The cerebral cortex is the thin, folded outer layer of the brain responsible for nearly everything you consciously experience and deliberately do: seeing, hearing, touching, speaking, planning, deciding, and moving. It is only about two to four millimeters thick, yet its surface area, wrinkled into ridges and grooves, is vast enough that it contains the majority of the brain’s neurons. This folding is not decorative. The expansion of the cortex across mammalian evolution, driven by prolonged proliferation of neural stem and progenitor cells, is what gave rise to the higher cognitive abilities that distinguish humans from other species.
Why the Cortex Is Folded and Layered
If you could flatten the human cortex, it would cover roughly the area of a large newspaper page. All those folds exist because the cortex expanded dramatically over evolutionary time while remaining tethered inside a skull of limited size. Species with smaller cortices, like mice, have smooth brains. Larger-brained mammals developed a gyrencephalic (folded) surface, which packs far more neural circuitry into the same volume.
The cortex is not a uniform sheet. It is organized into six distinct layers, each with different cell types and connectivity patterns. This six-layer architecture is a hallmark of the neocortex, the evolutionarily newest part. Older cortical structures like the hippocampus and the olfactory cortex have a simpler, three-layer design. The six-layer neocortex expanded that basic circuit by splitting into superficial layers (roughly layers one through three) and deep layers (four through six), each containing populations of pyramidal cells that interact through excitatory and inhibitory circuits.1Frontiers in Neuroanatomy. The Microcircuit Concept Applied to Cortical Evolution: from Three-Layer to Six-Layer Cortex This layered design lets the cortex handle more complex computations by separating different types of processing across distinct circuits within the same column of tissue.
How the Cortex Handles What You See
Vision occupies a surprisingly large share of cortical real estate. When light hits your retina, signals travel to the back of the brain, where the primary visual cortex (often called V1) begins the work of turning raw input into meaningful images. V1 contains a detailed map of your visual field: adjacent points in the world are processed by adjacent neurons in V1, a layout known as retinotopy. Functional MRI studies have confirmed this map in humans and shown that V1 neurons are selective for features like the orientation of edges and the contrast of patterns.2PubMed. Functional analysis of primary visual cortex (V1) in humans Research in other species has demonstrated that this spatial organization may be partly blueprinted by the structure of inputs arriving from the retina itself, before any visual experience occurs.3PubMed Central. Link between orientation and retinotopic maps in primary visual cortex
After V1 does its initial processing, visual information fans out along two broad pathways. One runs along the bottom of the brain toward the temporal lobe and is involved in recognizing what you are looking at: faces, objects, written words. The other runs upward toward the parietal lobe and handles where things are in space and how you might interact with them physically.4PubMed Central. Interactions between dorsal and ventral streams for controlling skilled grasp These are often called the “what” and “where” pathways, and while that shorthand is useful, the reality is messier. Both pathways contribute to shape perception, for instance, and the dorsal stream is not purely spatial. Brain imaging has shown that shape detection activates both streams about equally, while location detection is more strongly dorsal.5PubMed Central. Ventral and dorsal visual stream contributions to the perception of object shape and object location
A more recent framework proposes that there are actually four streams, not two. One ventrolateral pathway leads to the inferior temporal cortex for recognizing objects and faces and feeds identity information to memory and emotion systems. A separate superior temporal sulcus pathway responds to moving faces and bodies and is important for social perception. On the spatial side, a ventromedial pathway builds representations of scenes and landmarks for navigation, while the classic dorsal pathway to the parietal cortex handles actions in space and self-motion tracking.6PubMed. Two what, two where, visual cortical streams in humans The old two-stream model is not wrong so much as incomplete.
Touch and the Sensory Homunculus
The primary somatosensory cortex, located in a strip just behind the central sulcus, contains its own body map. If you could see this map drawn as a figure, it would look like a distorted human shape: the lips, tongue, and fingertips would be enormous, while the trunk and legs would be comparatively small. This distortion reflects how many neurons are devoted to each body part, and it matches functional reality. Your fingertips are exquisitely sensitive; your back much less so. Brain imaging studies have confirmed that stimulating different body parts activates distinct regions along this strip, arranged roughly in order, reproducing the classic “homunculus” on an individual person’s brain scan.7PubMed. Somatosensory homunculus as drawn by MEG
But the homunculus is not as neat as the textbook diagrams imply. More recent work using multivariate analysis of brain imaging data has found that information about specific body parts is distributed far beyond each part’s “home” region. For example, researchers could distinguish between hand and lip representations even in the part of the somatosensory cortex traditionally assigned to the foot, and could tell apart different hand movements in the face region.8PubMed Central. Beyond body maps: Information content of specific body parts is distributed across the somatosensory homunculus The body map is real, but the cortex also mixes information across regions in ways the simple cartoon does not capture.
Movement and the Corticospinal Tract
The primary motor cortex, sitting just in front of the central sulcus, is where voluntary movements are initiated. It has its own distorted body map, similar to the sensory one, with disproportionately large areas devoted to the hand, face, and tongue. When you decide to pick up a cup, neurons in your motor cortex fire in a coordinated pattern, sending signals down a long highway of nerve fibers called the corticospinal tract. In primates, the majority of these signals ultimately target motor neurons in the spinal cord that directly control muscles. Tracing studies in rhesus monkeys found that about 81% of the downstream targets of corticospinal neurons were the motor neurons controlling muscle fibers, with the remaining 19% being intermediary neurons in the spinal cord.9PubMed Central. Corticospinal Tract Development, Evolution, and Skilled Movements This direct corticospinal connection is part of what makes fine, dexterous movements possible.
The motor cortex does not work alone. Planning a movement involves the premotor cortex and supplementary motor area, which sit just forward of the primary motor cortex. These regions help sequence complex actions, like the series of finger movements needed to play a piano chord. And information constantly flows back from the somatosensory cortex, providing real-time feedback about limb position, pressure, and texture that the motor system uses to adjust on the fly.
Language Networks
For over a century, the standard model of language in the brain revolved around two famous regions: Broca’s area in the left inferior frontal gyrus, thought to handle speech production, and Wernicke’s area in the left superior temporal gyrus, thought to handle comprehension. Modern brain imaging has confirmed that these areas are important, but the picture is far more complex than the classic model suggested. Language processing involves parallel and interconnected streams running through both cortical and subcortical structures, with dorsal pathways mediating sound-based (phonological) processing and ventral pathways carrying meaning-based (semantic) processing.10Journal of Neurosurgery. Contemporary model of language organization: an overview for neurosurgeons
The white-matter fiber bundles connecting these cortical regions are critical. Along the dorsal route, one fiber tract connects the posterior temporal cortex to the premotor cortex for converting heard speech into motor commands for speaking, while another connects to the back part of Broca’s area and supports processing of complex grammar. Along the ventral route, one tract links part of Broca’s area with Wernicke’s area for semantic processing, and another connects frontal regions with the anterior temporal cortex for basic combinatorial operations in language.11Handbook of Clinical Neurology. White-matter pathways for speech and language processing Language, in other words, is not localized to a spot or two but is a distributed network that depends on reliable, fast communication between widely separated cortical zones.
These pathways mature at different rates during childhood. Brain imaging comparing children and adults has shown that the dorsal connection between Broca’s and Wernicke’s areas is less developed in children, with lower structural integrity suggesting incomplete insulation of the nerve fibers. Children compensate by relying more heavily on the ventral pathway, and their activation patterns in Broca’s area spread over a wider region than in adults.12Cerebral Cortex. Neuroanatomical Prerequisites for Language Functions in the Maturing Brain This is a vivid example of how the cortex adapts its functional organization to the structural hardware available at any given age.
Decision-Making and Executive Function
The prefrontal cortex, the large region at the very front of the brain, is the seat of what researchers call executive function: the ability to set goals, weigh options, hold information in mind, suppress impulses, and shift strategies when circumstances change. Among the many tasks the prefrontal cortex supports, decision-making under flexible conditions is one of the most studied. Brain imaging has shown that the prefrontal cortex is active during free-choice decisions, and that this activity is linked to working memory mechanisms that hold relevant information while a choice is being made.13PubMed Central. Prefrontal Contribution to Decision-Making under Free-Choice Conditions
One particularly interesting proposal about the medial part of the prefrontal cortex is that its involvement in both memory and decision-making reflects a single underlying function: learning associations between contexts, events, and the best emotional or behavioral response. When you remember that a particular restaurant was disappointing, and that memory steers your decision tonight, those are not separate memory and decision-making operations. They are one process, linking context to adaptive response, and the medial prefrontal cortex is at the center of it.14PubMed Central. The role of medial prefrontal cortex in memory and decision making
The Thalamus as Gatekeeper
The cortex does not operate in isolation. Nearly all sensory information passes through the thalamus, a pair of walnut-sized structures deep in the brain, before reaching the cortex. The thalamus is not a passive relay station. It actively filters and focuses what gets through. During states of arousal, thalamocortical suppression narrows the spread of sensory signals across the cortex, sharpening receptive fields so that neurons respond more selectively to the most relevant stimuli.15PubMed Central. Role of thalamocortical sensory suppression during arousal: focusing sensory inputs in neocortex In other words, when you are alert, the thalamus helps the cortex ignore distracting input and focus on what matters.
This thalamocortical loop also appears essential for consciousness itself. Computational modeling of what happens during sedation with anesthetics has shown that the loss of conscious awareness is associated with a decrease in excitatory thalamocortical resonance and an increase in inhibitory activity within the thalamus.16bioRxiv. Translating electrophysiological signatures of awareness into thalamocortical mechanisms by inverting systems-level computational models across arousal states When that loop is disrupted, awareness fades, even though the cortex itself is still physically intact.
Hemispheric Specialization and the Corpus Callosum
The cortex is split into two hemispheres, and they are not identical. The left hemisphere is typically dominant for language in right-handed people, while the right hemisphere tends to play a larger role in spatial attention and certain emotional processes. But the two halves communicate constantly through the corpus callosum, a massive bundle of roughly 200 million nerve fibers. One influential idea is that the corpus callosum actually enabled hemispheric specialization in the first place. Because it could relay information between the two sides, one hemisphere was free to develop a new function without the brain losing the old one: the other hemisphere simply kept performing the original task for both sides.17Brain. Cerebral specialization and interhemispheric communication: Does the corpus callosum enable the human condition? Split-brain patients, whose corpus callosum has been surgically severed, reveal how dependent normal cognition is on this bridge. Each hemisphere literally does not know what the other is doing.
Resting Networks and Background Activity
Even when you are doing nothing in particular, the cortex is far from quiet. Brain imaging studies have revealed a set of regions that are consistently active during rest and tend to deactivate when you focus on an external task. This “default mode network” includes the posterior cingulate cortex and parts of the medial prefrontal cortex, and its functional connectivity maps during rest are virtually identical to those seen during visual processing tasks.18PubMed Central. Functional connectivity in the resting brain: a network analysis of the default mode hypothesis The default mode network is now thought to be involved in mind-wandering, self-referential thinking, and mentally simulating past and future events. Its discovery overturned the old assumption that the brain idles between tasks.
Cortical activity is also coordinated by rhythmic oscillations. Gamma oscillations, synchronized neural firing in the range of roughly 30 to 90 Hz, are thought to regulate connectivity between different brain regions and are involved in perception, movement, memory, and emotion. These oscillations depend on networks of inhibitory interneurons, and disruptions caused by inflammation, oxidative stress, or metabolic imbalances can impair them.19PubMed Central. The role of gamma oscillations in central nervous system diseases: Mechanism and treatment Altered gamma rhythms have been observed in conditions from schizophrenia to Alzheimer’s disease, suggesting they are a useful marker of cortical health.
How the Cortex Builds and Reshapes Itself
During embryonic development, the cortex assembles itself from the inside out. New neurons are born in a zone lining the brain’s internal cavities and then migrate outward along temporary scaffolding provided by radial glial cells.20Neuron. Radial Glial Cells—Are They Really Glia? Later-born neurons migrate past earlier ones, so the deepest cortical layers form first and the superficial layers are added last. This orderly layering is what produces the six-layer architecture described earlier.
Once the basic structure is in place, the cortex enters a long period of refinement. Early in life, the brain produces far more synaptic connections than it will ultimately keep. During critical periods of development, a pruning process trims back the excess, partly mediated by the immune system’s complement proteins. This pruning is thought to sharpen neural circuits and is a normal, necessary part of brain maturation. When it goes awry, the consequences can be significant: abnormal complement-mediated synaptic pruning during early critical periods has been linked to increased risk for neurodevelopmental conditions including schizophrenia and autism.21PubMed Central. Complement Dependent Synaptic Reorganisation During Critical Periods of Brain Development and Risk for Psychiatric Disorder
Cortical plasticity does not end in childhood. In adults, the cortex can reorganize in response to injury or learning. After damage to the retina, for example, dendritic spines in the affected part of the visual cortex show an increased rate of turnover, suggesting the cortex is actively reconfiguring its wiring.22Neuron. Adult Cortical Plasticity Similar spine turnover increases during skill learning. Adult plasticity is more constrained than the dramatic rewiring possible in infancy, but the cortex retains a meaningful capacity to adapt throughout life.
When Cortical Balance Tips
The cortex depends on a careful balance between excitatory and inhibitory signaling. Excitatory neurons drive activity; inhibitory interneurons keep it in check. When this balance is disrupted, the effects ripple across the brain. In mouse experiments, chemogenetically increasing the ratio of excitation to inhibition within cortical circuits caused a significant reduction in long-range connectivity between brain regions, regardless of whether the shift was achieved by boosting excitation or suppressing inhibition.23Oxford Academic. Cortical Excitation:Inhibition Imbalance Causes Abnormal Brain Network Dynamics as Observed in Neurodevelopmental Disorders This pattern resembles connectivity disruptions seen in neurodevelopmental disorders, offering a possible mechanistic explanation for why conditions like autism and ADHD involve altered brain network dynamics.
In human infants, cortical markers of this excitation-inhibition balance, measured through EEG, have been linked to sensory behavior. Infants at five and ten months whose cortical signals suggested a particular balance profile were more likely to show heightened sensory responsiveness.24Translational Psychiatry. Cortical markers of excitation/inhibition balance are associated with sensory responsivity from infancy in longitudinal cohorts enriched for autism and ADHD Sensory over-responsiveness, where ordinary stimuli feel overwhelming, is common in autism and ADHD and may trace back to how well the cortex maintains this fundamental balance from very early in life.
The Cortex Meets the Blood Supply
The cortex is metabolically expensive. Although the brain represents only about 2% of body weight, it consumes roughly 20% of the body’s energy. The cortex achieves this through a mechanism called neurovascular coupling: when neurons in a particular region become active, local blood flow increases almost immediately to deliver oxygen and glucose. This tight linkage between neural activity and blood flow is what makes functional brain imaging possible, since techniques like fMRI detect changes in blood oxygenation rather than neural firing directly.25PubMed Central. Neurovascular coupling in humans: Physiology, methodological advances and clinical implications The coupling is mediated by a combination of chemical signals including nitric oxide and the activity of astrocytes, star-shaped support cells that sit between neurons and blood vessels and act as go-betweens.26PubMed Central. New insights into coupling and uncoupling of cerebral blood flow and metabolism in the brain
When neurovascular coupling breaks down, as it does in stroke, small-vessel disease, and some dementias, cortical regions can starve even when the neurons themselves are still alive. This is one reason why vascular health and brain health are so tightly linked: protecting your blood vessels is, in a very direct sense, protecting your cortex.
Brain-Computer Interfaces Built on the Motor Cortex
One of the most striking practical applications of our understanding of cortical function is the brain-computer interface. Researchers have been recording from the motor cortex of paralyzed individuals and using those signals to drive communication devices. A recent study demonstrated that inner speech, the silent words you hear in your head, is robustly represented in the motor cortex and that imagined sentences can be decoded in real time from multi-unit neural recordings.27Cell. Inner and attempted speech are robustly represented in motor cortex for real-time decoding This finding matters because many people who lose the ability to speak can still generate inner speech, meaning the cortical signals are there to be read even when the muscles are not cooperating.
In another study, a locked-in participant with no ability to produce speech had microelectrode arrays implanted in speech-related motor regions. Using recordings from 256 electrodes and a neural network decoder combined with a language model, researchers were able to translate attempted speech into text at the sentence level.28Cell Reports. Intracortical brain-computer interface decoding of speech in a locked-in person with anarthria These technologies are still experimental, but they illustrate how detailed our knowledge of cortical organization has become: we can identify which patches of cortex encode which phonemes and use that map to reconstruct language.
Growing Cortex in a Dish
Studying the living human cortex has always been difficult for obvious ethical reasons. Over the past decade, lab-grown brain organoids, three-dimensional clusters of human cells that self-organize into structures resembling the developing brain, have opened a new window. One challenge with early organoids was that their interiors would become oxygen-starved and die, limiting how mature they could grow. A technique called sliced neocortical organoids addressed this by cutting organoids to increase surface area for oxygen diffusion. The result was sustained growth and the formation of distinct upper and deep cortical layers for both neurons and support cells, resembling the human brain at roughly the third trimester of development.29PubMed Central. Sliced Human Cortical Organoids for Modeling Distinct Cortical Layer Formation
A further advance introduced blood vessel-like structures into organoids. These vascularized organoids recapitulated cortical cell types and layer organization more faithfully, with the proportion of upper-layer neurons increasing over time in a pattern that mirrors actual prenatal brain development.30PLOS Biology. Vascularized human cortical organoids (vOrganoids) model cortical development in vivo These organoids are not miniature brains and do not think or feel anything. But they are increasingly useful for studying how cortical layers form, how genetic mutations disrupt that process, and how potential drugs might affect developing cortical tissue without any risk to a human subject.