Primary Visual Cortex

Abstract

The primary visual cortex (V1) is the first cortical stage of vision and the single largest visual area in the brain. It lines the walls of the calcarine sulcus within the occipital lobe, where MeSH files it under the visual cortex, and it is also called the striate cortex for the stria of Gennari in layer 4. V1 receives the retina's output relayed through the lateral geniculate nucleus of the thalamus and re-represents it as an orderly, greatly magnified map of the contralateral visual field. Its neurons are the first in the pathway tuned to the orientation of an edge, and they are packed into orientation and ocular-dominance columns that tile the cortical surface. The discovery of this receptive-field architecture, and of its dependence on balanced visual experience during a critical period, defined modern visual neuroscience.

Keywords: primary visual cortex, striate cortex, retinotopy, orientation columns, ocular dominance

The primary visual cortex is where sight becomes cortical. Every conscious visual experience of shape, edge and texture begins with the transformation V1 performs on the thalamic signal, and almost everything the rest of the visual brain does is built on the map and the feature code it lays down. This article follows V1 from its laminar wiring and its retinotopic map, through the receptive fields and columns that make it the first orientation-analysing stage of vision, to the experience-dependent plasticity of its ocular-dominance columns and the syndromes — hemianopia and blindsight — that expose what it does and does not contribute to conscious sight.

## What the Primary Visual Cortex Is

The primary visual cortex is the cortical destination of the main visual pathway and the anatomical starting point of cortical vision. It occupies the occipital pole and the banks of the calcarine sulcus on the medial surface of each hemisphere, so most of it, like the sulcus itself, is hidden from a lateral view of the brain. Its classical names record two of its properties: striate cortex, because layer 4 carries the stria of Gennari, a myelinated band dense enough to see with the unaided eye; and Brodmann area 17, its number in the cytoarchitectonic map of the cortex. In the Medical Subject Headings vocabulary it is filed under the visual cortex, itself a subdivision of the occipital lobe.

Signals arrive by an orderly relay. The retina projects through the optic nerve and chiasm to the lateral geniculate nucleus of the thalamus, whose axons form the optic radiation and terminate densely in layer 4C of V1. Because the fibres from the two eyes cross partially at the chiasm, each V1 represents the contralateral half of the visual field seen by both eyes, and because the calcarine sulcus divides the map, the upper field is represented on its lower bank and the lower field on its upper bank. Gordon Holmes charted this correspondence in human cortex from the localised occipital wounds of the First World War, and Horton and Hoyt later revised his map with modern imaging, sharpening the estimate of how much cortex the central field commands (#ref-horton-hoyt-1991).

Figure 1

The primary visual cortex folded into the calcarine sulcus on the medial surface of the occipital lobe.

## Laminar Structure and Thalamic Input

V1 has the six-layered structure of all neocortex, but layer 4 is elaborated as nowhere else, which is why the layer is so conspicuous in stained sections. Geniculate axons carrying the two great retinal channels terminate in separate sublayers: the magnocellular stream, fast and achromatic, ends in layer 4Cα, and the parvocellular stream, slower and colour-opponent, ends in layer 4Cβ. From layer 4 the signal is distributed vertically to the other layers — layers 2 and 3 project onward to the extrastriate areas, layer 5 to the superior colliculus and pulvinar, and layer 6 back to the lateral geniculate nucleus, closing a corticothalamic loop.

This vertical wiring is the reason V1 is best understood as a stack of columns rather than a set of sheets. A signal entering layer 4C at a given point is processed through the full depth of cortex at that point before it is passed on, so the elementary unit of V1 is a narrow column spanning all six layers. The receptive-field properties and the columnar architecture described below are consequences of this laminar plan: orientation tuning is largely absent in the geniculate input to layer 4C and is constructed within the column, and the segregation of the two eyes established in layer 4 is elaborated into the ocular-dominance columns visible at the surface.

Table 1. The layers of V1 and their principal connections.

## Retinotopy and Cortical Magnification

V1 carries a retinotopic map: neighbouring points in the visual field drive neighbouring points on the cortical sheet, so the surface holds a distorted image of the retina. The distortion is systematic and severe. The central few degrees imaged on the fovea command a share of cortex out of all proportion to their angular size — cortical magnification — so that roughly half of V1 is devoted to the central ten degrees of a field that spans more than a hundred. Horton and Hoyt expressed this quantitatively for human V1, giving the millimetres of cortex per degree as a simple function of eccentricity (#ref-horton-hoyt-1991).

Functional imaging turned the map into a measurement. Sereno and colleagues used phase-encoded stimuli to reveal the borders of V1 and the neighbouring areas in the living human brain, each a mirror-reversed hemifield representation (#ref-sereno-1995); Wandell and colleagues catalogued the resulting field maps and their gradients (#ref-wandell-2007), and imaging them became routine (#ref-wandell-winawer-2011). Dumoulin and Wandell introduced the population receptive field method, estimating for each voxel the region of visual field that drives it and so recovering both the retinotopic map and the size of receptive fields as a function of eccentricity (#ref-dumoulin-wandell-2008). Probabilistic atlases now give each cortical location a likelihood of belonging to V1 across individuals (#ref-wang-2015), and template models predict a person's V1 retinotopy from anatomy alone (#ref-benson-winawer-2018), a prediction shown to be reproducible across independent datasets (#ref-himmelberg-2022).

## Receptive Fields and Orientation Columns

The retinotopic address tells only where a V1 neuron looks; the transformation V1 performs is revealed by what it responds to. Recording from single cells in the cat, Hubel and Wiesel found that most neurons in the striate cortex ignore diffuse light and fire instead to an edge or bar of a particular orientation crossing a small region of the field — the cell's receptive field (#ref-hubel-wiesel-1962). They distinguished simple cells, which require an edge in a fixed position and phase, from complex cells, which respond to the preferred orientation anywhere within their field, and they proposed that complex-cell selectivity is built by combining simple-cell inputs. The same organisation holds in the primate striate cortex (#ref-hubel-wiesel-1968).

The selectivity is laid out in orderly columns. As an electrode passes tangentially through V1, the preferred orientation shifts smoothly, cycling through the full range roughly every millimetre in an orientation column; interleaved with these are ocular-dominance columns favouring one eye or the other. A patch of cortex holding a full cycle of orientations for both eyes and covering one point of the visual field is a functional module Hubel and Wiesel called a hypercolumn, and in their Ferrier Lecture they set out this architecture as the general plan of the macaque striate cortex (#ref-hubel-wiesel-1977). This columnar order, recovered one cell at a time, is the microscopic counterpart of the retinotopic map. That V1 performs a specific analysis rather than relaying a picture is also clear downstream: the extrastriate areas it feeds are functionally specialised, one for colour and another for motion, as Zeki demonstrated in the monkey (#ref-zeki-1978), and V1's output is split into the ventral and dorsal streams for perception and for action (#ref-goodale-milner-1992).

## Ocular Dominance and the Critical Period

The ocular-dominance columns are the site of one of the clearest demonstrations of experience-dependent plasticity in the brain. In the normal animal the stripes serving the two eyes are of roughly equal width. Wiesel and Hubel closed one eye of a kitten for the first months of life and found that the deprived eye lost its ability to drive cortical cells almost entirely, while its columns shrank and the open eye's columns expanded to occupy the vacated territory (#ref-wiesel-hubel-1963). The change was not a simple consequence of disuse: it occurred only when deprivation fell within an early critical period, and closing the same eye in an adult had little effect.

The mechanism is competitive. The two eyes' inputs vie for cortical territory, and the balance is set by correlated activity during the sensitive window; depriving one eye tips the competition permanently in favour of the other. This is the neural basis of amblyopia, the reduced vision that follows an untreated squint or cataract in early childhood, and it established the general principle that cortical circuits are wired by patterned experience during defined developmental windows. The finding was central to the work for which Hubel and Wiesel shared the 1981 Nobel Prize.

## Damage: Hemianopia and Blindsight

Because V1 maps the visual field so faithfully, destroying it removes a corresponding region of sight, and the shape of the loss betrays the site of the lesion. Destruction of one entire striate cortex produces a homonymous hemianopia — blindness in the contralateral half of the field of both eyes. Damage confined to one bank of the calcarine sulcus spares the other and yields a quadrantanopia, and a small lesion carves out a scotoma at the retinotopic location the damaged patch represented. Central vision is frequently spared, a sign called macular sparing, in part because the large foveal representation at the occipital pole enjoys a dual blood supply.

The residual vision of the cortically blind field is among the most instructive findings in the field. Weiskrantz and colleagues showed that a patient with a restricted striate lesion could locate and discriminate stimuli in the blind hemifield well above chance while reporting no awareness of them — blindsight — implying that visual information reaches action systems by routes that bypass V1 (#ref-weiskrantz-1974). The phenomenon bears directly on what V1 contributes to conscious sight: Ffytche and Zeki argued that under some conditions the primary visual cortex, and feedback to it, are not strictly necessary for a conscious visual percept (#ref-ffytche-zeki-2011). Whether V1 activity is necessary for awareness, or merely usually accompanies it, remains contested.

## Worked Example

Cortical magnification can be made a single number. A widely used description of human V1, from Horton and Hoyt, gives the linear magnification factor — the millimetres of cortex per degree of visual angle at eccentricity E — in the inverse-linear form:

M(E) = A / (E + E₀)

with representative constants A = 17.3 mm·deg and E₀ = 0.75 deg. The cortical distance from the foveal representation out to eccentricity E is the integral of M:

d(E) = A · ln((E + E₀) / E₀)

Evaluate this at two and at twenty degrees. At E = 2°, d = 17.3 · ln(2.75 / 0.75) = 17.3 · ln(3.6667) = 17.3 · 1.2993 = 22.48 mm. At E = 20°, d = 17.3 · ln(20.75 / 0.75) = 17.3 · ln(27.6667) = 17.3 · 3.3200 = 57.44 mm. The central two degrees of the field therefore occupy about 22.48 mm of cortex, while the whole ring from two to twenty degrees occupies only 57.44 − 22.48 = 34.96 mm. The innermost two degrees — a tenth of the way out to twenty degrees — claim 22.48 / 57.44 ≈ 39 per cent of the cortical distance. That is cortical magnification in one figure: acuity is high at the fovea because so much of V1 is spent on it. The demonstration above computes d(E) from the same formula, so its readout agrees with the arithmetic here.

## Discussion

V1 is the region where the transparency of the visual brain is greatest. Its retinotopy makes the visual field legible on the cortical sheet, so a lesion's location can be read off a perimetry chart and a field defect predicted from a scan; its columnar architecture makes the elementary operation of cortical vision — the extraction of oriented edges — something that can be recorded, mapped and modelled. For these reasons V1 became the proving ground both of single-unit physiology and of functional imaging, and much of what is known about cortical microcircuitry in general was first established here.

Yet V1 is not merely an input stage. Its plasticity during the critical period shows that even the earliest cortical map is built by experience, not laid down fully by genes; its role in blindsight shows that the information it carries and the awareness that normally accompanies it can be dissociated. The open questions are correspondingly deep: what exactly V1 contributes to conscious perception as opposed to visual performance, how feedback from higher areas shapes its responses, and how its orderly map relates to the striking asymmetries in visual sensitivity around the field. V1 is thus both the best-understood piece of cortex and a region whose contribution to seeing is still argued over.

## Current Directions

The retinotopic map of V1 is now a computational, individualised object. Template models predict a person's field map from cortical folding alone, and Bayesian methods combine such templates with limited measurements to reconstruct a full map (#ref-benson-winawer-2018); the predictions have been shown to reproduce across independent datasets and scanners, putting the individual map on a firm empirical footing (#ref-himmelberg-2022). A second, active line concerns the asymmetries in the map. Sensitivity is not uniform around the visual field: performance is better along the horizontal than the vertical meridian and better in the lower than the upper field, and Benson and colleagues showed that the cortical magnification of V1 varies around the field in step with these behavioural performance fields (#ref-benson-2021). Himmelberg, Winawer and Carrasco have drawn the physiology and the psychophysics together, arguing that polar-angle asymmetries in perception are mirrored by asymmetries in the surface area and architecture of V1 itself (#ref-himmelberg-2023). The map, in other words, is being read not as a uniform grid but as a structure whose local distortions predict what an observer can and cannot see.

## Common Misconceptions

The primary visual cortex holds a picture of the world.

V1 does not represent an image; its neurons respond to oriented edges within small receptive fields, a decomposition of the retinal input rather than a copy of it. Seeing a picture would require a viewer, and there is none (#ref-hubel-wiesel-1962).

Each primary visual cortex serves one eye.

The division is by visual field, not by eye. Because of the optic chiasm, each V1 represents the contralateral half of the field seen by both eyes; the eyes are separated at a finer grain, in the ocular-dominance columns within V1 (#ref-hubel-wiesel-1968).

Losing V1 abolishes all vision in the affected field.

Some patients retain above-chance detection and localisation in a cortically blind field without conscious sight — blindsight — supported by visual pathways that bypass V1 (#ref-weiskrantz-1974).

## Glossary

Amblyopia.

Reduced vision in one eye from unbalanced binocular experience during the critical period, the clinical counterpart of ocular-dominance plasticity.

Blindsight.

Above-chance visual performance in a field made blind by V1 damage, in the absence of conscious visual experience.

Brodmann area 17.

The cytoarchitectonic designation of the primary visual cortex in Brodmann's map of the cortex.

Calcarine sulcus.

The deep horizontal fissure on the medial occipital surface along which V1 is folded.

Complex cell.

A V1 neuron tuned to edge orientation but responding to that orientation anywhere within its receptive field, independent of exact position.

Cortical magnification.

The over-representation of the central visual field on cortex; the millimetres of V1 per degree fall steeply with eccentricity.

Critical period.

An early developmental window during which cortical circuits, such as the ocular-dominance columns, are shaped by experience and after which they are relatively fixed.

Homonymous hemianopia.

Loss of the same contralateral half of the visual field in both eyes, typical of a complete unilateral V1 lesion.

Hypercolumn.

A patch of V1 containing a full cycle of orientation preferences for both eyes and covering one point of the visual field.

Lateral geniculate nucleus.

The thalamic relay that receives retinal input and projects, via the optic radiation, to layer 4C of V1.

Ocular-dominance column.

A stripe of V1 that responds preferentially to input from one eye; adjacent stripes alternate between the eyes.

Orientation column.

A column of V1 neurons sharing a preferred edge orientation; preferred orientation rotates smoothly across the surface.

Population receptive field.

The aggregate region of visual field driving a voxel or neuronal population, estimated from functional imaging to map retinotopy.

Primary visual cortex (V1).

The first cortical stage of vision, in the calcarine sulcus; also called striate cortex and Brodmann area 17.

Receptive field.

The region of the visual field within which a stimulus alters a neuron's firing.

Retinotopy.

The orderly mapping of visual-field position onto cortical position, preserved from the retina through V1 and beyond.

Simple cell.

A V1 neuron tuned to an edge of a particular orientation in a fixed position and phase within its receptive field.

Stria of Gennari.

The band of myelinated fibres in layer 4 that gives the striate cortex its name and is visible to the naked eye.

Striate cortex.

A synonym for V1, named for the stria of Gennari in its fourth layer.

## Key Researchers

Marisa Carrasco. Links V1 cortical architecture to visual performance, showing that cortical magnification around the visual field parallels behavioural performance fields and polar-angle asymmetries. Wikipedia - NYU profile - Google Scholar

David H. Hubel (1926–2013). With Wiesel, recorded from single cells in cat and monkey striate cortex, discovering simple and complex receptive fields and the orientation and ocular-dominance columns; shared the 1981 Nobel Prize in Physiology or Medicine. Wikipedia

Brian A. Wandell. Developed the population receptive field method and mapped human visual field maps with functional imaging; author of Foundations of Vision and a member of the National Academy of Sciences. Wikipedia - Stanford profile - Google Scholar

Torsten N. Wiesel (born 1924). Co-recipient of the 1981 Nobel Prize with Hubel; the monocular-deprivation experiments established the critical period for ocular dominance in V1. Wikipedia - Rockefeller faculty

Jonathan Winawer. Develops population-receptive-field and Bayesian methods for measuring human V1 retinotopy and cortical magnification. NYU profile - Lab - Google Scholar

Semir Zeki (born 1940). Mapped the functional specialisation of the extrastriate areas fed by V1 — colour in V4, motion in V5/MT — and argued that V1 is not strictly necessary for conscious vision. ORCID - Wikipedia - UCL profile

## Frequently Asked Questions

What is the primary visual cortex?
It is the first cortical stage of vision, also called V1, striate cortex or Brodmann area 17. It lines the calcarine sulcus in the occipital lobe and receives the retina's output relayed through the thalamus, re-representing it as a map of the visual field.

Why is V1 called the striate cortex?
Because layer 4 of V1 contains the stria of Gennari, a dense band of myelinated fibres that is visible to the naked eye in a sectioned brain. The stripe gives the region its name; no other cortical area shows it as clearly.

What is retinotopy?
The orderly mapping of visual-field position onto cortical position: neighbouring points in the field activate neighbouring points in V1. The central field is greatly over-represented, an effect called cortical magnification.

What do V1 neurons respond to?
Most V1 neurons respond best to an edge or bar of a particular orientation within a small receptive field, not to diffuse light. Hubel and Wiesel distinguished simple cells, tuned to a fixed edge position, from complex cells, tuned to orientation regardless of exact position.

What are ocular-dominance columns?
Alternating stripes of V1 that respond preferentially to one eye or the other. Their widths are set by balanced binocular experience during an early critical period; depriving one eye then shrinks its columns permanently.

What happens if the primary visual cortex is damaged?
Damage removes a region of vision whose shape reflects the lesion site. A complete unilateral lesion causes a homonymous hemianopia; damage to one bank of the calcarine sulcus causes a quadrantanopia. Central vision is often spared.

What is blindsight?
The ability of some people with V1 damage to detect or localise stimuli in their blind field above chance while reporting no conscious vision. It reflects visual pathways that reach the brain without passing through V1.

How is V1 different from the occipital lobe?
The occipital lobe is the whole posterior lobe of the cortex and contains many visual areas; V1 is the single first area within it, the cortical target of the thalamic visual relay and the source of the map the other areas build on.

References