The cerebral cortex, the folded outer sheet of the brain responsible for everything from vision to abstract thought, is built from six distinct layers of cells stacked from the surface inward. Each layer has a characteristic mix of cell types, connection patterns, and jobs, ranging from receiving raw sensory data to sending movement commands down the spinal cord. The layers are numbered I through VI starting at the brain’s surface, and they do not work in isolation: information flows up, down, and sideways across them in circuits that researchers are still mapping in detail.
Layer I, the Molecular Layer
Layer I is the outermost and thinnest of the six. It sits just beneath the meninges and is strikingly sparse in cell bodies compared with the layers below. Instead of being packed with neurons, it is filled with a dense web of fibers, including the long apical dendrites of pyramidal neurons whose cell bodies live deeper in the cortex. These dendrites branch into elaborate tufts in Layer I, and that is where something important happens: long-range projections arriving from distant brain regions land here, delivering what neuroscientists call “top-down” information. Signals about expectations, attention, and context from higher-order areas converge on these dendritic tufts, where they meet “bottom-up” sensory signals that travel up through the cell from below.1PubMed Central. Neocortical Layer 1: An Elegant Solution to Top-Down and Bottom-Up Integration
Think of Layer I as a mixing zone. A pyramidal neuron’s cell body might sit in Layer III or Layer V, but its antenna, the apical dendrite tuft, reaches all the way up into Layer I to sample feedback from the rest of the brain. This architecture means that a single neuron can integrate what it is seeing right now with what the brain expects it to see, all at different points along its own dendritic tree. Layer I also contains a small population of inhibitory interneurons that help regulate this integration, though these cells are far outnumbered by the passing fibers.
Layers II and III, the Supragranular Layers
Layers II and III are often discussed together because they share a major role: connecting one part of the cortex to another. The pyramidal neurons here send their axons horizontally or across to the opposite hemisphere, forming the cortico-cortical connections that let different brain areas talk to each other.2PubMed Central. Comparison of the Upper Marginal Neurons of Cortical Layer 2 with Layer 2/3 Pyramidal Neurons in Mouse Temporal Cortex When your visual cortex needs to coordinate with your frontal lobe during a complex task, much of that cross-talk runs through Layer II/III wiring.
In humans, these layers are dramatically expanded compared with other mammals, and that expansion comes with greater cell diversity. Recent electrophysiological work on human brain tissue has revealed multiple subtypes of pyramidal neurons within Layers II and III that differ in their electrical properties and synaptic connections, reflecting the increased cortico-cortical connectivity that distinguishes the human brain.3PubMed Central. Electrophysiological classification of human layer 2–3 pyramidal neurons reveals subtype-specific synaptic interactions This is one reason researchers suspect that the supragranular layers play a disproportionate role in the higher cognitive abilities that set humans apart: more cell types mean more computational possibilities in the circuits that link distant cortical areas.
Layer IV, the Internal Granular Layer
If Layer I is the mixing zone for feedback, Layer IV is the front door for incoming sensory information. It is the main cortical target for axons arriving from the thalamus, the relay station that funnels signals from your eyes, ears, and skin up to the cortex.4PubMed Central. Monosynaptic connections between pairs of spiny stellate cells in layer 4 and pyramidal cells in layer 5A indicate that lemniscal and paralemniscal afferent pathways converge in the infragranular somatosensory cortex The signature cell type here is the spiny stellate neuron, a small excitatory cell whose star-shaped dendrites are studded with spines that form synapses with thalamic axon terminals. Quantitative studies in mice have found that roughly 10 to 23 percent of all excitatory synapses on these cells come from thalamic inputs, a higher proportion than for other neuron types in the cortex, which underscores how strongly thalamic signals drive Layer IV activity.5PubMed. Quantification of thalamocortical synapses with spiny stellate neurons in layer IV of mouse somatosensory cortex
Layer IV’s thickness varies enormously across cortical regions, and that variation tells you something about each area’s function. In the primary visual cortex, which processes a massive stream of information from the eyes, Layer IV is conspicuously thick and subdivided into multiple sublayers. Spatial transcriptomic mapping has confirmed that Layer IV in the visual cortex is the largest relative to gray-matter thickness of any cortical area, reflecting the high density of RORB-expressing neurons there.6Nature Communications. Charting the spatial transcriptome of the human cerebral cortex at single-cell resolution In contrast, primary motor cortex has a barely detectable Layer IV, because the motor cortex is less about receiving sensory input and more about sending movement commands out. This is one of the clearest examples of how the six-layer template is adapted rather than rigidly copied across the brain.
Layer V, the Internal Pyramidal Layer
Layer V is the cortex’s main output hub for long-distance communication outside the cerebral hemispheres. Its large pyramidal neurons send their axons to the brainstem, spinal cord, basal ganglia, and other subcortical structures. In the motor cortex, this layer houses the Betz cells, giant pyramidal neurons that directly innervate motor neurons in the brainstem and spinal cord to initiate voluntary movement.7PubMed Central. Betz cells of the primary motor cortex Betz cells are among the largest neurons in the entire nervous system, and their presence is essentially confined to Layer V of the primary motor cortex, making them a defining anatomical landmark of that region.8PubMed. Organization of corticospinal neurons in the cat
Not all Layer V neurons are giant motor cells. In sensory areas, Layer V pyramidal neurons send projections to the superior colliculus, the pons, and other midbrain structures that help coordinate reflexes and relay information. The layer also contributes to thalamic feedback loops, though that role is shared with Layer VI. What unifies the layer is its function as a broadcaster: if Layers II and III talk to other cortical regions, Layer V talks to the rest of the brain and the spinal cord.
Layer VI, the Multiform Layer
Layer VI is the deepest and, as its name suggests, the most morphologically varied layer. Its primary job is modulating the thalamus, creating a feedback loop with the very relay station that feeds sensory information into the cortex through Layer IV. Layer VI corticothalamic neurons both excite thalamic relay cells directly and inhibit them indirectly by activating neurons in the thalamic reticular nucleus, producing a mix of excitation and inhibition that can sharpen or suppress the signals the thalamus sends up.9PubMed Central. Functional organization of the somatosensory cortical layer 6 feedback to the thalamus
Recent optogenetic experiments have revealed that these Layer VI neurons are not simply turning the thalamus up or down. They can either suppress or enhance thalamic spiking depending on how fast and how synchronously they fire, making them dynamic modulators that adjust the flow of sensory information moment to moment during wakefulness.10PubMed Central. Dynamic corticothalamic modulation of the somatosensory thalamocortical circuit during wakefulness On top of that, Layer VI corticothalamic neurons also project back into the cortex itself and help shape high-frequency brain oscillations through both cortico-cortical and cortico-thalamo-cortical pathways.11PubMed Central. Layer 6 corticothalamic neurons induce high gamma oscillations through cortico-cortical and cortico-thalamo-cortical pathways So Layer VI acts as a sophisticated volume knob for incoming sensation, and it does this through multiple parallel routes.
How Information Flows Through the Layers
Taken individually, each layer has a clear specialty. But in practice, the six layers operate as an integrated circuit often called the canonical cortical microcircuit. The classic flow works roughly like this: sensory signals from the thalamus arrive in Layer IV, which passes them up to Layers II/III for local processing and cross-cortical communication, and down to Layers V and VI for output and thalamic feedback. Layer I adds contextual top-down modulation from distant areas, and inhibitory interneurons distributed across every layer fine-tune the timing and balance of excitation throughout.
Laminar recordings in human parietal cortex have shown that this processing scheme operates beyond primary sensory areas. When people process numerical quantity, responses in the middle layers are broad and become increasingly selective as information moves outward toward the surface and inward toward the deep layers, mirroring what has been seen in primary visual cortex.12PubMed. Laminar processing of numerosity supports a canonical cortical microcircuit in human parietal cortex This suggests that the same laminar logic for sharpening and routing information is reused across very different cognitive tasks, not just early sensory processing.
The circuit is also shaped by chemical neuromodulators that are not evenly spread across layers. In monkey prefrontal cortex, neurons excited by acetylcholine cluster in Layers III and V, while noradrenaline-inhibited neurons concentrate in Layers III and IV, and dopamine sensitivity is highest in the deep layers.13PubMed. Laminar distributions of neurons sensitive to acetylcholine, noradrenaline and dopamine in the dorsolateral prefrontal cortex of the monkey These patterns mean that a burst of dopamine or noradrenaline does not affect all layers equally; it selectively alters processing at specific stages in the circuit. This laminar-specific neurochemistry helps explain why drugs that affect a single neurotransmitter system can have such varied cognitive and perceptual effects.
How the Layers Form During Development
The six-layer architecture is assembled during fetal development through a process called inside-out radial migration. New excitatory neurons are born from progenitor cells near the brain’s inner ventricles and migrate outward along the fibers of their parent cells. Each successive wave of newborn neurons migrates past the ones that arrived before it and settles at the top of the developing cortical plate. The result is that the deepest layers, V and VI, are the oldest, while the superficial layers, II and III, form last.14PubMed Central. Inside-Out Radial Migration Facilitates Lineage-Dependent Neocortical Microcircuit Assembly
A key molecular player in this process is Reelin, a protein secreted by cells in the outermost marginal zone. When a migrating neuron’s leading process reaches the marginal zone, Reelin triggers it to detach from its guiding glial fiber and pull its cell body to the top of the cortical plate. Without functional Reelin, as in the classic “Reeler” mutant mouse, layers form in roughly inverted order, and the cortex is severely disorganized.15Trends in Neurosciences. A unified model for inside-out neocortical lamination The precision of this layering matters because neurons that arrive in the same layer tend to wire together preferentially, so birth order directly shapes the circuitry that the adult cortex depends on.
Not Every Brain Region Has All Six Layers
The six-layer scheme is best thought of as a template. Different cortical regions modify it substantially. The primary visual cortex is “granular,” with a thick, subdivided Layer IV packed with small cells optimized for receiving dense thalamic input. The primary motor cortex is “agranular,” with Layer IV so reduced that it is nearly invisible and Layer V massively expanded to house the large output neurons. Between these extremes, most cortical regions fall on a gradient, sometimes called a cortical type spectrum, from agranular through dysgranular to fully granular.16Frontiers in Neuroanatomy. A Protocol for Cortical Type Analysis of the Human Neocortex Applied on Histological Samples, the Atlas of Von Economo and Koskinas, and Magnetic Resonance Imaging
This variation is not random. It reflects each region’s computational priorities. Areas that receive heavy sensory input from the thalamus need a robust Layer IV to handle the incoming traffic. Areas that prioritize output, like motor regions, invest in Layer V instead. Association cortex, which does higher-order integration rather than direct sensory or motor work, tends to land in the middle of this spectrum. Modern spatial transcriptomic techniques have begun to reveal that the molecular boundaries between layers do not always line up perfectly with the borders visible under a microscope, adding another dimension of complexity to the classic six-layer picture.17PubMed Central. A data-driven single-cell and spatial transcriptomic map of the human prefrontal cortex
Why the Upper Layers Are a Mammalian Innovation
Reptiles and birds have brain structures that perform cortex-like functions, but they lack the six-layer laminar arrangement seen in mammals. Comparative neuroanatomy suggests that the reptilian dorsal cortex is equivalent only to the deep layers (V and VI) of the mammalian neocortex, and that the upper layers, II and III in particular, are an evolutionary addition unique to mammals.18PubMed Central. A hypothesis for the evolution of the upper layers of the neocortex through co-option of the olfactory cortex developmental program Birds solve some of the same computational problems using clustered nuclei rather than stacked layers, and the molecular signatures of these nuclei show striking similarities to those of individual mammalian cortical layers, hinting at a shared deep ancestry even though the physical architecture diverged.19PubMed Central. Vertebrate brains and evolutionary connectomics: on the origins of the mammalian ‘neocortex’
The expansion of Layers II and III in mammalian evolution, and their further enlargement in primates and especially humans, appears to be linked to the explosion of cortico-cortical connectivity that supports complex cognition. This evolutionary perspective helps explain why the supragranular layers are so much more prominent in humans than in rodents, and why they contain cell types and circuit motifs that have no clear counterpart in smaller-brained species.
When Specific Layers Are Vulnerable to Disease
Because each layer has different cell types and connection patterns, neurological diseases tend to strike particular layers harder than others. In schizophrenia, postmortem studies consistently find reduced density of dendritic spines on Layer III pyramidal neurons, the very cells responsible for cortico-cortical communication. A recent study quantified this across multiple cortical regions and found spine density reductions ranging from about 13 percent in the prefrontal cortex to 38 percent for large spines in the posterior parietal cortex.20PubMed Central. Regional Specificity of Cortical Layer 3 Dendritic Spine Deficits in Schizophrenia Since these spines are where synapses form, fewer spines mean weaker connections between cortical areas, which aligns with the disconnection theories of schizophrenia.
Alzheimer’s disease offers a different pattern. The entorhinal cortex, one of the earliest regions affected, shows particular vulnerability in its Layer II neurons. These are Reelin-expressing cells, the same protein critical for building the cortex during development. In both human tissue and transgenic rat models, intracellular amyloid accumulates selectively in these Reelin-positive Layer II neurons during the earliest pre-plaque stages of Alzheimer’s pathology.21PubMed. Reelin-immunoreactive neurons in entorhinal cortex layer II selectively express intracellular amyloid in early Alzheimer’s disease In ALS, the damage is concentrated in Layer V: the Betz cells of the motor cortex undergo profound apical dendrite degeneration, a pathology not seen in Alzheimer’s patients or healthy controls even though the same brain region is examined.22PubMed Central. Apical dendrite degeneration, a novel cellular pathology for Betz cells in ALS Each disease, in other words, targets the layer whose circuitry is most relevant to the symptoms it produces.
Plasticity Is Not the Same in Every Layer
The brain’s ability to rewire in response to experience also has a laminar dimension. In developing animals, Layer IV is famously plastic during early “critical periods,” when sensory experience shapes thalamic connections. After the critical period closes, those synapses become harder to modify. But research has shown that peripheral sensory deprivation, such as removing whisker input in rodents, can reopen critical-period-like plasticity in the mature Layer IV barrel cortex, restoring features of the early developmental state to adult thalamocortical synapses.23Cell Reports. Peripheral Sensory Deprivation Restores Critical-Period-like Plasticity to Adult Somatosensory Thalamocortical Inputs
The supragranular layers follow different rules. Under normal adult conditions, sensory deprivation does not trigger much structural change in Layer II/III spine dynamics. However, mice engineered to carry a single functional copy of the gene SRGAP2, which constrains synapse density, show a striking increase in dendritic spine formation in Layer II/III after whisker trimming, well into adulthood.24PubMed Central. SRGAP2 limits experience-dependent structural synaptic plasticity in adult cortical circuits This finding suggests that the mammalian cortex has built-in molecular brakes on plasticity that are layer-specific, and that loosening those brakes in the upper layers could, in principle, extend the brain’s capacity to remodel connections in response to new experiences. Understanding which layers retain plasticity, and which molecular switches control it, is one of the more promising threads in current neuroscience for thinking about recovery from brain injury and neurodevelopmental disorders.