Skip to content
Robotic Brain Universe

The cerebral cortex and its maps

What the folded outer layer of the brain is made of, why it folds, and how scientists map it from Brodmann's microscope to petabyte connectomes.

Introductory · about 11 min · updated 2026-10-02 · awaiting clinical review

Illustrative simulation excitatory inhibitory

The cortex is a folded, layered sheet that holds a fifth of the brain's neurons. This reading explains gyri, sulci, layers, columns and areas; the Brodmann, Desikan–Killiany, Destrieux and HCP maps; the radial unit hypothesis; the universal folding law and how thickness is measured on MRI; malformations and Alzheimer's thinning; and new maps of genes, cell types and synapses.

Contents
  1. The thinking surface
  2. What the cortex is
  3. Why it folds, and why maps matter
  4. How the cortex is built and organised
  5. When the cortex matures and ages
  6. When the cortex goes wrong
  7. The geometry of a folded sheet
  8. Mapping at every scale
  9. A century of maps
  10. Frontiers: wiring diagrams
  11. Check yourself

The thinking surface

The wrinkled outer layer of the brain, the cerebral cortex, makes up 82% of the brain's mass but holds only 19% of its neurons. Within this folded sheet lie the areas for seeing, hearing, touch, movement, language and planning.[1]

Mapping it has taken more than a century. In 1909 Korbinian Brodmann divided the cortex into numbered areas by looking at its cells under the microscope. In 2016, combining several kinds of MRI from 210 young adults, the Human Connectome Project delineated 180 areas in each hemisphere, 97 of them new.[2,3]

The 3D model in this atlas carries three of these maps. This reading explains what they show, how the cortex is built and why it folds, how maps are made from MRI, and how new technologies are now charting the cortex down to single synapses.[4,5,6]

What the cortex is

The cortex is a folded sheet of grey matter: its ridges are gyri and its grooves are sulci. Most of it is neocortex, which Brodmann described as built on a basic plan of six layers; across the neocortex, the same basic laminar and tangential organisation of excitatory neurons is found wherever it has been looked for.[2,7]

A cortical area is a patch of cortex that differs from its neighbours. Glasser and colleagues drew area borders wherever there were sharp changes in cortical architecture, function, connectivity or topography, the four classic criteria. Areas with fewer distinguishable layers, such as many limbic areas, contrast with the complex lamination of sensory and association cortex.[3,8]

Three ways to map the cortex[2,3,4,5,6]
MapBased onIn this atlas
Brodmann areasCell architecture seen under the microscope after death12 areas, predicted on the template from its folding pattern
Desikan–KillianyGyral anatomy, labelled automatically on MRI34 regions per hemisphere
DestrieuxGyri and sulci with standard anatomical names74 regions per hemisphere
HCP multimodal (not shown)Architecture, function, connectivity and topography from MRI180 areas per hemisphere

The cortex in numbers

Share of the brain's mass in the cerebral cortex
82%[1]
Share of the brain's neurons in the cerebral cortex
19%[1]
Areas per hemisphere in the HCP multimodal map
180 (97 newly described)[3]
Cells in one cubic millimetre of human temporal cortex
about 57,000[9]
Synapses in the same cubic millimetre
about 150 million[9]

Why it folds, and why maps matter

Larger brains tend to have more folded cortex, but the number of neurons does not decide how much a cortex folds. Mota and Herculano-Houzel found that across species with smooth and folded brains alike, and across individuals, the degree of folding follows one function of the cortex's surface area and the square root of its thickness, as a simple physical model predicts. The same law describes how a ball of crumpled paper folds.[10]

Precise maps matter because function is local. A shared parcellation lets studies of development, ageing and disease compare like with like, and Glasser's machine-learning classifier can find the same areas in a new person's scan, even when that person's areas are arranged atypically.[3]

Maps also line up with genes. In a transcriptional atlas of about 900 brain subdivisions, the neocortex looked relatively uniform but with distinct features in primary sensorimotor cortex, and the closer two cortical regions lay to each other, the more similar their gene expression.[11]

How the cortex is built and organised

Columns. Recording single neurons in the cat's somatosensory cortex, Vernon Mountcastle found that cells stacked vertically through the depth of the cortex respond to the same kind of stimulus from the same place on the body, the first evidence that the cortex is organised in columns.[12]

Built from the inside out. Pasko Rakic's radial unit hypothesis explains how. Proliferating cells lining the embryonic ventricles form units that act as a proto-map of the future areas; their offspring climb along radial glial fibres into the growing cortex as ontogenetic columns, and input arriving from elsewhere can then change how many columns an area gets.[13]

A repeated circuit. Because the same basic circuit recurs across the neocortex, Douglas and Martin proposed a canonical model: neurons in the superficial layers of a local patch cooperate to weigh possible interpretations of their inputs and select one consistent with all of them.[7]

How a cortical area is built (radial unit hypothesis)Proliferative unitslining the embryonic ventricle;a proto-map of areasRadial glial guidesnew neurons migrate outwardsalong themOntogenetic columnsone per unit, in the growingcortexCytoarchitectonic areaits final size set by itscolumnsAfferent inputconnections arriving fromelsewheremodifies the number
How a cortical area is built (radial unit hypothesis). In the radial unit hypothesis, the layout of areas is first specified near the ventricles and carried outwards by migrating neurons; incoming connections then adjust how large each area becomes.[13]
Text version of the diagram
  1. Proliferative units: lining the embryonic ventricle; a proto-map of areas. Leads to Radial glial guides.
  2. Radial glial guides: new neurons migrate outwards along them. Leads to Ontogenetic columns.
  3. Ontogenetic columns: one per unit, in the growing cortex. Leads to Cytoarchitectonic area.
  4. Cytoarchitectonic area: its final size set by its columns.
  5. Afferent input: connections arriving from elsewhere. Leads to Ontogenetic columns (modifies the number).

Maps from MRI. Software such as FreeSurfer reconstructs the boundary between white and grey matter and the outer (pial) surface of the cortex from an MRI scan, inflates and registers it to a common spherical coordinate system, and labels it with an atlas. Measuring cortical thickness this way, automatically across the whole brain, replaced manual measurements that took a trained anatomist several days.[14,15,16]

When the cortex matures and ages

Synapses in the cortex form before birth and are overproduced in infancy, then pruned through childhood, earlier in auditory cortex than in prefrontal cortex, where pruning lasts into mid-adolescence.[17]

Cortical thickness also changes on a schedule. From 764 MRI scans of 375 children and young adults, Shaw and colleagues found simpler growth trajectories in areas with simple layering, such as most limbic areas, and the most complex trajectories in the polysensory and high-order association areas; the age at which a region reaches its peak thickness differs from region to region, so maturation moves across the cortex over time.[8]

In adulthood every lobe ages at a similar rate in its folding parameters, and in over 1,500 healthy people different regions of the same cortex obeyed the same folding law.[18]

When the cortex goes wrong

Built wrongly. Malformations of cortical development, in which neurons proliferate, migrate or organise abnormally, are common causes of neurodevelopmental delay and epilepsy. Their classification has been revised repeatedly as genetics and imaging have revealed new types.[19]

Too smooth. In lissencephaly ('smooth brain') the human cortex fails to fold normally. The folding law places smooth and folded cortices on one continuum, which may help explain how such pathologies arise.[10,19]

Thinning with disease. Alzheimer's disease thins the cortex in a reliable pattern, a 'cortical signature' that follows the regions known to be vulnerable to the disease's pathology. The thinning tracks symptom severity from the earliest stages and is already subtly present in older people without symptoms who have amyloid in the brain.[20]

The geometry of a folded sheet

A handful of simple quantities describe the cortex's shape: its total area, the area of a tight wrapping around it, and its thickness.[10,16]

Folding index[10,18]
g=AtAeg = \frac{A_t}{A_e}

How much more cortex there is than would be needed to cover the brain's outer envelope. A smooth cortex has gg close to 1; the more deeply folded, the larger gg.

Symbols in Folding index
SymbolMeaningUnit
AtA_ttotal area of the cortical surface, including the walls of every sulcusmm²
AeA_eexposed area: the area of a smooth envelope wrapped around the cortexmm²
Universal folding law[10,18]
At T=k Ae α,α=54A_t\,\sqrt{T} = k\,A_e^{\,\alpha}, \qquad \alpha = \tfrac{5}{4}

Total area times the square root of thickness grows as the exposed area to the power 5/4. The exponent comes from minimising an effective free energy in a simple physical model and fits data across mammalian species, across people and across the lobes of one brain; only the offset kk varies, and it decreases with age.

Symbols in Universal folding law
SymbolMeaningUnit
TTaverage cortical thicknessmm
kkoffset, a dimensionless constant related to the pressure term in the model—
α\alphascaling exponent, predicted as 5/4 = 1.25—
Cortical thickness at a point[16]
T(v)=12[ d(v, pial)+d(p∗(v), white)]T(v) = \tfrac{1}{2}\left[\, d\big(v,\ \text{pial}\big) + d\big(p^{*}(v),\ \text{white}\big) \right]

FreeSurfer measures thickness at each vertex vv of the white-matter surface as the average of two distances: from vv to the nearest point p∗p^{*} on the pial surface, and from that point back to the nearest point on the white surface. Using both directions makes the measure symmetric.

Symbols in Cortical thickness at a point
SymbolMeaningUnit
vva vertex on the white-matter surface—
p∗(v)p^{*}(v)the closest point to v on the pial surface—
d(⋅,⋅)d(\cdot,\cdot)shortest distance between a point and a surfacemm
Dice coefficient (how well two maps agree)[21]
D=2 ∣X∩Y∣∣X∣+∣Y∣D = \frac{2\,|X \cap Y|}{|X| + |Y|}

Compares two labellings of the same region, for example an automatic atlas and an expert's drawing: 1 means identical, 0 means no overlap. The measure was introduced in ecology in 1945 to compare where two species occur.

Symbols in Dice coefficient (how well two maps agree)
SymbolMeaningUnit
X,YX, Ythe two sets of labelled points (vertices or voxels)—
∣X∩Y∣|X \cap Y|the number of points labelled the same in both—

Mapping at every scale

From MRI to a parcellation. The HCP map combined several kinds of MRI from each person with a semi-automated method to find borders. A machine-learning classifier then learned each area's multi-modal 'fingerprint' and detected 96.6% of the areas in new people.[3]

A brain at near-cellular resolution. BigBrain reconstructed an entire human brain in 3D at 20 micrometres from 7,404 histological sections, allowing microscopic detail to be extracted anywhere and redefining classical maps such as Brodmann's.[22]

Genes and cell types. The Allen Human Brain Atlas profiled gene expression in about 900 precisely dissected subdivisions of two brains. A later census sequenced more than three million cell nuclei from about 100 dissections of three donors and found 461 clusters and 3,313 subclusters of cell types.[11,23]

From a whole brain to a single synapse[3,9,11,22,23]
ScaleMethodExample
Whole cortex, millimetresMulti-modal MRI and machine learningHCP map, 180 areas per hemisphere
Whole brain, 20 µm3D reconstruction of histological sectionsBigBrain
Molecules and cell typesGene expression and single-nucleus sequencingAllen atlas; human brain cell census
One cubic millimetre, nanometresElectron microscopy and computational reconstructionH01 human cortex fragment

A century of maps

Milestones

  1. 1909Brodmann publishes his cytoarchitectonic map of numbered cortical areas.[2]
  2. 1945Dice introduces the overlap measure now used to compare brain maps.[21]
  3. 1957Mountcastle's recordings reveal columnar organisation in somatosensory cortex.[12]
  4. 1988Rakic sets out the radial unit hypothesis of how areas form.[13]
  5. 1999A surface-based coordinate system lets cortices be averaged across people (fsaverage).[15]
  6. 2000Automated measurement of cortical thickness across the whole brain.[16]
  7. 2006The Desikan–Killiany atlas labels gyral regions automatically.[4]
  8. 2010The Destrieux atlas labels gyri and sulci with standard names.[5]
  9. 2012An anatomically comprehensive atlas of gene expression in the adult human brain.[11]
  10. 2013BigBrain: a whole human brain at 20 µm.[22]
  11. 2015A universal law of cortical folding.[10]
  12. 2016The HCP multi-modal parcellation: 180 areas per hemisphere.[3]
  13. 2024A cubic millimetre of human cortex reconstructed at nanoscale.[9]
  14. 2025MICrONS links function and wiring for about 75,000 neurons of mouse visual cortex.[24]

Frontiers: wiring diagrams

A cubic millimetre of human temporal cortex, removed during epilepsy surgery to reach the focus beneath it, has been reconstructed at nanoscale resolution: about 57,000 cells, 230 millimetres of blood vessels and 150 million synapses, in 1.4 petabytes of data. Glia outnumbered neurons two to one, and among thousands of weak connections onto each neuron were rare, powerful inputs of up to 50 synapses from a single axon.[9]

In the mouse, the MICrONS project imaged the activity of about 75,000 neurons in visual cortex while the animal watched natural and synthetic scenes, then reconstructed the same tissue by electron microscopy, more than 200,000 cells and half a billion synapses, linking what neurons do to how they are wired.[24]

Check yourself

Check yourself

  1. What share of the brain's mass and of its neurons is in the cerebral cortex?
    Show answer

    About 82% of the mass but only about 19% of the neurons.

  2. What four kinds of change did Glasser and colleagues use to draw area borders?
    Show answer

    Changes in cortical architecture, function, connectivity and topography.

  3. What does the Destrieux atlas label that the Desikan–Killiany atlas does not?
    Show answer

    Sulci as well as gyri (74 regions per hemisphere against 34 mainly gyral regions).

  4. According to the folding law, what decides how much a cortex folds?
    Show answer

    Its surface area and the square root of its thickness, not its number of neurons.

  5. What is the radial unit hypothesis?
    Show answer

    Proliferative units near the embryonic ventricles form a proto-map; their neurons migrate outwards along radial glia as columns that build the areas.

  6. How does Alzheimer's disease show up in cortical thickness?
    Show answer

    As a regional pattern of thinning that tracks symptom severity and is subtly present even in amyloid-positive people without symptoms.

  7. What does a Dice coefficient of 1 mean?
    Show answer

    The two labellings being compared are identical.

Glossary[2,3,9,10,12,13]

Cerebral cortex
The folded outer sheet of grey matter of the cerebral hemispheres.
Neocortex
The six-layered type of cortex that makes up most of the human cerebral cortex.
Gyrus and sulcus
A ridge and a groove of the folded cortical surface.
Cortical area
A region distinguished from its neighbours by architecture, function, connectivity or topography.
Cytoarchitecture
The arrangement, size and density of cells in the cortical layers, as seen under the microscope.
Parcellation
A division of the cortex into labelled areas or regions.
Cortical column
A vertical group of neurons through the cortical depth that share response properties.
Radial glia
Embryonic support cells whose long fibres guide newborn neurons to the cortex.
Pial surface
The outer surface of the cortex, just under the pia mater.
Folding index
Total cortical area divided by the area of its smooth outer envelope.
Connectome
A map of the connections between neurons or brain regions.

References

  1. Azevedo FAC, Carvalho LRB, Grinberg LT, Farfel JM, Ferretti REL, Leite REP, Jacob Filho W, Lent R, Herculano-Houzel S. Equal numbers of neuronal and nonneuronal cells make the human brain an isometrically scaled-up primate brain. Journal of Comparative Neurology 2009;513(5):532-541. doi:10.1002/cne.21974
  2. Brodmann K, Garey LJ (translator). Brodmann's Localisation in the Cerebral Cortex. Springer (English translation of the 1909 original) 2005. doi:10.1007/b138298
  3. Glasser MF, Coalson TS, Robinson EC, Hacker CD, Harwell J, Yacoub E, et al.. A multi-modal parcellation of human cerebral cortex. Nature 2016;536(7615):171-178. doi:10.1038/nature18933
  4. Desikan RS, Ségonne F, Fischl B, Quinn BT, Dickerson BC, Blacker D, et al.. An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. NeuroImage 2006;31(3):968-980. doi:10.1016/j.neuroimage.2006.01.021
  5. Destrieux C, Fischl B, Dale A, Halgren E. Automatic parcellation of human cortical gyri and sulci using standard anatomical nomenclature. NeuroImage 2010;53(1):1-15. doi:10.1016/j.neuroimage.2010.06.010
  6. Fischl B, Rajendran N, Busa E, Augustinack J, Hinds O, Yeo BTT, et al.. Cortical Folding Patterns and Predicting Cytoarchitecture. Cerebral Cortex 2008;18(8):1973-1980. doi:10.1093/cercor/bhm225
  7. Douglas RJ, Martin KAC. Neuronal circuits of the neocortex. Annual Review of Neuroscience 2004;27:419-451. doi:10.1146/annurev.neuro.27.070203.144152
  8. Shaw P, Kabani NJ, Lerch JP, Eckstrand K, Lenroot R, Gogtay N, et al.. Neurodevelopmental trajectories of the human cerebral cortex. The Journal of Neuroscience 2008;28(14):3586-3594. doi:10.1523/JNEUROSCI.5309-07.2008
  9. Shapson-Coe A, Januszewski M, Berger DR, Pope A, Wu Y, Blakely T, et al.. A petavoxel fragment of human cerebral cortex reconstructed at nanoscale resolution. Science 2024;384(6696):eadk4858. doi:10.1126/science.adk4858
  10. Mota B, Herculano-Houzel S. Cortical folding scales universally with surface area and thickness, not number of neurons. Science 2015;349(6243):74-77. doi:10.1126/science.aaa9101
  11. Hawrylycz MJ, Lein ES, Guillozet-Bongaarts AL, Shen EH, Ng L, Miller JA, et al.. An anatomically comprehensive atlas of the adult human brain transcriptome. Nature 2012;489(7416):391-399. doi:10.1038/nature11405
  12. Mountcastle VB. Modality and topographic properties of single neurons of cat's somatic sensory cortex. Journal of Neurophysiology 1957;20(4):408-434. doi:10.1152/jn.1957.20.4.408
  13. Rakic P. Specification of cerebral cortical areas. Science 1988;241(4862):170-176. doi:10.1126/science.3291116
  14. Fischl B. FreeSurfer. NeuroImage 2012;62(2):774-781. doi:10.1016/j.neuroimage.2012.01.021
  15. Fischl B, Sereno MI, Tootell RBH, Dale AM. High-resolution intersubject averaging and a coordinate system for the cortical surface. Human Brain Mapping 1999;8(4):272-284. doi:10.1002/(SICI)1097-0193(1999)8:4<272::AID-HBM10>3.0.CO;2-4
  16. Fischl B, Dale AM. Measuring the thickness of the human cerebral cortex from magnetic resonance images. Proceedings of the National Academy of Sciences of the USA 2000;97(20):11050-11055. doi:10.1073/pnas.200033797
  17. Huttenlocher PR, Dabholkar AS. Regional differences in synaptogenesis in human cerebral cortex. The Journal of Comparative Neurology 1997;387(2):167-178. doi:10.1002/(SICI)1096-9861(19971020)387:2<167::AID-CNE1>3.0.CO;2-Z
  18. Wang Y, Necus J, Rodriguez LP, Taylor PN, Mota B. Human cortical folding across regions within individual brains follows universal scaling law. Communications Biology 2019;2:191. doi:10.1038/s42003-019-0421-7
  19. Barkovich AJ, Guerrini R, Kuzniecky RI, Jackson GD, Dobyns WB. A developmental and genetic classification for malformations of cortical development: update 2012. Brain 2012;135(5):1348-1369. doi:10.1093/brain/aws019
  20. Dickerson BC, Bakkour A, Salat DH, Feczko E, Pacheco J, Greve DN, et al.. The cortical signature of Alzheimer's disease: regionally specific cortical thinning relates to symptom severity in very mild to mild AD dementia and is detectable in asymptomatic amyloid-positive individuals. Cerebral Cortex 2009;19(3):497-510. doi:10.1093/cercor/bhn113
  21. Dice LR. Measures of the amount of ecologic association between species. Ecology 1945;26(3):297-302. doi:10.2307/1932409
  22. Amunts K, Lepage C, Borgeat L, Mohlberg H, Dickscheid T, Rousseau MÉ, et al.. BigBrain: an ultrahigh-resolution 3D human brain model. Science 2013;340(6139):1472-1475. doi:10.1126/science.1235381
  23. Siletti K, Hodge R, Mossi Albiach A, Lee KW, Ding SL, Hu L, et al.. Transcriptomic diversity of cell types across the adult human brain. Science 2023;382(6667):eadd7046. doi:10.1126/science.add7046
  24. The MICrONS Consortium, Bae JA, Baptiste M, et al.. Functional connectomics spanning multiple areas of mouse visual cortex. Nature 2025;640(8058):435-447. doi:10.1038/s41586-025-08790-w

Related readings

  • Neurons and electrical signals

    How a nerve cell makes a 100-millivolt spike, the equations that describe it, and why one neuron may be a deep network in disguise.

    Introductory

  • Neural networks, biological and artificial

    How brains and machines learn from experience, the maths both share, and why AI and neuroscience keep borrowing from each other.

    Intermediate

  • How the brain is organised

    A map of the whole brain: its divisions, cells and energy budget, how it develops from a tube, how its network is wired, and how it changes over a lifetime.

    Introductory

  • Mapping the brain: the breakthroughs

    From Golgi and Brodmann to whole-brain connectomes, cell atlases and AI foundation models: how the brain is mapped and where mapping is heading.

    Introductory

Template anatomy for education. Not patient-specific. Not for clinical decision-making.