Visual Fields

Abstract

MeSH classifies the visual field under visual perception, though it names not a kind of perception but the region of space seen by a steadily fixating eye. Its sensitivity is not uniform: vision is sharpest at the fovea and falls toward the periphery, tracing an island of vision whose slopes perimetry measures point by point and whose one true gap is the optic-disc blind spot. That falloff mirrors the retina, where photoreceptor and ganglion-cell density peak centrally, and it is inherited by the cortex, whose retinotopic maps devote disproportionate tissue to the central field. Characteristic patterns of field loss localize disease along the visual pathway, which is why perimetry remains central to glaucoma and neurology. This article sets out the field's extent, its measurement, its neural map, and its polar asymmetries.

Keywords: visual field, perimetry, retinotopy, blind spot

Consider a single word held steadily in view: without any movement of the eye, a great deal of the surrounding room is still registered around it. That whole expanse — the fixated point and everything taken in at once around it, out to the edges — is the visual field, the total area visible to an eye whose gaze is held straight ahead. It is not a uniform window. The centre resolves fine print while the far periphery catches little more than motion and coarse form, and hidden in the temporal field of each eye is a small region that sees nothing at all. Charting the shape of that field, explaining why it has the profile it does, and reading its losses as signs of disease are among the oldest and most practical projects in vision science (Phu et al., 2017).

What the Visual Field Is

The visual field is defined by fixation. With the eye pointed steadily at one place, the field is the entire region of external space from which light reaches the retina and is seen — everything taken in at that instant without moving the eye. The MeSH scope note states it at its barest, as the total area or space visible in a person's peripheral vision with the eye looking straight forward, and that phrasing captures two facts at once: the field is anchored to a fixation point, and it extends well beyond the small central zone usually thought of as looking directly at something (Phu et al., 2017).

The monocular field of a single eye is not circular. Cut off above and nasally by the brow and the nose, it reaches only about 60 degrees upward and 60 degrees toward the nose, but sweeps out to roughly 75 degrees downward and about 100 degrees temporally, toward the ear, where nothing obstructs it, as Table 1 sets out. A useful convention runs throughout the study of the field: a point in the world is named by its eccentricity, the angle between it and the fixation point, and by its meridian, the clock direction around fixation. The fovea sits at zero eccentricity, and every measurement of the field is in the end a statement about how vision changes as eccentricity grows.

Because the two eyes are set side by side, their monocular fields overlap across a wide central band — the binocular field that supports stereoscopic depth — while each eye retains a temporal crescent seen by it alone, the arrangement Figure 1 lays out.

The single most striking fact about the field is not its outer border but a gap near its centre. Where the optic nerve and retinal vessels pierce the back of the eye, the retina has no photoreceptors, and the corresponding patch of the visual field — about 15 degrees temporal to fixation in each eye, a few degrees across — is completely blind. It goes unnoticed because the brain fills the gap from its surroundings and because the other eye covers it, but it is an absolute scotoma present in every healthy eye, and it is the anchor landmark of every perimetric chart.

The Island of Vision

Sensitivity is not spread evenly over the field. A classic image, due to the perimetrist Traquair, describes the field as an island of vision in a sea of blindness: a landmass whose height at each point is the eye's sensitivity there. The island rises to a sharp peak at the fovea, slopes down in all directions toward the periphery, and is punctured by one bottomless well — the blind spot. Measuring the field is surveying this island, tracing the contours of equal sensitivity, called isopters, and locating any pits, or scotomas, that disease has sunk into its surface.

The shape of the island is not arbitrary; it is a direct readout of retinal anatomy. Sensitivity and acuity are highest at the fovea because the retina's sampling elements are densest there. Cone photoreceptors reach an extreme peak of density at the very centre of the fovea and fall steeply within the first few degrees of eccentricity, a topography quantified across whole human retinas by Curcio and colleagues, whose maps remain the reference for how the mosaic thins from centre to edge (Curcio, Sloan, Kalina, & Hendrickson, 1990). The retinal ganglion cells that carry the signal to the brain show the same organization: their density peaks in a ring around the fovea and drops with eccentricity, so that the number of ganglion cells devoted to each degree of the field falls precipitously outward (Curcio & Allen, 1990). Because each ganglion cell effectively samples a patch of the field, this declining density is why the periphery is coarse, why fine detail must be brought to the fovea by an eye movement, and why the slope of the island of vision follows the slope of the cell density beneath it. Reduced sampling is not the only limit on peripheral vision, however: even when a peripheral target is large enough to resolve in isolation, nearby clutter prevents its identification — the phenomenon of crowding, whose spatial extent grows in proportion to eccentricity (Bouma's law) and which, more than acuity alone, sets the ceiling on recognizing form and objects away from the fovea (Bouma, 1970; Whitney & Levi, 2011). The first demonstration walks out along a meridian, tracing sensitivity as it falls from the foveal peak and drops into the blind spot on the way.

Measuring the Field: Perimetry

Perimetry is the measurement of the visual field, and modern clinical perimetry is static and automated: the fixating eye is presented with spots of light at many fixed locations across the field, and at each location the instrument finds the threshold — the dimmest spot the eye can just detect. Sensitivity is reported in decibels, a logarithmic scale on which a high number means a dim spot was seen (good sensitivity) and a low number means only a bright spot was detected (poor sensitivity). The resulting grid of thresholds is a numerical survey of the island of vision, one height reading per location, from which the whole terrain is reconstructed (Phu et al., 2017).

Reading a field requires knowing how much variation is normal, because a threshold that would be alarming at the fovea may be ordinary in the periphery. The normal eye is not equally reliable everywhere: threshold sensitivity varies more from test to test, and more between healthy people, as eccentricity increases, so the same decibel deviation carries different weight at different locations. Heijl and colleagues quantified this normal variability point by point across the central field, providing the statistical baseline against which a real defect is distinguished from ordinary scatter (Heijl, Lindgren, & Olsson, 1987). A practical obstacle for decades was time: measuring dozens of thresholds by bracketing each one took so long that patients tired and reliability suffered. The breakthrough was to treat thresholds as a statistical estimation problem — using the known correlations between neighbouring locations and a model of normal and glaucomatous fields to choose each next stimulus efficiently — which is the logic of the SITA algorithms that made accurate threshold perimetry fast enough for routine care (Bengtsson, Olsson, Heijl, & Rootzén, 1997). The second demonstration presents a schematic threshold chart on which the classic patterns of loss can be imposed, so the numbers and the greyscale that clinicians read become concrete.

From Retina to Cortex

The visual field does not stop at the retina; it is carried, map by map, into the brain, and the geometry of that carriage explains much of how the field behaves. The founding observation was clinical. Studying soldiers with occipital gunshot wounds in the First World War, Gordon Holmes showed that a localized wound of the striate cortex produced a localized blind area in a predictable part of the visual field, and by collecting many such cases he reconstructed an orderly map — the first human retinotopic map — in which the visual field is laid out across the primary visual cortex in a regular way (Holmes, 1918). Two features of that map are decisive. The field is split down the vertical midline, each half-field projecting to the opposite hemisphere, and the central field is hugely over-represented: a small central patch of field commands a large expanse of cortex while the whole vast periphery is compressed into a little.

Holmes's map, drawn from lesions, was refined quantitatively over the following century. Reanalysis of clinical cases sharpened the constants of cortical magnification — the millimetres of cortex devoted to each degree of field as a function of eccentricity — into a simple formula that captures how steeply central vision dominates (Horton & Hoyt, 1991). Functional MRI then made the maps visible in the living brain: by driving the cortex with stimuli that swept through the field, the borders between the multiple retinotopic areas tiling human visual cortex could be drawn non-invasively, confirming and extending the lesion-based picture (Sereno et al., 1995). These methods established that human cortex holds not one map but a mosaic of visual field maps, each a complete or partial representation of the field, whose organization is now a foundation of visual neuroscience (Wandell, Dumoulin, & Brewer, 2007). Measuring the maps precisely became possible with population receptive field modelling, which estimates, for each patch of cortex, the region of visual field that drives it, turning the qualitative map into a quantitative one (Dumoulin & Wandell, 2008). Because the maps are so regular across people, an individual's cortical map can even be predicted from anatomy alone by a Bayesian atlas fitted to many measured brains, a striking demonstration of how stereotyped the field-to-cortex transform is (Benson & Winawer, 2018). The third demonstration makes cortical magnification tangible, mapping a ring of the visual field onto the strip of cortex it commands and tracking what fraction of the primary map the central field claims.

Performance Fields and Their Asymmetries

The island of vision is often drawn as if it were rotationally symmetric — as if all that mattered was eccentricity, with every direction around fixation equivalent. It is not. At a fixed eccentricity, visual performance depends systematically on the polar angle, the clock direction of the location around fixation, and the pattern of that dependence is called a performance field. Two anisotropies are robust. Performance is better along the horizontal meridian than the vertical — the horizontal-vertical anisotropy — and, along the vertical, better below fixation than above it, the vertical-meridian asymmetry. Carrasco and colleagues mapped these differences psychophysically and showed they hold across tasks, spatial frequencies and set sizes, establishing that the field is inhomogeneous in a lawful, direction-dependent way rather than only radially (Carrasco, Talgar, & Cameron, 2001). These are large effects: contrast sensitivity can differ by a substantial factor between the horizontal and the upper vertical meridian at the same eccentricity, and the asymmetries grow with eccentricity and with spatial frequency, a stimulus dependence charted in detail more recently (Himmelberg, Winawer, & Carrasco, 2020). Notably, these perceptual asymmetries have cortical correlates in the amount of retinotopic surface area devoted to the corresponding parts of the field, tying the performance field back to the maps of the previous section.

The field is functional in another sense that a threshold map misses. The area over which a person can take in and use information in a single fixation — without moving the eyes — can shrink under load or with age, and this useful field of view, measured with divided-attention search tasks, predicts real-world outcomes such as driving safety better than ordinary acuity does (Ball, Beard, Roenker, Miller, & Griggs, 1988). The static island of sensitivity, the direction-dependent performance field, and the attentionally governed useful field are three complementary descriptions of the same region of space, each capturing what the others leave out.

Reading Field Loss: Localizing Disease

The clinical power of the visual field comes from a simple fact: because every part of the visual pathway maps to a definite part of the field, the shape of a field defect reveals where along the pathway the damage lies. Damage confined to one eye, before the optic chiasm, produces a defect in that eye alone. Damage at the chiasm, where the nasal fibres of the two eyes cross, classically knocks out both temporal half-fields — a bitemporal hemianopia — the signature of a pituitary tumour pressing on the crossing fibres. Damage behind the chiasm, in the optic tract, radiations or cortex, produces a homonymous defect: the same side of the field lost in both eyes, because past the chiasm each hemisphere carries the opposite half-field of both eyes. It was exactly this logic, worked out from cortical wounds, that let Holmes read the location of a lesion from the position of the blind area it produced (Holmes, 1918).

Within the eye, the arrangement of the retinal nerve fibres stamps its own signatures on the field. Because ganglion-cell axons arc around the fovea and respect the horizontal raphe, glaucomatous damage to bundles of them yields the characteristic arcuate scotoma sweeping from the blind spot, and a nasal step where defects above and below the horizontal midline fail to match. Detecting such losses early, and above all measuring whether they are progressing, is the central task of glaucoma management, and doing it reliably against the noise of normal variability is a hard measurement problem that continues to drive refinements in how fields are tested and analysed (De Moraes, Liebmann, & Levin, 2017). Because progression is subtle and perimetry is variable, the field has become a natural target for machine analysis: deep-learning models trained on large numbers of fields and optic-nerve images can help screen for glaucoma and flag progression, an active frontier in turning the raw field into a decision (Thompson, Jammal, & Medeiros, 2020). Across all of these uses, the visual field remains an indispensable functional measure precisely because its structure is so tightly tied to the anatomy that produces it (Phu et al., 2017).

Worked Example

Two calculations make the central themes concrete, and the retinotopy demonstration reproduces the first. Take the cortical magnification of primary visual cortex in the linear form fitted to human data, M(E) = 17.3 / (E + 0.75), where E is eccentricity in degrees and M is millimetres of cortex per degree of field (Horton & Hoyt, 1991). At the fovea, M(0) = 17.3 / 0.75 = 23.1 millimetres per degree; at 20 degrees out, M(20) = 17.3 / 20.75 = 0.83 millimetres per degree. The central degree of the field therefore commands about 23.1 / 0.83 = 28 times as much cortex as a degree at 20 degrees eccentricity. Integrating M along a meridian gives the cortical distance from the foveal representation out to eccentricity E as d(E) = 17.3 x ln((E + 0.75) / 0.75). Out to the edge of the mapped field near 90 degrees, d(90) = 17.3 x ln(121) = 83.0 millimetres, while the central 10 degrees reach d(10) = 17.3 x ln(14.33) = 46.1 millimetres. So the central 10 degrees — a small disc of the visual world — occupy 46.1 / 83.0, about 55 percent, of the primary map, and the central 2.5 degrees alone claim d(2.5) / d(90) = 25.4 / 83.0, roughly 31 percent. The periphery, vast in the world, is a thin rind on the cortex.

The second calculation reads the perimetric scale. Sensitivity in decibels is S = 10 x log10(L_max / L), where L_max is the perimeter's brightest stimulus, defined as 0 decibels, and L is the threshold luminance at a location. A point whose threshold is one hundredth of maximum has S = 10 x log10(100) = 20 decibels; a healthier point seen at one thousandth of maximum has S = 10 x log10(1000) = 30 decibels. Suppose disease deepens a location from 30 to 24 decibels, a 6-decibel loss. Because the scale is logarithmic, that location now needs a stimulus 10^(6/10) = 10^0.6 = 4.0 times brighter to be seen — a fourfold loss of sensitivity behind a modest-looking change in the number, which is exactly why the decibel scale, and a firm sense of what counts as normal variation, are needed to read a field.

Discussion

The visual field looks at first like a plain fact of anatomy — the patch of world an eye can see — and turns out to be a layered construction. Its outer border is set by facial structure, but its interior shape, the island of vision, is a direct projection of the retinal mosaic, whose photoreceptor and ganglion-cell densities peak at the fovea and fall with eccentricity (Curcio et al., 1990; Curcio & Allen, 1990). That same declining sampling is carried forward, and amplified, in the cortical maps that Holmes first read from war wounds and that functional imaging and receptive-field modelling have since made quantitative, in which the central field commands a hugely disproportionate share of tissue (Holmes, 1918; Horton & Hoyt, 1991; Dumoulin & Wandell, 2008). The field is thus not one thing but a stack of registered maps, from mosaic to cortex, each preserving the topography of the last.

Reading the field this way explains both its scientific and its clinical value. Because performance depends on polar angle as well as eccentricity, the field is a sensitive probe of how vision is organized, and its asymmetries have become a testbed for linking perception to cortical surface area (Carrasco, Talgar, & Cameron, 2001; Himmelberg, Winawer, & Carrasco, 2020). Because each region of the field traces to a definite place on the pathway, a defect is a map reference to a lesion, which is what makes perimetry indispensable in glaucoma and neurology and a natural target for automated analysis (De Moraes et al., 2017; Thompson et al., 2020). The open problems are ones of measurement as much as biology: separating true progression from the field's own considerable variability, extending reliable perimetry into the far periphery and to populations who test poorly, and connecting the perceptual field, the anatomical field and the useful field of view into a single account (Heijl et al., 1987; Ball et al., 1988). The region of space an eye sees, examined closely, is one of the clearest windows there is onto how the visual system is built.

Common Misconceptions

The visual field is uniform, like a camera frame.

Sensitivity and acuity fall steeply from the fovea outward, tracing an island of vision rather than a flat frame, because the retinal sampling density beneath the field peaks at the centre and drops with eccentricity (Curcio et al., 1990).

The blind spot is a defect that only some people have.

Every healthy eye has an absolute blind spot about 15 degrees temporal to fixation, where the optic nerve leaves the retina and there are no photoreceptors; it goes unnoticed because the other eye covers it and the brain fills it in. It is a landmark on every perimetric chart, not a disease (Phu et al., 2017).

Every point at the same distance from fixation sees equally well.

At a fixed eccentricity, performance still depends on direction: it is better on the horizontal meridian than the vertical, and better below fixation than above. These performance-field asymmetries are lawful and sizable, with correlates in cortical surface area (Carrasco et al., 2001; Himmelberg et al., 2020).

Glossary

Arcuate scotoma.

An arc-shaped region of depressed sensitivity sweeping from the blind spot and respecting the horizontal midline, the classic glaucomatous field defect, reflecting damage to an arcing bundle of retinal nerve fibres.

Blind spot.

The absolute gap in each monocular field, about 15 degrees temporal to fixation, corresponding to the optic disc where the nerve exits and the retina has no photoreceptors.

Cortical magnification.

The amount of cortical tissue, in millimetres per degree, devoted to each part of the field; it is highest at the fovea and falls with eccentricity, so central vision dominates the map.

Crowding.

The failure to identify a peripheral target flanked by nearby clutter, even when it is large enough to resolve in isolation; its spatial extent scales with eccentricity (Bouma's law), and it limits form and object recognition away from the fovea.

Decibel (perimetry).

The logarithmic unit of perimetric sensitivity, S = 10 log10(L_max / L); a high value means a dim stimulus was detected and thus good sensitivity, a low value the reverse.

Eccentricity.

The angular distance of a location from the fixation point; the fovea is at zero eccentricity, and most properties of the field vary with it.

Fovea.

The central retinal pit of peak cone density that serves the point of fixation, the summit of the island of vision and the site of highest acuity.

Hemianopia.

Loss of half the visual field; bitemporal when the chiasm is damaged, homonymous (the same side in both eyes) when the lesion lies behind it.

Island of vision.

Traquair's metaphor for the field as a hill of sensitivity in a sea of blindness, peaking at the fovea, sloping to the periphery and pierced by the blind spot.

Isopter.

A contour of equal sensitivity on the field, the perimetric analogue of a contour line on a map of the island of vision.

Meridian.

The polar direction of a location around fixation; the horizontal and vertical meridians are reference axes along which performance-field asymmetries are defined.

Performance field.

The pattern by which visual performance varies with polar angle at fixed eccentricity, including the horizontal-vertical anisotropy and the vertical-meridian asymmetry.

Perimetry.

The measurement of the visual field; in static automated perimetry the threshold at each of many fixed locations is found and reported in decibels.

Population receptive field.

The region of visual field that drives a patch of cortex, estimated from functional imaging; the tool that turns a qualitative retinotopic map into a quantitative one.

Retinotopy.

The orderly mapping of the visual field onto a neural surface, so that neighbouring field locations project to neighbouring cortical locations across a mosaic of visual field maps.

Scotoma.

A localized region of reduced or absent sensitivity within the field; a relative scotoma is a depression, an absolute scotoma a total gap like the blind spot.

Visual field.

The total region of space visible to a steadily fixating eye; monocular for one eye, binocular where the two overlap.

Key Researchers

Marisa Carrasco. Julius Silver Professor of Psychology and Neural Science, New York University; she established that visual performance varies systematically with polar angle around the field — the horizontal-vertical anisotropy and the vertical-meridian asymmetry — and showed how covert spatial attention reshapes these performance fields. Faculty Page - Google Scholar - ORCID

Christine A. Curcio. Professor Emeritus of Ophthalmology and Visual Sciences, University of Alabama at Birmingham; her quantitative maps of human photoreceptor and retinal ganglion cell topography give the visual field its anatomical substrate, explaining why sensitivity falls with eccentricity and why the field has a blind spot. Faculty Page - Google Scholar - ORCID

Anders Heijl. Senior Professor of Ophthalmology, Lund University (Malmö), Sweden; a founder of modern automated static perimetry who quantified the normal variability of threshold sensitivity across the field and co-developed the SITA algorithms that made threshold perimetry fast enough for routine glaucoma care. Faculty Page - Google Scholar - ORCID

Sir Gordon Morgan Holmes (1876-1965). National Hospital for Nervous Diseases, Queen Square, London; from the localized field defects of soldiers with occipital gunshot wounds he mapped the orderly representation of the visual field onto the human striate cortex, giving neurology its first cortical retinotopic map and the localizing logic that still reads a field defect as a lesion reference. Biographical Memoir - Wikipedia

Brian A. Wandell. Isaac and Madeline Stein Family Professor, Stanford University; a leader in computational neuroimaging of the human visual system who charted the retinotopic visual field maps tiling human visual cortex and helped develop population-receptive-field modelling to measure them. Faculty Page - Google Scholar - ORCID

Jonathan Winawer. Professor of Psychology and Neural Science, New York University; he uses computational neuroimaging to model the retinotopic field maps of human cortex, including Bayesian atlases that predict an individual's cortical map and the polar-angle asymmetries that link cortex to perception. Faculty Page - Google Scholar - ORCID

Frequently Asked Questions

What is the visual field? The visual field is the total region of space an eye sees while held steady on one point. It takes in far more than the small central zone at fixation, reaching about 60 degrees up, 75 down, 60 toward the nose and 100 toward the ear in one eye, and it is the surface that perimetry measures and that field defects disturb (Phu et al., 2017).

Why is the centre of the field so much sharper than the edges? Because the retina samples the centre far more densely. Cone photoreceptors and retinal ganglion cells peak in density at the fovea and fall steeply with eccentricity, so each degree of central field is served by many more cells than a degree of periphery, tracing the island of vision that is high at the centre and low at the edge (Curcio et al., 1990; Curcio & Allen, 1990).

What is the blind spot and why does it go unnoticed? The blind spot is a small region about 15 degrees toward the ear from fixation where the optic nerve leaves the eye and there are no photoreceptors, so it is truly blind. It goes unnoticed because the fellow eye covers that part of the field and the brain fills the gap from its surroundings; it is present in every healthy eye (Phu et al., 2017).

How is the visual field measured? By perimetry. In static automated perimetry the fixating eye is shown dim spots of light at many fixed locations, and at each the instrument finds the dimmest spot that can be seen, reporting sensitivity in decibels. Modern algorithms such as SITA make this fast and reliable enough for routine clinical use (Bengtsson et al., 1997; Heijl et al., 1987).

How is the visual field represented in the brain? As retinotopic maps. The field is laid out in an orderly way across the primary visual cortex and a mosaic of higher maps, split at the vertical midline between the hemispheres, with the central field commanding a hugely disproportionate share of cortex. This organization was first read from wounds by Holmes and later made quantitative by functional imaging and receptive-field modelling (Holmes, 1918; Wandell et al., 2007; Dumoulin & Wandell, 2008).

Do all directions in the field see equally well at the same eccentricity? No. At a fixed distance from fixation, performance is better along the horizontal meridian than the vertical, and better below fixation than above. These performance-field asymmetries are lawful and sizable and have correlates in the amount of cortex devoted to each direction (Carrasco et al., 2001; Himmelberg et al., 2020).

What can a visual field defect reveal about disease? A great deal, because each part of the visual pathway maps to a definite part of the field. A bitemporal hemianopia points to the chiasm, a homonymous defect to a lesion behind it, and an arcuate scotoma with a nasal step to glaucomatous nerve-fibre loss. The shape of the defect localizes the damage (Holmes, 1918; De Moraes et al., 2017).

How is the visual field different from visual acuity? Acuity measures the finest detail resolved at the point of fixation, a single number for the centre; the visual field describes vision across the whole seen region, its extent and its sensitivity everywhere. An eye can have normal central acuity while a large part of its field is lost, which is why perimetry adds information that an acuity chart cannot (Phu et al., 2017).

References