmitter release at all that neuron’s pre-synaptic terminals along the way.

Although the rising phase on an action potential propagates through the axon, it is very brief, followed very quickly by a plunge back down to a negative potential. One factor limiting the rising phase is the inactivation of the voltage-gated Na⁺ channels. Each channel has a “tail” of positively charged amino acids on the intracellular side of the membrane. When these channels open and drive a neuron to a positive potential, the tail is repelled, and this repulsion actually pushes it into the channel, clogging it! Even though the voltage-sensor is still pulling the channel open (because the neuron has a positive membrane potential), the inactivating tails temporarily eliminates the Na⁺ conductance through these channels (step 3 of Action potential propagation).

The clogging of the inactivating voltage-gated Na⁺ channels is reversible. When a neuron returns to a negative resting potential, that negative charge attracts the positive charges in the inactivating tails, pulling them out of the Na⁺ channels so that they can work again. This amazing system helps prevent excessive activity in the nervous system: once a neuron generates an action potential, it cannot generate another until it returns to rest and the majority of its voltage-gated Na⁺ channels are unclogged. We observe this as a brief refractory period that occurs immediately after each action potential (recall, again, step 4 of this figure).

Action potential generation recaps these steps in the action potential, showing how a neuron goes from rest (step 1) to the rising phase (step 2) as the inactivating voltage-gated Na⁺ channels open, but then clog (step 3). With the clogging of the voltage-gated Na⁺ channels, Na⁺ can no longer enter the neuron. This means the leak K⁺ channels can now begin restoring the resting potential, allowing K⁺ to leave (it is more concentrated outside than in, and during the rising phase is also repulsed by the neuron) to bring the membrane potential back down to a negative potential. We observe this as the falling phase of the action potential.

The falling phase is “turbo charged” by voltage-gated K⁺ channels. These channels are somewhat similar to the voltage-gated Na⁺ channels: they have a sensor that detects increases in the membrane potential, swiveling them open when there is sufficient excitation. When these channels are open, they provide additional conductance for K⁺ above and beyond what is provided by the leak K⁺ channels. This allows K⁺ to be more rapidly pushed out of the neuron by diffusion, accelerating the falling phase (step 4 of Action potential generation). You can think of K⁺ as soccer fans trying to leave a stadium after a match. Everyone could be anxious to leave, but if there are only a few exits it will take a long time for the stadium to empty. The opening of the voltage-gated K⁺ channels is like adding a bunch of new exits to the stadium, allowing a much more rapid departure of the crowd. Without the voltage-gated K⁺ channels, it would take a typical neuron about 2-3 times as long to return to rest after an action potential, and that would greatly limit the frequency at which neurons can send messages.

Although the voltage-gated K⁺ channels respond to excitation in the same way that Na⁺ channels do, they are different in two key ways. First, the K⁺ channels do not inactivate—they stay open as long as the neuron has a positive charge, working continuously to allow K⁺ to depart to drive the neuron back down to its negative resting potential. A second distinctive feature is that the voltage-gated K⁺ channels are slow—their voltage sensor takes longer to open and to close the channels than the sensor on the voltage-gated Na⁺ channels. This “tardiness” is important, causing the K⁺ channels to only begin opening just as the Na⁺ channels are beginning to clog. This remarkable feat of timing makes the action potential extremely efficient, limiting the overlap between the increases in Na⁺ and K⁺ conductance that happen during the rising and falling phases. If the K⁺ channels opened earlier, K⁺ would be able to diffuse out of the neuron as quickly as the voltage-gated Na⁺ channels were allowing Na⁺ to diffuse in, an offsetting current that would produce no signal in the membrane potential!

A second consequence of the slow operation of the voltage-gated K⁺ channels is the undershoot (step 5 of Action potential generation). Even after the neuron begins to return to a negative charge, it takes a few milliseconds for the K⁺ channels to swivel closed. During this time, K⁺ continues to be able to leave the neuron, making the neuron temporarily more negative than usual. This is what we observe as the undershoot phase (step 6). The enhanced negative charge of the undershoot phase helps to unclog the voltage-gated Na⁺ channels, ending the refractory period (step 6). Ending the refractory period means the neuron is now capable of firing another action potential (the voltage-gated Na⁺ channels have been reset), and yet the undershoot makes this a bit less likely than usual by keeping the neuron a bit further from threshold for a few milliseconds. Thus, the undershoot both prepares the neuron for sending the next message while also helping to ensure the neuron isn’t hyper-active. What a clever system!

this figure provides another way of organizing all this information, providing an overview table of what is happening in each phase of the action potential.

We see that each action potential is a sequential opening of voltage-gated Na⁺ channels and voltage-gated K⁺ channels, progressing down an axon and into all its synaptic terminals. The rapid opening of the voltage-gated Na⁺ channels allows the pressure of diffusion to rapidly push Na⁺ into the neuron, driving it from its negative resting potential all the way up to a positive potential that then opens the next voltage-gated Na⁺ channel and the next and the next, all the way down the axon. The voltage-gated Na⁺ channels clog, however, stopping the rising phase, and just as this happens the voltage-gated K⁺ channels finally open, allowing the pressure of diffusion to push K⁺ out of the neuron, a negative current that rapidly brings the neuron back down to its negative resting potential and even temporarily making the neuron more negative than usual. Each action potential, then, lets a bit of Na⁺ into the neuron and a bit of K⁺ out, undoing some of the hard work of the ion pumps. It’s a surprisingly small amount of Na⁺ and K⁺ that actually shuffles around, producing only a fractional change in concentrations. Still, ion pumps have to continuously operate to undo the Na⁺ entry and K⁺ departure that occurs with each action potential, otherwise the pressures of diffusion would gradually dissolve away and action potentials could no longer be generated (see Our Deep but Still Incomplete Understanding of Neural Signaling for an interesting example of how this can happen).

Action potential propagation is sped up by myelin

When you connect a battery to a wire, it pushes current directly through the circuit. As long as the wire is properly insulated, almost no current leaks out as it travels, so you don’t need additional batteries along the length of the wire, and current covers the distance of the circuit at very close to the speed of light. That’s why flipping a light switch instantly turns on your lights, even though the power station might be located several miles away.

As we have seen, that’s not exactly how an action potential propagates. Instead, neuronal axons are “leaky”, expressing leak K⁺ channels. This means that when the first segment of an axon experiences the rising phase, the Na⁺ that rushes in can quickly be offset by K⁺ rushing out. Due to this leak, the action potential is regenerated down the length of the axon, with each segment undergoing the precise opening and closing of voltage-gated channels that triggers the next segment, and so on. This takes time, and it means that the action potential does not travel down the axon at anything near the speed of light.

Many animals, including all mammals, have evolved ways to speed up action potential propagation by having support cells wrap axons in myelin, a fatty substance that insulates the axon, greatly limiting the operation of the leak channels (Saltatory conduction).

A segment of axon wrapped in myelin works more like the wires we’re used to dealing with in electronics: a current applied to one end can spread rapidly, at almost the speed of light, with fairly little leakage. That’s a great speedup in the spread of an action potential. Unfortunately, myelin isn’t perfect, and there is still some leakage, so it can only effectively deliver currents over relatively short distances, around 1mm. Because of this, myelin is wrapped around axons in bands and between these bands are nodes of Ranvier, bare patches of axon crowded with voltage-gated channels that can regenerate the action potential and push current through the next band of myelin. The mix of relatively slow regeneration (at each node of Ranvier) and exceptionally fast current spread (through each segment of myelin) is called saltatory conduction, and it lets action potentials travel fast, at speeds of up to 150 meters per second (333 miles per hour). That’s very slow compared to the currents in your cell phone, which approach the speed of light (almost 300 million meters per second or 670 million miles per hour). But it’s still much faster than action potentials can travel in bare, unmyelinated neurons, which typically ranges between 0.5 to 10 meters per second (1 to 22 miles per hour).

Wrapping axons in myelin requires support cells and energy, so not all axons in the CNS are myelinated. But most tracts, where axons from many neurons carry information over long distances feature almost exclusively myelinated axons. The high density of fatty myelin in these tracts is what gives white matter its distinctive appearance (see Introduction).

Neural signaling is complex

While we can feel triumph at the way neuroscience is helping to illuminate medical mysteries, it is important to note that this chapter only scratches the surface of what neuroscientists have discovered about neural signaling. For both space and clarity, some topics have been greatly simplified. That’s fine: we needed to start somewhere, and this chapter was already quite long and complicated, right? But it’s worth at least a peak behind the curtain at some of the additional complexities of neural signaling:

• Chemical synapses are not only one-way. In this chapter we’ve emphasized communication from the pre-synaptic to the post-synaptic neuron, noting that the pre-synaptic terminal is loaded with vesicles of transmitter to send messages and the post-synaptic terminal is studded with receptors to receive messages. It turns out this is only part of the story. The pre-synaptic terminal also has receptors (often called autoreceptors) and the post-synaptic terminal releases messages back to the post-synaptic neuron (these are often called retrograde messengers). So while there are clear specializations for communication from pre- to post, it is more accurate to think of chemical synapses as a point of interaction between neurons. This will be discussed more in Introduction
• Action potentials aren’t always one-way either! In this chapter we emphasized that action potentials are initiated at the initial segment of the axon and are then propagated down the axon. This is also just part of the story! In some neurons, action potentials also backpropagate, meaning that when they reach the end of the axon they then propagate back up the axon to the cell body and sometimes also into the dendritic tree. In some experiments, blocking back-propagation has impaired plasticity, suggesting an important role in fine-tuning connectivity (Stuart et al., 1997).
• Synaptic messages are often quite complex. We’ve explained how neurons release transmitter to partners to produce EPSPs (when binding to a ligand-gated Na⁺ channel), IPSPs (when binding to a ligand-gated Cl^– channel), or neuromodulation (explained in Introduction). These are, indeed, the fundamental types of messages that can be communicated at chemical synapses. Things become more complex, however, when we realize that a pre-synaptic neuron can actually release multiple transmitters (called co-transmitters) and that the post-synaptic neuron can express multiple types of receptors. This means that what one neuron “says” to another at a chemical synapse can be very complex, often involving a blend of excitation, inhibition, and modulation!
• It’s not just neural signaling taking place in the nervous system; glia are involved, too. Glia have long been considered mere support cells in the nervous system, but there is increasing evidence that they play important roles in nervous system communication as well (Allen and Lyons, 2018). Glia express many different types of receptors, they can both absorb and release transmitters, and they exchange signals with neurons that seem to play important roles in determining which synapses a neuron forms and maintains. We will learn much more about microglia specifically in Introduction.

Neurons are astonishingly diverse. This chapter has tried to describe signaling in a “typical” neuron. This leaves out the tremendous diversity of neurons. First, there is incredible variety across species. Not all species have myelin. Not all species have a clear distinction between dendrites and axons. In fact, some species even lack inactivating voltage-gated Na⁺ channels and instead have Ca²⁺-based action potentials! Even within a species, there is incredible variety in size, shape, and signaling (Diversity of neurons).

There is still so much to learn about neural signaling

We shouldn’t leave this chapter with the impression that neuroscientists have figured it all out. There is still so much about neural signaling that we don’t know, and there is still tremendous promise for applying the bits we do know to improve our lives. So, let’s end this chapter by highlighting just a few of the many mysteries still to be resolved. That way we can close inspired by the possibility of helping to solve these mysteries, and with dreams of the better world we might build with that knowledge.

• How do neuronal circuits maintain their function despite changing circumstances? Most humans learn to walk early in life, at about 1 year of age. By adulthood, however, you are about twice as tall as when you learned to walk. That means the walking circuits in your spine have had to adapt throughout your lifetime. As you grew, the axons and dendrites in your walking circuits had to be extended over longer and longer distances. As your mass and center of gravity changed, inputs and outputs had to be adjusted to adapt the muscle commands for walking. Through it all, you experienced no major disruptions: you just kept on being able to walk! This is just one example of homeostasis in neural circuits: their ability to maintain the same function despite tremendous change (Marder and Goaillard, 2006). We don’t fully understand how this works (does each neuron have an ‘ideal’ level of activity it is striving for?) or why some changes are easy for a neural circuit to cope with while others cause dysfunction.
• How do neuronal circuits self-organize? Each of your neurons carries your entire genome, about 6.4 billion base-pairs of DNA. That’s an impressive amount of DNA, but it is simply not enough DNA to provide a complete wiring diagram for each of the 86 billion neurons in your brain. And yet, most human brains end up with striking and recognizable similarities in organization, with axons of sensory neurons finding their way to the thalamus, axons of the primary motor cortex descending down to the lumbar spine, etc. How does the human brain self-assemble? And is the assembly program fixed or does it incorporate feedback from the developmental environment? Introduction discusses in more depth what we know and don’t know about the self-organization of neural circuits.
• How many types of neurons are in the human brain? Neuroscientists have long sought to create a classification system for neurons: to develop a complete list of the different types of neurons in the human brain (Bakken et al., 2021). While some differences seem clear (compare some of the different neuronal shapes in Diversity of neurons), it has proven remarkably difficult to come up with a clear and consistent classification system. Molecular analysis has shown that two neurons that look alike might express markedly different channels and transmitters. Meanwhile, neurons that look different can serve very similar functions. As an added layer of complication, neurons are constantly adapting to new circumstances, so what seems like one type of neuron might adapt over time in ways that make it very difficult to classify. Maybe we’re not thinking about neuron identity correctly? Or maybe we just need to crunch more data to see the patterns? Or, perhaps, neurons are so adaptable that detailed typologies aren’t possible. This is a sobering gap in our knowledge: despite all we do know about the human brain we still don’t have a master parts list! There really is so much more to learn.

How do neurons produce neurotransmitters?

Neurotransmitters are produced inside neurons. Their production depends upon the neuron obtaining an adequate supply of precursors from the blood. This is not as easy as it sounds. The production of neurotransmitters requires a variety of amino acids, sugar, fats, vitamins, and minerals. These nutrients float around in the blood after each meal. Every cell in the body competes with the brain for access to these nutrients. The brain is at a distinct disadvantage because it sits behind a firewall called the blood-brain barrier (see Introduction). However, fortunately, the blood-brain barrier possesses a variety of specialized transporting mechanisms that can actively transfer nutrients into the brain (Step 1 in Neurotransmitter lifecycle). Once a nutrient crosses the blood-brain barrier it is usually transported to neurons with the assistance of astrocytes.

After an amino acid, fatty acid or sugar molecule is inside a neuron, it is generally acted upon by synthesizing enzymes (Step 2 in Neurotransmitter lifecycle). Enzymes are proteins that help build, breakdown or otherwise modify substances in our cells. The action of many of these enzymes uses metal ions as co-factors to assist with the conversion process. For example, neurons in the substantia nigra contain iron, neurons in the hippocampus contain zinc, neurons in the locus coeruleus contain copper, and neurons in the red nucleus contain cobalt. The presence of these metals is critical for normal neurotransmitter metabolism.

Brains demonstrate a degree of energy efficiency in terms of which of these transmitters it utilizes most extensively. For example, GABA and glutamate are simple amino acids and are the most abundant neurotransmitters in the human brain. They require only single-step enzymatic modification. Four amine neurotransmitters (dopamine, norepinephrine, epinephrine, and serotonin) require enzymatic modification of dietary amino acids to produce and inactivate. The brain contains 1000-fold fewer of these amine-releasing neurons, as compared to glutamate and GABA, probably because they require more energy to function. Finally, the other, far less abundant neurotransmitters, such as the neuropeptides (which are a string of amino acids) require a considerable amount of energy by the neuron to produce and inactivate. Neuropeptides occur roughly one-million-fold lower concentration in the brain than glutamate and GABA. These neurotransmitters, as well as the fat-derived neurotransmitters endocannabinoids and prostaglandins, all utilize the DNA in the nucleus to produce the necessary enzymes and precursors. The endogenous levels of these two fat-derived are kept quite low and are produced only when needed.

How are neurotransmitters stored and released?

Once synthesized, most neurotransmitter molecules are then actively transported into synaptic vesicles (step 3 Neurotransmitter lifecycle); these are very tiny spheres clustered at the presynaptic terminal with hollow centers into which approximately 10,000 molecules of a typical neurotransmitter can be stored for later release from a neuron. Neurons pay attention to the shelf life of their neurotransmitters; they prefer to release the most recently produced neurotransmitter molecules first. This means that the freshest products are released first, thus guaranteeing that the communication between neurons is successful. Many neurons produce two (and rarely three) different neurotransmitters; each is stored inside different vesicles and are released independently of each other.

The arrival of the action potential at the end of the axon opens a group of voltage-dependent calcium channels that allows the entry of calcium ions (see Introduction). The elevated intracellular concentration of calcium ions initiates the next step in the communication of one neuron with the next: A synaptic vesicle either merges into its cell wall (both are made of lipids so imagine two soap bubbles coming together), or opens a small pore into the neuronal membrane, and releases the neurotransmitter into the small space between neurons, called a synaptic cleft (step 4 Neurotransmitter lifecycle). Vesicle fusion shows the process of vesicle fusion in more detail. The junction at which two neurons communicate via the release of a neurotransmitter molecule is called a synapse. Because the synapse is such a small space, the concentration of neurotransmitter molecules in the synapse following release becomes very high for a brief period.

Post-synaptic actions of neurotransmitters

Once released into the synapse, neurotransmitter molecules briefly interact or bind with a protein, called a receptor, that is (usually) located on the surface of the neuron on the other side of the synapse (step 5 in Neurotransmitter lifecycle). These receptors are divided into two general types: either ionotropic or metabotropic receptors (Metabotropic vs ionotropic receptors). Ionotropic receptors are transmembrane molecules that can “open” or “close” a channel that would allow smaller particles to travel in and out of the cell. Ionotropic receptors allow different kinds of ions to travel in and out of the cell. Ionotropic receptors are not opened (or closed) all the time. They are generally closed until another small molecule (called a ligand — in our case, a neurotransmitter) binds to the receptor. As soon as the ligand binds to the receptor (step 1 in the left side of Metabotropic vs ionotropic receptors), the receptor changes conformation (the protein that makes up the channel changes shape), and, as they do so, they create a small opening that is big enough for ions to travel through (step 2 in the left side of Metabotropic vs ionotropic receptors). Therefore, ionotropic receptors are called “ligand-gated transmembrane ion channels”. The abrupt change in concentration of these ions induces secondary biochemical processes which may have either short-term or long-term consequences on the post-synaptic neuron’s behavior.

Metabotropic receptors do not have a “channel” that opens or closes. Instead, they are linked to another small chemical called a “G-protein.” As soon as a ligand binds the metabotropic receptor (step 1 in the right side of Metabotropic vs ionotropic receptors), the receptor “activates” the G-protein (it basically changes the G-protein shape). Once activated, the G-protein itself goes on and activates another molecule (step 2 in the right side of Metabotropic vs ionotropic receptors). This new molecule is called a “secondary messenger.” A secondary messenger is a chemical whose function is to go and activate other particles. So far, this process is common to all metabotropic receptors. What happens from there on is different for every metabotropic receptor. In some cases, the secondary messenger travel until it binds to and opens ion channels located somewhere else on the membrane (step 3 in the right side of Metabotropic vs ionotropic receptors). In some cases, the secondary messenger will go and activate other intermediate molecules inside the cell. The important thing to remember is that metabotropic receptors do not have ion channels, and binding of a ligand may or may not result in the opening of ion channels at different sites on the membrane. But they will always activate a G-protein that will in turn activate secondary messengers. G-protein coupled receptors are a large group of evolutionarily related proteins that are very common in the brain and body; over eight hundred human genes are devoted to producing these proteins.

Despite many safeguards, the synaptic communication processes described above often fail because the vesicles are empty or do not fuse properly, or the neurotransmitters are oxidized and therefore inactive (Linden, 2007). Neural systems have a built-in redundancy that allows them to compensate for failures in single neurons. Brains try to be energy efficient and not waste neurotransmitters. Many, but not all, neurotransmitters, or components of them after enzymatic inactivation, are re-absorbed by the axon terminal and re-used. Finally, although neurons utilize components of the diet to produce neurotransmitters, the production of neurotransmitters is rarely significantly enhanced by dietary supplements.

Ending neurotransmitter action

After most neurotransmitters are released into the synaptic cleft they need to be inactivated. The primary mechanism for inactivation is reuptake (step 7 in Neurotransmitter lifecycle). Reuptake is the reabsorption of a neurotransmitter via a protein transporter located on the (pre-synaptic) membrane of the axon terminal. Reuptake allows for the recycling of neurotransmitters. Because neurotransmitter molecules are large and hydrophilic, they cannot simply diffuse through the membrane; this is why specific transporter proteins are necessary.

Neurons benefit from feedback on how much neurotransmitter they are releasing at their axonal terminals. This feedback is provided by a special type of protein receptor called an autoreceptor (step 6 in Neurotransmitter lifecycle). Autoreceptors are located on the membranes of nerve cells. They may be located on the cell body or the axonal terminal. These receptors are only sensitive to the neurotransmitters released by the neuron on which the autoreceptor sits. Activation of an autoreceptor usually leads to a reduction in the rate of release of the cell's neurotransmitter.

Dopamine, norepinephrine, and epinephrine.

Dopamine (often abbreviated DA), norepinephrine (often abbreviated NE), and epinephrine are sequentially synthesized from the same dietary amino acid precursor, tyrosine. Norepinephrine and epinephrine are also sometimes called noradrenaline and adrenaline, and the neurons that produce them are sometimes called noradrenergic and adrenergic. Understanding the sequence of biochemical steps that lead to the production, storage, release, and receptor interactions of these three neurotransmitters has allowed neuropharmacologists to design drugs that can either enhance or inhibit specific processes to treat neurological and psychiatric diseases.

Synthesis

Tyrosine is actively transported from the blood across the blood-brain barrier (Step 1 of this figure). Tyrosine is utilized by all the cells in the brain for a variety of purposes, not just for making these neurotransmitters. Dopamine, norepinephrine, and epinephrine neurons produce a series of enzymes that convert tyrosine into each of these neurotransmitters. The process begins inside the cytoplasm of the neuron with the conversion of tyrosine into L-DOPA by the addition of a hydroxy molecule to the benzene ring via the enzyme tyrosine hydroxylase (Step 2 of this figure). This enzyme requires the presence of iron ions to function properly. This enzyme does not work well in humans with severe anemia who have difficulty absorbing enough iron from their diet. Severe anemia is typically associated with feelings of depression and mental fatigue. These symptoms offer insight into the role of these neurotransmitters in normal brain function.

L-DOPA is converted into dopamine by the removal of a molecule of carbon dioxide via the enzyme aromatic amino acid decarboxylase, also called DOPA decarboxylase (Step 3 of this figure). This enzyme is extremely efficient, which may explain why brain levels of L-DOPA tend to be very low and why providing exogenous L-DOPA to patients who lack sufficient dopamine—such as in patients with Parkinson’s disease—leads to a dramatic increase in the production of dopamine. After synthesis, the dopamine is transported into a synaptic vesicle and stored until it is released from the axonal terminal (Step 4 of this figure).

In norepinephrine and epinephrine neurons, once dopamine is transported into the vesicle, it can be converted to norepinephrine by the vesicular enzyme, dopamine-beta-hydroxylase (Step 5 of this figure). In addition to dopamine-beta-hydroxylase, the vesicles contain copper. Copper is required for dopamine-beta-hydroxylase to function appropriately.

A very small group of neurons in the brainstem also express an enzyme, called phenolethanolamine-N-methyltransferase, that converts norepinephrine into the neurotransmitter epinephrine by adding a methyl group to the nitrogen atom (Step 6 of this figure). Epinephrine should be seen more as a hormone than intraneuronal signal. In addition to being released by neurons, it is also secreted by the medulla of the adrenal glands and does not enter the brain. It produces the peripheral manifestations of strong emotions such as fear or anger when it is released into the bloodstream.

Receptors

Neurotransmitters relay their messages by traveling between cells and attaching to specific receptors on target cells. When a neurotransmitter attaches to a receptor, it triggers an action in the target cells. Receptors, even for a single neurotransmitter, exist as many different types that are differentially distributed and produce unique post-synaptic effects. Norepinephrine binds to three main G-protein-linked receptors: alpha1, alpha-2, and beta receptors that are located throughout the brain. Epinephrine can also stimulate all the adrenergic receptors including alpha and beta subtypes.

Dopamine binds to five different types of receptors, labelled D1-D5. Receptors have different, but often overlapping, functions. Dopamine receptors shows two different dopamine receptors that act via G-protein-linked receptors: one activates the enzyme adenylyl cyclase while the other inactivates this enzyme.

Once again, notice that the function of any receptor depends upon which region of the brain the receptor is located and the nature of its second messenger, G-protein-linked systems.

The functions of dopamine, norepinephrine, and epinephrine

The functions of dopamine and norepinephrine depend entirely on the function of the structures in which they are located. Consider the basal ganglia, a collection of nuclei that are responsible for producing normal movement (see Introduction). The level of the neurotransmitter dopamine in these nuclei is one hundred times higher than in the surrounding brain regions. Therefore, scientists have concluded that dopamine within the basal ganglia is involved in the control of movement. Furthermore, if we expose your brain to a drug that impairs the function of dopamine or the neurons that produce and release it, then your ability to move will be impaired. However, it would be incorrect to assume that dopamine is only involved with movement—it is not. You can also find dopamine in the retina of your eye and in your hypothalamus, structures that have nothing to do with movement. The nucleus accumbens receives an input of dopamine axonal projections that originate in the midbrain. Drugs that enhance the release of dopamine in the nucleus accumbens induce considerable motor stimulation always toward the rewarding stimulus. Thus, dopamine sometimes has mixed actions that influence movement. The multiple roles played by dopamine in the brain have contributed to difficulty in developing drugs to treat specific psychological and neurological conditions due to unintended effects on other brain circuits that use the same neurotransmitter.

Similar to dopamine, the neurotransmitter norepinephrine influences the function of multiple brain systems. It can be found in the hippocampus, a structure critical for forming new memories. Thus, norepinephrine influences the formation of memories. However, norepinephrine also plays a role in other brain regions that have nothing to do with making memories. The take-away point is that there is no such thing as a specifically unique “dopamine function” or an exclusively distinct “norepinephrine function.” The brain region that the neurotransmitter is found within defines its function, not the neurotransmitter itself. This statement is true for all neurotransmitters. In fact, neurotransmitters exhibit a complex array of actions in different brain regions, and so we can rarely make a single universal statement about their role in brain function.

We know a lot about dopamine and norepinephrine primarily because so many drugs and nutrients have been discovered that can modify their function. Norepinephrine underlies the major components of arousal and behaviors that arise in association with increased arousal; dopamine is intimately related to the control of movement, aspects of consciousness, hormone release, visual image processing and the experience of reward.

In humans, almost all norepinephrine neurons are located within a region called the locus coeruleus (Latin for “blue area” due to the presence of so many copper ions) that lies in the floor of the 4th ventricle (Norepinephrine system anatomy).

The name of this region is related to the fact that these neurons concentrate copper into the pigmented polymer neuromelanin. Neuromelanin is synthesized from the precursor L-DOPA. Although copper is required for the synthesis of norepinephrine, the concentration of the copper in the locus coeruleus far exceeds what is necessary for neurotransmitter synthesis. Unfortunately, the presence of this metal makes these neurons vulnerable to oxygen, leading to oxidative stress, which poses a particular risk for these neurons. Norepinephrine neurons in the locus coeruleus project throughout the brain, mostly unilaterally. Their diffuse widespread projection into virtually all brain regions allows them to influence your level of arousal and almost every aspect of behavior.

Dopamine neurons are more numerous (by about five times) than norepinephrine neurons. However, they do not project as widely throughout the brain. Instead, these neurons, which originate in the ventral midbrain, send axonal projections forward primarily into basal ganglia and frontal lobes (Dopamine system anatomy).

One major dopamine pathway originates within the substantia nigra, or dark substance, so called because this region concentrates iron into neuromelanin. The oxidation of iron in the neuromelanin (you know this process as rusting) contributes a significant degree to the vulnerability of dopamine neurons to oxygen. Dopamine-containing neurons may be vulnerable to the presence of oxygen due to its original role in plants as an antioxidant; the dopamine molecule may sacrifice its molecular integrity during oxidative stress. Due to the presence of various toxins, mostly insecticides, in our environment and the oxygen that we require for mitochondrial oxidative metabolism, these neurons are gradually lost with aging. In the US today, the greatest risk for developing Parkinson’s disease is growing up in a rural environment around pesticides (Tanner et al., 2011).

Another pair of dopamine pathways originates in a region of the midbrain near the substantia nigra and ascends upward into the brain. One pathway projects to the limbic system, brain regions that are associated with the control of emotion. The other dopamine pathway projects to the frontal lobes and may play a critical role in the generation of pleasure and consciousness. For more than 50 years, scientists have speculated that both too little and too much activity in these pathways underlies the symptoms associated with psychosis (see Introduction). Currently, though, dysfunction of dopamine neurons is no longer considered the direct cause of psychosis. Instead, some recent studies suggest that dysfunction of glutamate NMDA receptors also play a role (Dong et al., 2023).

Generally, in a normal healthy person, the production of dopamine and norepinephrine is not easily affected by dietary supplements but can be negatively affected by a lack of nutrients. The reason is that the first enzyme in the production of dopamine and norepinephrine, tyrosine hydroxylase, is a rate-limiting step in the synthesis pathway. For example, consuming more tyrosine will not induce this enzyme to produce more dopamine.

Serotonin

This section will describe the lifecycle and functions of serotonin (often abbreviated 5HT).

Synthesis

The neurotransmitter serotonin is built from a molecule of the amino acid tryptophan. Tryptophan is actively transported across the blood-brain barrier (step 1 in Serotonin synthesis). Tryptophan is utilized by all the cells in the brain to produce proteins for a variety of purposes that have nothing to do with neurotransmission. Serotonin neurons produce a series of enzymes that convert a small percentage (less than one tenth of one percent) of the tryptophan absorbed by the brain into serotonin. The process begins inside the cytoplasm of the serotonin neuron with the conversion of tryptophan into 5-hydroxytryptophan by the addition of a hydroxy molecule to the benzene ring via the enzyme tryptophan hydroxylase (step 2 in Serotonin synthesis). This enzyme is the rate-limiting step in the production of serotonin (Fernstrom, 2013). Unlike tyrosine hydroxylase, which is close to saturated with tyrosine under normal conditions, tryptophan hydroxylase is only half saturated with tryptophan (Young and Gauthier, 1981). Therefore, the entry of additional dietary tryptophan into the brain will lead to the production of more serotonin. Next, also occurring in the cytoplasm, 5-hydroxytryptophan is converted into serotonin by the removal of a carbon dioxide molecule via the enzyme aromatic amino acid decarboxylase, also often called 5-hydroxy-tryptophan decarboxylase (step 3 in Serotonin synthesis). Notice that the production of serotonin involves the same two chemical reactions, the addition of a hydroxy molecule and removal of a carbon dioxide molecule, that were used to produce the neurotransmitters dopamine and norepinephrine. The enzymes involved are also genetically similar to each other and likely evolved from genetic duplication.

Once produced in the cytoplasm, serotonin can either be stored in synaptic vesicles and released a fashion similar to dopamine and norepinephrine, or, despite having a positive charge, can cross cell membranes through a diffusion-like process (step 4 in Serotonin synthesis). Multiple low-affinity, high-capacity, sodium-independent transporters, widely expressed in the brain, allow the carrier-mediated diffusion of serotonin into forebrain neurons. The amount of serotonin crossing cell membranes through this mechanism under physiological conditions is considerable (Andrews et al., 2022).

Functions of serotonin

To understand the role of serotonin in the brain, one should first consider the entire body. Ninety percent of the total serotonin in the body is contained within the neurons of the gut. About 8% of the body’s serotonin is found in the blood and is localized inside platelets and mast cells; in fact, serotonin was initially discovered in the serum and determined to have tonic (or constricting) effects on the vascular system—hence its name. About two percent of the body’s serotonin is found in the pineal gland, which is located inside the skull but not considered part of the brain. The remaining one-half of one percent of the body’s serotonin is found in the brain. The nuclei that contain serotonin neurons are in roughly the same anatomical location in the human brain as in every other vertebrate or invertebrate brain which implies a conservation of purpose across time (Berger and Gray, 2009). Despite the relative scarcity of serotonin neurons in your brain, drugs that alter serotonin function can produce profound changes in how you experience the world around you. Serotonin regulates various activities, including behavior, mood, memory, and gastrointestinal homeostasis.

Neurons that produce and release serotonin in the brain are organized into a series of nuclei that lie in a chain along the midline, or seam, of the brainstem; these are called the Raphe nuclei (raphe means “seam” in Latin) (Serotonin synthesis).

These neurons project their axons to every part of the brain, and some of these axons form synaptic connections with small blood vessels. Serotonin controls the availability of blood to regions of the brain associated with its function. Another group of serotonin neurons project down-ward into the spinal cord to provide control over the autonomic nervous system and incoming pain signals.

If you were able to insert a recording device into the major raphe nuclei and “listen” to the activity of your serotonin neurons, you would discover that they have a regular, slow spontaneous level of activity that varies little while you are awake. When you fall asleep, the activity of these neurons slows. When you start to dream, these neurons temporarily cease their activity (see Introduction).

Neuropeptides

The neuropeptide neurotransmitters can be quite small, only four or five amino acids strung together, or quite large, containing hundreds of amino acids folded into complex three-dimensional structures. Neurons that produce and release these neuropeptides are found throughout the brain. Neuropeptides may also be released along with the classical neurotransmitters, such as acetylcholine, dopamine and serotonin, from the same neuron. Which neurotransmitter is released is determined by the firing rate of the neuron; faster firing rates release neuropeptides.

Synthesis

Neuropeptide neurotransmitters are synthesized via a complicated, time consuming (about two to four weeks), and energy-demanding process of reading the DNA code to produce mRNA that can be translated into a long string of amino acids (step 1 in Synthesis of neuropeptides). The process utilizes the rough endoplasmic reticulum and Golgi apparatus where the long string of amino acids, called a pre-pro-peptide at this point, is converted enzymatically into intermediate shorter versions called pro-peptides, until finally being enzymatically converted into an active neuropeptide that is then packaged in synaptic vesicles that are transported down the axon to await release (step 2 in Synthesis of neuropeptides).

After being released from the terminal axon (step 3 in Synthesis of neuropeptides), the neuropeptide binds to its specific post-synaptic receptor and is then quickly inactivated by enzymatic degradation that generates not only biologically inactive fragments but also biologically active fragments that can modulate or even counteract the response of their parent peptides. Let’s consider two of the best studied neuropeptides with considerable relevance, the endogenous morphine-like peptides called endorphins and oxytocin.

Oxytocin

Oxytocin is a neuropeptide that is synthesized in magnocellular neurons in the paraventricular and supraoptic nuclei of the hypothalamus. It is synthesized as a large inactive precursor protein and progressively hydrolyzed into smaller fragments and stored in a vesicle. Oxytocin neurons send axons into the posterior pituitary gland as well as throughout the limbic system, olfactory bulb, nucleus accumbens (the brain’s principal pleasure center), brainstem (to regulate pain), and cortex (Oxytocin networks). Oxytocin receptors are G-protein-linked and are found on the presynaptic and postsynaptic membranes. Oxytocin also acts partially on vasopressin receptors, which complicates understanding its functions. It also indirectly modulates neurogenesis (the birth of new neurons) and synaptic plasticity via these receptors.

Oxytocin has several behavioral functions. In highly social animals, oxytocin, supported by actions of the endocannabinoid system (discussed in Neurotransmitters Made from Fats), intensifies social attachment by conveying the social salience of environmental stimuli. Elevated levels of oxytocin via intranasal administration have been linked to a range of cognitive and behavioral effects including within-group conformity, trust, affiliation, and cooperation. Oxytocin also promotes aggression toward threatening out-group rivals and may underlie aspects of racism (Zhang et al.,2019). It appears to play a role in stimulating bonding between mother and child. Other studies have linked exposure to oxytocin with reduced social anxiety. Despite the many popular claims, there is no evidence that oxytocin has an aphrodisiac effect. Experiments with human couples sniffing oxytocin have not shown any increased tendency to fall in love.

Glutamate and GABA

These two amino acids are utilized by more neurons as a neurotransmitter than any other. Glutamate is the principal producer of excitation while GABA (gamma-amino butyric acid) is the principal producer of inhibition. The lifecycle of glutamate is shown in Glutamate lifecycle. The lifecycle of GABA is shown in GABA lifecycle. Because these lifecycles overlap substantially, we will discuss them in tandem.

Receptors

Glutamate signaling activates a family of receptors consisting of metabotropic glutamate receptors and ionotropic glutamate receptors (Glutamate receptors). Three of these receptors are ligand-gated ionotropic channels: NMDA receptors, AMPA receptors, or kainate receptors. These glutamate receptors were named after the agonists that activate them: NMDA (N-methyl-d-aspartate), AMPA (α-amino-3-hydroxyl-5-methyl-4-isoxazole-propionate), and kainic acid. These ionotropic glutamate receptors are nonselective cation channels; they will allow the passage of Na+ and K+, and small amounts of Ca2+. AMPA, kainate, and NMDA receptor activation always produce excitatory postsynaptic responses. All three of these ionotropic receptors are formed from the association of several protein subunits. The metabotropic glutamate receptor, in contrast, is formed from a single protein and functions as a G-protein linked receptor that provides a mechanism for glutamate to modulate cell excitability and synaptic transmission via second messenger signaling pathways.

GABA receptors are divided into GABA-A and GABA-B (GABA receptors). Recently a GABA-C receptor was defined. Less is known about its function, but it appears, based on its sequence homology and structure, to have evolved from the nicotinic acetylcholine receptor (see Neurotransmitters Made from Fats). GABA-A is a fast-acting ligand-gated ion channel/inotropic receptor. The binding of GABA opens an ion pore to allow negatively charged chloride ions to move across the cell membrane into the cell, increasing the resting negative potential inside the cell. This is also known as hyperpolarization. GABA-A receptors are located throughout the central nervous system with a high concentration in the limbic system and cortex.

The GABA-B receptor is a G-protein linked receptor that leads to the opening of potassium channels, allowing the efflux of positive ions down their concentration gradient leading to the hyperpolarization of the cell membrane. Adenylyl cyclase is also activated, which prevents calcium entry; this action inhibits presynaptic release of other neurotransmitters. GABA-B receptors are slow-acting synaptic inhibitors and are in cortex and thalamic efferent pathways.

The functions of GABA and glutamate

GABA- and glutamate-releasing neurons are densely located throughout the brain. They tend to function competitively to induce a balance of excitation and inhibition. Neurons spontaneously fire off action potentials due to their tendency to constantly leak potassium ions. The brain takes advantage of this tendency and processes information primarily via the actions of GABA-induced inhibition.

Glutamate neurotransmission is critical for probably the most important physiological process in the brain: learning. This process is mediated through a two-step process shown in AMPA and NMDA functional interaction that begins with the glutamate-induced depolarization of the post-synaptic membrane via the activation of the AMPA ionotropic receptors. At resting membrane potentials, external magnesium (Mg2+) ions enter the NMDAR channel and bind tightly, preventing the influx of any ions. Mg2+ ions are present at millimolar concentrations in the external milieu of neurons, while intracellular Mg2+ concentrations are in the micromolar range, resulting in a net inward driving force for Mg2+ ions at negative membrane potentials. A depolarization of sufficient amplitude and duration by activation of nearby AMPA receptors is required to dislodge the Mg2+ ions from the channel, thereby allowing the influx of sodium and calcium ions. The increase in intracellular calcium ion concentration activates a complex cascade of biochemical changes that ultimately involve the genes of the neuron and that may change how the neuron behaves for the rest of your life. This process underlies the brain’s ability to learn (see Introduction).

Histamine

Histamine neurons originate from the tuberomamillary nucleus of the posterior hypothalamus and send projections throughout the brain (Histamine system anatomy).

Histamine influences arousal, control of pituitary hormone secretion, suppression of eating and cognitive functions. I have included a discussion of this neurotransmitter because it is the target of so many different over the counter medications that are commonly overused.

Receptors

Histamine binds to G protein-coupled receptors, designated H1 through H4 (step 5 in Histamine life cycle). The H1 and H2 receptors produce neuronal depolarization. The H3 receptor functions as an auto-receptor; it downregulates release of histamine, norepinephrine, acetylcholine, and serotonin (step 6 in Histamine life cycle).

Endocannabinoids

This section will describe the lifecycle and functions of endocannabinoids.

Synthesis

The two currently best studied endocannabinoid neurotransmitters (there are at least seven) are anandamide and 2-AG (2-arachidonoyl-glycerol); 2-AG occurs at much higher levels in the brain than anandamide (Lu and Mackie, 2016). The production, synthesis and release of anandamide and 2-AG occur independently within the cell, suggesting that 2-AG and anandamide can be recruited differentially from the same postsynaptic neuron depending on the type of presynaptic signaling. The synthesis of both endocannabinoid neurotransmitters begins with phospholipid components of the neural membrane (Synthesis of arachidonic acid and prostaglandins). Diacylglycerol (DAG) is produced first followed by enzymatic conversion into 2-AG via the enzyme diacylglycerol lipase (DAGL). A different membrane phospholipid is converted into N-arachidonoyl phosphatidylethanolamine (NAPE) which is then enzymatically converted into anandamide by the enzyme phospholipase D (PLD).

Unlike the other neurotransmitters discussed thus far, these two endocannabinoids are not stored in synaptic vesicles. The depolarization of a postsynaptic neuron leads to Ca2+ influx through voltage-gated channels and causes the enzymatic de novo generation and the release of endocannabinoids such as anandamide. Once produced they are released to flow in reverse direction (retrograde neurotransmission) across the synapse to find their receptors on the pre-synaptic side of the synapse (Retrograde neurotransmission).

Receptors

Two different types of endocannabinoid receptors are located throughout the brain and body. Anandamide binds best to CB1 receptors, while 2-AG binds best with the CB2 receptors (Function of endocannabinoid receptors).

The type 1 receptors are densely distributed in several brain regions; the type 2 receptor is found at much lower levels in the brain. The type 1 receptor is among the most abundant G protein–coupled receptors in the brain. Anandamide and 2-AG also prevent the release of both glutamate and GABA, thus altering the balance of excitation and inhibition throughout the brain.

Prostaglandins

This section will describe the lifecycle and functions of prostaglandins.

Acetylcholine

This section will describe the lifecycle and functions of acetylcholine.

Synthesis

Acetylcholine (commonly abbreviated ACh) is made by transferring a molecule of acetic acid on to a molecule of choline that is derived from the fatty acid lecithin (steps 1 and 2 in Acetylcholine lifecycle). This conversion occurs via the action of an enzyme called choline acetyltransferase (Picciotto et al., 2012). The acetyl is derived from a molecule of sugar, usually glucose, that is initially modified inside the mitochondria. The brain needs sugar (usually in the form of glucose) to function normally. Once inside the brain, only a very small percentage of the dietary sugar is used to produce acetylcholine. The brain typically has a constant and ample supply of choline. Choline can be easily obtained from the diet, such as from eggs and baked goods. Despite this fact, many health foods stores sell choline powder to gullible customers, claiming that consuming more choline will somehow enable their brains to make more acetylcholine. Given the vital role of acetylcholine in learning and memory, this is an appealing claim. Regrettably, it has no basis in fact. For adults, the brain responds only to deficits in choline, not surpluses, in the diet. It has a ready source of choline in the diet or stored in the liver and, in fact, never develops a deficit in choline, even in patients with Alzheimer’s disease. Thus, consuming extra choline does not induce your brain to make more acetylcholine. Instead, it only results in a gaseous by-product that you exhale and that smells like rotting fish. Rather than enhancing your cognitive abilities, choline supplements merely generate a terrible case of bad breath.

Once acetylcholine is produced, it is transported into synaptic vesicles (step 3 in Acetylcholine lifecycle). The vesicles containing acetylcholine fuse with the presynaptic membrane of the axon terminal in response to an action potential (step 4 in Acetylcholine lifecycle). Each vesicle releases about 10,000 acetylcholine molecules into the synapse. Some of these acetylcholine molecules will bind to protein receptors on the surface of the nearby neuron (step 5 in Acetylcholine lifecycle). Ultimately, all of them are inactivated by an enzyme.

Receptors

Once released into the synapse, the neurotransmitter acetylcholine can act on two quite different protein receptors that have been designated, as have most receptors, according to the compounds that were originally used to study them—in this case, muscarine and nicotine (ACh channels).

Most of the acetylcholine receptors in the brain are the muscarinic subtype, whereas less than 10% are nicotinic. They differ in size, structure, and mechanism of action; yet they both respond to acetylcholine. Nicotinic acetylcholine receptors are simple ligand-gated fast-opening ion channels that are also responsive to nicotine. Most of the nicotine receptors in the body live in the neuromuscular junction (see Introduction). Nicotine receptors allow the passage of sodium, potassium or calcium ions and typically produce depolarization. Muscarinic acetylcholine receptors are G-protein-linked receptors that activate other ionic channels via a second messenger cascade. These receptors are responsive to muscarine and can produce either depolarization or hyperpolarization depending on the nature of the ionic channel linked to the G-protein complex.

The functions of acetylcholine

Within the human brain are numerous acetylcholine pathways that influence the function of the cortex, hippocampus, and many other subcortical regions (Acetylcholine system anatomy).

Within these various regions, the actions of acetylcholine enable you to learn and remember, to regulate your attention and mood, and to control how well you can move. Thus, anything that affects the function of acetylcholine neurons has the potential to affect all these brain functions. Sometimes we can learn much about the role of a particular neurotransmitter system by investigating what happens when it is injured or diseased. In the brains of people with Alzheimer’s disease, for example, acetylcholine neurons that project into the hippocampus and cortex degenerate. The loss of normal acetylcholine function in the cortex may be why patients with Alzheimer’s disease have difficulty paying attention to their environment. The loss of acetylcholine projections into the hippocampus may underlie the profoundly debilitating memory loss that is the hallmark of this disease (see Introduction).

History of Neuroscience: Otto Loewi and dreams of neuroscience

In 1902, a young German scientist named Otto Loewi at University College London began working with Henry Dale on how neurons communicate with each other. Loewi began thinking about ways that he might prove, or disprove, the chemical transmission hypothesis. According to Loewi, the idea of a viable experiment came to him in a dream a couple days after Easter Sunday in 1921. Unfortunately, his notes that night were illegible. Fortunately, the following night he had the dream again and he rushed off to the lab to test his ideas. Loewi determined that a substance was released by the vagus nerve that could communicate with other nervous tissues. Several years later this substance was isolated by Dale, who named it acetylcholine. Loewi and Dale shared the Nobel Prize for Physiology or Medicine in 1936 for their work on chemical neurotransmission.

Neurosteroids

Neurosteroids are produced from their precursor steroids by astrocytes and neurons (Maguire and Mennerick, 2024). Dietary cholesterol is converted into pregnenolone, an intermediate necessary for the synthesis of neurosteroids. Pregnenolone is then enzymatically converted into progesterone. Overall, neurosteroids are not themselves active at typical intracellular steroid receptors. They modulate brain excitability primarily by interaction with neuronal membrane receptors and ion channels. For example, progesterone is converted into allopregnanolone, which can enhance the function of GABA receptors and reduce seizure activity, anxiety and stress. Many neurosteroids are broad-spectrum anticonvulsant agents. In contrast, another neurosteroid, pregnenolone sulfate, is a negative GABA receptor modulator and tends to activate general brain activity and induce seizures. Neurosteroids are found at high levels in the cortex, hippocampus, and amygdala. Within these brain regions, neurosteroid synthetic enzymes are localized to glutamatergic principal neurons. Given their role in the control of neuronal excitability, neurosteroid analogs are now being considered for treatment of epilepsy, anxiety, depression, and stress-sensitive conditions.

Selecting a model system

A model organism or model system is an animal that is studied to gain understanding about biological phenomena, with the expectation that what is learned will apply broadly across other species. The choice of model system depends on several key criteria. In behavioral neuroscience, it is primarily based on the behavior or dysfunction we aim to understand.

Scientists sometimes select closely related species with very different behaviors so that they may understand how differences in neural systems support differences in behaviors. For example, we will see in subsequent sections that several species of voles are conducive to studying social bonding because they are closely related to each other but show highly divergent social behaviors. These studies have helped researchers understand how neural organization of voles underlie social behaviors.

The selection of a model system can also be guided by its extraordinary ability. For example, bats use echolocation to navigate their environment. They send out sonic signals, which they then use to interpret the shape of their environment based on what sound bounces back. Bats integrate this information with visual information, making them good models to understand the neural basis of multi-sensory integration (Moss and Sinha, 2003).

Finally, phylogenetic proximity to humans is a key benefit when we want to use model organisms to understand human biology. For example, we will discuss how macaques have helped us understand human neurodevelopment because they are closely related to us. There are many other examples that illustrate the importance of letting behavior guide the choice of a model system. While researchers may each have a preferred model system for studying their question, the integration of information across model systems provides a comprehensive understanding of the neurobiological mechanisms of a behavior.

In addition to having the behavior drive model selection, there are practical concerns associated with the use of model species. In many cases, scientists are concerned about how easily the model organism can breed and be housed in a laboratory. While experiments in the wild can be informative, laboratory environments provide much greater control over what animals experience and how data are collected. Mice are by far the most used laboratory species in biomedical research because they are sufficiently similar to humans, and because they are readily (and cheaply) maintained in a controlled lab setting (de Sousa et al., 2023). They are also amenable to study because scientists can obtain many of them. Other model systems include rats, voles, hamsters, guinea pigs, naked-mole rats, rabbits, opossums, and cats (Clancy et al., 2001). Non-human primates are also used in research and include macaques and marmosets. Each of these organisms has practical drawbacks and benefits in their use.

History and regulation of animal research

Humans have been using animal models for centuries. For example, researchers in ancient Greece dissected animals with the goal of better understanding human anatomy. This is because there were taboos of studying human anatomy. Prominent physicians also performed surgeries on live animals called vivisections to learn about living anatomy in ways they could not in humans (Allen Shotwell, 2013). Today, we have different ethical standards than we did in the past, and scientists must follow a strict code of conduct when it comes to experimenting on animals or humans. That is, researchers must follow strict guidelines to perform research on animals.

Any research on animals must first be approved by an Institutional animal care and use committee (IACUC) at the researcher’s home institution. The IACUC is composed of scientists, as well as members of the general public. The first step in the process is for a researcher to write a proposal that articulates the necessity of using animals in research and explains the benefit to society. The IACUC reviews this proposal and evaluates whether the proposed study is sufficiently justified to warrant the use of animals in a research study (Prentice et al., 1992). A major focus for the IACUC is to ensure that measures are used to minimize pain. In addition, the IACUC considers how many animals the researcher proposes to use to make sure they are not proposing either more or less than are needed to yield reliable results. Therefore, IACUC plays a critical role in fostering ethical and responsible animal research, which benefits researchers as well as the community at large.

The use of animals in research has a proven track record of saving human lives. There are emerging alternatives that may reduce the need for animal models. Later in this chapter, we will discuss one exciting new frontier which has much potential to reduce the need for animals. We are making reference to organoids. Organoids are derived from stem cells and are grown into clumps that share many features with actual organs. Theoretically, these organoids can be grown from stem cells of any species. Organoids can be used to study biological processes in rare and endangered species, and to test possible treatments for disease. We can take skin cells, reprogram them to stem cells, grow organoids, and test possible treatments for disease (Quadrato et al., 2016). Testing treatments in organoids is a viable option, especially in cases where testing treatments directly in humans would be unethical. Therefore, these stem cell-based approaches open up new and exciting possibilities for research. They also provide insights about systems that would otherwise be unethical or impractical to study in a model system.

Comparing neural structures

In this section, we discuss homology and how it is used to compare neural structures in different species. As mentioned above, homologous structures share an evolutionary ancestry. Organ systems can be considered homologous regardless as to their function. One well known example concerns the mammalian forelimbs, which are homologous across mammals. These structures are considered homologous across species because they have similar patterns in bone structure, which reflect their shared ancestry. Yet, forelimbs take on many functions. For example, bats have evolved wings to fly, whales have evolved fins to swim, and horses have long hoofed legs to walk and gallop. Although forelimbs support different functions in different animals, forelimbs are homologous across mammals because the anatomical structures that make up forelimbs are recognizable across mammals and arise from a common ancestor.

Comparative neuroscience typically relies on identifying corresponding brain parts across species. Major anatomical divisions and brain nuclei are considered homologous if these structures emerge from shared ancestry (i.e., homology). While this concept might appear relatively straightforward at first, identifying corresponding brain parts becomes a thorny problem when applied to distantly related species (Striedter, 2005). Scientists seeking to define homologous structures search for shared similarities in traits present in the last common ancestor. But which similarities they focus on can vary. Homologous characters may be neuroanatomical features (e.g., connectivity patterns) or genes expressed by different cell populations in adulthood, but they can also be based on developmental processes such as spatiotemporal patterns of gene expression. Importantly, focusing on developmental or adult phenotypes as a basis for homology can yield different conclusions (Medina and Reiner, 2000; Briscoe, 2019).

In the next section, we will discuss the issue of defining homologies across the brains of birds and mammals as one controversial example. We will discuss how the use of different characters to define homologous structures may lead to different conclusions. We will also discuss this challenging case because it is the investigation of challenging cases that deepens our understanding of homology.

Marked differences in anatomical organization between birds and mammals

As discussed previously, the brain is typically divided into three major regions that can be defined early in development. These are the forebrain, midbrain, and hindbrain. The forebrain is composed of the telencephalon and diencephalon. The telencephalon includes the cerebral cortex and several subcortical structures (e.g., basal ganglia, amygdala, hippocampus) (see Introduction). In some species, the cerebral cortex makes up much of the telencephalon. We first discuss the basic features of the mammalian telencephalon before discussing how this basic organization of the mammalian telencephalon differs from that of birds.

The mammalian cerebral cortex consists of grey and white matter (see Introduction). The grey matter largely contains cell bodies of neurons and glial cells. The white matter contains glial cells and axons coursing within and across cortical and subcortical structures. In the white matter, many of these axons are surrounded by a fatty sheath called myelin. It is myelin that gives the white matter its white color. In mammals, there is a clear distinction between the white and grey matter with cell bodies of neurons separated from the white matter, which consists of axons. The top half of White and gray matter in birds and mammals show this segregated organization in a human and a mouse brain, on the left and right respectively.

The grey matter also has a specific organization of interest. The mammalian cerebral cortex grey matter is typically organized into six layers (layers I-VI), shown in a mouse brain on the left side of Cortical cell organization in birds vs mammals. Each cortical layer has a characteristic distribution of neuronal and glial populations with each layer possessing stereotypical patterns of efferent (outgoing) and afferent (incoming) projections.

The basic organization of the mammalian cerebral cortex stands in sharp contrast to the structural organization of the avian telencephalon. At first glance, the telencephalon of birds appears drastically different from mammals. First, whereas the mammalian cerebral cortex consists of grey and white matter, there is no such distinction between grey versus white matter in the avian telencephalon, shown on the bottom of White and gray matter in birds and mammals. Rather, the axons of bird brains course through cell bodies with no segregation between cell bodies and axons. Second, whereas the mammalian cerebral cortex is organized into layers, neurons and glial cells in birds are primarily clustered into nuclei, shown on the right side of Cortical cell organization in birds vs mammals.

The nomenclature to define regions of the avian telencephalon are also distinct from those used in mammals. The avian telencephalon is subdivided into the striatum, and the hyper, meso, and nidopallium. The meso- and nidopallium are organized into nuclei whereas the dorsal pallium has a layered organization, but only consists of at most three layers. These major differences in anatomical structures have contributed to challenges in defining homologies across the telencephalon of birds and mammals.

The differences in structural organization between birds and mammals are accompanied by differences in connectivity patterns, which further complicates the search for homologies. The hyper, meso, and nidopallium of the bird cortex receive input from the thalamus and process sensory information, including visual, somatosensory and auditory information. For example, the hyperpallium receives extensive visual input and the nidopallium processes somatosensory information. In mammals, in contrast, these kinds of sensory inputs project primarily to layer IV neurons spread along the cortical surface. The bottom of Cortical cell organization in birds vs mammals exemplifies these two organizations.

Some researchers have used connectivity patterns as a tool to define homologies, but others have used other criteria to define homologous structures across the telencephalon of birds and mammals. Depending on what criteria a researcher uses, they can come up with completely different ideas of which structures are homologous in birds and mammals. In the next section, we will learn more about one specific area of uncertainty in this field.

Definition and challenges: comparing bird brains to those of mammals

In this section, we detail conflicting definitions of homologous structures across the mammalian telencephalic structures in birds. Two opposing hypotheses in this debate are illustrated in Bird vs mammalian telencephalon. As we will see below, some of the conflicting views here arise depending on which scale of investigation a researcher uses. Recall that scale refers to the biological level of organization. Here, we will see conflicting results between what the macro-scale (connectivity) suggests are homologous structures in birds and mammals and what the micro-scale (developmental gene expression) suggests are homologous structures.

Some researchers have focused especially on patterns of afferent and efferent projections in the avian pallium to define homologous structures between the avian and mammalian telencephalon (Karten, 1991). As we discussed above, in mammals, much of the cerebral cortex consists of primary cortical areas that receive extensive thalamic input and process sensory information. Those areas include the primary visual cortex, the primary somatosensory cortex, and the primary auditory cortex (see Introduction). The avian dorsal ventricular ridge (DVR) contains discrete nuclei (including those of the mesopallium and nidopallium) that receive extensive thalamic input (Butler, 1995; Butler et al., 2000). Researchers have used these insights to argue that the dorsal ventricular ridge of birds is homologous to the mammalian cerebral cortex. This is known as the isocortex-DVR hypothesis.

Other researchers have reminded the community that homologous structures can take on different functions across species. That is, homology between structures arises from shared ancestry, which may or may not serve the same function across species. Think back to how bat wings fly, whale fins swim, etc. In this case, we are not looking at function. Rather, we are considering the evolutionary origins of populations of cells, an area called field homology. Field homology refers to populations of cells that derive from evolutionary conserved regions but may be populated across diverse morphological structures. Indeed, there are a number of cases where cells are generated in one location and migrate over long distances in the central nervous system (Corbin et al., 2001). Many researchers focus on developmental field homologies to define homologous cell populations. Typically, looking at developmental fields means that researchers focus on transcriptional (gene expression) landscapes in the embryo as a basis for defining homologies. These fields thus represent the basic units for comparisons.

To better understand what we mean when we say we are defining a developmental field, let’s consider transcriptionally defined subdivisions as it applies to defining the homologous structure for the bird DVR in mammals. In this case, researchers have used a combination of genes to transcriptionally define fields (populations of cells that are similar to each other) and define comparable brain regions across species. For example, there are several genes, including Pax6, Tbr1, Nkx-2.1, and Dlx-2 that are expressed in specific locations during development. These genes are expressed in clusters of cells and can be used to define homologous areas because we know of their migratory patterns so we know where they will end up in adulthood. Evaluating spatiotemporal patterns of gene expression during development has been used to argue that the precursors in the DVR give rise to the amygdala in mammals and give rise to nido- and mesopallium in birds (Puelles et al., 2002). Therefore, large regions of the avian telencephalon that might be considered homologous to the cerebral cortex in mammals based on adult connectivity could be considered homologous to the mammalian amygdala when considered based on developmental fields (Striedter, 1994). In this case, consideration of homology based on developmental fields is called the claustroamygdala-DVR hypothesis, which stipulates that mammalian homologues of the DVR consist of the claustrum, endopiriform nuclei, and amygdala (Butler et al., 2012).

Defining homologous structures between species is a crucial effort, as this effort forms the foundation for any comparison between species. We have focused on developmental processes and connectivity patterns as two different ways to identify homologous structures. Researchers who focus on adult patterns of connectivity consider the lateral and ventral pallium to be homologous to the mammalian cortex because the lateral and ventral pallium of birds and the cortex of mammals both receive inputs from sensory thalamus. Researchers who focus on gene expression from a developmental perspective place an emphasis on developmental patterns. In contrast, other researchers may support the notion that the meso and nidopallium in birds are homologous to the mammalian amygdala. This challenging case illustrates the difficulty in defining which features are relevant and should be used as a basis to compare highly divergent brains.

The human brain is very big in comparison with many other species

One noticeable feature about humans is that our brains are large relative to many other mammals. However, the human brain is not the largest of mammals. For example, elephants and whales have larger brains than those of humans (Diversity of brains). Nevertheless, the observation that the human brain is approximately 3 times larger than that of great apes (i.e., chimpanzees, orangutans, gorillas) and any other nonhuman primates has led many researchers to suggest that it is the size of the human brain that confers humans with enhanced cognitive capacities. Indeed, there is a positive relationship between brain size and cognitive abilities across species. Yet, brain size accounts for only a small percentage of the variance in cognitive capacities across species. Importantly, these correlative analyses are unreliable indicators of causative relationships because correlations do not mean causation. It is therefore clear that size of the brain does not explain everything about human cognitive capacities, and that other factors must play a role in cognitive capacities.

Relative brain regions vary across species

One salient feature about the human brain is that the human cerebral cortex is very big, and that it occupies a large proportion of our brain, more so than in many other species. For example, the cerebral cortex occupies approximately 80% of the brain in humans but the cerebral cortex only occupies 20% of brain volume in mice. It might therefore seem reasonable to assume that our proportionally enlarged cerebral cortex sets humans apart from other mammals. But, it is well known that the size of brain parts, including the cerebral cortex, varies allometrically. Allometry means that proportions change with overall size (Dial et al., 2008). We will explain the concept of allometry in greater detail.

Brain region volumes vary across species. Brain regions vary allometrically depicts brains of mice, rhesus macaques and humans.

As can be seen from the to-scale images in the far right, these brains have very different total sizes. One noticeable observation from Brain regions vary allometrically is that larger brains are preferentially composed of the cerebral cortex. This observation is true when we consider these species, but it also holds when we consider other mammals. Brain allometry means that as overall brains vary in size, some brain regions come to occupy a relatively large proportion of the brain while other brain regions occupy a relatively small proportion of the brain.

A formal way to describe proportional changes in the relative size of brain parts is to say that brain regions vary allometrically (Stephan et al., 1981; Reep et al., 2007; Yopak et al., 2010; Bush and Allman, 2004). Saying that the cerebral cortex has a positive allometry means that the cerebral cortex becomes relatively enlarged in bigger brains. Species such as humans, whales, and elephants possess a big brain and a relatively enlarged cortex. In contrast, small-brained mammals such as mice, rats, hamsters, and mole rats have relatively small cerebral cortices by virtue of their small brain.

The human brain is by far the biggest of all primates and the human cortex occupies the largest proportion of the brain among primates. There are many debates as to whether the size of human brain regions, like the cortex, is unusually small, large or within the expected range considering the size of the human brain. It is possible our large cerebral cortex is exactly the size one would expect of a primate based on the total human brain size. The controversy as to whether human brain parts are small or large based on our brain size arises because we have no nonhuman primate brain of the same size for comparison, which fuels debate as to whether human cerebral cortices are unusual in size or the consequence of allometric changes. Given these debates, it is still a mystery as to whether the human cortex-to-brain ratio is large relative to other species.

Some brain regions vary across species as allometry would predict. That is, proportional differences in brain region volume are explained by changes in total brain volume. In other cases, we observe grade shifts. Grade shifts are relative changes in size of a brain region after controlling for brain size (Barton and Harvey, 2000; Charvet et al., 2012). In other words, brain regions may be larger or smaller than they should be based on what we would expect for a given brain size. What a grade shift means in terms of brain function, though, is still difficult to predict based on brain region size. We will discuss this topic below.

Grade shifts are evident in different groups of mammals (including primates), and they are observed in different regions. Some primate species, including great apes, and monkeys possess a disproportionately enlarged cerebral cortex relative to other mammals (Barton and Harvey, 2000, Finlay et al., 1998, Stephan et al., 1981). In the case of the cerebral cortex, the cerebral cortex is much bigger than that of other mammals of similar brain size. What is interesting is that we see grade shifts multiple times across different groups of mammals. For example, the cerebral cortex is bigger in monkey species relative to other rodents of similar brain size. For example, monkeys have relatively enlarged cortices even though these groups live in diverse environments each with species-specific life histories. The presence of these grade shifts shows that there is no clear one-to-one connection between the size of brain parts and behavior or environmental demands. Rather, changes in relative size of brain regions apply to whole groups of species.

Sensory and motor specializations across mammals

A standard approach to testing whether a brain region is unusually large or small is to ask whether a brain region is expanded or reduced after controlling for its overall brain size. Our example of the cerebral cortex in primates above suggests that just seeing an enlarged brain region does not necessarily tell us its function. That is not to say that brain region sizes have no relationship to behavior. On the contrary, there are several prominent examples where brain region size tells us a lot about the behaviors of a species. The link between size and function is especially evident when considering animals with a loss of sensory capability or if they show a specialization. In those cases, there is generally an association between cortical area size and sensory ability after accounting for allometric variation in brain region size. For instance, the platypus detects vibrations in the water, and possesses an expanded primary somatosensory cortex, which is responsible for processing somatosensory information from the bill (Krubitzer, 1995; Krubitzer et al., 1995; Krubitzer and Prescott, 2018). The top of Evolution of cortical areas shows some representative brains from several species, which show how relative proportions of cortical areas vary across species.

This variation is shaped by allometry as well as species-specific behavioral adaptations. If we compare the relative size of the rodent somatosensory area with that of primates, we see that rodents have expanded somatosensory cortex dedicated to input from their whiskers. This situation is similar to that observed in the platypus, which show an expanded somatosensory cortex linked to their bill (as described above). Primates, in contrast, rely far less on somatosensory input and have a much smaller relative somatosensory cortex. The expansion of somatosensory areas in rodents typically comes at a cost to visual and other sensory areas. Rodent species with reduced or no visual capabilities provide a particularly notable example of this kind of specialization. This is the case in blind mole rats, which are nearly blind (Evolution of cortical areas). These blind mole rats have reduced cortical territory devoted to processing visual information and a massively expanded somatosensory cortex to process sensory information from the body.

A relevant question at this point is where these variations in cortical size come from? Are they “hard-wired” into each species or do they arise from experience and therefore can change? The answer is a bit of both. Much of the size differences in brain regions across species is encoded in the genome. That is, a human cannot increase their exposure to touch sensation and develop a somatosensory cortex to rival a platypus. Yet, experience can shape the size of brain regions to some extent. Indeed, a range of studies focused on the neural basis of experience-dependent plasticity and cross-species comparisons have systematically shown a link between experience and size in cortical territories. For example, musicians have an expanded auditory cortex, and this is thought to be because of their sustained attention to listening and playing music (described below; Strait et al., 2014; Moreno and Bidelman, 2014).

Studies from experimental models further demonstrate that cortical areas change with experience, and others show the importance of attention in mediating changes in cortical territory (Dooley et al., 2017; Ramamurthy and Krubitzer, 2018; Dooley and Krubitzer, 2019). Animals may be exposed to stimuli or be deprived from stimuli, and their cortices reorganize in response to this input. See Introduction for an interesting example of how the auditory cortex reorganizes based on the frequencies of sound during development. Experimental results combined with cross-species analysis of sensory and motor specialization demonstrate that cortical plasticity is determined by the environment.

Techniques to study the evolution of connections in model systems

The toolbox available in model systems typically relies on studying specific populations of neurons to identify afferent (incoming) and efferent (outgoing) connections. Methods available to trace pathways in model systems include tract-tracers and diffusible dyes (see Inorganic Dyes). We first discuss neural tract tracing techniques because these methods have been used for decades and are a major source of what we know about projection patterns.

Tract-tracers are injected in a region of interest and actively diffuse across neurons. Traditionally, tract-tracers are injected in the living brain, and post-mortem brains are cut into micron-thick sections to visualize labeled axons from histological material (Oh et al., 2014). There are various kinds of tract-tracers. Those include anterograde, retrograde, monosynaptic and trans-synaptic tracers (Retrograde and anterograde tracers). Anterograde tracers are taken up by neuronal cell bodies at the injection site and diffuse along the axon towards terminal processes. In contrast, retrograde tracers are taken up by a neuron’s terminals and diffuse back towards (i.e., retrogradely) the soma. Mono-synaptic tract-tracers remain within the neuron they penetrated and do not diffuse to neighboring neurons. Other tract-tracers spread trans-synaptically (across synapses) to diffuse throughout the network. Therefore, there are many variant forms of tract-tracers that are available for study, though they only permit visualizing a few projections in the brain at a time.

Tract-tracers have been used extensively to trace pathways across model systems and in different species. The bottom of Retrograde and anterograde tracers shows three views of an example 3D reconstruction of tract-tracing performed using a trans-synaptic anterograde tracer injected in the left retina of a mouse. The tracer is seen extending from the eye, through the lateral geniculate nuclei of the thalamus, and on to the primary visual cortices (see Introduction). These methods enable the identification of anatomical pathways between brain regions.

Tract-tracers are invasive and permit only studying a few pathways at a given time. Combining the work of many researchers over many years has enhanced our understanding of how connections have evolved, and these methods have heavily influenced our understanding of vertebrate brain evolution. For example, researchers used this method to show connections from the thalamus to the DVR in birds, and these observations were used to support the notion that the DVR is homologous to the mammalian cortex (as discussed above).

As another example, this compendium of information on tract-tracers has yielded insights about pathway properties, and principles that explain how axons target other cells (Modha et al., 2010; Liu et al., 2020). One principle to emerge from these comprehensive analyses is that local projections are more common than long-range projections, and this is true whether we consider small or large brains. Brain networks follow a small-world organization, in that they are composed of highly connected cells coupled with sparse long-range projections (Liao et al., 2017). The example of connections from retina to the thalamus to the occipital cortex in Retrograde and anterograde tracers makes striking 3D images, but these kinds of long-range connections are the minority of connections in the brain. Mostly, connections happen between neurons that are close neighbors.

As we will learn more in the next section, the direct study of human connections derives from non-invasive neuroimaging techniques, which are open to interpretation. We therefore use the principle of conservation to make inferences about connectivity patterns in the human brain.

Techniques to study the evolution of connections in humans

Neuroimaging techniques are widely used in neuroscience to visualize neural activity and connections across the lifespan in humans. The non-invasive nature of neuroimaging has revolutionized the field of neuroscience. Here, we discuss a subset of neuroimaging techniques, which are used to make inferences about connectivity patterns from brain activity. Imaging the live human brain provides an overview of several broad classes of neuroimaging techniques.

We will cover structural and functional magnetic resonance imaging (MRI) (see Methods: fMRI). Structural MRI evaluates the tissue structure from the soft tissue structure of the brain. Functional magnetic resonance imaging relies on blood fluctuations (measured as BOLD or blood-oxygen-level-dependent signal) to infer neuronal activity levels. Two notable techniques used to study connectivity in the live human brain that rely on MRI imaging are resting states fMRI and diffusion MRI (right side of Imaging the live human brain). We will discuss how these methods work, how they have contributed to mapping pathways in humans, and what uncertainties they have. We will highlight the need for information from model systems to trace pathways in the human brain.

fMRI and Resting states: Many fMRI studies focus on BOLD signals that occur while the subject is performing certain tasks. Resting-state fMRI, in contrast, measures spontaneous low-frequency fluctuations in the BOLD signal at rest so that it can be used to investigate the functional architecture of the brain. Comparing the timing of BOLD activations at rest across regions can give clues about connections (Lee et al., 2013; Murphy et al., 2013). In this method, observations of covarying fluctuations in brain activity are used as an indirect means to infer connections between regions. Covariance of activity patterns (increases or decreases at the same time in two brain areas) suggests that the areas are in some way connected. A benefit of resting state fMRI is that it broadly surveys patterns of blood flow across the brain and can be studied in many individuals. This broad survey is in contrast with tract-tracers, which provide detailed but sparse information about projection patterns in model systems. One major limitation of resting states fMRI is that coordinated patterns of connectivity do not address the underlying pathways. A finding that regions co-activate doesn’t necessarily mean that there are actual physical pathways that underlie the covariance in activity. That is, coordinated activity may arise without any direct connections (Murphy et al., 2013).

Diffusion MRI relies on the diffusion of water molecules to make inferences about the composition of neural structures. More specifically, diffusion MRI detects movement of water and relies on the principles that water diffuses preferentially along the axons. By measuring the direction of water diffusion, which correlates with the direction of axon fibers, we can reconstruct the location and orientation of fibers coursing through the adult and developing brain (Wedeen et al., 2005, 2008). Diffusion MRI tractography can be used to document similarities and differences in the projection patterns of different species.

There are certain limitations imposed by diffusion MRI tractography that have precluded its use at the level of detail needed to map some connections in the developing and adult brain. First, scans capture information at millimeters resolution (i.e., macroscale) and therefore can be used to average biological information that spans micro- and cellular scales. This limitation in resolution means large amounts of critical biological information is lost. Another major issue is that the pathway reconstructions from diffusion MRI are not always accurate. In many cases, there is much uncertainty in what the correct pathway reconstructions are in the human brain. There is a lack of alternative methods to evaluate the accuracy of neuroimaging techniques in humans. Therefore, scientists use the principle of conservation from model systems to make inferences about human pathway reconstructions. There are a number of cases where reconstructed fibers from diffusion MRI tractography in the human brain appear very improbable (based on what we know about model systems). Overcoming these issues is an active area of research.

Despite these limitations, diffusion MRI tractography has yielded some interesting findings. For example, tractography studies have shown that some pathways connecting cortical areas are enlarged in humans relative to other species. For example, the arcuate fasciculus, which is a pathway that connects the lateral frontal and temporal lobe, is very large in humans compared to nonhuman primates. The arcuate fasciculus is of particular interest because it is a pathway thought to be associated with language comprehension (mediated by the parietal and temporal lobes) and production (mediated by lateral frontal cortical areas). Tractography studies have revealed a disproportionately expanded arcuate fasciculus in humans compared with both chimpanzees and macaques (Rilling et al., 2008). These observations have been used to argue that the expansion of the arcuate fasciculus may be linked to language in humans. While it is clear that the arcuate fasciculus is big in big-brained primates such as humans, it is still possible that the arcuate fasciculus has an allometry. That is, the arcuate fasciculus may have become disproportionately bigger in big brained species. The tractography has yet to be validated and we have very few tools available to ensure the accuracy of these projection reconstructions in humans versus great apes.

Monkeys as model systems to understand human fetal development

Much of what we know about development in humans, and especially human fetal brain development, is inferred from model systems. Human fetal development is an especially challenging phase to study because there are limited methods for study. Accordingly, there is only so much we can study in prenatal humans (Byrston et al., 2008). Macaque monkeys have been one important model system to understand human development. This monkey is a popular choice for this kind of research because of its close phylogenetic proximity to humans, and because monkeys share many similar features with humans (Passingham, 2009; Rakic, 2009). More recently, researchers are turning their attention to a small monkey called the marmoset because it is an easier animal to handle and breed than macaques. Macaques have been used for some time and they have yielded insights into human neurodevelopment (Mitchell and Leopold, 2015; Okano et al., 2012).

This section articulates the basic outline of fetal and early postnatal development in macaques and how it can be used to make inferences about human development. We discuss how the study of model systems can enhance our understanding of neurogenesis (the process of neuron production), and synaptogenesis (the process of generating synapses) in the human brain (Clancy et al., 2001). More detail on these and surrounding neurodevelopmental processes is in Introduction.

Neurogenesis is the term used to describe this entire process of generating neurons. Early in development, the brain is composed of proliferating cells (i.e., cells that are in the process of multiplying), which have yet to adopt a final, functional form (such as being a neuron or glial cell). These early, undifferentiated cells are called neural stem cells, and they have a unique shape, with a long fiber or process extending from their cell body to the edge of the developing brain tissue (Neurogenesis in mammalian cortex). Neural stem cells divide to create daughter cells that turn on to mature, or differentiate, into neurons or glial cells (step 1 in Neurogenesis in mammalian cortex). The maturing cells wrap around elongated fibers and migrate radially towards the outer surface of the cortex (step 2 in Neurogenesis in mammalian cortex).

Neurogenesis largely occurs during prenatal development (Byrston et al., 2008), and extends to postnatal stages of development. In humans, we have limited tools available to visualize this process. We can make inferences about the process of neurogenesis from anatomy and histology (thin slices of post-mortem brains), from gene expression of individual cells or from tissue, but these methods are only indirect means to probe neurogenesis, and we have to map findings from model systems onto humans (Clancy et al., 2001; Charvet and Finlay, 2018; Boldrini et al., 2018; Kozareva et al., 2019).

The use of model systems, such as macaques, has enabled us to better understand the process of neurogenesis in humans (Rakic, 2002, 2009). For example, researchers have used cell birth-dating techniques to see where proliferative cells migrate to their final destination. Essentially, birth-dating involves the administration of a compound that gets taken into proliferating cells. The compound marks the cell, which enables researchers to visualize where the cell ends up at a later stage in development. Once the compound gets incorporated into the cell, we can track it over time. We can, for example, use this method to track migratory patterns of cells. These techniques obviously cannot be done in humans for ethical reasons, but they can be performed in macaques and other species.

Birth-dating methods in macaques have revealed basic patterns of migration from the proliferative pool to the cortical plate (upper edge of the tissue, where the neural stem cell fiber ends). For example, proliferative cells exit the cell cycle and migrate according to an inside-out fashion such that recently generated neurons migrate past older ones in the cortical plate (Rakic, 1974, 2002). Cell birth dating and inside-out formation of the mammalian cortex shows how these birth-dating experiments trace the birth of different cell types. Here, we use mice as an example. Mother mice are injected with the tracer at different days of pregnancy and the compound is taken into only those cells that are in the process of proliferating at that particular time (step 1 in Cell birth dating and inside-out formation of the mammalian cortex). The brains of their offspring are then examined after birth (step 2 in Cell birth dating and inside-out formation of the mammalian cortex). Offspring exposed to the compound early in development show labeled cells in the inner layers of the cortex (step 3 in Cell birth dating and inside-out formation of the mammalian cortex). Offspring exposed to the compound later in development show cells in the outer layers of the cortex. Based on these observations, we can say the cortex forms inside-out, or the first-born cells form the inner layers with later born cells migrating past them to form the outer layers.

Similar experiments to the one diagrammed in mice have been performed in macaques and other primates. These experiments revealed wide conservation in the inside-out pattern of neurogenesis across mammals. Early in development, recently generated neurons preferentially populate lower/inner layers of the cortex. Later born neurons preferentially populate the outer layers. Conservation of this pattern goes beyond primates and rodents. The inside-out sequence of cell birth specification has been observed in every mammalian isocortex studied (primates, rodents, marsupial mammals; Clancy et al., 2001). Because we can’t inject pregnant humans or their fetuses with tracers, we use observations such as these from model systems to make inferences about human brain development.

Synaptogenesis(i.e., the production of synapses) is yet another process that researchers have relied on model systems, particularly macaques and cats, to map the steps and the expected time course in humans. In macaques and cats, synaptogenesis occurs as neurogenesis wanes (Cragg, 1972; 1975; Bourgeois et al., 1994; Clancy et al., 2001). The number of synapses peaks in early infancy and subsequently declines in early adulthood. Many researchers believe that this surge in synapse numbers early in infancy provides an important critical period where young individuals can learn and adapt to their postnatal environment. Human synaptogenesis is to a large extent inferred from model systems.

In this section, we have considered how we make inferences about basic developmental processes from model systems, but not every developmental process is similar across humans and other species. For example, the overall duration of brain development is extended in humans compared with macaques (Clancy et al., 2001; Charvet and Finlay, 2018). The longer overall duration in neurogenesis, synapse development and myelination may be contributing factors to the evolution of the human brain structure and function. We can determine what is unique from a structural perspective in humans compared to other species, but we can’t explain the reasons why those differences exist. For example, humans and other non-human primates may differ in their length of gestation, but we can't necessarily conclude that this is the reason why humans have some structural difference relative to non-human primates. We next discuss how we can use closely related species to learn basic concepts in neurobiology.

Studying closely related species to make inferences about the neural basis of social bonds

Thus far, we have focused on looking for similarities between species to allow us to investigate the neural bases of behavior and extrapolate to difficult-to-study species like humans. Investigating differences between closely related species is also a useful approach to understand the neural basis of behaviors. In species that are close in phylogenetic proximity, one would expect relatively few differences in brain and behavior. These relatively few differences across closely related species permit pinpointing what neural structures form the basis of select behaviors. We discuss an example where phylogenetic proximity has been used to make inferences about the neurobiological basis of social bonds.

Throughout their lives, humans form selective attachments to family, friends, and romantic partners, which are critical for both mental and physical well-being. The pair bonding of Microtine rodents such as prairie versus meadow or montane voles show variation that mirrors the variation in pair bonding in humans (see Introduction). Prairie voles are monogamous; they form stable pairs. Other vole species, such as montane voles, are polygamous with males and females forming multiple pairs. That is, individuals will copulate with multiple partners. In these studies, montane and meadow voles are used as the non-monogamous counterpart to the monogamous prairie vole. The affiliative behavior can be measured with a partner preference assay. In short, a male vole is given a choice of where to spend his time: with a female with whom he previously mated or with a new female. Pair-bonded prairie voles will spend the majority of their time with their partner when given the opportunity to do so, whereas meadow voles show no preference toward either an established partner or a novel opposite-sex partner. The diverse pair bonding behaviors across vole species mirrors the diverse behaviors we see among humans with some humans engaging in monogamous relationships whereas others form polygamous relationships.

These two vole species are closely related so that their brain structure and function is highly similar with only a few differences. Studies of the basis of social bonding using these species therefore started with the idea that there should be relatively few differences across species since they are closely related. Brain region differences, then, could be related to variation in pair-bonding between the two species (Winslow et al., 1993; Insel et al., 1994). Researchers compared the brains of these meadow and prairie voles, and they found interesting species differences in vasopressin receptors (i.e., vasopressin 1a receptors, V1aRs) in the basal ganglia. Specifically, expression of V1aR in a region called the ventral pallidum was higher in the pair-bonding voles than in the solitary voles. This difference is shown in the autoradiograms in the top two rows of Pair-bonding in voles.

While suggestive, these observed differences in V1aR expression do not by themselves demonstrate that the expression of vasopressin receptors is causally related to species differences in pair bonding. Researchers therefore went one step further with these cross-species investigations. Specifically, researchers experimentally increased the expression of vasopressin receptors in the normally non-pair bonding meadow voles to test whether the expression of vasopressin receptors is causally related to pair bonding. They then quantified partner preference as an index of pair bonding (Young et al., 1999). The bottom graph in Pair-bonding in voles shows the results of this experiment. First, untreated prairie voles showed a preference for their partner and untreated meadow voles did not, as we would expect. Increasing vasopressin receptors in meadow voles changed this, though. Meadow voles with V1aR overexpressed behaved much more like the pair-bonding prairie voles, spending more time with a previous partner than with a stranger (Winslow et al., 1993; Insel et al., 1994).

These findings point to vasopressin receptor expression in the ventral pallidum area of the basal ganglia as a major player in specifying social bonds. Other molecules (e.g., oxytocin) are also important in dictating social bonds (Anacker and Beery AK, 2013; Romero et al., 2014; Ahern et al., 2021). While V1aR expression is clearly important for partner bonding in prairie voles, can we apply this knowledge to human behavior? Interestingly, the results of these vole studies motivated studies in humans, which have identified individual differences in vasopressin receptors, specifically finding genetic variants of this receptor that predict different partner bonding behaviors (Walum et al., 2018). That is, similar genetic variants account for behavioral variation amongst humans. Examples such as this one demonstrate that we can understand the neurobiology of social bonds in humans by focusing our efforts on comparing the neurobiological basis of social bonds.

How to grow stem cells into brain organoids

Organoids are generated from embryonic stem cells or induced pluripotent stem cells, which are a kind of stem cell made from a mature cell (like a skin cell) in the laboratory. As diagrammed in Organoids, the stem cells are grown in a dish in the lab (step 1) and allowed to form small embryoid bodies (or collections of cells that stick to each other in a small ball) (step 2). Those balls of cells are then moved into a droplet of Matrigel, a substance filled with proteins that will support the further growth of the sphere and encourage the differentiation of cells into cell-specific fates such as the neuroectoderm lineage, which are developmental precursors of our nervous system (step 3 in Organoids) (see Introduction). Those bits of neuroectoderm are then spun at high speed (step 4 in Organoids). The spinning forces make them stick to each other as cells grow. Remarkably, these cells self-assemble into a collection of progenitor cells (cells that can divide but are not as uncommitted as stem cells) that can give rise to neurons and glial cells that in many ways resemble pieces of a real brain.

Brain organoids express patterns of gene expression reminiscent of the developing nervous system, and form functional networks, thus resembling miniature versions of brains. They also show cellular organization similar to developing brains. For example, the bottom of Organoids shows how the inside of a brain organoid has a ring of progenitor cells that give rise to neurons that migrate outwards. This arrangement strongly resembles the migration of developing neurons in the cortex that we discussed in How Can Diverse Species Help Us Make Inferences about Human Neurobiology?.

Because of their similarity to real brain networks, brain organoids are providing exciting tools that can be geared for personalized medicine. For example, characterizing and testing therapies on organoids provides a new opportunity to define and evaluate the effectiveness of treatments to help individuals suffering from disorders (Schörnig et al., 2021). Mature cells from a given individual can be induced into pluripotency (i.e., a proliferative state), and grown into a brain organoid. In theory, this kind of model could allow screening of therapies for individual patients using tiny brains generated from their own cells.

Below, we will provide a few examples to show how brain organoids have provided new insights into the evolution of the brain. Many of these studies rely on gene expression to investigate species differences in organoids. We will therefore briefly describe how RNA is measured from sequencing methods in the brain before discussing how they are used to identify species differences from brain organoids.

Measuring RNA from sequencing technologies

The emergence of sequencing techniques has created many new and exciting opportunities to study transcriptional variation across species. Though there are multiple techniques to study gene expression, here we will focus on RNA sequencing, which is a method that probes gene expression very broadly in cells. RNA sequencing is a technique to measure the abundance of RNA transcripts. While many techniques for measuring RNA require that you pick which RNAs to measure ahead of time, RNA sequencing measures RNAs from many thousands of genes.

The basic process of RNA sequencing is shown in Sequencing technologies.

Tissue is extracted and then RNA is isolated via a series of steps that capture and purify it (step 1 and 2 in Sequencing technologies). Unlike DNA, which is double-stranded and stable, RNA is single stranded so that it easily degrades. Researchers turn their RNA into double stranded DNA to keep it stable (step 3 in Sequencing technologies). The process of turning RNA back into DNA is called reverse transcription and we call the resulting DNA “cDNA” because it is copied (or complimentary) from the RNA it was made from. These cDNA fragments are sequenced (step 4 and 5 in Sequencing technologies). This means that each nucleotide (adenine, uracil, guanine, cytosine) is read by a sequencing machine. Relatively short strings of nucleotides read at a time. The transcript reads are then aligned to the genome, which involves matching up sequences of nucleotides from the transcripts to the genome template (step 6 in Sequencing technologies). This effectively aligns the transcripts to the corresponding gene. In a typical experiment, the process of generating read alignment will cause some genes to have many aligned transcripts. Whereas other genes will have very few or no aligned transcripts. The number of reads aligned to each gene is used as an index of gene expression (Finotello and Di Camillo, 2015; Van den Berge et al., 2019). Using these techniques, we can compare gene expression between a diseased state and a healthy control, but we can also use these data to compare transcription across development and across the brains of different species. We can relate variation in gene expression with behavior in healthy and in diseased states.

Originally, researchers sequenced RNA from whole organs, or a piece of tissue, but this approach has limitations. Using whole pieces of tissue is a limitation because tissue is composed of a heterogeneous population of cells (e.g., neuron types, glial cells), and there is no way to tell which cells or cell populations express genes of interest. More recently, researchers developed novel methods to capture RNA sequencing from individual cells. We can now identify the transcriptional profiles from individual cells within a tissue rather than the whole piece of tissue as is the case with RNA sequencing from bulk samples. Various methods exist to measure gene expression from individual cells. Single cell RNA sequencing of primate brain shows one common way to approach single cell RNA sequencing.

In this workflow, tissue is dissected (step 1 in Single cell RNA sequencing of primate brain) and broken apart into a suspension where cells float separate from one another (step 2 in Single cell RNA sequencing of primate brain). Sometimes, specific subpopulations of cells are selected using fluorescence activated cell sorting (FACS) (step 3 in Single cell RNA sequencing of primate brain). FACS uses a machine to sort cells into separate tubes based on their fluorescence. We can make specific cells fluoresce using antibody labeling (see Methods: Immunohistochemistry) or by using transgenic model organisms that express fluorescent reporters in specific cell classes (see Methods: Transgenic Organisms). Once the cells of interest are collected, RNA sequencing is performed from individual cells, keeping transcripts from each cell separate (step 4 in Single cell RNA sequencing of primate brain). This approach permits capturing transcriptional information from the individual cells that comprise a particular tissue. Researchers can now consider all of the transcriptional profiles from different cell populations (step 5 in Single cell RNA sequencing of primate brain).

The data from single cell RNA sequencing can be visualized in several ways but one common way is in the kind of plots shown in the bottom of Single cell RNA sequencing of primate brain (also step 6 in Single cell RNA sequencing of primate brain). On these plots, individual cells are represented as dots. Cells (dots) that are close to others are considered similar to each other. If cells are similar enough, they are considered a cluster. Clusters are shown as the color coding of the dots in Single cell RNA sequencing of primate brain. These clusters are thought to represent different cell populations. For instance, glial cell subpopulations cluster together as do different neuronal populations. These methods have been used to investigate species differences in cell composition of different brain regions. Much of this work has been unfolding in the very recent years and we are in the early stages of using these techniques to learn about species differences in cellular composition. There are several interesting findings to emerge from this work already. For example, researchers have identified conservation but also modifications in cortical cell types across species (Bakken et al., 2021; Campagnola et al., 2022; Kim et al., 2023), and they have even identified novel cell types in the primate lineage (Krienen et al., 2020). Much work is ongoing to uncover species variations from single cell RNA sequencing. We now return to brain organoids and how these sequencing technologies have been used to assess species differences in brain organoids.

Species differences revealed from brain organoids

Comparative analyses of brain organoids at different stages of maturity have revealed interesting insights about the evolution of brains. In these studies, researchers selected mature cells from humans and great apes, induced them into pluripotency, and grew them in the laboratory as brain organoids. They then sampled these brain organoids at different ages in different species and measured variables like gene expression and anatomy. For example, one study compared transcriptional profiles using RNA sequencing across the brain organoids of humans and chimpanzees (Pollen et al., 2019). They found that the maturation of brain organoids was similar in the early phases of organoid development, but later diverged between human and non-human primates. Another study relied on single cell RNA sequencing data in humans, chimpanzees, and macaques and found that the maturation of brain organoids in humans was slower than in other non-human primates (Kanton et al., 2019). Studies such as these have converged on the finding that the maturation of brain organoids is protracted in humans relative to other primates (Schörnig et al.. 2021). Yet, an unresolved question is how observations made from brain organoids translate to the full organism, and whether media used to grow these primate brain organoids are appropriate for other species.

Why is brain development important?

CNS abnormalities that cause functional impairment or suffering occur with a frequency of approximately 0.1-0.2% of all live births (Cater et al., 2020). Congenital (those that occur at birth) defects can occur due to environmental toxins (teratogens), as well as other prenatal experiences, such as poor maternal diet, maternal stressors, or genetic mutations. Abnormal brain development that impairs quality of life can also occur postnatally due to the presence or absence of environmental stimuli. For example, a lack of or poor environmental influences postnatally can severely impair cognition years later (Ornoy, 2006). These types of detrimental changes in brain function result because both the prenatal and the postnatal brain have a tremendous ability to sense and respond to the environment by reorganizing its structure, function, or connections. Simply put, the brain is very sensitive to environmental stimulation both pre- and postnatally. The fact that early life experience and brain development can lead to functional deficits well into adulthood underscores a need for research into how the brain develops, how genetics modulates pre-and postnatal development, and how the environment positively or negatively affects overall brain development and function.

How does development begin?

Development of the nervous system requires specific patterning events in the early embryo. These events are summarized in Overview of early neurodevelopment, which we will elaborate on throughout the sections of this chapter. These events organize the fertilized egg and partition specific units for brain development. Thus, to understand brain development in detail, we must first begin by understanding the mechanisms in place that set up the original partitions.

Gastrulation

Development of the nervous system begins with the process of gastrulation, one of the earliest developmental events in the embryo. The goal of gastrulation is to start the subdivision and specialization of cells in the developing embryo. When complete, gastrulation results in the production of 3 primary germ layers. The creation of germ layers can be thought of as the earliest step towards successfully partitioning an organism. Each of the 3 germ layers are made of cells that will give rise to specific tissue types depending on their location.

Gastrulation begins several days after the fertilization of an egg in species like mice and humans. In those intervening days after fertilization (but before gastrulation), the egg undergoes several rounds of cleavage (cell division) events to produce a sphere of cells called the blastocyst. The sphere of cells surrounds a fluid-filled cavity called the blastocoel (step 1 in Gastrulation). Reorganization of the blastula will produce two early layers of cells, the inner cell mass and the outer trophoblast layer (step 2 in Gastrulation). The trophoblast layer will become the mammalian placenta. Through cellular movements, migration, and invagination, the cells of the inner cell mass will produce the 3 germ layers (step 2 in Gastrulation).

Gastrulation begins after the formation of the primitive streak, which is a long groove that forms across the developing embryo. At the anterior region of the primitive streak is the node (De Robertis, 2006). This node represents the organizing center of gastrulation and is the site where invagination of cells begins. Gastrulation step 3 demonstrates an example of early invagination. After the cellular movements are complete, the gastrula contains 3 germ layers: the endoderm, mesoderm, and ectoderm (Gastrulation). They are named based on where they are located within the developing embryo: the endoderm is the innermost layer, the mesoderm is in the middle layer, and the ectoderm is the outermost layer. The CNS is produced from the ectoderm, although all layers communicate with one another to coordinate development.

Neurulation

Completion of gastrulation lays the foundation for a process called neurulation, which will generate the early brain and spinal cord tissues via the formation of a structure called the neural tube (summarized in Neurulation). Of the 3 germ layers, the ectoderm, or outer layer, will generate cells that become neural tissue or epidermis: two vastly different tissues. To make neural tissue (instead of epidermis), a portion of the ectoderm called the neural plate must undergo neural induction.

The neural plate is a single layer of cells located in a small, central region of the ectoderm (Neurulation, step 1). It is induced to become neural tissue by the expression of molecules we call neural inducers. These are secreted factors that direct the surrounding tissues through neurulation (Harland and Gerhart, 1997). One example of a neural inducer is noggin (Valenzuela et al., 1995). Localized expression of noggin induces the neural plate to invaginate (step 2 of Neurulation). Invagination can be described as the process of folding cells inward, to produce a small cavity. In step 2 of Neurulation, the cells of the neural plate that are pushing down into the tissue beneath them are migrating and growing towards cells that are secreting noggin (and other factors). This migration of neural plate cells into the underlying tissue results in the formation of the neural groove (Neurulation, step 2), a center region that is surrounded by indentations that form the neural folds. Ultimately by cellular movements, the two neural folds connect producing a neural tube that sits ventral to the outer epidermis (Neurulation steps 3 and 4). The inner region of the newly closed neural tube develops into the central canal of the spinal cord and what will eventually form the ventricles of the brain.

Environment and Bidirectional communication: Neural tube defects and folic acid

Neurulation is a critical step in early brain development and if this process does not occur properly a class of disorders called neural tube defects arise. The most common neural tube defects are the result of abnormal neural tube closure, which can be affected by environmental and genetic influences. These defects fall into different classes based on type and severity. In the severest forms such as anencephaly, the neural tube does not close at the anterior end and causes missing regions of the skull and brain. Other examples include encephalocele, when the brain protrudes through the skull, and hydrocephalus, which is characterized by an accumulation of cerebrospinal fluid in the brain. Spina bifida is one of the most well-known examples of a neural tube defect and occurs when the posterior end of the neural tube does not close fully. Spina bifida can range in severity, causing either a complete opening along the vertebrae of the back or a minor separation in the bones of the vertebrae (Defects of neurulation).

Interestingly, the different types of neural tube defects observed occur at the extreme ends of the neural tube (not in the middle) as a result of how the neural tube closes. The neural tube begins closure in the middle and slowly closes on either end. Spina bifida is localized to the posterior regions of the neural tube, while anencephaly, encephalocele, and hydrocephalus occur in the anterior regions of the neural tube.

It is predicted that approximately 70% of neural tube defects occur due to genetic factors. There have been over 240 genes associated with neural tube defects (Harris and Juriloff, 2010), which strongly supports the link between genetic predisposition and neural tube abnormalities. However, genetics do not solely determine the presence of neural tube defects, as a huge proportion of neural tube defects can be prevented with the nutritional supplementation of folic acid in the mother’s diet. Folic acid is a water-soluble B-vitamin and deficiencies in folic acid have been associated with an increased risk of neural tube defects (Wallingford et al., 2013). Many foods including orange juice and some cereal grains are fortified with folic acid. The addition of folic acid to specific food products of general consumption has had an immediate impact on the prevalence of neural tube defects. Despite the efficacy of folic acid supplementation, it is still unclear how folic acid directly prevents neural tube defects.

Segmentation of the early brain

The enclosed neural tube contains almost all the components needed to produce the brain and spinal cord. The tube itself will now morph into a region-specific structure that includes the forebrain, midbrain, hindbrain, and spinal cord (Early segmentation of the brain). (See Introduction.)

This process is called segmentation (Hirth et al., 1995; Millet et al., 1996; Nomura et al., 1998; Krumlauf and Wilkinson, 2021) and includes the process of bending and folding regions of the neural tube. The anterior region of the neural tube will bend forward with the anterior most region forming the prosencephalon (forebrain) and the mesencephalon (midbrain), which is adjacent to the cephalic flexure. A flexure is a term used to describe a bend or fold. The cephalic flexure is distinguished as a curvature located at the junction between the brainstem and spinal cord. It is a critical anatomical adaptation that helps support the body to walk on two legs, opposed to four like cats or dogs. The body is able to do this by balancing the weight of our heads on top of the spinal cord, while keeping our eyes forward to gather spatial awareness as we walk on two feet.

The cephalic flexure is anterior to the rhombencephalon (hindbrain), which is located in between the cephalic flexure and the cervical flexure, the bend that will divide the hindbrain and spinal cord (Early segmentation of the brain). The prosencephalon is further subdivided into the telencephalon and the diencephalon. The rhombencephalon is divided into the metencephalon and myelencephalon. The telencephalon will produce the cerebral cortex, while the diencephalon will generate the thalamus, and hypothalamus. The midbrain and hindbrain will produce parts of the brainstem, pons, and cerebellum, collectively (see Introduction). Posterior to the hindbrain region, the neural tube becomes the tissue of the spinal cord. These divisions are all specified early in development, even though their full functional maturation will take months (and sometimes years).

Environment and Bidirectional communication: Genetic diseases of development with environmental influence

Disorders that lead to structural damage or malformations which impair lifelong function are the most common forms of congenital brain disorders. While there are a variety of environmental teratogens that can alter brain development, a significant number of congenital brain malformations are associated with a genetic mutation. For many of these disorders, a gene of interest has been assigned or associated with the disease, but how a specific mutation disrupts overall brain development is still a puzzle. These mysteries persist most likely because, in many cases, genes interact with environmental influences to determine their final effect on the brain. For example, holoprosencephaly is a congenital brain disorder characterized by abnormal separation of the forebrain (Weiss et al., 2018). This occurs prenatally and mutations in various genes have been shown to cause holoprosencephaly (Geng and Oliver, 2009). However, the degree of illness is often variable amongst individuals (Holoprosencephaly). Severe forms of the disease exhibit cyclopia (a single eye) and the brain does not divide into two hemispheres whereas some individuals diagnosed with holoprosencephaly can present with semilobar holoprosencephaly, in which the brain has a small separation between each hemisphere (Holoprosencephaly). It is generally accepted that environmental exposures in combination with genetic mutations ultimately dictate the presence and severity of holoprosencephaly. For example, some environmental teratogens such as ethanol are known to exacerbate the phenotypes of holoprosencephaly (Hong and Krauss, 2012). Additional environmental interactions include maternal diabetes and prenatal exposure to specific pharmaceutical drugs (Cohen and Shiota, 2002).

Proliferation and differentiation: how are neurons and glia formed?

This section will cover the generation of neurons and glia in development.

Neural stem cells and proliferation

As the neural tube forms, we see the emergence of a special class of cells that will proliferate (divide) to create numerous daughter cells that go on to become neurons and glia of the CNS, except for microglia which are formed in the peripheral yolk sac. These cells are committed to become nervous system cells and are referred to as neural epithelial cells or neural stem cells. In the very early stages of neural tube formation (Step 1 in Embryonic corgical neurogenesis), these cells divide rapidly to create identical daughter cells, thereby expanding the neural tissue and growing the neural tube. Some of these daughter cells go on to become a type of cell called a radial glial cell (RGC).

RGCs are a class of neural stem cells located in a proliferative region called the ventricular zone, the thin layer of tissue surrounding the fluid-filled central cavity of the neural tube (i.e., the future ventricular system). RGCs have a unique, bipolar shape, shown in Embryonic corgical neurogenesis, where they extend one process medially, to contact the ventricular space, and another laterally, to contract the edge of the developing neural tube (known as the pial surface). Within the ventricular zone, RGCs divide (proliferate). The goal of this expansion is to create enough RGCs to produce the neurons and glia required for a fully formed brain.

RGCs are unique in that they have the ability to self-renew, which means they produce more of themselves through symmetric cell divisions where one cell produces two independent identical cells. RGCs also have the ability to divide asymmetrically, which means they can produce one cell that is identical to themselves, but at the same time produce a different, more committed cell that we call an intermediate progenitor (IP) (Step 2). IPs migrate along the radial processes of the RGCs into the subventricular zone (Step 3 and detailed location on the y-axis) where they will undergo symmetric cell divisions to produce neurons (Step 4 and 5). We call this process of creating new neurons neurogenesis. Later, RGCs will switch towards the development of daughter cells that go on to become astrocytes and oligodendrocytes, processes referred to as gliogenesis (Step 6). Gliogenesis occurs well after neurogenesis and closer to birth, with production of astrocytes occurring first followed by production of oligodendrocytes.

As neurogenesis and gliogenesis progress, the brain, of course, grows. For humans and several other highly intelligent mammals, our brain growth presents an exceptional challenge. In short, we need to fit a lot of brain, particularly a lot of cortex, into a small space, otherwise our heads would be too big for our necks to support. All mammalian species have a cortex, but it differs in size and texture (see Introduction). In smaller mammals such as mice, the cortex is smooth and smaller relative to skull size. However, in higher mammals, the cortex is much larger. It has a surface area the size of a pillowcase, approximately 40 by 62.5 centimeters. This size difference can be attributed to increased proliferation and production of neurons. To accommodate the increase in cell number, the human brain produces cortical folding, which is the formation of gyri and sulci. In rare instances, the human brain lacks the formation of gyri and sulci, this condition causes lissencephaly, or smooth brains. Lissencephaly is associated with intellectual disability.

As the human brain grows during development, we see it go from a relatively smooth structure to one with many sulci and gyri. Brain growth and folding shows how during development the large cortical lobes curl over on top of the lower brain regions, expanding to cover and surround mesencephalic regions. This curling over during development helps explain why many of our subcortical structures (hippocampus, thalamus, basal ganglia, and lateral ventricles) have a curved shape. To get an idea of how brain folding occurs during brain development, take out four sheets of paper. First take one sheet and crumple it tightly. The paper will slightly expand then settle into a crumpled state. Now get the remaining three sheets of paper and crumple all three together. Notice how thicker paper leads to fewer folds. The crumpled pieces will be much larger and harder to fold. Now apply that logic to the growing brain. Thicker cortices fold less and end up with less cortical material hiding below the surface.

Differentiation of neurons and glia

As mentioned previously, asymmetric cell divisions produce neurons first, and later in development they produce glial cells. Thus, the same neural stem cell has the capacity to produce neurons and glia, two vastly different cell types (this figure).

This capacity to make more than 1 cell type is referred to as multipotency. The process by which the daughter cells mature in their final functional form (neuronal or glial) is called differentiation. Differentiation of neurons and glia are separated by time. Their formation is regulated by the promotion or repression of gene expression. For example, the expression of proneural genes promotes the differentiation of daughter cells into neurons. During this time, the expression of proglial genes are inhibited. Thus, proneural genes have the capacity to suppress gliogenic signals (Bertrand and Guillemot, 2002). Later, proneural genes are repressed and proglial genes turn on, allowing daughter cells to mature into glia.

But what causes these genes to turn on and off? Expression of proneural or proglial genes, and therefore differentiation, is highly influenced by neighboring cells in the environment. Cells usually interact and communicate by secreting soluble signals that bind receptors expressed on the cell surface, or by cell contact, which allows cell-surface proteins on each cell to bind each other directly. Either kind of interaction (soluble, cell contact) can initiate intracellular signaling that drives changes in expression of the neural/glial driving genes. As the brain matures, these environmental signals change, different genes are turned on and off, and the balance between neuro- and glio-genesis therefore shifts.

Neuronal Migration Patterns

Neurogenesis occurs whilst the neural tube is expanding into multiple layers. A key step in neurogenesis is for these new cells to migrate to the right spot, where they can eventually become part of functional circuits. The neural tube contains a mixture of neural stem cells and IPs. Dividing IPs are produced in a sub-region called the subventricular zone, from an asymmetrically dividing population of RGCs. IPs will eventually produce neurons that travel along the two outstretched processes of the RGC. Newly born neurons migrate along the glial projections of RGCs towards the pial surface (Step 3 in Embryonic corgical neurogenesis). As development proceeds, subsequent waves of neurons will migrate past the existing newly differentiated neurons. This process results in the neural tube becoming partitioned into different layers as it expands. In the mammalian cortex, the remnants of this process can be seen in the distinct cell layers that contain different subtypes of neurons (Rakic, 1974). Not all neurons migrate along RGCs. Inhibitory GABAergic interneurons migrate tangentially from zones called ganglionic eminences. There are 3 ganglionic eminences from which many different neurons arise and undergo tangential migration. Radial versus tangengial migration diagrams tangential versus radial migration, showing how the two differ.

Communication by electrical impulse

The anatomist Santiago Ramón y Cajal was the first to describe fibers radiating from individual neurons (Cajal, 1995). He postulated that neurons were separate entities and used long fibers and connections to communicate. How the brain communicates through synapses was previously discussed in Introduction. Here we will provide a short review of this process because it is integral to the formation of synaptic connections early in development. Neurons communicate with each other via the synapse which can be seen in Reminder of synaptic transmission.

The presynaptic neuron (sending the signal) is in close proximity with the postsynaptic neuron (receiving the signal) and can use signaling molecules and neurotransmitters to either increase or decrease the probability that the postsynaptic neuron fires an action potential. This communication occurs at the synaptic cleft (Reminder of synaptic transmission). Complex mechanisms are required to ensure that the correct presynaptic neuron pairs with the appropriate postsynaptic neuron/tissue during development. Neurons extend and grow their axon using multiple cues and once they encounter their partner, they will test their match using temporary junctions, cell signaling and adhesion molecules, combined with electrical activity. Such signals, if positive, will then promote formation of a mature synapse.

During development, neurons extend a specialized structure from the tip of a growing axon or dendrite. This structure was first described by Ramon y Cajal and was coined the growth cone (Growth cones). Growth cones are dynamic, in that they sense and respond to the environment in search of a target tissue to innervate (connect to via a synapse). Growth cones change shape by extending outward a leading edge (Growth cones), which interacts with the environment. The leading edge of the growth cone appears as a hand with fingertips. The outstretched fingertips are called filopodia (plural). Each filopodium (singular) samples the environment for attractive and repulsive signals that help to determine the direction of growth. In between each filopodium is a flat structure called the lamellipodium (Growth cones). Growth cones express various molecules on their surface that interact with the environment. Therefore, two unique growth cones may respond differently to the same environment.

The growth cone is dynamic, and when it meets attractive cues, a structural protein inside the budding growth cone called actin facilitates changes in the direction of the growth cone. Actin subunits can come together in a process called polymerization to move the growth cone. Actin filaments are primarily located in the filopodia.If a repulsive signal is detected, in contrast, actin subunits detach from each other, which is commonly referred to as depolymerization. Actin subunits are essential for growth cone movements, but the axon itself is not made of actin. It is produced from a more rigid protein called tubulin. Tubulin subunits come together at the opposite end of the leading edge of the growth cone to elongate the axon.

Types of receptors expressed on the growth cone

There are several different classes of molecules expressed on the surface of a growth cone. Signals guiding the growth cone shows three of these major classes.

Cell adhesion molecules (CAMs) such as neural cell adhesion molecule (NCAM) bind to other CAMs on neighboring cells. These molecules can bind to each other in a process called homophilic binding (bind to the same receptor on another cell) or can bind to other CAMs through heterophilic binding (binding to a different receptor on another cell). CAMs are unique from calcium dependent cell adhesion molecules (cadherins) because cadherins require calcium for an interaction and bind via homophilic binding. Cadherins are similar to CAMs because they are receptors present on adjacent cells which interact. Both CAMs and cadherins are unique from a third class of membrane bound receptors called integrins. Integrins bind and interact with components of the extracellular matrix. These components include laminin, fibronectin, and collagen.

Additional cues can be received through ephrins, semaphorins, and netrin. Ephrins bind to Eph receptors while semaphorins bind to plexin or neuropilin receptors. Netrins are secreted factors that bind to the Unc-5/DCC receptor. Semaphorins are mostly inhibitory but under some circumstances can present attractive signals to the growing axon. There are two classes of ephrins, A and B, class A ephrins are anchored into the membrane and class B are transmembrane receptors that have an intracellular and extracellular domain. In vertebrates, netrin binds to the receptor DCC, which is expressed on the growth cone. The interaction between DCC and netrin is an attractive signal for the growth cone. Interestingly, netrin can also bind to the UNC5 receptor and when it does, it provides a repulsive cue. Thus, netrin can provide both attractive and repulsive cues.

Synaptogenesis

Ultimately, the growth cone is directed by those attractive and repulsive signals to a putative target cell to form a mature synapse. The mature synapse has basic structural components that facilitate neuronal activity. The mature synapse contains vesicles of neurotransmitters, ionic channels (calcium channels), components to release neurotransmitters, a synaptic cleft, and a postsynaptic density (PSD) (Reminder of synaptic transmission). These are all features that develop as a result of the pre- and postsynaptic cells interacting to help each other mature. When the presynaptic cell first meets the postsynaptic target, the synapse has immature characteristics, exemplified in Immature v. mature synapses.

There are two major classes of synapses: electrical and chemical (see Introduction). Chemical synapses communicate through the release of neurotransmitters, which are located in vesicles within the presynaptic neuron. The early immature chemical synapse lacks a PSD and has only very few vesicles in the presynaptic neuron. This immature chemical synapse has the ability to signal via neurotransmitter release, though it is not as robust as a mature synapse will eventually be (Immature v. mature synapses). Rudimentary neurotransmitter release is required for chemical synapse maturation. Electrical synapses are fewer in number relative to chemical synapses and function through gap junctions that allow the electrical current to move from one cell to the next. Recent studies have shown that the components of an electrical synapse (i.e. gap junctions) are required for the development of chemical synapses (Todd et al., 2010). Thus, there is evidence that neurons first form transient electrical synapses before chemical synapses develop.

Developing a synapse entails 3 main steps. The growing axon must 1) detect a postsynaptic target and defasciculate, 2) form temporary contact with the target cells, and 3) undergo matching between the neurotransmitters of the presynaptic cell with the receptors on the target postsynaptic cell. During axon outgrowth, like-minded axons, those with similar growth trajectories, form bundles called fascicles (Stoeckli and Landmesser, 1995). Therefore, when an axon recognizes a potential postsynaptic match, the axon must undergo defasciculation. Defasciculation is when the axon breaks away from the bundle of axons to encounter its target cell. After detachment, the axon will undergo a matchmaking process that relies on cell adhesion. Cell adhesion is a process of connecting to neighboring cells in the environment using cell surface receptors. Cell contacts between the pre and postsynaptic neuron hold the two cells together long enough to initiate synaptogenesis. Subsequent signaling between the two cells ensures that the presynaptic cell synthesizes, packages, and releases an appropriate neurotransmitter that the postsynaptic cell can respond to via expression of appropriate receptors in its PSD.

Neuronal survival and cell death

The developing brain undergoes significant expansion of neural stem cells, which are required for the development of neurons and glia. Interestingly, the developing brain initially produces too many neurons, meaning that the number of cells produced is more than what is needed for adult function. How does the brain compensate for this overproduction of neurons? As it turns out, the developing brain undergoes a process called programmed cell death, or apoptosis. Through apoptosis, excess neurons die, and the developing brain is refined.

Neuronal survival is regulated by multiple mechanisms, one of which is the size of the target tissue which a neuron will innervate. In 1909, Marian Lydia Shorey first demonstrated that there was a correlation between target tissue size and the number of surviving neurons (Shorey, 1909). Removal of the budding limb from chick embryos led to a decrease in the total number of neurons innervating the tissue whereas adding an extra limb tissue increased the total number of neurons. Subsequent experiments by Viktor Hamburger (1934) later revealed that neurons underwent cell death after the target tissue was removed (Hamburger, 1934). Tissue size and neuronal survival in development shows an example of this kind of experiment and the resulting loss of motor neurons in the lumbar spine of developing chicks.

These studies show that neuronal survival is not a random process and is influenced by the ability of tissues to form connections. How might the embryo decide which neurons die and which should survive? Neuronal survival is mediated by a class of molecules called neurotrophins that bind to a family of cell surface receptors called Trk receptors. Neurons that have inadequate signaling between neurotrophins and Trk receptors ultimately undergo apoptosis. Neurons have to compete with each other for access to these neurotrophins (Neurotrophin and cell survival). Tissues producing less neurotrophins will support the survival of fewer neurons. This competition will also favor the survival of neurons that start to make better connections with sources of neurotrophin. In the chick limb bud removal experiments, the loss of the limb bud tissue removed a major source of neurotrophins, causing motor neurons headed towards that tissue to die off while others that found more neurotrophin-abundant targets in other tissues survived. This selection is one of the reasons we refer to this step as part of nervous system refinement. It is not random which neurons die and which survive, but rather driven by the connections the neurons are starting to form.

Synapse refinement

Early prenatal brain development combines coordinated control of various signals to produce an elaborate foundational unit capable of precise control of the human body. However, it is important to note that this unit represents only a foundation and is subject to refinement. We discussed refinement above in terms of which neurons survive. Synapse refinement is also a prominent feature of the developing nervous system. Synapse refinement is the process by which the early synapses are modified or eliminated because of specific signaling events.

The neuromuscular junction (NMJ) is a well-defined example of synapse refinement after birth (Bishop et al., 2004). We have learned a great deal about the capacity for refinement by studying the NMJ. The NMJ represents the synapse between the axon of a motor neuron and a muscle fiber (Introduction). Muscle fibers are regulated by a single motor neuron in adults, but it has been established that in newborn rats, muscle fibers are innervated by a minimum of 2 different motor neurons. This type of innervation is called polyneuronal innervation. After birth, motor neuron innervations are reduced, but not because of neuronal cell death. Rather, the branches are eliminated or refined to target only a single muscle fiber (Muscle innervation development) (Brown et al., 1976).

The process of transitioning from polyneuronal innervation to mono-neuronal innervation is dependent on electrical activity and competition between two innervating neurons. It is critical for the precise recruitment of the motor unit during force and contraction generation (Lee, 2020). Should it be dysfunctional, more than one neuron would control the activity of a muscle cell. This phenomenon is not limited to the NMJ and has been demonstrated in the CNS. For example, Purkinje cells of the cerebellum that are innervated by climbing fiber inputs of the inferior olive are poly-innervated at birth, but by 2 weeks of age, the synapses are refined to a single innervation per Purkinje cell (Mariani and Changeux, 1981).

Synapse refinement in the CNS

We have discussed synapse refinement at the NMJ as a model for how synaptic connections are reorganized postnatally. Similar processes are occurring in the CNS, but what drives this process in the CNS? It turns out that, similar to how neurons must compete for neurotrophin survival cues, competition is also critical to synapse survival.In short, synapses that get used more get strengthened and those that are not used are lost. The most effective synapses are those that have a strong association between presynaptic neurotransmitter release and postsynaptic firing. Presynaptic neurons that fire in unison are especially effective at competing as a team to maintain strong connections that cause the postsynaptic neuron to fire. Synaptic correlation shows a representation of this principle. Early in development, a postsynaptic neuron may receive multiple inputs (# 1), some of which fire together (# 2) and others of which do not (# 3). The correlated inputs are much more likely to cause the postsynaptic cell to fire, because they are working together. As development proceeds, those correlated connections get stronger (# 5) and uncorrelated ones get weaker.

What happens to those “weaker” synapses? Rival synapses that cannot fire synchronously with a neuron or a target cell and receive little input eventually lose synaptic connection and are eliminated through pruning (# 4 in Synaptic correlation). Why do we need to eliminate synapses? Consider it like how tree roots compete for water in the rainforest. To survive and strengthen, the roots that branch out towards areas which fortify the livelihood of the tree will grow stronger, and those that fail in this sense weaken and die. Synapse numbers tend to peak during childhood in humans, with pruning then extending from late childhood through adolescence (Synapse formation and pruning).

Pruning requires a balance; too little or too much can reduce the efficiency of brain function. Aberrant pruning can result in cognitive disorders due to faulty wiring in early life stages (Johnston, 2004; Sakai, 2020). For example, patients diagnosed with schizophrenia, a debilitating psychiatric illness with a myriad of hallucinogenic symptoms, show fewer synapses (over pruning) in brain scans relative to healthy individuals (see Introduction). The loss of synaptic connections is pronounced in the prefrontal cortex, a region where pruning peaks during adolescence, an active period of synapse elimination, coinciding with the onset of the disorder. In contrast, disorders such as autism have shown an excess in connections, suggesting failure to eliminate synapses. Findings like these suggest that CNS disorders can result from excessive or insufficient pruning during childhood and adolescence.

Myelination and plasticity

Myelin is a sheath that surrounds axons to promote the rate of electrical signaling. It is made of proteins and lipids and deposition of myelin on axons begins 1-2 months before birth. This makes myelination one of the later steps in nervous system development (Primate neurodevelopment).

Myelin is produced by oligodendrocytes in the CNS and Schwann cells in the peripheral nervous system (see Introduction). Both of these are glial cells produced during gliogenesis. Myelination does not occur all at once, but rather in a progression. Generally, axons in the peripheral nervous system are myelinated first, followed by spinal cord axons and finally axons in the brain. Myelination in the developing brain shows MRI images of the brain during the first year of life, showing rapid accumulation of myelin (pseudocolored red) in the brain. This first year is when the peak of myelination occurs in humans. A parallel in the first month of life in rodents can also be seen in the bottom panel of Myelination in the developing brain, where myelin has been stained dark in brains from rats aged postnatal day 7 to 35.

Though the first year is when most myelination happens in humans, myelination continues in the forebrain well into childhood and adolescence and accounts for some degree of plasticity during these time periods. Interestingly, extensive piano playing is associated with increased myelination (Fields, 2005). In adolescence, increased risk-taking behavior can be the result of incomplete development of the forebrain, which includes reduced white matter (myelination) development (Beckman, 2004). But how does the environment affect myelin deposition? For the CNS, the answer probably lies in how sensitive developing oligodendrocytes are to environmental stimuli. Oligodendrocytes are glial cells of the CNS that produce and deposit myelin. For example, when rats were raised in an enriching environment, with toys and other rats to play with, the number of oligodendrocytes in their cortex actually increased (Szeligo and Leblond, 1977). Consistent with the increase in the number of oligodendrocytes, an enriching environment is also associated with an increase in the number of myelinated axons (Juraska, 1988). These studies were performed in rats, but there is also evidence for white matter developmental plasticity in humans. For example, neglect during childhood has been shown to decrease white matter by 17% in the corpus callosum (Teicher et al., 2004). We will discuss more about experience-dependent plasticity in neurodevelopment, but these examples help to highlight how development is not a simple series of steps but a constant interplay between developing systems and their environment.

Embryonic and fetal period

Developmental sensitivity to defects defines critical periods in prenatal development. These periods are timepoints in prenatal development when specific organ systems of the embryo or fetus are particularly susceptible to environmental stimuli, meaning that negative stimuli can result in structural or functional brain malformations that impair function lifelong. CNS development begins embryonically at 3 weeks post-fertilization. The embryonic period lasts until 10 weeks, and encompasses neurulation, a critical period we have previously discussed as it relates to neural tube formation and closure. CNS development continues from 10 weeks to 38 weeks, which is defined as the period of the fetus.

The pioneering work of Charles Stockard and Hans Spemann has shown that there are stages that are particularly sensitive to environmental exposure and/or manipulation (Stockard, 1921; De Robertis, 2009). Stockard is known for his work with fish embryos, in which he found that introducing chemical agents at a specific stage would create what he called “monsters”. Stockard's drawings shows a drawing by Stockard of one of these fish after exposure to a teratogen during the critical embryonic period. This damage was permanent, and when replicated at other stages of development, there was essentially no similar defect seen, suggesting that there are specific prenatal stages that are very sensitive to teratogens. The reason for this susceptibility is that the fetus is experiencing rapid growth in size, weight, and muscle at early stages. Below we detail two common environmental exposures during fetal development that result in abnormal nervous system development, causing lifelong impairment in function.

Fetal alcohol syndrome

Exposure to alcohol during the fetal period can result in fetal alcohol syndrome (FAD) and fetal alcohol spectrum disorders (FASD). Alcohol is a teratogen, and its effects can lead to growth retardation, abnormal brain development, intellectual disability, and craniofacial abnormalities, which include a narrow forehead, flat midface, narrow eyelid openings, shortened nose, a long upper lip, and an absent philtrum (Fetal alcohol syndrome; Sulik et al., 1981).

FASD is the most common non-genetic form of intellectual disability, with broad symptoms, including poor cognition and behavioral phenotypes (Denny et al., 2017). The effects of FAD on growth and development can be observed in Fetal alcohol syndrome. While the term “fetal” in FAD/FASD implies the effects of alcohol are focused only on the fetal period, alcohol can also have a negative impact earlier in pregnancy. When mice had a brief exposure to alcohol during embryonic development, they presented with severe facial malformations (Sulik et al., 1981). This period of embryonic development in mice is equivalent to approximately the third week of human gestation, a time point when many women aren’t aware of their pregnancy (Sulik et al., 1981). This is why it is recommended that individuals planning to become pregnant refrain from alcohol consumption.

2: Perinatal period

The perinatal period encompasses the time frame from childbirth up to 24 months. During this time, the CNS possesses tremendous plasticity, defined as the ability to remodel or reorganize based on experience. Such plasticity is very robust from birth to 3 years and gives rise to developmental critical periods. While the early brain maintains a high degree of plasticity, it is important to note that plasticity is maintained into adulthood, although it does decrease as we age. As discussed previously, significant synapse refinement occurs in the form of loss of poly-innervation and pruning during this period of development. Below, we will discuss several examples of how long-term brain function can be shaped by experiences during this perinatal period.

Perinatal brain development and parental care

One of the most tragic consequences of the perinatal critical period for CNS development is the long-term and severe impact that neglect or abuse during this time can have. Brain symptoms of neglect highlights some of the major brain regions where disrupted growth is evident years after childhood abuse/neglect. They include regions important for emotion (amygdala), memory (hippocampus), and executive function (prefrontal cortex). These kinds of extreme examples show how experience in early life can disrupt the structure of the brain long-term.

Perinatal critical period in visual development

Perinatal sensory experience is required for the development of the visual system at several levels. One prominent example of this is development of ocular dominance in the primary visual cortex (called V1 for short). At birth, the layers of the primary visual cortex are innervated by axons from both eyes. However, by adulthood through visual stimulation, the innervations are remodeled and segregated between layers to produce ocular dominance columns (Stryker and Harris, 1986; Levelt and Hübener, 2012; Berry and Nedivi, 2016). Within an ocular dominance column, the cells respond preferentially to light being shown in one eye over the other. This process of refinement to create preference for input from one eye requires visual input, as was first demonstrated by studies in cats.

David Hubel and Torsten Wiesel performed the first key studies to establish critical periods of vision development (Wiesel and Hubel, 1963; Hubel and Wiesel, 1970; Levelt and Hübener, 2012; Berry and Nedivi, 2016). To study how exposure to light influences formation of cortical cell preference for responding to input from one eye or the other, Hubel and Wiesel used the simple model of suturing shut a single eye of feline subjects at different times. They measured ocular preference of cortical cells by recording from individual cells in the cat visual cortex while shining a light on one eye or the other. Ocular dominance critical period describes their findings.

In an adult cat that never had an eye sutured shut (i.e., a normal cat), they found a mixture of visual cortical cells that responded preferentially to the right eye, left eye, or a bit of both. However, if a cat experienced just 3 to 4 days of visual deprivation when they were a kitten (perinatally), almost all their visual cortical cells responded only to light from the unsutured eye (Hubel and Wiesel, 1970). The formerly sutured eye was effectively blind because it did not stimulate any cortical cells. Importantly, if the eye was sutured shut during adulthood, there was no effect (Ocular dominance critical period), effectively confirming a critical period for vision development. The first few weeks of life are a critical period for this part of visual development because this is when inputs to the visual cortex are forming synapses on V1 cortical cells. Cortical cells end up only connected to inputs deriving from the open eye because of a lack of competition from the axons in the sutured eye. Opening the eye later can’t fix that as the critical period for synapse formation has passed.

The significance of visual stimulation and critical periods has been further substantiated by a famous case study in Oliver Sacks’ renowned publication (1985). There he described a human patient by the surname Fergal, who was born blind and received surgery to regain vision after 50 years without it. Despite the success of regaining vision, Fergal lacked the early experiences necessary for synaptic reorganization and proper vision development (Sacks, 1985). This resulted in an influx of stimuli, as his other senses were heightened. As Sacks (1985) describes, “There were no visual memories to support a perception; there was no world of experience and meaning awaiting him. He saw, but what he saw had no coherence. His retina and optic nerve were active, transmitting impulses, but his brain could make no sense of them.” A second example of sensory critical periods and intervention includes the usage of cochlear implants (see Introduction). The efficacy of cochlear implants is widely affected by the age of the patient. For example, the efficacy associated with cochlear implants in children at a young age, with shorter periods of auditory deprivation, is much higher than those with prolonged auditory deprivation (Pisoni et al. 1999).

Plasticity during adolescence

Adolescence refers to the stage of development in between childhood and adulthood. Thus far, we have discussed robust examples of plasticity during infancy and early development. However, the adolescent brain exhibits tremendous neural development, which makes it highly plastic and susceptible to environmental influence. Adolescence is particularly characterized by refinement in connectivity between major brain regions. For example, the connections between the frontal lobe and the amygdala are subject to extensive experience dependent refinement during adolescence. The prefrontal cortex (PFC) is responsible for higher order function and decision making called executive function and the amygdala regulates fear, aggression, and emotion. Adolescent decision making is also influenced by the immaturity of the reward centers in the brain, which include the nucleus accumbens (Spear, 2000), which contains dopaminergic input. The relative immaturity of these 3 brain regions during adolescence leads to more impulsive, reward based, and risk like behaviors relative to adults (Ernst, 2014). It also underlies their enhanced vulnerability/plasticity during adolescence. Studies have shown that chronic stress and psychological distress can have brain region specific impacts during adolescence.For example, abnormal development of the PFC can affect working memory, cognitive flexibility, and self-control.

Adolescence is also a unique critical period for the effects of substance misuse on specific regions of the brain. For instance, substance misuse that occurs during early adolescence, around ages 14-16, is associated with reduction in the volume of frontal cortex gray matter, whereas misuse in childhood selectively impacts the hippocampal gray matter volume (Larsen & Luna, 2018). Such differences can be explained by the different time points in development where growth is more prevalent for each of these two brain regions.

We Capture Photons of Light Reflected from Objects Around Us

Light is a form of electromagnetic radiation, energy packaged in particles called photons. Physicists understand that light is both a particle and an electromagnetic wave, and although light can be described by the energy in each photon, light is more typically described by its wavelength on the electromagnetic spectrum. Electromagnetic wavelengths stretch from long wavelength, low energy radio waves to short wavelength, high energy gamma rays. (Light is a component of the electromagnetic spectrum)

Sunlight includes wavelengths from infrared to ultraviolet (UV) that penetrate the atmosphere, but only the middle of that range penetrates water, which is where vertebrate vision first evolved. This is the visible spectrum that our eyes can detect. Different components (wavelengths) of the visible spectrum are absorbed or reflected by the surfaces of objects and our visual system interprets the reflected wavelengths as an object’s color and shape. Shorter wavelengths within the visible spectrum we call blue while longer ones look red. White light combines wavelengths from the entire visible spectrum. Bright lights produce many photons per second, while dim lights emit few photons per second.

Anatomy of the Eye

The eye is the primary sensory structure that intercepts electromagnetic waves in the visible spectrum, eventually allowing us to perceive light. Anatomy of the human eye shows the major parts of the eye.

Correcting Optical Flaws: Myopia, Hyperopia and Presbyopia

Sharp vision requires the image projected on the retina to be in good focus, but for many people, the focus is in front of the retina (myopia) or behind the retina (hyperopia) (Correcting optical flaws). Corrective lenses restore sharp focus on the retina. Although in some cases people are born with an eye that is too short or too long for good focus, and thus need to wear corrective lenses from an early age, there is also some indication that constant close-up work can affect the continued development of the eye’s shape and lead to myopia.

Even people who do not need corrective lenses when they are young eventually require reading glasses as they age. This condition is called “presbyopia,” literally “old eyes,” and it results from the lens of the eye becoming less flexible as people age. The lens normally becomes rounder to focus on nearby objects, but as people age, the lens becomes more rigid and is unable to change its shape to focus on nearby objects such as pages of text. At that point, corrective lenses are needed to supplement the lens’s ability to focus.

The Retina Is Composed of Several Neural Cell Types

The retina (Retina structure) arranges five types of nerve cells in distinct layers. The retinal processing chain for conscious vision begins with the receptors (rods and cones) that absorb light. The rods and cones make synapses with bipolar cells, which connect in turn to ganglion cells. Horizontal cells support visual processing with lateral synaptic connections to receptors and bipolar cells. Amacrine cells provide lateral connections in the synaptic layer between bipolar and ganglion cells. Ganglion cell axons spread across the inner retinal surface, gathering as a bundle to leave the eye as the optic nerve. The figure does not show the pigment epithelium above the receptors, or the extensive array of blood vessels that spread over the retina next to the ganglion cells.

Embryologically, the retina is an outpocketing of the brain, and because it is more accessible than the developed brain and arranged in clear layers of defined neuronal cell types, it has been the subject of many experimental studies. This began in the 1890s when Cajal published images of retinas stained by the Golgi method, and identified many of the cell types. Since all vertebrate retinas have the same general structure, retinas with relatively large cells have been the most intensively studied, such as the retina of the mudpuppy, a large amphibian related to frogs and salamanders.

The human retina shown in Fovea structure has relatively small cells, but it is clear they are arranged in distinct layers. Although the eye is sometimes cited as evidence for “intelligent design,” the retina is actually a poor optical design because light must pass through several layers of neurons to reach the photoreceptors. These layers scatter the light and blur the image. It is never possible to know exactly why evolution produced a particular structural outcome, but two theories have suggested why the receptors are placed in an optically poor position, facing away from the light. One possibility is that because the receptors are metabolically very active, they require a dense blood supply, which they receive by facing the blood vessels in the choroid layer. A second possibility is that receptors evolved from ciliated cells that lined the embryonic neural tube (the “connecting cilium” is a group of microtubules arranged like a cilium that are still present in photoreceptors). As additional retinal layers are added during development, the receptors remain in an optically poor position. The fovea compensates for this “unintelligent” optical design by pushing aside the neural layers at our center of gaze, where tightly packed cones provide our sharpest visual acuity.

Rods and Cones Capture Photons of Light

In The Retina we introduced photoreceptors, the specialized neurons that respond to light. Photoreceptors (Morphology of photoreceptors) have an outer segment of discs, or infolded membrane in which millions of visual pigment molecules are embedded. These pigment molecules give photoreceptors the ability to capture light and will be discussed more below. Photoreceptors also have an inner segment with mitochondria, the nucleus, and other structures typical of all cells (see Introduction). The synaptic terminal is the photoreceptor’s output region, where it connects to bipolar and horizontal cells.

Our photoreceptors come in two types, each with a unique function in vision: cones and rods. The human eye has about 6 million cones, concentrated in the fovea and central retina but also distributed sparsely in the retinal periphery. Cones are responsible for high-acuity color vision in daylight. The human retina also has about 120 million rods, which are absent from the fovea but distributed across the rest of the retina. Rods serve low-acuity vision in dim light (such as moonlight), and under optimal dark-adapted conditions a rod can respond to as little as a single photon of light. In daylight and typical artificial light, the rods no longer respond because the bright light saturates their response, and only the cones signal differences in brightness.

Rods and Cones Depolarize in the Dark and Hyperpolarize in the Light

When neurobiologists first succeeded in recording from rods and cones with intracellular electrodes, they were puzzled by what they found. Most sensory receptors, including invertebrate photoreceptors, depolarize in response to a stimulus, but vertebrate photoreceptors hyperpolarize (become more negative) in response to light. Rod and cone membrane potential changes shows an example of the kind of data one of these experiments generated, where a photoreceptor’s intracellular electrical response was recorded while light was flashed on it. Dim flashes elicited a small hyperpolarization, and bright flashes evoked a large hyperpolarization. The receptors in the dark also had a surprisingly depolarized resting potential, about -40 mV instead of the -60 mV more typical of nerve cells (see Introduction). Everything seemed backward!

Later work unraveled the mystery. In the dark, channels in the outer segment are open and admit positive Na+ ions, creating a depolarized resting potential. Capturing photons of light leads to closing some of these sodium channels, which diminishes the resting depolarization, an apparent hyperpolarization. With brighter lights and more photons captured, more of the channels close, leading to greater degrees of hyperpolarization. In a way, we can think of the response to light as allowing the photoreceptors to return to their “real” resting potential (the potential when the sodium channels are closed, about -60 mV like other neurons). Light interrupts the depolarization that occurs in the dark. Like all other neurons, photoreceptors release their transmitter (glutamate) when they are depolarized, so photoreceptors continually release transmitter in the dark, and light interrupts transmitter release. In the next section, we will discuss how photoreceptors keep Na+ channels open to become depolarized in the dark and how light exposure cancels that.

Phototransduction

Eventually, researchers worked out how the capture of photons by visual pigment molecules in the disk membranes of photoreceptors led to closing sodium ion channels in the cell membrane. We call this process of turning photons into an electrical signal "phototransduction" (Phototransduction). Here, we will describe this process in rods. Cones use similar processes but differ in the wavelengths of light that they absorb, which will be discussed in the next section.

In rods, rhodopsin molecules are the visual pigment packed into the disk membranes. Rhodopsin molecules consist of two components: opsin, a large protein, and 11-cis retinal, a small subunit that is derived from Vitamin A. In the dark, the 11-cis retinal has a bent side chain. Capturing a single photon provides the energy to straighten the side chain to its straight, “all-trans” shape. This in turn activates the opsin protein, which leads to a series of steps that reduce the level of cGMP, a messenger molecule in the cytoplasm. While the intermediate steps may seem detailed, the important point is that light leads to closure of sodium channels that are open in the dark.

As Phototransduction shows, cGMP is normally abundant in the dark and attaches to Na+ channels in the cell membrane, keeping them open. In the dark, Na+ ions enter the cell and create a depolarized resting potential. The action of light reduces the level of cGMP, which leads to closure of Na+ channels and less depolarization (which appears as hyperpolarization). The more photons that each photoreceptor catches, the lower its level of cGMP and the more Na+ channels that close. In bright light, when most or all of the Na+ channels are closed, the cell is essentially at its “real” resting potential: strongly hyperpolarized from its resting state in the dark.

Color Vision in the Retina: Three Kinds of Cones

Rhodopsin, the visual pigment that catches photons and begins the process of phototransduction, is found in rods (“rod opsin”), but our retinas have three other visual pigments found in three types of cones. These pigments differ slightly in their opsin protein but work in similar ways. The different opsins give the pigments different absorption spectra, making them effective in catching photons at different peak wavelengths. Absorption spectra of the four visual pigments in the human eye shows the peak absorbance for the 4 major visual pigments found in human photoreceptors. For example, rhodopsin in rods has its absorption peak at 498 nm, in the middle of the spectrum, while the three cone pigments have peaks in the short wavelengths (S or “blue cones,” peak at 420 nm), middle wavelengths (M or green cones, 534 nm), or long wavelengths (L or red cones, 564 nm). The middle and long wavelength pigments differ only slightly because they diverged only recently in evolution, giving humans and old-world monkeys three cone types. Most mammals such as dogs, cats and new-world monkeys have only two types of cones, one for short wavelengths (blue) and one for long wavelengths (yellow). Thus, the color vision of most mammals is different from ours, and your pet dog or cat does not see colors the same way you do.

Although each absorption spectrum indicates that a single pigment can catch photons in a range of wavelengths, one pigment cannot distinguish between a small number of photons at the pigment’s best wavelength and many photons at a wavelength that the pigment does not absorb well. Both stimuli could result in the same total number of photons being caught, and any photon that is caught initiates phototransduction in the same way. (One visual pigment cannot discriminate colors)

To distinguish light of two wavelengths, it is necessary to compare the responses of nearby cones with different pigments. For example, if the blue-absorbing cones in a patch of the retina catch many photons but nearby green or red cones catch only a few, then the light is at the blue end of the spectrum.

Since we have three types of cones, our color vision has three primary colors. Together, they can generate the appearance of any possible color. Looking at a TV or computer screen with a magnifying glass will reveal blue, green and red pixels (Primary colors). By controlling the brightness of each pixel, the screen simulates any real-world color. Red and green pixels together give us the sensation of yellow, and all three types together appear white. It’s as if each pixel type is activating one of the three cone types. In combination, the RGB pixels evoke responses in the cones that are equivalent to their response to any real-world color.

For lights and TV screens, where photons are added together to make the final mixture, red, green and blue are the primary colors. But surfaces covered in paints and inks subtract photons from the illuminating light, reflecting the non-absorbed photons. For example, a yellow paint patch in white light absorbs blue photons and reflects red and green, which appear to us as yellow. The three primary colors for subtractive surface colors are cyan, magenta and yellow (CMY). As explained in our example, yellow absorbs blue, reflecting green and red, while magenta absorbs green and reflects red and blue, and cyan absorbs red and reflects blue and green. For printing inks (Primary colors), the three primary inks together should appear black, but in practice the mixture often looks muddy, and a fourth ink, a true black, is usually added. Test strips of the four primary printing inks in different densities can often be found on printed materials, including the hidden flaps of cookie packages.

As shown in the bottom of Primary colors, color allows us to distinguish objects that have equiluminant surfaces (surfaces that would appear equally bright in a grayscale image) but reflect different wavelengths. The relatively recent evolution of separate red and green cones in primates (including humans) from a single yellow-absorbing cone further improves discrimination. Detecting red apples in the midst of green leaves would be difficult without our trichromatic color vision.

Color Blindness

Some people carry a genetic variation that gives them only two cone types instead of three, or they produce an anomalous version of one of the cone pigments. As a result, they make color matches that differ from those made by typical trichromats, who have three types of cones. They are not actually “blind” to a color, but rather can’t distinguish some colors that trichromats see as different. Color blindness affecting the red or green genes is the most common, occurring in about 8% of males and only 0.4% of females (Neitz & Neitz, 2011). The red and green genes are both on the X chromosome (further evidence that they were duplicated recently in evolution). Since XY biological males have only one X chromosome, they cannot compensate for a defective gene, unlike XX females who have two different X chromosomes, one from each parent. Blue blindness is very rare, since the blue gene is on a chromosome for which everyone has two copies.

Bipolar Cells

Take a moment to remind yourself of the cellular architecture in the retina, from Retina structure in the previous section. From the photoreceptors, the next synaptic partners in the pathway out of the eye are bipolar cells. This is where detecting borders begins. Bipolar cells gather information from the photoreceptors that synapse directly onto them and also from additional photoreceptors in a surrounding region, creating a circular region with a center and a surround that together are the bipolar cell’s receptive field (see Feature Box on the concept of receptive fields). This arrangement is diagrammed on the left side of Bipolar cell types. As mentioned, photoreceptors in the center of the receptive field synapse directly onto the underlying bipolar cell, while receptors that contribute to the surround connect to horizontal cells that modulate the synapses from central photoreceptors, thereby canceling the bipolar cell’s response to the center. (Photoreceptors make multiple synapses, so one receptor can contribute to the center of one bipolar cell’s receptive field but also help form the surround for other bipolar cells.) Contrast between the light falling on the center and the surround is required for a bipolar cell to respond. Uniform illumination is not effective.

We are actually not very good at perceiving brightness. In the gray rectangle in the top of Bipolar cell types, we don't see that the inner rectangle is a uniform gray, but we see its edges accurately. Similarly, we make poor judgments of absolute brightness (if you are now in a room lit by artificial light, is the room brighter than the light outdoors on a cloudy day? it’s hard to know the answer.). On the other hand, we are very good at detecting and remembering edges. A cartoon drawn with just a few lines can easily represent a face or an object for us.

The center-surround receptive fields of bipolar cells provide the first processing step that ignores absolute brightness in favor of detecting contrasting borders between light and dark. This arrangement is sensitive to contrast because bipolar cells respond in opposite ways to input from photoreceptors in the center of their receptive field compared to photoreceptors in the receptive field surround. To better understand how this works, we need to consider the two types of bipolar cells, on-center and off-center bipolar cells.

About half of bipolar cells are on-center, where the bipolar cell depolarizes to light covering the center of the receptive field if a darker region surrounds the center, or off-center, which is the reverse. In an on-center cell, the bipolar cell is directly affected by responses of the photoreceptors that synapse on it directly. The photoreceptors in the surround do not synapse directly on the bipolar cell. Instead, they synapse on horizontal cells, a type of interneuron that serves as a bridge between photoreceptors and bipolar cells. Activating horizontal cells cancels the bipolar cell’s response to the center stimulus.

Although the complete neural circuit is more than we will explore here, it still may surprise you that light in the center of the receptive field depolarizes on-center bipolar cells but has the opposite effect on off-center bipolar cells. How can there be two types of bipolar cells, both of which receive direct synapses from photoreceptors in the center of their receptive fields? The cells differ because they have different molecular receptors for the neurotransmitter glutamate that is released by rods and cones. Recall that neurons release transmitter only when they are depolarized (see Introduction), and also that rods and cones are depolarized in the dark. This means that in the dark, the depolarized rods and cones release glutamate continually onto the bipolar cells they connect to. For an off-center bipolar cell, a dark patch of a visual scene falling in the center of the receptive field will depolarize the receptors and lead to the release of glutamate. The receptor for glutamate in off-center bipolar cells depolarizes the cell, so if the photoreceptors in the center are in the dark, the off-center bipolar cell receives transmitter and is depolarized: an off response.

On-center bipolar cells have a different receptor for glutamate, one that leads to activation of a second messenger that closes depolarizing channels. Light on the center of the receptive field interrupts the release of transmitter from photoreceptors, which stops activation of the messenger that closes depolarizing channels in the bipolar cells. Those channels then open, and the bipolar cell depolarizes, producing an on-response. Thus, because they have different receptors for the neurotransmitter glutamate, on-center and off-center bipolar cells have different responses to light in the center of their receptive field.

While it’s helpful to understand the mechanisms for on and off responses, let’s return to the important task that bipolar cells accomplish: responding to contrast between light and dark regions of the image while ignoring uniform illumination. If we imagine the image background as a medium gray, the two types of bipolar cells would depolarize in response to either dark spots or bright spots covering the center of their receptive fields. Although uniform illumination is not a good stimulus, a border between light and dark falling on the receptive field works well if it puts light or dark on the center and the opposite illumination on at least part of the surround. The bipolar cells’ response to contrast is conveyed to the next stage of retinal processing, the ganglion cells. Since the distances are very short, action potentials are not necessary to transmit signals, and bipolar cells are “non-spiking,” making only graded changes in potential. The bipolar cells release varying amounts of their transmitter, which is also glutamate, depending on how depolarized they are.

Retinal Ganglion Cells

Like the bipolar cells that drive them, retinal ganglion cells have center-surround receptive fields and are either on-center or off-center, characteristics they inherit from the bipolar cells that excite them. On-center ganglion cells are driven by on-center bipolar cells, and off-center bipolar cells drive off-center ganglion cells. Because the ganglion cell axons transmit signals a significant distance to the brain, retinal ganglion cells do make action potentials, and they are the first cells in the retinal processing chain that do so. In Retinal ganglion cell receptive fields, an on-center ganglion cell on the left responds with a burst of action potentials to a bright spot on its receptive field center, and also to a dark ring on the surround. Both patterns represent contrast between the center and the surround. The ganglion cell also responds to reverse patterns, but they inhibit the cell’s spontaneous firing instead of exciting firing. A dark spot on the center or a bright ring on the surround inhibits an on-center ganglion cell. Since uniform light does not activate bipolar cells, it also does not activate ganglion cells.

The off-center ganglion cell on the right in Retinal ganglion cell receptive fields has the opposite responses. A dark spot on the center or a bright surround excites the cell, while the opposite patterns inhibit it. In all cases, a boundary between light and dark is necessary to generate a response in ganglion cells. It is that information about borders and edges, not brightness, that is carried to the brain in the optic nerve.

The optic nerve projects to several different destinations in the brain because it carries axons from several types of ganglion cells. The on- and off-center ganglion cells that are responsible for conscious visual perception provide more than 80% of the axons in the optic nerve. Those important ganglion cells can be further divided into two groups, relatively few large ganglion cells with large receptive fields and large axons that conduct signals rapidly (the “magnocellular” cells), and very abundant smaller cells with small receptive fields and medium diameter axons that conduct more slowly (the “parvocellular” ganglion cells). Magnocellular and parvocellular axons connect to neurons in different layers of the lateral geniculate nucleus of the thalamus, which we will discuss in The Thalamus and Primary Visual Cortex.

The difference in receptive field sizes of magnocellular and parvocellular ganglion cells gives them different roles in conscious visual perception. The abundant parvocellular cells with small receptive fields provide detailed, high-acuity vision, including details of the image that falls on the fovea. If you look straight ahead at a scene, the center will appear sharp, but if without moving your eyes you pay attention to objects near the edge of your visual field, you will discover that peripheral vision is quite blurry. If you wish to inspect peripheral details, you move your eyes to project those details on the fovea, where vision is sharp. Our brains create the illusion that the entire scene is sharply focused, but in fact, peripheral regions served by ganglion cells with large receptive fields are not sharp at all. However, magnocellular cells are very sensitive to motion. If you wiggle your fingers near the edge of your peripheral vision, you will easily detect their motion even though they are blurry. The large diameter axons and rapid conduction velocity of magnocellular ganglion cells helps convey information about motion quickly to the brain.

Two other types of retinal ganglion cells convey responses to light but do not contribute to conscious visual perception. One evolutionarily older class of ganglion cells, less than 10% of all ganglion cells, has wide receptive fields and narrow axons, and projects to the superior colliculus. These ganglion cells contribute to generating eye movements as we watch a scene. The highest visual area in the brains of fish and amphibians, the optic tectum, is the precursor of the superior colliculus in mammals, but it has been superseded by the evolutionarily new visual pathway for visual perception that leads to the thalamus and the visual cortex. Another category of specialized ganglion cells, about 1% of the total, are the intrinsically photosensitive ganglion cells (ipRGCs). These cells contain the pigment melanopsin, which makes them directly responsive to light. They do not contribute to visual perception (and are sometimes called “non image-forming” cells, NIFs), but most of them project to the suprachiasmatic nucleus and help control circadian rhythms (see Introduction). A few ipRGCs project instead to the brain nucleus that controls the pupillary reflex, which constricts the iris in response to bright light.

The retina is carpeted in these multiple types of ganglion cells, and the receptive fields overlap within and between types. Any visual pattern falling on the retina will activate many ganglion cells, with the specific responses depending on the portions of the image that fall within each ganglion cell’s receptive field. There are between one and two million retinal ganglion cells in each human retina. Receptive field centers can be as small as a single cone in the fovea, where the parvocellular ganglion cells are dense, or much larger in the periphery where the ganglion cells are spread out and have large receptive fields.

Color Vision in the Retina: Opponent Color RGCs

Many experiments to record electrical activity from retinal cells were first performed in fish and amphibians because their retinal cells are much larger than in mammals, making them easier to penetrate with sharp microelectrodes. The goldfish retina was also the site of the first measurements of absorption spectra for individual cones, revealing three types as in our retinas. Meanwhile, research on color vision in people mostly involved a controversy among psychologists about whether three primary colors explained color vision, or whether vision involved pairs of complementary colors. This controversy continued for many years because both sides turned out to be right. At the receptor level, three cone types account for the three primary colors, but retinal ganglion cells receive input from cones representing opposing pairs of color and thereby respond to pairs of complementary colors.

Early recordings from retinal ganglion cells in the goldfish retina showed excitation to one color but inhibition to its complementary color, establishing the concept of opponent colors. Soon after, recordings were made from lateral geniculate neurons in monkeys, which have receptive fields like ganglion cells, and this settled the debate about color vision in humans. Opponent color retinal ganglion cells firing in goldfish retina shows responses of a goldfish retinal ganglion cell to blue-green light, which excited bursts of action potentials, and to red light, which inhibited the cell’s spontaneous firing. Other ganglion cells were the reverse, excited by red and inhibited by green. Later, other ganglion cells were found that had opponent responses to blue and yellow. This corresponded to observations about human vision, where we can’t imagine a reddish-green or a yellowish-blue because those pairs are opponent colors.

Across many vertebrate retinas, including fish, color vision at the ganglion cell level was found to fall into the three categories shown in Opponent color retinal ganglion cells receptive fields.

One class of ganglion cells pools the responses of different cone types, responding to light or dark on their receptive fields but not to differences in color. These are the achromatic ganglion cells where a spot of white light in the center of an on-center receptive field excites the cell, while a white spot in the surround inhibits the cell and produces a brief burst of activity when the light goes off, an “off response.” The example in Opponent color retinal ganglion cells receptive fields is an on-center, off-surround ganglion cell like the neurons we discussed in the previous section.

Another class, red/green ganglion cells, are excited by a red spot in the receptive field center, while a green spot in the center inhibits the cell and produces an off response. The responses in the surround are reversed: red inhibits and green excites. These are opponent-color cells, and the strongest response for this example would be to a red stimulus in the center surrounded by a green ring. Other red/green cells would have the reverse organization, excited by a green spot in the receptive field center against a red surround. These cells would respond well to a border between red and green areas.

A third class, blue/yellow ganglion cells, is organized in a similar way, but now the best stimulus for the example shown would be a yellow spot against a blue background. (Recall that yellow excites both red and green cones.) The reverse type also exists, responding best to a blue spot against a yellow background. These are all examples of “double-opponent” cells, where the center and surround have opposite responses to color spots. In the human retina, the ganglion cells are red/green or blue/yellow, but the receptive fields are usually not divided into a center and surround. Instead they are “single opponent” cells, still excited by one color and inhibited by its complementary color, but across the entire receptive field.

The number of cones contributing to the response of a single retinal ganglion cell varies with the cell’s location in the retina. In the fovea, where we have the highest visual acuity, a ganglion cell’s receptive field center may be as small as a single red or green cone. In the retinal periphery, ganglion cell receptive fields are larger, driven by clusters of cones, and visual acuity is much lower.

You can experience opponent color organization in your own visual system by staring at Visual after-effects of opponent color retinal ganglion cells and then looking at a blank page or a white wall. An afterimage will appear with colors that are the complements of the colors in Visual after-effects of opponent color retinal ganglion cells. Black will appear white, cyan will appear red, and yellow areas will be blue. This is an indication of opponent color organization, which begins at the ganglion cell level. Afterimages occur when prolonged exposure to one color causes an opponent-color ganglion cell to adapt, becoming less sensitive to the exposed color while retaining full sensitivity to the complementary color. Flooding the receptive field with white light would ordinarily activate both opponent colors, but because one color system has adapted, the balance is shifted and the ganglion cell responds as if it is being exposed to the complementary color—which is what we see in the afterimage.

Visual Fields and the Visual Pathway

Visual pathways shows the first stages of the visual pathway, going from the eye through the thalamus to the primary visual cortex.

Each eye is turned to project the fixation point between the left and right visual fields onto the fovea in the center of the retina. Because lenses invert images, the left visual field (shown in green) is imaged on the right side of each retina, and the right visual field (purple/blue) is imaged on the left side. Ganglion cell axons from both sides of the retina bundle together to form the optic nerve, which leads to the optic chiasm in the midline. There, the nasal (inner) half of the axons from each retina cross over, while the temporal (outer) half remains on its side of origin. Consequently, the right side of the brain (green) receives axons from the right half of each retina, and the left brain (purple) receives axons from the left retinas. Because the left retina views the right visual field, each hemisphere of the brain receives information about the opposite side of the visual world from both eyes. Lesions in the visual pathway on one side of the brain lead to deficits in seeing the opposite side of the visual world.

Lateral Geniculate Nucleus

The main cerebral target for the retinal ganglion cell axons is the lateral geniculate nucleus (LGN), a structure in the thalamus on the path to the primary visual cortex. The LGN is a six-layered structure. The left side of Layers of neurons in the lateral geniculate nucleus shows the LGN in a post-mortem monkey brain. The dark purple dots are cell bodies and show how distinct these six layers are. Each of these layers is unique in the input it receives. Two layers are composed of large neurons, magnocellular layers 1 and 2, and four layers are composed of smaller neurons, the parvocellular layers 3-6. Magnocellular retinal ganglion cells, which convey low resolution information and respond well to motion, send their axons to the magnocellular LGN layers. Parvocellular ganglion cells with small receptive fields, red/green opponent color responses, and high resolution connect to the parvocellular layers. There are also small intermediate koniocellular LGN layers, which were overlooked in early studies. They are the target for axons from blue/yellow ganglion cells. The neurons in each LGN layer are arranged in a two-dimensional array that matches the ganglion cell locations in the retina. This preserves a map of receptive field positions, a retinotopic map.

Although ganglion cells from the right half of each eye project to the LGN on the right side of the brain, the ganglion cells from each eye connect to different layers of the LGN. Thus, information from the two eyes remains separate in the LGN, and neurons driven by both eyes will not appear in the visual pathway until the next stage, the primary visual cortex.

V1 Simple, Complex and “Hypercomplex” Neurons

LGN neurons that receive input from retinal ganglion cells send their axons to the primary visual cortex (V1), a large visual area in the posterior occipital lobe and midline. This area is also called “striate” cortex because cross sections appear striated, or striped, and it is also sometimes referred to as “area 17” in reference to labels for different cortical areas based on their appearance. V1 is the first stage of visual processing in the cortex. In a collaboration that began in 1958, David Hubel and Torsten Wiesel conducted an extensive research program to discover how V1 cells respond to visual stimuli (more about them and their experiments can be found later in this section). By recording from individual cortical neurons in experiments that often lasted for more than 24 hours, they showed that the cortex could be understood at the level of individual neurons, and that the neurons had receptive fields responsive to edges. This continued earlier stages in the perception of objects and scenes. Hubel and Wiesel’s experimental method was to advance an insulated needle electrode into the visual cortex to pick up extracellular action potentials from neurons along the electrode’s path (see Methods: Electrophysiology). They projected visual stimuli on a screen facing the anesthetized animal (initially cats and later monkeys), while they searched for the most effective visual stimulus. By listening to the neuron’s action potentials as they altered the projected stimulus, they mapped the neuron’s receptive field: the visual pattern that optimally excited the cell. Once they characterized a neuron’s receptive field, they advanced the electrode until it encountered another neuron, which they then mapped. In this way, they gathered data on the receptive fields of hundreds of V1 neurons in each experiment.

In addition to mapping receptive fields, they also kept data on the location of each neuron along the electrode’s track, so they could later reconstruct the positions of the mapped neurons on slices of the post-mortem brain. In this way, they not only were able to classify V1 neurons into functional categories, but they also revealed the functional architecture of the primary visual cortex.

The earliest experiments recording from individual neurons in V1 used stimuli that had been effective in exciting or inhibiting retinal ganglion and LGN cells: small light or dark spots. Many cortical cells could be stimulated and mapped with small spots. But unlike the earlier neurons in the visual pathway, where receptive fields were circular with opposing center and surround regions, the cortical receptive fields had elongated excitatory and inhibitory areas organized along straight-line edges. Hubel and Wiesel found that bright bars, dark bars, or an edge between light and dark were even more effective than spots as stimuli. For a vigorous response, the straight-line stimulus had to be at an appropriate angle of orientation aligned with the receptive field edge, and positioned exactly within the receptive field (V1 simple cortical cells). These first cortical cells with receptive fields that could be mapped with small spots they called “simple cells.”

Hubel and Wiesel also discovered a second group of cortical cells that could not be stimulated with small spots. They named these cells complex cells. At first, they were a puzzle to the experimenters. Hubel has told the story of how they accidentally discovered what was special about complex cells. After struggling for hours without finding an effective stimulus for the neuron they were recording from, they slid a glass slide out of the projector, and as the faint dark edge of the slide moved across the screen, the neuron suddenly fired vigorously. It soon became clear that complex cells would respond to edges but not to spots. Their receptive fields could not be divided into static excitatory and inhibitory areas. Instead, complex cells responded to an appropriately oriented straight-line edge at any position in the receptive field. If the angled light or dark bar swept across the receptive field, the cell would respond continuously. (V1 complex cells respond to orientation and direction) Many complex cells were also directionally selective, responding to an edge moving in one direction but not the reverse. This was different from simple cells, where an appropriately oriented edge would be effective at only one exact position; complex cells would respond to the stimulus at any position. This seemed to be a step beyond simple cells, hence the name “complex.”

Hubel and Wiesel recorded videos that showed how they mapped receptive fields of simple and complex cells. The videos show the screen in front of the animal on which stimuli are projected, and we hear the action potentials (as clicks) produced by the neuron they are mapping. Although the video quality is not the best, these historically important videos show how V1 neurons respond selectively to projected visual stimuli.

Other neuroscientists explored responses of V1 neurons to other visual stimuli in addition to the projected bright or dark bars and edges that Hubel and Wiesel used. One group of stimuli, which are generated on a computer screen, are sinusoidal luminance gratings. These look like soft-edged stripes, and are discussed in the Feature Box on spatial frequency selectivity. Theoretically, any image can be broken down into its component spatial frequencies, which makes these stimuli of particular interest. V1 neurons actually respond more vigorously to appropriately oriented spatial frequency gratings than they do to simple bars or edges, and a neuron’s preferred spatial frequency is now considered a significant characteristic of its receptive field.

Building Simple and Complex Receptive Fields

Hubel and Wiesel were interested in how information from LGN neurons with center-surround receptive fields could be transformed into the elongated receptive fields of cortical cells. They proposed that a simple cell in V1 receives excitatory synaptic input from a group of LGN neurons that have overlapping receptive fields located in a staggered straight line. Activating any one of the LGN cells with an appropriate light or dark spot would slightly excite the simple cell, but activating all of the LGN neurons with an oriented edge that covered all their receptive field centers would excite all of the LGN cells. This would elicit a vigorous response from the simple cell. Later experiments that recorded simultaneously from LGN and V1 neurons confirmed this scheme. Simple cell connections to LGN neurons shows the proposed cellular anatomy of these connections on the left, with multiple LGN neurons converging to excite a single V1 simple cell. The receptive field perspective on the right shows how a bar of light would provide the most effective stimulus for a V1 simple cell receptive field.

Hubel and Wiesel further proposed that complex cells were excited by the activity of a group of simple cells that had overlapping but similar receptive fields (V1 complex cells respond to simple cells). An edge at an appropriate angle and position would excite one or more of the simple cells, thereby exciting the complex cell. But the stimulus could be at any position in the receptive field and still excite the complex cell. A moving edge crossing the receptive field would lead to continuous excitation. Directional selectivity could be imposed by additional circuitry that enhanced the complex cell’s response to movement in one direction but suppressed the response to the opposite direction.

As Hubel and Wiesel continued their exploration of complex cells’ responses, they found an additional characteristic. Making an appropriate stimulus bar longer so it spilled beyond the receptive field had no effect on the responses of many complex cells, but it decreased the response of other complex cells (Endstopping). This effect, where a stimulus bar extending beyond the receptive field actually inhibits a cortical cell’s response, is referred to as endstopping. They named the new category “hypercomplex cells.” Later research suggested that endstopping was a variable characteristic of every complex cell, and most neuroscientists no longer regard hypercomplex cells as a separate category. Endstopping shows examples of complex cells with weak endstopping (top) and strong endstopping (bottom). Notice how both examples respond more and more to a stimulus as it elongates to fill the receptive field, but one stops responding when the stimulus extends beyond the receptive field. That complex cell shows endstopping in action.

The functional purpose of endstopping has been puzzling, but one possible explanation is that it would make complex cells respond selectively to curved edges. A long curve falling in the receptive field would excite an end-stopped neuron if the segment within the receptive field had an appropriate orientation angle, but the part of the curve falling outside the receptive field would have a different angle that would not trigger endstopping. In contrast, long straight lines would retain the same orientation angle outside the receptive field and would activate inhibitory endstopping.

Binocular Units

We noted earlier that information from the right side of each retina ends up on the right side of the brain, but axons from the two retinas synapse in separate layers of the lateral geniculate nucleus. This means that LGN neurons projecting to the primary visual cortex will be driven by one eye or the other but not both. In V1, however, binocular neurons are found that are driven by both eyes (V1 biocular units). Covering one eye and then the other while the animal views the stimulus screen shows whether the V1 neuron responds best to the left eye, the right eye, or often to both eyes together. If one or the other eye is more effective than the other alone, this is classified as “ocular dominance,” yet another characteristic of a V1 neuron’s receptive field.

It was also possible to demonstrate that these binocular neurons contribute to stereo vision. For some neurons, moving an appropriately angled edge in space in front of the animal generated a strong response only if the stimulus was at a particular distance. Moving it nearer or farther decreased the neuron’s response. Such neurons would report on the depth of edges in the visual scene. You can demonstrate stereoscopic depth perception in your own vision by moving a finger on your outstretched arm close to you and far away while closing one eye and then the other. You automatically perceive the distance of your finger as its relative position on your two retinas changes, affecting depth-sensitive neurons in your visual system. But automatic stereo vision operates only for near distances. For outdoor scenes, for example, there are other cues to depth such as occlusion (which objects are in front of others) and atmospheric haze, while relative positions of distant objects on our two retinas are no longer significant.

Functional Anatomy of V1

The orderly layout of the retina is preserved in the layers of the lateral geniculate nucleus and in the surface of the primary visual cortex. This is the retinotopic map, shown in the left side of V1 visuotopic organization>. The right visual world is represented in the left visual cortex, with an orderly arrangement of receptive fields from the central retina to the periphery. Because there are many more retinal ganglion cells and LGN cells serving the fovea and central retina compared to the periphery, a disproportionately large area of cortex is devoted to the center of the visual field.

The retinotopic map was established using several methods, including locating visual deficits (scotomas) for soldiers in World War II who had suffered head wounds that lodged shrapnel in the visual cortex. This established the overall layout of the human retinotopic map by presenting the average position in the visual world of the receptive fields of neurons from different cortical locations

Hubel and Wiesel added fine detail to the map by recording from individual neurons. A representation of their findings is shown on the right side of V1 visuotopic organization>. The receptive fields of nearby neurons overlap around their average position in visual space, but if the electrode is moved about 3 mm across the cortical surface, the receptive fields of neurons in the new location no longer overlap with receptive fields for the earlier location.

In addition to mapping hundreds of receptive fields in their experiments on V1, Hubel and Wiesel also recorded data on the location of each neuron along the electrode’s track. This allowed them to later reconstruct the positions of the mapped neurons on slices of the post-mortem brain, which helped reveal the functional anatomy of the primary visual cortex. They noticed that as their electrode was advanced vertically (radially) deeper into the cortex, receptive fields of neurons along the electrode’s track shared the same orientation preference. A nearby radial penetration encountered neurons that shared a different orientation preference, indicating that the cortex was organized in vertical orientation columns. They also saw broad regions that shared ocular dominance, where one eye drove a binocular neuron more strongly than the other eye. This changed from one eye to the other if the electrode was advanced horizontally (tangentially) across the cortex. They concluded that the cortex was organized in vertical columns of neurons that shared orientation and ocular dominance preferences. Later researchers used a different technique, optical recording, to reveal the two-dimensional organization of orientation and ocular dominance columns in V1. Optical recording detects differences in blood supply that reflect the level of activity of nearby neurons. In this way it resembles fMRI, but it provides finer detail and requires creating a window in the skull to permit imaging of the cortical surface. By imaging the cortex while systematically presenting a series of oriented stimuli delivered to one eye or the other, regions of maximal activity could be identified. Using computational techniques to give each orientation preference a false color, the orientation columns were revealed to be arranged like pinwheels, while the ocular dominance columns were long stripes (V1 orientation pinwheels and ocular dominance columns).

Another structural feature in the cortex had been identified earlier: cytochrome oxidase blobs in the upper layers of the cortex. Cytochrome oxidase is an enzyme associated with metabolic activity, and by using a substrate that leaves a colored reaction product, the cortex was shown to have a regular array of “blobs," small patches of neurons with high metabolic activity. The array of blobs across V1 seen in stained microscope sections became another architectural element to add to V1’s orientation and ocular dominance columns. Optical recording using achromatic (non-colored) striped gratings at various angles or gratings with alternating red and green stripes demonstrated that the cytochrome oxidase blobs coincide with regions that respond most strongly to color stimuli.

Putting all these features together created an overlapping map of orientation pinwheels, cytochrome oxidase blobs, and ocular dominance columns (shown in the bottom of V1 orientation pinwheels and ocular dominance columns). The cortex has a repeating modular structure. Within each module, the blobs and the pinwheel centers are both aligned along the center of the ocular dominance columns, but they do not have a specific relation to each other. The overall implication is that as the visual cortex is built in development, a repeating subunit structure is genetically specified to organize the cortex.

The final aspect of cortical architecture to consider is cortical layers. V1 is named “striate” (striped) cortex because of the prominent layering seen in cross sections of stained cortex. There are six anatomical layers, numbered from 1 at the cortical surface to 6 adjoining the white matter. The layers turn out to have functional differences (Layers of the visual cortex). Axons from the LGN arrive at the cortex in layer 4 and also in layer 6. Layer 4 is where simple cells are found (red dots), supporting the theory that simple cells receive direct synapses from LGN axons. Complex cells (blue) are found in Layers 2-3 and Layers 5 and 6, where they could receive connections from simple cells in the same column, again consistent with the theory that simple cells drive complex cells. Complex cells in the upper layers project to more advanced cortical areas, leading to the next stages of visual processing. Complex cells in layer 5 project to the superior colliculus, and cells in layer 6 project back to the LGN, modifying the LGN’s responses to signals from the retina.

Color Vision in the Cortex

Most neurons in V1 receive their input from the LGN’s parvocellular, color-selective layers, but surprisingly, relatively few V1 neurons respond selectively to colors. In monkey V1, one type of color-selective receptive field has a center-surround, double opponent receptive field. This is shown in the top of Many neurons in V1 respond selectively to colored stimuli. The example in the figure shows a neuron that is excited by red in the center and inhibited by the opponent color in the center (green). In the surround, the responses are reversed: green excites and red inhibits. A pattern where the two colors meet at an edge is the strongest stimulus.

Other primate V1 neurons include simple and complex cells that are selective for color stimuli. The simple cell in the middle of Many neurons in V1 respond selectively to colored stimuli is excited by flashing a properly aligned and positioned red bar on a green background (top), but reversing the colors elicits an off response after the stimulus ends (bottom). The complex cell in the middle of Many neurons in V1 respond selectively to colored stimuli is excited by a red-green border moving to the left, but not if the border moves to the right, and not if the colors are reversed.

An interesting complex cell (bottom of Many neurons in V1 respond selectively to colored stimuli) is strongly excited by an angled bar of light moving upward if the light is yellow, but not if the bar moves downward or if it is a color other than yellow. Note that white light does not excite this neuron, which means it would be overlooked in an experiment that did not test colored stimuli. These complex cells have all of the stimulus selectivities that we met earlier, such as orientation angle and movement direction, but in addition they are selective for color. Color selectivity in V1 is not well understood, and some aspects remain controversial. These and other V1 neurons provide the beginning of our visual perception of the form, color and location of objects, but V1 is just the first stage of visual processing. Processing continues in areas outside of the striate cortex: the extrastriate cortex.

Extrastriate Visual Areas Overview

The extrastriate cortex beyond V1 divides into two major pathways: a dorsal pathway concerned with stimulus position and movement, and a ventral pathway leading to the perception of objects, faces, bodies, and scenes. The two pathways were first identified through experiments with monkeys trained on tasks that required distinguishing the location of an object (“where”) or the appearance of an object (“what”). Lesions in the dorsal and ventral cortex selectively interfered with one task or the other. Later studies recording from neurons confirmed the two pathways. Anatomy and functional connections of extrastriate areas shows many of the extrastriate visual areas that have been discovered. The relative size of each cortical area is represented by the size of the labeled boxes, and the number of axons connecting areas is represented by the thickness of the lines between them. These connections go in both directions, with axons projecting forward from a lower to a higher area, and also feedback connections from higher areas to lower ones. In fact, except for the one-way optic nerve from the retina to the LGN, all other visual areas have both ascending and descending connections, and the descending (feedback) connections have been shown to modify the activity of neurons in lower areas. For example, V1 projects strongly to V2 and V4, but those areas also project back to V1.

Dorsal Stream

The dorsal stream visual pathway processing areas have received less attention than the ventral stream, but one area that has received interesting experimental attention is area MT, “middle temporal.” The MT cortex is a motion-selective area organized in columns where all the neurons in a column are selective for a single direction of movement. Experiments at Stanford University showed that monkeys trained to report the direction of movement of subtle regions of random dots moving amidst stationary dots could be led to see sub-threshold movements if a column in MT was electrically stimulated (Salzman et al., 1990). The stimulation caused the neurons in that column to fire, which led the monkey to perceive movement that did not actually exist. The experiment was a direct demonstration that activity in an extrastriate area can lead to a visual perception.

Ventral Stream

The ventral “what” pathway for conscious visual perception leads from V4 to the lower (“inferior”) surface of the temporal lobe, the inferotemporal cortex (IT). Receptive fields of neurons in IT are significantly different from the earlier visual areas. They are not organized in a retinotopic map, but instead are grouped by the similarity of their most effective stimuli, such as the shapes of objects or the components of faces. IT receptive fields are very large (from 25 to 70 degrees of visual angle, a substantial span of the visual field’s 180 degrees), they always include the fovea, and they span the midline to include portions of both the left and right visual fields. Effective stimuli for an IT neuron can include objects, faces, or hands, independent of the size or exact position of the stimulus.

The selectivity of IT neurons represents major steps toward conscious visual perception. Neurons in the first stage of cortical processing, the simple cells of V1, have small receptive fields (about 1 degree) that require exact positioning of an oriented edge. They will detect small components of an object’s outline, with thousands of V1 neurons responding to even a small object. Shifting the stimulus’s position or size in the retinal image will activate a completely different population of V1 neurons. In contrast, IT neurons are flexible in their response to the position and size of the retinal image, a necessary step in perceiving objects.

Extensive experiments to record from IT neurons began in the 1980s. Among the earliest findings were that posterior IT areas, the ones closest to extrastriate cortex, had optimal stimuli that were more elaborate than V4 neuron receptive fields. Actual objects such as toy animals were effective, although a tiger’s face, for example, could be reduced to a simpler image of a rectangle with attached circles. More anterior regions of the IT cortex had more complex receptive fields, with patches of neurons that responded selectively to faces, colors, and scenes.

If a visual area like IT has neurons with selective responses to complicated images, a methodological problem arises. How do the experimenters know if they have found the optimal stimulus? It is feasible to show a test array of hundreds of photos of different images, flashing each photo briefly while detecting if an IT neuron fires action potentials, but the most effective stimulus may not be in the test set. The best that can be accomplished is to identify a general category that the neurons respond to, and then narrow down the important details in the stimulus. This strategy has been employed to characterize neurons that respond to objects and faces.

Another approach, especially for intermediate extrastriate areas like V4, has been to use computer-generated patterns that are systematically and randomly varied. Variations that lead to more vigorous firing are retained, while less effective variations are discarded. Repeating the cycle and continuing to choose more effective variations eventually generates apparently optimal stimuli, but they are often complicated and do not resemble simple geometric shapes that are easily described in words. This remains a puzzling aspect of visual processing.

Face-Selective Units

In the 1980s, exploratory recordings made by Charles Gross and his colleagues at Princeton University discovered neurons in monkey IT that responded vigorously to images of faces. The faces could be of monkeys, humans, or even cartoons, but if the eyes were omitted or the image was scrambled, the response was reduced or eliminated (Face-selective units). Faces are an important element of communication for monkeys and humans. Faces support identification of familiar individuals, and they also reveal emotional expressions from unfamiliar strangers. In recording from randomly encountered neurons in electrode penetrations of IT, the Princeton group found neurons that were selective for a variety of images, not just faces.

Face patches have also been identified by fMRI in the human brain, a region called the fusiform face area (FFA). Research on IT face patches continues, with a goal of revealing how advanced visual processing occurs in sequential stages for a well-defined image category. These studies may help explain how a scene that activates millions of neurons in multiple cortical areas leads to our conscious perception of the components of the visual world around us.

The Binding Problem

Any visual image elicits simultaneous responses from millions of neurons in V1 and the extrastriate areas that respond to aspects of the stimulus. For example, a bluebird flying across the sky will activate color-selective neurons in V1, V2, and V4 corresponding to the bird’s color. The bird’s motion will activate movement-selective neurons in MT. Other objects in the visual scene will activate other neurons in visual cortical areas. If many visual neurons are firing simultaneously to multiple components of the image, how are the ones activated by the bluebird identified as belonging together? This is the binding problem, the need to link activity in different neurons that are stimulated by the same stimulus object. We assume that the activity of these neurons is somehow combined to generate our perception of that object, but how?

One suggested mechanism for binding responses is rhythmic firing, where the timing of synchronous action potentials identifies neurons that are responding to a particular stimulus. Where and how that synchrony would be detected, and where conscious perception is ultimately localized, remain unanswered questions, and synchrony as the answer to the binding problem remains controversial.

Top-Down Processing

Top-down processing is a famous picture, a high-contrast assortment of shapes and splotches that leads most people to perceive a spotted dog in leafy shade.

It seems inconceivable that seeing the dog could be accomplished just from the activity of V1 neurons responding to the components of the image that fall within their receptive fields. That would be a “bottom-up” approach, where elements of the image are combined in successive stages of cortical processing to generate the perception of the object. Instead, some prior knowledge of the shape of a dog seems necessary for organizing the elements of the scene. That would be a “top-down” contribution to visual perception.

Computer scientists working on automatic image interpretation have described top-down processing as “hallucinating” the scene and then calculating how well the actual elements match the “hallucinated” image. After revising the imagined scene in areas where the match is poor, the calculation would repeat until a good match is found. Our visual system may work in a somewhat similar way. Bottom-up processing detects edges and activates neurons in V1 and extrastriate cortex, but top-down processing organizes the edges that are detected in the lower visual areas.

Visual illusions further support the existence of top-down organization of the elements of a scene. The image on the left of Visual perception is either a white vase or profiles of two faces. We can instantly switch from one perception to the other, indicating that although the elements of the image falling on the retina that elicit bottom-up activity are unchanged, a top-down process determines how the components are grouped and perceived. Other visual illusions and ambiguous images support the same point.

These examples make it clear that top-down processes influence the activity of neurons in visual areas, and thus that visual perception cannot be an exclusively bottom-up process. Other experiments recording from individual neurons in lower visual areas also show that their firing rates are modulated by activity in higher visual areas. Extensive axonal feedback connections from higher to lower areas offer an anatomical pathway for top-down modulation. However, how top-down processes modulate lower visual areas to contribute to visual perception is not generally known. Progress has been made, but this remains a largely unsolved puzzle in visual perception.

Grandmother Cells

The first stages of visual processing suggest that activity converges from one stage to the next. Axons from multiple LGN neurons converge to create the receptive fields of individual simple cells. Activity from multiple simple cells excites single complex cells, and so on. Very large receptive fields in extrastriate and IT cortex suggest the convergence of signals from earlier neurons that have small receptive fields. How high could this convergence reach? Could there be advanced neurons with extremely selective responses that combine visual and possibly other information? This idea was proposed, not seriously, as the existence of “grandmother cells”: cells that would respond to any view of your grandmother, and by extension, other neurons that would respond uniquely to other specific people or objects.

Although the term was presented mockingly, recordings of hippocampal neurons from a person undergoing brain surgery for epilepsy did reveal a neuron that seemed to respond to images and also the written name of a well-known actress, Jennifer Aniston (Quiroga et al., 2005). This opened two questions: first, could recognition depend on single neurons responsible for identifying a particular stimulus, and second, could neurons represent not just a visual image, but also other information about the subject?

The first possibility, that single neurons are responsible for recognizing an image, seems unlikely. Any visual stimulus evokes activity in an ensemble of hundreds or thousands of neurons, even at high levels. Similar visual stimuli would activate overlapping but distinct ensembles. There is no indication that potential loss of a single neuron suddenly eliminates recognition of a particular object. Loss of neurons does occur from strokes, for example, and the loss does eliminate function. Prosopagnosia is a neurological deficit where the patient no longer recognizes faces, but it involves the loss of whole regions of neurons and problems with many faces.

The second idea, that ensembles of neurons might encode information associated with a visual stimulus, seems more likely. “Gnostic neurons” have been proposed, and the neuron that responded to Jennifer Aniston’s image and written name seems to be an example of a gnostic neuron. It’s clear that our brains somehow generate our unified conscious experience, but like other aspects of consciousness, the later stages of visual perception remain an unsolved and challenging puzzle.

Why is the sense of hearing important?

Like every other sensory system, the auditory system has evolved because it provides useful information about the objects and events in an animal’s environment. Animals that can accurately sense what is happening around them can respond in a way that increases their chances of survival and of successfully passing on their genes. Hearing is especially valuable to survival because sound can travel over long distances and around obstacles, allowing animals to make timely responses, for example by fleeing from a predator before it gets too close.

To further illustrate why hearing is such an important sense, consider the female bird pictured perching in the center of Acoustic signals. Sounds are coming from every direction. Some are not especially relevant, like the splash of a turtle falling into the water or the rippling of water flowing in the stream. Other sounds carry important information, like the rustle in the nearby bushes that could be coming from a predator, or the song of a nearby male advertising himself as a potential mate. Sounds that are produced unintentionally are called cues, whereas sounds produced intentionally, with the goal of communicating something, are called signals.

Some of the most common and important cues are the sounds animals make as they move. Predators and prey alike have evolved to move as quietly as possible to minimize the cues they give about their presence and location. It is almost impossible to avoid making any sounds, however. Sticks break, grasses rustle, and even simply moving through the air can create turbulence. As we will see, all of these events can generate acoustic waves that travel for tens of meters or even further. Just as there is selective pressure for animals to move quietly, many species have evolved extremely sophisticated and sensitive auditory systems. For example, the common house cat (Felis domesticus) can hear a mosquito flying in a quiet room from as far as 30 m away — no wonder my cat seems to appear from nowhere the moment I open a can of food!

In contrast with cues, acoustic signals have evolved to allow one individual, called a sender, to influence the behavior of another individual, the receiver. Birdsong, for example, is a signal that communicates to female songbirds that a potential mate is nearby. The male sings with the intent of influencing females to approach and mate. In many species, the song also informs the female of how desirable the male is, because only healthy males can produce songs with a certain level of complexity (Nowicki and Searcy 2004). Acoustic mating displays are common throughout the animal kingdom, including many insects, fishes, amphibians, birds, and mammals.

Acoustic signals can also be used to broadcast the identity of the individual producing the signal, as long as receivers can distinguish the unique features of each individual’s signal. Individual recognition is particularly important to animals that live in large social groups where visual cues are not easily distinguished, as in a flock of nesting seagulls or a herd of elephants (McComb et al., 2000) or where individuals need to find each other over long distances, as in dolphins and other marine mammals (Sayigh et al., 1999).

Animals use acoustic signals to communicate to each other about more than just themselves. When social animals are foraging in large groups, a few individuals may keep watch and signal to the group about sources of danger. Acoustic signals have a particular advantage here because they can travel long distances and around obstacles that would block sight. For example, vervet monkeys have different kinds of alarm calls that indicate the presence of specific kinds of predators (Seyfarth and Cheney 1990). When one monkey sees a leopard, it makes the “leopard” call, and the other members of its troupe will climb into trees, even without themselves seeing the leopard. In contrast, when a monkey makes the “eagle” call, the troupe will hide in nearby bushes that are too thick for an aerial predator to enter. The monkeys’ auditory system must be able to discriminate between these signals for them to respond appropriately.

Humans have evolved a much more elaborate system of acoustic communication called speech, in which words are formed from combinations of distinct vocal elements called phonemes. For example, the vowels in the English words “bat” and “pat” are the same phoneme (/a/), whereas the initial consonants (/b/ and /p/) are different. One of the most important tasks of the human auditory system is to identify and discriminate between similar-sounding phonemes, and as we will see in How Does the Brain Process Acoustic Information?, this ability is shaped by the sounds an infant hears in the first year of life.

Some species produce acoustic signals for themselves to hear, with the purpose of detecting obstacles and prey in their environment. Because sound waves reflect from solids and liquids, they can be used for echolocation. The ability to echolocate has evolved separately in bats and in some aquatic mammals. Echolocating animals produce short vocalizations that reflect off nearby objects. Specialized neural circuits use the delay between the signal and its echo to determine distances to those objects, while other circuits compute the direction of the object (see How Does the Brain Process Acoustic Information?) and even some of its physical characteristics (Moss and Sinha 2003). A bat hunting by echolocation can easily discriminate between a solid obstacle and a delicious insect, and it can rapidly alter its course of flight to intercept the insect while avoiding the obstacles (Surlykke et al., 2009).

In summary, the sense of hearing serves many important functions for humans and other animals. Acoustic cues and signals are a critical source of information about predators, prey, obstacles, and dangers in the nearby environment, as well as the presence and intentions of conspecifics (other animals of the same species). In the next section, we will see how the physical nature of sound allows it to transmit this information in a way that complements the other senses.

How is sound produced?

What is sound? From the perspective of physics, sound consists of pressure waves moving through air or some other physical material. Waves are found throughout nature and share a common mathematical description. At the same time, acoustic waves have special properties that make them distinct from other kinds of waves.

To illustrate what a wave is, imagine throwing a rock into a still pool of water. When it hits the surface, the rock will displace some of the water, causing it to rise in a circle around the point of impact. Gravity pulls the displaced water back down, which in turn pushes up the water a little further out. This process repeats again and again to produce what we observe: a ripple traveling outward from the point of impact. An acoustic wave moves in much the same way, except that instead of a two-dimensional surface moving up and down, the molecules of the air move closer together and further apart, and the wave spreads in three dimensions.

Now instead of throwing a rock into a pool, imagine clapping your hands together. As your hands meet, they push air out of the space between them. Where does this air go? There is air all around your hands, but air is mostly empty space, so there is plenty of room for the displaced molecules of air to crowd in with the molecules that are already there. Now there is a region around your hands where the air is at a higher pressure, with an increased density of air molecules. The compressed molecules in the region of higher pressure push out into the surrounding space, which is less dense, or rarefied. This creates a new region of higher pressure, which in turn spreads further out from your hands. Just as the water ripple radiates out in a circle from a central point, the pressure wave created by your hands clapping radiates out in a sphere. If a microphone or some other pressure sensor is placed along the path of the wave, it will measure successive increases and decreases in the pressure at that location as the wave travels through space (Physical properties of acoustic waves).

It is important to notice that the individual air molecules do not move nearly as fast or as far as the acoustic wave. The air itself is not flowing, as in a wind. Instead, neighboring molecules are bumping into each other, and an individual molecule pushed out from a region of higher pressure may only travel a few nanometers before it collides with another molecule. Sound waves move through air at around 343 m/s, but this varies depending on temperature and elevation.

Acoustic waves can be described in terms of three principal quantities: amplitude, frequency, and phase. Amplitude (also called intensity or level) is a measure of how much the air pressure changes between compression and rarefaction. Clapping your hands more forcefully causes a larger increase in pressure and a wave with a higher amplitude. Amplitude is measured in units of pressure (in the metric system, Pascals); however, because the perception of acoustic amplitude follows a logarithmic scale, it is more often reported in units of decibels of sound pressure level (dB SPL). As shown in this figure, each step of 20 on the decibel scale corresponds to a proportional, ten-fold increase in the pressure.

The frequency of a wave is a measure of how rapidly the air at one location changes from compressed to rarefied and back (Physical properties of acoustic waves). Frequency is measured in Hertz (Hz), the number of cycles that occur each second. The frequency of a sound wave is perceived as pitch. Low-frequency waves sound deep and rich, whereas high-frequency waves are perceived as sharp and thin.

The phase of a wave is a measure of time relative to the cycles of compression and rarefaction. Phase is not perceived directly, but it affects how different waves interact with each other.

Almost any process that displaces air molecules will result in an acoustic wave. One of the most common sources of sound is periodic vibration. You can generate periodic motion by plucking or striking a string that is under tension (for example, on a guitar, piano, or harp). Pluck it right in the middle, and listen to the sound while watching the string. The vibration will probably be too fast for your eyes to track, so the string will appear wide in the middle and narrow at the ends, as illustrated in Physical properties of acoustic waves. As it vibrates, the string compresses the air while moving in one direction and rarefies it while moving in the other. The vibrations tend to occur at a specific frequency that depends on the mass of the string, its length, and how tightly it is tensioned. The motion of the string and the sound wave it produces are described by a simple mathematical function called a sinusoid (or sine wave) with a single frequency and amplitude. The form of this function and its relationship to the physical properties of the string will be covered in most elementary physics texts.

Now press the string down or pinch it at its midpoint before plucking it. The half that you pluck will vibrate twice as fast as the whole string did, producing a sound wave with twice the frequency. You might notice that the higher note sounds similar even though the frequency is higher. If you pinch the string three-quarters of the way and pluck the short part, the frequency will be four times as high, but the note will still sound similar. This is an illustration of how the perception of frequency is logarithmic, as we’ll discuss more later.

What happens if you pluck the whole string, but nearer to one end? You might notice that the note sounds richer and more complex, and that the movement of the string looks more complicated, without one clear wide part in the middle. Just as white light is composed of electromagnetic waves with frequencies spanning from red to violet, sounds can be composed of multiple waves with different frequencies. When the guitar string is plucked nearer the end, it will vibrate at the frequency corresponding to the full length of the string but also at twice that frequency (corresponding to half the length), three times that frequency (corresponding to one third of the length), and so on. A series of integer multiples like this is called a harmonic series. The lower-frequency and higher-frequency sound waves add together to produce complex harmonic motion. The specific combination of frequencies in a complex sound is called its spectrum. Most simple and complex periodic sounds are perceived as tones with a defined pitch that corresponds to the lowest frequency in the harmonic series.

Take a moment to think about this: you hear the string on the guitar as a single note, yet it is composed of several waves with different frequencies. As we will see later, each of these waves activates a different set of receptors in the inner ear and a different set of neurons that transmit information into the brain. In order for us to understand how audition works, we will need to understand how these different channels of information are combined by the brain to produce a single unified percept or split apart to distinguish between different sources, as when a person is singing along to the guitar.

Many species of animals generate sounds through periodic motion. The chirp of a cricket is produced by rubbing two textured surfaces on their wings together. The vocal folds in the mammalian larynx and the avian syrinx vibrate periodically as air is exhaled, and the frequency of the vibrations can be controlled by muscles that increase or relax the tension. In human speech (and even more so in singing), the vowels are produced by periodic movement of the vocal folds.

Sounds that do not have a regularly repeating pattern are not perceived as having pitch and are called aperiodic. The main sources of aperiodic sounds are transient collisions between surfaces and continuous, turbulent air flow. In turbulent air flow, the molecules are not all moving at the same speed and direction. This causes eddies and ripples to form as faster-moving air collides with slower-moving air, and the pressure variations in the ripples propagate outward as acoustic waves. Turbulent flow tends to occur whenever air is forced through narrow openings or disorderly obstacles like trees. You can make turbulent air flow by lightly pursing your lips and exhaling forcefully (but not forcefully enough to make a whistle). Aperiodic sounds do not contain a single frequency or even a regular combination of frequencies as in complex harmonic sounds. Instead, the pressure fluctuates at random across a broad range of frequencies. This is perceived as noise. A continuous sound that has equal amplitude across all frequencies is called white noise. Aperiodic sounds are also commonly used in acoustic communication. In human speech, many of the consonants are produced by aperiodic sounds.

How does sound travel and interact with objects?

Sound’s ability to travel through the air is what allows the auditory system to sense cues and signals coming from a distance. In this respect, the sense of hearing shares certain similarities with vision, but acoustic waves are much slower and have longer wavelengths than visible light. Humans can hear sounds with frequencies between approximately 20 and 20,000 Hz. In air, these frequencies correspond to wavelengths between 17 m and 17 mm. In contrast, visible light has wavelengths between 400 and 780 nm. In other words, whereas light waves are much, much smaller than most behaviorally relevant objects, sound waves are around the same size. As a consequence, acoustic waves exhibit a number of phenomena that are important to understanding how sound works.

First, sound waves can diffract around common objects. You can illustrate diffraction by placing a stone about 1 cm in diameter in the center of a shallow bowl of water. Tap the water near the edge of the bowl and watch how the ripples seem to pass right through the stone. This is diffraction. Waves can diffract around objects that are around the same size or smaller than the wavelength, which means that lower frequencies (longer wavelengths) are less likely to be blocked. Thus, a predator may be impossible to see when it hides behind a rock or a tree, but most of the noises it makes will simply pass around the object.

Second, solid objects are more likely to transmit sound waves than visible light waves. Low-frequency sounds in particular are more easily transmitted through solids, which is why you can hear construction noises through solid walls and the bass line from your downstairs neighbor’s music. At higher frequencies, solids tend to reflect sound. Reflections cause an acoustical phenomenon called reverberation, in which echoes of a sound reach the same point at slightly different times. For example, if you are in a large room with a concrete floor and metal walls, and someone a few meters away from you claps, you will hear not only the wave coming directly from the source to your ears, but all the other waves that reflect off the floor and the walls. The reflected waves have longer to travel, and because the speed of sound is around 343 m/s, there can be a perceptible delay between the time when the direct wave and its reflections arrive. Reverberations with long delays are heard as distinct echoes. Reverberations with short delays tend to fuse with the main sound, making it sound more “alive”. Surfaces that are soft or uneven are better able to absorb sound, reducing the intensity of the reverberation, and rooms that have little reverberation sound muffled or “dead”. With practice, people can learn to use the acoustical characteristics of different materials as cues to their location within a room, an ability akin to the specialized echolocation of bats and dolphins.

The long wavelengths and reflectiveness of sound cause another phenomenon called resonance. If sound waves enter a space where they can reflect back and forth between two surfaces, they will interact with each other through interference. If the sound has a wavelength the same as the distance between the two surfaces (or an integral fraction of the distance), the peaks and valleys of the wave going in one direction will overlap with the wave going in the other direction, leading to constructive interference. Wavelengths that are not an integral fraction of the difference will not overlap, leading to destructive interference. For complex sounds comprising waves of multiple frequencies, resonance boosts certain frequencies (wavelengths) while attenuating others. This effect is also known as filtering. When you blow across the mouth of a bottle, the aperiodic noise produced by turbulence resonates in the bottle, boosting a single frequency (and its harmonics) to produce a tone.

Filtering is important to the production of speech. Recall that vowels are produced by periodic motion of the vocal folds. But how do we make different vowels? You can illustrate this by singing “aah” and then changing to “ooh” while holding the same pitch. If you pay close attention to your tongue and lips, you’ll notice that the shape of your mouth changes as you transition between different vowels. What’s happening as you move those muscles is that the sizes of the oral and pharyngeal cavities in your airway are changing to create different resonances that boost and suppress specific frequencies in the sound coming from your larynx.

A consequence of these phenomena is that it is common for a listener to hear many different sound sources simultaneously. This is an advantage in that it is possible to hear sounds without having to look at them, to take in an entire auditory scene in full 360-degree surround, and to hear multiple instruments or singers making music together. It also makes a difficult task for the auditory system, which has to separate out multiple sound sources and determine their locations. As we shall see, the structure of the ear performs a part of this task, separating complex sounds into their component frequencies, but the central nervous system still has an enormous amount of work to do to transform auditory scenes into coherent perception.

External ear

When someone is asked to point to their ears, they will usually indicate their auricles, structures of skin and cartilage on either side of their head. The shape of the auricles (also called pinna), differs greatly among species. Some species, like cats, can manipulate the shape of their pinna using muscles on the skull, while many other species, including most birds, amphibians, and reptiles, have no pinna at all.

The auricle serves two functions. The first is to funnel acoustic waves into the external auditory meatus (ear canal). This amplifies the waves by compressing them into a smaller opening, increasing the sensitivity of hearing by up to 10 dB. The second function is to aid in localizing sound sources. The ridges and folds in the auricle create resonant cavities that filter out specific frequencies that depend on the direction from which the sound is coming. As a simple experiment, filling the cavities of the auricles with modeling clay makes it considerably more difficult to tell where a sound is coming from with your eyes closed (Oldfield and Parker 1984).

Middle ear

Separating the external auditory meatus from the middle ear cavity is the tympanic membrane, a thin sheet of tissue. Like the tightly stretched surface of a drum, the tympanic membrane vibrates in response to acoustic waves in the ear canal. As compressed air reaches the membrane, it pushes it inward, because the pressure in the ear canal is slightly higher than the pressure in the middle ear. Similarly, when the air in the ear canal is at a lower pressure (rarefied), the tympanic membrane is pushed outward by the air in the middle ear.

In order for the tympanic membrane to vibrate effectively, the static pressure in the ear canal and the middle ear cavity need to be equal. The external atmospheric pressure depends on altitude, and a rapid change of altitude, as in a descending or ascending airplane, creates an imbalance that reduces hearing sensitivity, particularly for high frequencies. The Eustachian tube, which connects the middle ear to the oral cavity, allows the balance to be restored. Sometimes it is necessary to encourage the Eustachian tube to open by “popping your ears”. A bad cold or sinus infection can cause the Eustachian tube to close. This results in a persistent imbalance that can be acutely painful and even cause damage to the tympanic membrane. Infections of the middle ear, called otitis media, can cause the middle ear cavity to fill with fluid, causing a pressure imbalance if the fluid is unable to drain through the Eustachian tube.

In mammals, the middle ear cavity contains three bones: the malleus (hammer), the incus (anvil) and the stapes (stirrup). These bones are the smallest in the human body. Their function is to transmit the vibrations of the tympanic membrane to the inner ear, specifically the oval window of the cochlea. The middle ear bones form a mechanical lever that compensates for the large difference in the impedance of the air in the ear canal and the fluid inside the cochlea. As illustrated in Acoustic impedance, when air is on both sides of a barrier like the tympanic membrane, acoustic energy is efficiently transferred from one side to another because the molecules have the same size and density. They collide like billiard balls: one ball comes to a stop while the other absorbs its momentum and moves on with the same speed. However, when an acoustic wave in air collides with a fluid, the water molecules are densely packed, and much of the energy is reflected, like ping-pong balls bouncing off a bowling ball. The ear compensates for this impedance mismatch through mechanical advantage. The tympanic membrane has a surface area around ten times the size of the oval window. Its vibrations are captured by the stiff middle ear bones, which add further mechanical advantage by acting as levers. Without the middle ear, only a thousandth of the energy in airborne acoustic waves would make its way into the fluid filling the cochlea (Wever and Lawrence 1948).

The impedance-matching of the middle ear can be compromised by buildup of scar tissue from repeated middle ear infections or by immobilization of the ligaments supporting the middle ear bones. This condition can result in a severe loss of hearing.

The middle ear also contains two muscles that attach to the middle ear bones, the tensor tympani and the stapedius (Acoustic reflex). These muscles adjust the tension of the tympanic membrane and the stiffness of the connection between the incus and stapes, respectively. These muscles allow the brain to control the gain of the middle ear. The stapedius reflexively contracts in response to high-intensity sounds and during self-vocalizations, causing a reduction in transmission through the inner ear by up to 25 dB (Pang and Peake 1986; Rosowski 1991). This reflex is an important protection against the damage that intense sounds can do to the inner ear, but like every reflex, there is some delay. Very sudden sounds like gunshots can therefore do significant harm.

Inner ear

The inner ear is a hollow bony structure that contains epithelial tissue and fluid. It comprises the cochlea, which is involved in hearing, along with the semicircular canals, the utricle, and the saccule, which are part of the vestibular system. The cochlea gets its name from the Greek for “snail” because it is shaped like a spiral (in mammals).

Cochlea

The interior of the cochlea (Inner ear and cochlea) has a circular cross-section, divided into three fluid-filled compartments that run the length of the cochlear duct: the scala vestibuli, the scala media, and the scala tympani. The scala vestibuli begins at the base of the cochlea at the oval window, a membrane-covered opening that contacts the foot of the stapes. It connects to the scala tympani at the apex, the tip of the innermost spiral of the cochlea. The scala tympani terminates at the base of the cochlea in the round window, another membrane-covered opening. The fluid found in the scala vestibuli and scala tympani is called the perilymph. The scala media is a separate compartment filled with a different fluid called endolymph.

Basilar membrane

Acoustic waves enter the cochlea through the oval window via the stapes. As the stapes vibrates in and out, it pushes and pulls on the perilymph in the scala vestibuli and scala tympani, which in turn displaces the membrane covering the round window. Some of the energy of this wave, however, is absorbed by the basilar membrane, which stretches between the two walls of the cochlear duct and is able to move up and down. The basilar membrane varies systematically in thickness and width over its length. It is thickest and narrowest at the base and thinnest and widest at the apex (Unrolled cochlea). These physical properties translate into tuning for different frequencies, a phenomenon called resonance. Near the base, the basilar membrane resonates at high frequencies, around 20,000 Hz. Vibrations in the acoustic wave at these frequencies are absorbed into vertical displacements of the basilar membrane, while the lower-frequency components continue to travel down the scala vestibuli. The resonance of the basilar membrane becomes progressively lower, to about 20 Hz at the apex, so as the acoustic wave travels toward the apex, each segment of the basilar membrane absorbs energy at its resonant frequency (von Békésy G 1960; Rhode 1971).

Take a moment to notice in Unrolled cochlea how the frequencies are spaced along the basilar membrane. About half of the length is dedicated to the frequencies between 25 and 1600 Hz; the other half goes from 1600 to 20,000 Hz, a much larger range. In other words, the relationship between frequency and position is not linear but (approximately) logarithmic. This organization may be related to how frequencies in a harmonic series are spaced logarithmically (i.e., integer multiples of the fundamental frequency). In humans, it may also reflect the fact that most of the information in speech is carried by relatively low frequencies (see How Does the Brain Process Acoustic Information?).

Although the physical principles are different, the way the basilar membrane separates complex sounds into waves of different frequencies is analogous to the way a prism can separate white light into its component frequencies (or colors). It is an essential first step in how the brain analyzes complex sounds. The organization of frequency (a non-spatial property of the sound) along the axis of the basilar membrane (a spatial dimension) is an example of topography. Because the ear contains a map of frequency, this topography is also known as tonotopy.

On top of the basilar membrane sits the organ of Corti, the epithelial membrane that transduces mechanical movements of the basilar membrane into neural impulses. The organ of Corti comprises four rows of hair cells along with a large number of various kinds of support cells. Hair cells are characterized by stereocilia arranged in a stairstep bundle that protrudes into the endolymph filling the scala media (SEM of cilia; . The organ of Corti is attached at one side to the tectorial membrane, forming a hinge. The tectorial membrane is a gelatinous structure that floats just above the tips of the hair cells’ stereocilia.

There are two types of hair cells in the organ of Corti. The inner hair cells form a single row nearest the hinge with the tectorial membrane. They are responsible for sensing the movement of the basilar membrane and transducing those movements into neural impulses. The other three rows are the outer hair cells. Their function is not to sense movement, but to help amplify low-intensity sounds.

Inner hair cells

The inner hair cells are in many ways the most important cells in the auditory system, because they are the sole point at which the physical energy of acoustic waves is converted into the electrochemical signals used by the nervous system. On one end, called the apical end, are the stereocilia, whereas the basal end makes a synapse with an afferent terminal from a single bipolar neuron in the spiral ganglion. There are about 3500 inner hair cells in the human cochlea.

As the basilar membrane vibrates up and down, the tectorial membrane slides back and forth in the transverse direction. The fluid between the organ of Corti and the tectorial membrane tends to remain stationary because of inertia, causing the stereocilia of the inner hair cells to deflect toward and away from the hinge. The tips of the inner hair cell stereocilia contain ion channels that are physically attached to the next-tallest cilium by a tip link. As illustrated in Tip link channel transduction, deflection of the stereocilia toward the taller edge of the bundle puts tension on the tip links, opening the ion channels and permitting potassium and calcium ions to enter the cell from the endolymph. This depolarizes the cell, causing voltage-gated calcium channels at the basal end of the hair cell to open, which in turn causes synaptic vesicles to fuse with the plasma membrane and release neurotransmitter into the synapse. Deflection of the stereocilia in the other direction relieves tension on the tip link, closing the ion channels, hyperpolarizing the cell, and stopping neurotransmitter release (Hudspeth 1985; Fettiplace 2017).

The transduction process in inner hair cells is exceptionally sensitive and fast. Sounds at the threshold of hearing produce basilar membrane movements of about 0.1 nm, about the diameter of a hydrogen atom, and movements of the stereocilia may be even smaller. The membrane potential of the hair cell responds to stereocilia deflection in as little as 10 µs, and a specialized ribbon synapse (Tip link channel transduction) allows neurotransmitter to be rapidly released onto the afferent fiber. The speed and fidelity of neurotransmission allows auditory nerve fibers to track frequencies as high as 1000 Hz or more, firing only at a specific phase of a periodic sound (Dynes and Delgutte 1992). Phase-locking is critical to one of the mechanisms the brain uses to determine the direction of a sound.

Ascending auditory pathways

After the inner hair cells of the organ of Corti have transduced the mechanical deflection of their hair bundles into the release of neurotransmitter, the auditory stimulus propagates as neural impulses through a number of brain regions and fiber tracts on its way to the cerebral cortex. This ascending pathway is more complex than the pathways for other sensory systems, so there are more names and connections to learn. One reason for this complexity may be that information about the location of sound sources, arguably one of the most important tasks for the auditory system’s function in predator avoidance, needs to be extracted from the stimulus early on.

This section will focus on the anatomy of the classical ascending auditory pathway, the most prominent and best studied route by which auditory information enters the brain (Central auditory pathways). Although we may briefly note the function of some brain regions and pathways, a full discussion of the physiology and neural computations will be deferred to How Does the Brain Process Acoustic Information?.

Auditory nerve

The inner hair cells form synapses with the dendrites of auditory nerve cells, which are bipolar neurons with their cell bodies in the spiral ganglion, a thin strip of gray matter within the spiraling interior wall of the cochlea. The axons of these cells course together in the center of the cochlea to form the auditory branch of cranial nerve VIII. The auditory nerve also contains many efferent axons descending from the brain to the cochlea; these primarily make synapses with the outer hair cells.

Cochlear nucleus

All of the afferent axons in the auditory nerve terminate in the cochlear nucleus, located right at the junction between the medulla and the pons. The cochlear nucleus has three subdivisions: dorsal, posterior ventral, and anterior ventral. Each auditory axon branches twice to make synapses in all three subdivisions. The tonotopic organization of the cochlea is preserved in these projections, resulting in three complete maps of frequency within the cochlear nucleus, one per subdivision (Tonotopic oragnization auditory cortex). Almost all of the stations in the classical auditory pathway have a complete tonotopic map.

The neurons within the cochlear nucleus give rise to three major output pathways that cross the midline and then ascend up the brainstem through a bundle of axons called the lateral lemniscus.

Location

One of the most important functions of the auditory system is to locate nearby animals. If the other animal is a predator, knowing its location may enable it to be escaped. If the other animal is prey, knowing its location may allow it to be captured. In behavioral tests, humans can reliably pinpoint the location of sound sources to within several degrees (Oldfield and Parker 1984). The acuity of sound localization is highest in front of the head along the horizontal axis. In our own experience, the reflexive turning of the head to look at an unexpected sound is so natural and automatic, it may be difficult to appreciate how difficult the underlying calculations are. Yet evolution has solved this problem with incredible precision.

Interaural timing and level differences

The fact that animals have two ears is not simply a consequence of bilateral symmetry. Depending on the location of a sound source, the two ears will receive slightly different inputs. By comparing the inputs between the ears, the brain is able to determine the location of the source.

When we discuss auditory spatial cues, we are referring to location relative to the head (Locating sources of sound). More specifically, what the brain needs to determine is the angle of the source relative to some reference direction, which by convention is usually taken to be the front of the head. The angle of the source has two components: the azimuth, which is the angle to the left or right on a horizontal plane, and the elevation, which is the angle up or down from that plane. Because the ears are separated from each other by the head along the plane of the azimuth, the azimuth of a sound source will affect the relative level and timing at which the sound reaches each ear.

Level differences between the ears arise because the head casts an acoustic shadow. Incoming sound waves tend to reflect off the head or to be absorbed by it, so the ear that is closer to the sound source will receive pressure waves with a higher amplitude. This effect is more pronounced for higher frequencies, because low-frequency sounds can more easily transmit through the head and diffract around it. The interaural level difference is greatest when the sound source is 90 degrees to the left or the right, and zero when it is directly in front of (0 degrees) or behind (180 degrees) the head.

Timing differences between the ears arise because the speed of sound is finite (343 m/s at sea level). At 90 degrees azimuth, the distance between the ears in humans is around 0.2 m, corresponding to a delay of about 600 µs. By comparing the timing at which sound arrives at the two ears, it is possible to determine its azimuth. For transient sounds, this comparison involves the difference between the times when the pressure wave’s onset arrives at either ear. This difference is called the interaural time delay. Interaural time differences are more relevant cues for the location of low-frequency sounds, whereas interaural level differences are more important at high frequencies.

The elevation of a sound source does not produce interaural differences in level or timing. Nevertheless, humans and many other species can determine the elevation of sound sources thanks to the irregular shape of the auricles. The cavities and ridges of the auricle create resonances that selectively amplify some frequencies while dampening others, creating “notches” in the sound spectrum. These resonances depend on both the elevation and azimuth of the sound source (Oldfield and Parker 1984). Thus, by determining where there are notches in the sound spectrum, it is possible for the brain to determine elevation.

The cues for distance are subtle, and much less is known about how this percept is formed (Kolarik et al., 2016). Without any obstacles, sound attenuates with the square of the distance, so there is no difference in intensity between a source two meters away from one that is twice the distance but four times the amplitude. However, sounds coming from greater distances have the opportunity to interact with more objects, resulting in greater reverberance. It is likely that the perception of distance relies on such cues.

Neural circuits involved in decoding location

The superior olivary complex is the primary brain structure involved in determining azimuthal location. The medial superior olive compares timing between the two ears, whereas the lateral superior olive compares levels between the ears (Neural circuits for decoding sound location).

One of the earliest models for how the brain might calculate timing differences is the coincidence detector. As proposed by Jeffress (1948), if a neuron receives input from both ears, neither of which alone is enough to make it spike, then it will only fire when the two inputs are active coincidently (i.e., at the same time) (see Introduction). If the pathways from the ipsilateral and contralateral ear are the same length, then coincidence will only occur when the sound arrives at both ears simultaneously, from directly in front (0 degrees) or behind (180 degrees). If the ipsilateral pathway is slightly shorter, then the neuron will respond best when the input arrives at the contralateral ear slightly before the ipsilateral one. By systematically varying the relative lengths of the axons, an array of such coincidence detectors would be able to represent timing differences as a place code, much as different frequencies are represented in a tonotopic map. In birds, the medial superior olive appears to implement just such a detector (A model of coincidence detection for sound localization). It receives inputs from the contralateral and ipsilateral cochlear nucleus, and the incoming axons, in many species, branch to form an orderly array, with the shortest paths from the ipsilateral side at one end and the longest paths at the other. In further support of this model, electrophysiological recordings from the medial superior olive reveal that neurons in this area respond preferentially to specific interaural time differences (Joris et al., 1998). The calculation of interaural time differences in mammals uses a similar principle, but is implemented differently (Lesica et al., 2010).

Differences in level between the two ears are computed by the lateral superior olive, which receives excitatory input from the ipsilateral ear and inhibitory input from the contralateral ear (via the nucleus of the trapezoidal body). The firing rates of these inputs are proportional to the sound level in each ear. Neurons in the lateral superior olive add together the excitatory (positive) and inhibitory (negative) inputs, so their firing rate represents the difference in the sound level between the ears.

Spatial information from the superior olivary complex propagates forward through the auditory pathway, and as it does, inputs from the medial and lateral superior olive combine to integrate information about timing and level differences. This results in a coherent, unitary perception of sound location.

Perceptual contents

In addition to determining the locations where sounds originate, the auditory system also has to determine what is making the sounds. These complex identifications are based on patterns of simpler spectral and temporal features of the sounds. In other words, we can describe complex sounds like “bird singing” or “lion roaring” in terms of simpler perceptual features. In this section, we will examine some of these perceptual features and how they relate to the physical properties of the sound, then describe some of the neural circuits involved in decoding these features.

Speech and other communication signals

Acoustic communication signals are often complex, consisting of periodic and aperiodic sounds. The frequency and amplitude of these components are often dynamic, changing over time. To illustrate, Spectrograms of birdsong and human speech shows a spectrogram of a song sparrow singing (top) and the author speaking a phrase (middle). Spectrograms are two-dimensional plots of how the spectrum of a complex sound changes over time. Frequency is shown on the y-axis, time on the x-axis, and the relative amplitude by the intensity or color of the plot. The sparrow song consists of a series of different kinds of tonal (periodic) sounds in rapid succession. Some of the tonal sounds have clear harmonics, indicated by parallel lines at integer multiples of the lowest frequency. In other sounds, there is only a single frequency, but it is rapidly modulated up or down to produce chirps and trills. The human speech consists of long tonal components, corresponding to vowels, transients spanning the whole range of frequencies, corresponding to consonants like “t”, “b”, and “k” where the flow of air is stopped by the tongue or lips, and broadband aperiodic noises, corresponding to consonants like “s” where turbulent air is passing through the teeth.

In this section, we will focus on human speech, one of the most important stimuli our auditory system has to process.

The acoustic and phonetic structure of speech

Human speech consists of a series of periodic and aperiodic sounds. Speech sounds are produced by the movement of air through the vocal tract. The parts of the vocal tract that are moved to shape the air stream and produce speech sounds are called articulators.

The main articulator for speech is the larynx, and more specifically, the vocal folds, bands of muscle tissue in the center of the airway. In normal breathing, the vocal folds are relaxed, and air is able to pass through freely. In speaking or singing, the folds are tightened and the air pressure from the lungs increases, causing them to vibrate open and closed. This vibration, also called phonation, produces a periodic wave that forms the basis for vowels.

In most languages, vowels are distinguished not by the fundamental frequency of the vocal fold vibrations, but by how this sound, which is rich in harmonics, is filtered in the upper vocal tract. Positioning the tongue and lips creates cavities of different sizes and shapes. The resonances of these cavities act as filters, producing peaks in the vocal spectrum called formants.

Consonants are phonemes that occur at the beginnings and ends of vowels. They can be produced by starting and stopping air flow through the vocal tract, or by restricting its flow through a narrow opening, producing turbulence.

As you might imagine, the human vocal system can produce a great variety of different vowels and consonants. Not every speech sound, or phone, is used in every language, however, and different phones may be used interchangeably. A group of phones that can be used interchangeably is called a phoneme. In English, the phones [r] and [l] are distinguished from each other: switching [r] for [l] causes the meaning of a word to change (for example “right” and “light”). The same is not true in Japanese: there is only a single phoneme that includes the whole range of sounds between [r] and [l].

Phonetic perception is categorical: acoustically dissimilar stimuli can be perceived as identical. This phenomenon was demonstrated by Alvin Liberman and Ignatius Mattingly, two pioneers of speech perception, using synthetically generated speech (Liberman et al., 1957; Liberman and Mattingly 1985). As illustrated in Categorical perception of phonemes, when the shape of the onset of a consonant-vowel syllable is systematically shifted, English-speaking listeners report hearing either “ba”, “da”, or “ga”. Even though the acoustic difference in each step is the same, this linear acoustic relationship is not reflected in what listeners perceive. Rather, listeners hear sudden transitions from one phoneme to another (there is no percept midway between “ba” and “da”), and they are unable to distinguish sounds within the same category from each other.

Categorical perception results in a many-to-one mapping between the acoustic structure of speech and phonetic perception. At some level of the auditory pathway, the brain must be responding the same way to different stimuli (Holt and Lotto 2010). This is a critical feature of vocal communication, because it allows us to understand the same words and phonemes produced by different individuals despite wide variations in pronunciation. It also allows us to understand the same phoneme produced by the same individual in different contexts, at different frequencies, and with different emphases.

Developmental Perspective: Normal development of phonetic perception

Different languages have different sets of phonemes, and categorical perception tends to follow the phonetic structure of the listener’s native language. In other words, individuals who grow up hearing English may not be able to hear the difference between sounds that form two distinct phonemes in another language. This observation suggests that phonetic perception is shaped by experience. Moreover, the older people get, the more difficult it is for them to learn to perceive the phonemes of a new language. They hear the new language through the lens, so to speak, of their native language, which makes it more difficult to rapidly process speech or correct an accent. These observations imply that the effect of experience on perception is limited to a narrow window during development, i.e. a critical period (see Introduction).

Work by Patricia Kuhl and Janet Werker has revealed how phonetic perception is established early in life (Werker and Tees 1984; Kuhl et al., 2006). Prior to about 6 months of age, infants from all over the world, raised in environments where different languages are spoken, can discriminate between all of the phonemes found in all languages. However, by 9–12 months of age, infants become better at discriminating phonemes in the language of their parents or caregivers and worse at discriminating phonetic contrasts not found in that language (Infant perception of phonetic contrasts). Children raised hearing Japanese lose the ability to distinguish [r] from [l]. In contrast, children raised in English-speaking environments stop being able to tell apart “dental” [t], made by placing the tongue against the teeth, from “retroflex” [T], formed by placing the tongue against the roof of the mouth; these sounds are distinct phonemes in Hindi but not in English. This shift, which commits the infant to a certain way of hearing speech, is echoed by changes in how the brain responds to native and nonnative phonemes (Bosseler et al., 2013). This commitment is necessary for normal language development (Tsao et al., 2004), but it makes it increasingly difficult for children to hear other languages as native speakers do. This could explain why children who begin learning a language later in life are more likely to speak with an accent (Piske et al., 2001), because they would have a harder time hearing their own mispronunciations and correcting their errors.

Sensing angular and linear movements

As animals move around in the world, they undergo many changes in position and orientation. These movements occur within the earth’s gravitational field, which exerts a force that must be resisted to avoid falling over. To understand how the vestibular system tracks these movements and forces, it is necessary to consider some elementary physics.

Movements in space can be linear or angular (see Angular and linear movement). Linear movements are changes in the position of the animal’s center of mass. These can occur along any of the three dimensions of space, or some combination of directions. From the animal’s frame of reference, these are up and down, left and right, and forward and backwards. Angular movements are changes in the orientation of the animal around the center of mass. These can also occur along three dimensions: pitch, roll, and yaw. Pitch corresponds to nodding the head, yaw to shaking it from side to side (like indicating “no”), and roll to tipping it sideways (like getting water out of your ear). If the animal is at rest, any change in linear or angular position requires a change in linear or angular velocity. A change in velocity is called acceleration. According to Newton’s first law, acceleration is proportional to net force. The forces produced by acceleration are dynamic. Bodies are also subject to the static force produced by gravity. Much as the cochlea has evolved to sense the forces produced by acoustic pressure, the vestibular labyrinth has evolved to sense the forces produced by acceleration of the head and by the earth’s gravitational field.

The vestibular labyrinth is a fluid-filled structure connected to the cochlea and sharing the same endolymph (Otolith organs). In humans, the labyrinth consists of the anterior, posterior, and lateral semicircular canals, three swellings at the base of the canals called ampullae, and two larger swellings called the utricle and the saccule, which together are called the otolith organs.

The semicircular canals and ampullae are responsible for sensing the dynamic forces created by angular movements. They are arranged at right angles to each other. The anterior and posterior canals sense a combination of pitch and roll, and the lateral canal senses yaw. When the head is rotated along a plane parallel to the canal, the inertia of the endolymph causes it to stay still while the surrounding bony and membranous structure moves. This relative motion between the fluid and surrounding tissue exerts a force on the cupula, a gelatinous mass found in each of the three ampullae, atop the ampullary crest (Crista ampullaris). The ampullary crest contains hair cells with stereocilia embedded in the cupula. Deflection of the cupula by the force of fluid motion produces depolarization of the hair cells by a similar mechanism to the one seen in the inner hair cells of the organ of Corti. The hair bundles within each ampulla are aligned, so that rotation of the head in one direction causes excitation while movement in the other direction causes inhibition.

The otolith organs are responsible for sensing linear movements and the direction of gravity. They consist of swellings similar to the ampullae. The floor of each swelling, called the macula, contains hair cells covered in a gelatinous mass. Unlike the cupula, this mass is embedded with crystals of calcium carbonate called otoconia (“ear dust”). The utricular macula is parallel to the horizontal plane of the head. When the head is tilted or linearly accelerated along the horizontal plane, a force is exerted on the otoliths, which in turn causes the hair cells to bend (Otolith organs). The saccular macula is oriented along the vertical plane of the head and senses linear movement up and down or to the left and right. Whereas the stereocilia in the ampullae are all aligned in the same direction, the macular hair cells have stereocilia that point in all directions. Tilting or linear movement of the head therefore depolarizes some hair cells while hyperpolarizing others.

Vestibular hair cells are innervated by afferent nerve fibers from bipolar neurons with their cell bodies in the vestibular ganglion. As in the cochlea, they also receive efferent connections from neurons in the brainstem, but the function of this system remains poorly understood. The ascending axons from vestibular ganglion neurons course through cranial nerve VIII and synapse in the vestibular nuclear complex, a group of nuclei located in the medulla and the pons. The main outputs of the vestibular nuclei are to the cerebellum, the spinal cord, and the reticular formation, involved in postural control, and to the abducens nucleus and the oculomotor nucleus, which are responsible for maintaining gaze direction while the head is rotated (Vestibuloocular reflex). The vestibular nuclei also project to the cortex via the thalamus, and this pathway is responsible for the conscious perception of tilt and acceleration. The cerebellum contributes a major descending connection to the vestibular nuclei that is responsible for modulating the gain of postural and gaze reflexes.

Vestibular control of gaze and balance

Maintaining balance, especially in bipedal species like humans, requires rapid, reflexive muscular movements to keep the body’s center of mass in alignment with gravity. Furthermore, locomotory activities like running, walking, and flying cause displacements of the head, and animals need to correct for these movements to maintain a clear visual image. The vestibular system plays a critical role in both of these processes.

The central concept in understanding how gaze and balance are maintained is feedback: movements that displace the body or the eyes from the desired position produce signals in the vestibular system that output to the specific muscles that generate a corrective or compensatory movement. The vestibular reflexes are some of the fastest in the body, and to achieve that speed, the underlying circuitry is relatively simple.

The vestibulo-ocular reflex (VOR) is the simplest and most easily demonstrated vestibular reflex. Hold a finger up at arm’s length directly in front of your eyes, and then rotate your head from side to side while maintaining your gaze on the finger. Notice how the image of the finger remains sharp and stable. Now try moving your finger back and forth while holding your head still and notice how the image starts to blur. The blurring occurs because the brain is not fast enough to adjust the position of the eyes in response to a rapidly changing visual stimulus: there are many synapses between the retina and the muscles of the eyes, and computing the position of a moving object from a visual image is not trivial.

The VOR is fast because the hair cells in the semicircular canals provide direct information about the movement of the head, and they are connected to the ocular muscles by only 2–3 synapses. Vestibuloocular reflex illustrates how the VOR works for rotations along the horizontal plane (yaw). As the head moves counterclockwise, the fluid in the lateral semicircular canals remains stationary due to inertia. This deflects the hair cells in the lateral ampullae, producing excitation on the left side and inhibition on the right. Excitation of neurons in the left vestibular ganglion in turn excites neurons in the left vestibular nucleus and then the left abducens nucleus. The abducens excites the contralateral rectus via cranial nerve VI, which adducts the right eye, pulling it away from the midline. The abducens also excites the contralateral oculomotor nucleus, which activates the medial rectus muscle of the left eye, abducting it towards the midline. The output from the left lateral ampulla is inhibited, which in turn causes the opposing muscles (the left lateral rectus and the right medial rectus) to relax. The net effect is for the eyes to move in the opposite direction from the head, directly compensating for its motion without the need for any visual processing. There are similar circuits that mediate ocular reflexes in response to angular and linear movements along the other directions, thereby allowing animals to maintain steady gaze during all kinds of tasks.

For the VOR to maintain gaze, the angular distance of the eye movement needs to be matched to the angular displacement of the head. The scaling between the sensory input (head movement) and the motor output (eye movement) is called the gain of the reflex, and it depends on the distance of the object from the eyes. To illustrate, repeat the experiment in the previous paragraph, holding your finger still while rotating your head from side to side. Keep your focus on your finger, but notice how objects in the background do not remain steady. Now focus on the background, and notice how your finger now appears to move. This modulation in gain depends on descending input from the cerebellum to the vestibular nucleus. Purkinje neurons in the flocculus of the cerebellum integrate information from many sources, including optic flow from the superior colliculus, neck movements from proprioceptors, and motor commands from the cortex to predict how eye movements will affect the visual image. The output of these neurons can increase or decrease the gain of the vestibular nucleus neurons to appropriately scale eye movements. The cerebellum can learn to adjust its predictions when visual or vestibular inputs are perturbed, for example if the vestibular hair cells on one side of the head are damaged or if glasses are worn with prisms that distort or reverse the visual image (Blazquez et al., 2004; Day and Fitzpatrick 2005).

The other two major vestibular reflexes are the vestibulocollic reflex and the vestibulospinal reflex. The vestibulocollic reflex generates neck muscle contractions to stabilize the head’s position relative to the body, whereas the vestibulospinal reflex generates limb and core muscle movements to maintain balance. When the head and trunk tilt to one side, signals from the vestibular nucleus to motor neurons in the spinal cord cause the ipsilateral limbs to extend while the contralateral limbs contract, which has the net effect of pushing the body back towards the vertical. As with the VOR, the cerebellum is responsible for modulating the gain of the vestibulocollic and vestibulospinal reflexes during volitional movements.

Taste-mediated behaviors

The primary role of the sense of gustation is to determine the chemical composition of substances we are about to ingest. The gustatory system is capable of detecting only a few specific types of molecules when in high concentration at close proximity with receptor cells, and it triggers only a handful of sensations or modalities (Sweet, Umami, Salty, Sour, Bitter, etc.). While this is relatively few modalities compared to olfaction, gustation is essential for survival. It is the final mechanism by which animals decide if materials are full of life-giving nutrients or life-threatening poisons.

Another example of a taste that potentially signals danger is bitterness. Humans perceive several classes of molecules as being bitter. These molecules are usually found in plants and are likely to have physiological or psychoactive effects on animals. Analogous to “acquiring a taste” for fermented foods, many humans also acquire a taste for bitter compounds that alter their physiology; alcohol, caffeine, nicotine, and cocaine are all perceived as being intensely bitter, but humans commonly will learn to tolerate the bitter taste of these drugs in pursuit of the psychoactive effects. Conversely, the bitter taste of many pharmaceuticals can cause individuals to avoid ingesting therapeutic drugs, and some green vegetables that have health benefits are also avoided because of their bitter taste. Despite these last two examples, perceiving compounds as bitter likely warns of a substance that could detrimentally alter an animal's health.

The primary behavior mediated by the gustatory system is the acceptance or rejection of particular foods. However, it is the integration of these signals with olfactory signals and even the somatosensory system (see The Gustatory System) by the brain that creates the sensation we refer to as “taste” or more precisely flavor.

Olfactory-mediated behaviors

The function of the olfactory system stands in contrast to the gustatory system. While the gustatory system detects only a handful of molecules, the olfactory system can detect almost any type of volatile (e.g. air-borne) organic molecule, producing a seemingly inexhaustible collection of different types of smells by detecting specific motifs in those molecules. The olfactory system is also sensitive to incredibly small concentrations of molecules and can therefore detect sources of volatile molecules at great distances. Our sense of smell does more than just add sensory complexity to the flavor profile of a meal or allow us to enjoy the scent of a spring flower. Across the animal kingdom, the olfactory system is an important sensory modality for numerous sensory-guided behaviors. Below, we describe three olfaction-dependent behaviors: navigation, nutrient finding, and mate selection/communication.

Chemesthesis-related behaviors

Many of the culinary ingredients we call spices contain chemicals that stimulate the somatosensory system to produce warming and cooling sensations. The technical term for the chemical stimulation of the somatosensory system is chemesthesis. Many plants produce chemicals that simulate thermal stimuli to alter the behavior of predatory animals. However, because our brains integrate chemesthetic with olfactory and gustatory signals to produce the perception of flavor, human beings cultivate these plants to use as flavor enhancers. The sense of chemesthesis is essential to describing the distinct culinary traditions around the world, but increasingly research indicates that the sense of chemesthesis is important for monitoring parts the body for growing bacteria (see Chemethesis, Spices, and Solitary Chemosensory Cells).

These differences highlight that stimulation of the gustatory system is only capable of producing a few very specific sensations that correspond to a few specific classes of molecules; conversely the olfactory system can detect the presence of an almost limitless number of different compounds and the language we use to describe those sensations is similarly diverse. Additionally, you can begin to determine the different sensations mediated by the gustatory and olfactory systems by comparing how similar your description of sniffing the jellybean was to the other experiences. Likely, the sniffing description was very similar to what was recorded for the “Jellybean/Open” experience. There are some foods, however, that may produce different sensations from orthonasal olfaction (e.g. sniffing) and retronasal olfaction. For example, some cheeses (e.g. parmesan cheese or blue cheeses) are famously stinky when perceived by orthonasal olfaction but produce a different sensation when perceived retronasally.

When you compare the description of the mint to the jellybean, you will likely notice the addition of words like burning, cool, pungent, or painful in the description of the mint. These sensations are due to chemicals in the mint stimulating the free nerve endings of the trigeminal nerve—part of the somatosensory system, your sense of chemesthesis.

The jellybean experiment is easy to do with most foods, and, by practicing it, you can become better at perceiving the difference between gustation, olfaction, and chemethesis. Eventually you may even be able to tell the difference without pinching your nose closed!

Gustation and peripheral anatomy to receptors

The primary role of the gustatory system is to assess the quality and concentration of chemicals before they are allowed entrance into the digestive tract. This process begins with bringing chemicals into the mouth where they can interact with specialized receptors on sensory cells.

The tongue and oral cavity

The sensory cells of the gustatory system are typically concentrated in the oral cavity. These sensory cells are found in clusters called taste buds. The exact distribution of taste buds will vary among species. In most mammals, taste buds are concentrated on the tongue (Gross anatomy of gustatory and olfactory systems), but are also found on the soft palate, epiglottis, larynx, upper esophagus, and associated with salivary glands behind the molars.

Many species of fish, in contrast, possess taste buds in the oral cavity but also on their lips and on the skin of their bodies—indeed, taste buds are so prevalent on the bodies of catfish that they are sometimes described as being “swimming tongues.” Conversely, marine mammals (e.g. dolphins, seals) that feed by swallowing prey whole typically have a reduced number of taste buds throughout the oral cavity and tongue. These anatomical adaptations all facilitate the specific feeding behaviors of each species. Despite the observed anatomical variation, chemicals in food are detected by sensory cells in taste buds, and this information is subsequently passed on to homologous regions of the central nervous system.

While there is great variation in the anatomy of the oral cavity between species, the tongue of most vertebrates is a sensory organ specialized for both chemosensation and somatosensation. If you have ever looked at your tongue closely in the mirror, you might have noticed that its anterior surface appears rough, except for scattered bumps which have a smooth texture (Tongue and papillae anatomy).

Both the rough areas and smooth bumps are made up of structures called papillae, which enhance somatosensation on the tongue. The rough areas of the tongue are covered in filiform papillae, which are cone-shaped, come to a point, and do not contain taste buds. The three other types of papillae contain taste buds: the smooth round bumps on the tongue are called fungiform papillae and often have 1 to 5 taste buds at their apex. On either side of your tongue there are 4 to 6 groves lined with hundreds of taste buds called foliate papillae. And at the very back of your tongue are 4 to 8 circular troughs containing taste buds called circumvallate papillae. The exact number of papillae varies across vertebrate animals, but most have each type of papillae to facilitate exploring the textures of food and channeling the biomolecules towards taste buds.

Behind the tongue is a space called the pharynx. Gross anatomy of gustatory and olfactory systems shows the three major divisions of the pharynx. In most terrestrial animals, this space connects to the oral cavity, esophagus, trachea, nasal cavity, and even the inner ear (see Introduction. This anatomical arrangement allows for the chemicals in food to be sampled with both the gustatory and olfactory systems simultaneously. The co-simulation of these systems when eating strengthens and facilitates our perceived integration of these senses.

Taste buds and taste receptor cells

Papillae are visible with the naked eye, while taste buds are so small they require a microscope to see. People without formal anatomical training sometimes mistakenly equate fungiform papillae (0.5 to 0.8 millimeters or 500 to 800 micrometers across) with taste buds (30 to 100 micrometers across). Taste buds resemble an onion or bulb of garlic in cross-section (Taste bud anatomy).

Each taste bud is a collection of 50 to 150 neuroepithelial cells and associated nerve fibers. These neuroepithelial cells are spindle-shaped or bottle-shaped with apical microvillar tufts that open into the lumen of the oral cavity in a structure called a taste pore. For the gustatory system to detect particular chemical stimuli, a chemical must first enter the taste pore so it can interact with proteins in the cell membrane of the microvillar process of these cells. Because these cells are the first element of the gustatory system that responds to chemicals, they are referred to as taste receptor cells (TRCs).

Anatomists studying the cells of taste buds have grouped them into three types based on characteristics observed by transmission electron micrography (Farbman, 1965) (see Methods: Transmission Electron Microscopy). These classifications based on micro-anatomy turn out to be functional, as different classes of chemicals activate Type I, Type II and Type III taste receptor cells. Type II and Type III cells will make occasional cytoplasmic contact with each other or nerve fibers but Type I cells enwrap the other cell types supporting them and helping degrade neurotransmitters released by the other cells. These functions have caused scientists to describe Type I cells as “glial-like.” Type IV cell is a designation given to basal cells in the bud that divide to replace the other taste receptor cell types that are lost to damage. The specialized anatomical arrangement of taste receptor cells allows taste buds to detect chemicals and selectively modulate either other cells in the taste bud or the associated nerve fibers to ultimately produce the familiar taste qualities of sweet, umami, bitter, salty, and sour.

Taste receptor cells and taste receptor proteins

But how does a taste receptor cell in a taste bud turn a chemical in your mouth into a signal that your brain can understand? Decades of research have helped us understand that the mechanisms for detecting and transducing a chemical into a neural signal are different for each of the half dozen or so taste qualities. This stands in dramatic contrast to the mammalian olfactory system, where hundreds of genes belonging to the same receptor family using the same basic transduction paradigm allow you to perceive an incomprehensible number of olfactory qualities. In the mammalian taste system, at least 4 different families of receptors are implicated in detecting the chemicals we perceive as sweet, umami, bitter, sour, and salty. An overall summary of these receptors and how those signals are transduced is provided in Taste receptor cell comparison.

Three distinct groups of Type II taste receptor cells are responsible for detecting sweet, umami, and bitter chemicals using receptors from the GPCR super-family. The sweet and umami-sensitive taste receptor cells detect chemicals using taste receptors from subfamily 1, which is made up of 3 individual receptor genes designated Tas1R1, Tas1R2, and Tas1R3. To produce a functioning receptor protein from this family of receptors—either sweet or umami—two different receptor genes must combine to make a heterodimer: The sweet receptor requires both Tas1R2 and Tas1R3; while an umami receptor requires Tas1R1 and Tas1R3. The third group of type II taste receptor cells detects bitter compounds via taste receptor subfamily 2 (Tas2Rs), which consists of many more genes than subfamily 1 (e.g. humans have 25 Tas2R genes). Thus, sweet-sensitive type II receptor cells will express Tas1R2+Tas1R3, umami-sensitive type II receptor cells will express Tas1R1+Tas1R3, and bitter-sensitive type II receptor cells will express some combination of Tas2Rs. Activation of any of these receptors triggers lipid secondary messenger signaling that ultimately results in the release of ATP to activate gustatory nerve fibers.

Ion channels are responsible for the detection of sour and salty chemicals. Most sour chemicals are acids that dissociate to produce a proton (e.g. hydrogen ion) in solution. Type III taste receptor cells that are sensitive to sour chemicals express a proton channel Otopetrin 1 (Otop1). Protons flow into these cells through Otop1 and ultimately trigger a change in membrane potential resulting in the vesicular release of serotonin (5HT).

While scientists are still searching for the precise mechanism of salt detection, Epithelial Sodium (Na) Channels (ENaC) are widely considered to play an essential role in the depolarization of Type I taste receptor cells in the presence of sodium ions.

Identifying the receptor proteins TRCs utilize to detect chemicals is a prerequisite for a complete understanding of the taste modalities the gustatory system is capable of detecting. The involvement of at least four different families of receptor proteins and distinct transduction mechanisms in taste receptor cells highlights the inherently molecular nature of gustatory perception.

Gustation - neurotransmission and the central nervous system

After taste receptor cells detect chemical stimuli in the oral cavity, this signal must be transmitted to the nerves innervating the taste bud. These intragemmal (from the Latin gemma meaning bud, gem, or cup) nerve fibers enter the taste buds (shown in Taste bud anatomy), detect neurotransmitters released from taste receptor cells and conduct this information into the brain regions where the perception of umami, sweet, bitter, sour, and salty are represented. Before reaching gustatory cortex, taste information passes through 2 relays along the way: the nucleus tractus solitarius in the medulla and the posteroventral nucleus of the thalamus (Neural pathways of taste). Once this information has arrived in the gustatory cortex, an animal can make decisions about the palatability of the substances in the oral cavity. How the primary taste qualities are represented between the taste buds and brain remains one of the great open questions in sensory neuroscience.

While the broad pathway of gustatory information is known, exactly how that pathway encodes different tastes remains uncertain. Historically, most hypothesized models for how the nervous system encodes primary taste qualities can be classified as either labeled line coding or across-fiber pattern coding. In labeled line coding, there is a discrete pathway or circuit from the receptor cells to the afferent neurons, to higher brain regions for each of the primary taste qualities. In this coding model, a sweet chemical of adequate concentration would result in activity through the sweet “line” and would not trigger substantial activity in the circuits for other primary tastes. A classic example of a sensory system that uses label line coding is the somatosensory system (see Introduction).

In across-fiber pattern coding, any given pathway (e.g. receptor cells, nerve fibers, higher brain regions) is not associated with any individual primary taste quality; rather, a specific pattern of activity among overlapping neural pathways codes for each of the primary tastes. Across-fiber pattern coding allows the vertebrate olfactory system to represent an overwhelming number of olfactory modalities.

Labeled line coding diagrams these two opposing models, using circles to represent neurons that might respond to sour, salt or both. As technological advancements allow scientists to examine the activity in gustatory circuits at an ever-increasing resolution, new refinements are constantly being made to theories of gustatory system coding.

Neuroscience in the lab: Studying taste receptor cell to neurotransmission

Recording action potentials from a nerve that primarily makes connections with taste buds provides an important assay for determining if chemicals are activating the gustatory system. The chorda tympani branch of the facial (VII) nerve and the glossopharyngeal (IX) nerve innervate the taste buds on the rostral ⅔ and caudal ½ of the tongue, respectively. In contrast to behavioral data, chorda tympani and glossopharyngeal nerve recordings in ATP receptor knockouts or animals treated with an ATP receptor blocker show decreased activity not just for umami, sweet and bitter tastants, but also for salty and sour compounds (Chorda tympani nerve recording). Conversely, while knockout or pharmacological blockade of serotonin receptors results in no taste-related behavioral deficits in mice, it does decrease chorda tympani nerve response to MSG, sucrose, salts, and both weak and strong acids. Taken together, these data suggest that ATP release is necessary for taste receptor cell nerve neurotransmission for all taste modalities, while serotonin acts synergistically to intensify the neural response to certain taste qualities.

However, this interpretation is difficult to resolve with the observation regarding transmitter release from taste receptor cells. Specifically, what is the source of ATP, if only type III cells are activated by sour tastants and those cells do not appear to release ATP or possess the proteins necessary for ATP’s release? One hypothesis, which would explain these results, would be the existence of a currently unknown mechanism by which low levels of ATP could be packaged alongside serotonin. Another explanation is that the limited cell-to-cell contacts between type II and type III cells allow for serotonin signaling directly between taste receptor cells to modulate ATP release from type II cells in the presence of sour stimuli. However, this model must account for the serotonin’s apparent inhibitory action on type II cells making ATP release from these cells less likely in the presence of sour stimuli.

Science as a process: Taste nerves and the nucleus of the solitary tract

As intergemmal nerve fibers exit taste buds, they merge into larger bundles until ultimately joining cranial nerves, which conduct taste information to the brainstem. A recently published connectome (a complete map of cellular and neural connections) of taste buds determined that 97% of intergemmal fibers innervate one taste receptor cell or the same taste receptor cell type (Wilson et al., 2022). Thus, each intergemmal nerve fiber is passing information from only one taste receptor cell type to the nucleus of the solitary tract in the brainstem.

Three cranial nerves innervate taste buds in distinct regions of the oral cavity: Taste buds on the rostral or anterior 2/3 of the tongue are innervated by the chorda tympani branch of the facial (VII) nerve, taste buds on the caudal 1/3 of the tongue and palate are innervated by the glossopharyngeal (IX) nerve and taste buds found on the epiglottis are innervated by the vagus (X) nerve. Chemosensory information from all three of these nerves is conducted to a region of the brainstem called the nucleus of the solitary tract.

The nucleus of the solitary tract (NTS, from the Latin nucleus tractus solitarii) is the primary gustatory and visceral chemosensory nucleus of the central nervous system. Gustatory information to the central nervous system projects to the most rostral portion of the NTS (Nucleus of the solitary tract). Within the rostral NTS (rNTS) or gustatory nucleus, projections from the facial, glossopharyngeal, and vagal nerves are arranged rostrally to caudally reflecting their innervation of taste buds in the oral cavity. Chemosensory afferents projecting to the caudal NTS (cNTS) carry information from visceral chemoreceptors. This anatomy creates a rostral to caudal topographic map of chemosensory inputs from the mouth to the alimentary canal (i.e. digestive tract) and other internal organs along the NTS.

In addition to its anatomical map, investigators continue to define the chemotopic map or coding logic for taste qualities within the NTS. For example, most chemicals perceived as bitter trigger activity in neurons located in the medial rNTS. MSG and NaCl, which taste savory and salty, both activate the intermediate zones of the rNTS. Sweet chemicals produce diffuse activity along the nucleus. Sour chemicals, like citric acid, trigger activity in lateral-rostral and mid-NTS. One major challenge that researchers attempting to define the chemotopic map of the NTS face is determining the influence that non-gustatory inputs play in the coding logic of these taste qualities.

While the rNTS is the first relay of gustatory information in the central nervous system, it also receives and integrates other signals that can influence our perception of flavor. For example, contrary to its moniker as the gustatory nucleus, the rNTS also receives oral somatosensory information from the trigeminal nerve (see How Do Connections Differ Across Species?). The temperature of solutions with identical chemical composition can modulate the firing neurons of the NTS, presumably through these somatosensory inputs. The result is that a particular rNTS neuron that fires action potentials at a specific rate to a specific chemical solution may increase or decrease its activity as the temperature of that solution changes. Mouse NTS response to tastants at different temperatures shows an example of this phenomenon.

In this experiment, the activity of neurons in the NTS was recorded while fluids of different temperatures, containing different tastants were flowed over a mouse’s tongue. The top of the figure shows firing of a sucrose-oriented NTS neuron in response to room temperature sucrose and warm sucrose, which results in more firing than the room temperature solution. The responses of NTS neurons to temperature are not all the same, though. The salt-oriented NTS neuron fires less to warm salt than room temperature salt, for example. The bottom on the figure shows more evidence of this variability across NTS neurons, with some tastants leading to more firing when warm, some leading to less and some seeming relatively insensitive to this temperature change. This anatomical and physiological arrangement is one biological explanation for the observation that temperature and other sensory information can have a profound impact on flavor perception.

Gustatory pathway and cortical processing

Ultimately the gustatory information from the NTS is relayed to the primary gustatory cortex (GC), where the conscious perception of taste occurs. In humans and other primates, fibers from the NTS project directly into the parvocellular division of the ventral posterior medial nucleus of the thalamus (VPMpc), and then into the GC.

The conscious perception of taste occurs in the GC. This cortical region can be viewed by separating the lateral sulcus and is found along the anterior insular lobe and frontal operculum (see Neural pathways of taste). In a variety of vertebrates, techniques from electrophysiology to cellular and functional imaging have established that neurons of the gustatory cortex demonstrate differential activity in responding to chemicals with different tastes. Additionally, some neurons in the region fire different patterns of action potentials in response to changing concentrations of tastants. Thus, both the identity and intensity of the chemical stimuli detected by the taste receptors of the tongue can be represented in the GC. From the GC, a large portion of gustatory information passed on to the secondary gustatory cortex, located in the caudolateral section of the orbitofrontal cortex and ultimately integrates gustatory signals with other sensations, including those of satiety.

Gustatory information is not just a one-way street to the primary and secondary GC, though. Interconnection with other sensory modalities, and even other non-sensory systems, occurs throughout the gustatory neural pathways. The gustatory connections to the limbic system are one, particularly important example of this cross-talk. Along the gustatory neural pathway, collateral and descending fibers pass gustatory signals back to the NTS and into the limbic system, modulating the limbic system’s regulation of feeding behaviors (see Introduction). For example, the neural connections between the GC and the lateral hypothalamus allows for gustatory signals to immediately modulate feelings of satiety.

Additionally, neural inputs to the amygdala mediate negative associations with foods that might have poisoned an animal. For example, it is common for people who have suffered from food poisoning to avoid the food that has made them sick for a long time. This is a phenomenon called conditioned taste aversion and is observed in a variety of animals. In the lab, rodents will avoid a novel tastant if it is paired with a subcutaneous injection of KCl (which makes the animal feel sick) just a single time. A conditioned taste aversion is an important tool for designing experiments to assess if an animal perceives two chemicals as having similar tastes; specifically, if an animal that has a conditioned taste aversion to one chemical also avoids a different chemical, then the animal likely perceives the two chemicals as having the same or a very similar taste. This behavioral assay has been an essential tool for studying taste perception in animals.