Hubel & Wiesel's Visual Cortex Experiments Explained
In 1958, in a basement laboratory at Johns Hopkins University, two young researchers spent hours trying to make a single neuron in a cat's visual cortex respond to anything at all. They projected bright spots and dark spots onto a screen in front of the animal. Nothing much happened. Then, as one of them slid a glass slide into the projector, the cell suddenly erupted. In his 1981 Nobel lecture, David Hubel recalled that the audio monitor "went off like a machine gun." The trigger was not the spot on the slide. It was the faint shadow of the slide's edge, sweeping across the screen at one particular angle.
That accident started a 25-year collaboration between Hubel and Torsten Wiesel that changed how scientists understand the brain. They showed that neurons in the primary visual cortex do not simply relay a pixel-by-pixel picture. Each one responds to a specific feature, most famously a line or edge at a specific orientation, and these cells are arranged in an orderly map across the cortical surface. They also showed that this machinery depends on early visual experience. In 1981 they shared half of the Nobel Prize in Physiology or Medicine "for their discoveries concerning information processing in the visual system." The other half went to Roger Sperry for his split-brain research.
Before Hubel and Wiesel: what was known about vision
By the 1950s, neurophysiologists could record from single cells in the eye. Stephen Kuffler, working at Johns Hopkins, had shown in 1953 that retinal ganglion cells in the cat have circular receptive fields with a center-surround structure: some cells fire when light falls on a small central spot and are suppressed when it falls on the ring around it ("on-center"), while others do the opposite ("off-center"). These cells care about local contrast, not just raw brightness.
What happened after the signal left the eye was much less clear. The pathway runs from the retina to the lateral geniculate nucleus (LGN) of the thalamus and from there to the primary visual cortex, also called the striate cortex, area 17 in the cat, or V1. Earlier attempts to record from visual cortex had produced weak and puzzling responses, partly because researchers were using diffuse flashes of light that most cortical cells barely notice.
Hubel, a Canadian-born physician trained at McGill, had built a better tool. In 1957 he published a design for a varnish-insulated tungsten microelectrode in Science. Sharpened tungsten wire was sturdy enough to push through brain tissue and fine enough to pick up the electrical spikes of one neuron at a time, and it could be used over long recording sessions. Wiesel, a Swedish physician trained at the Karolinska Institute, had come to Kuffler's lab to study the retina. Kuffler paired them, and the two began working on the cortex together in 1958. In 1959 the group moved with Kuffler to Harvard Medical School, where the department of neurobiology would later be founded.
The experimental setup
The basic preparation is described in their 1959 paper in the Journal of Physiology, "Receptive fields of single neurones in the cat's striate cortex." Recordings came from 24 cats that were lightly anesthetized. The animal faced a screen, its eyes held open, and the researchers projected spots, bars and edges of light onto the screen while a tungsten electrode was advanced slowly into the visual cortex.
The electrode signal was fed both to an oscilloscope and to a loudspeaker. Each time the neuron fired an action potential, the speaker produced a click. This audio monitor turned out to be essential: the researchers could move a stimulus by hand and hear, in real time, whether a cell liked it. The receptive field of each cell was mapped by marking on the screen where light excited the cell and where it inhibited it.
Two other details mattered. First, they worked one cell at a time, often for hours, which allowed careful mapping. Second, they used small electrical lesions at the electrode tip to mark recording sites, so after the experiment they could reconstruct in tissue sections exactly which layer and region each cell had come from. Combining physiology with anatomy in this way is what later let them describe columns.
The discovery: cells that respond to lines, not spots
The core finding of the 1959 paper was that most cells in cat striate cortex responded poorly or not at all to diffuse light and only weakly to small spots. They responded strongly to elongated stimuli, a slit of light, a dark bar or an edge between light and dark, but only when the stimulus was presented at the right orientation. Rotate the bar away from the preferred angle and the response dropped off; at right angles to the preferred orientation, the cell typically fell silent.
Different cells preferred different orientations, and all orientations were represented. Many cells also preferred movement in one direction across the receptive field over the opposite direction. This property, orientation selectivity, had no counterpart in the retina or the LGN, whose cells have round, center-surround fields. Something in the cortex was building a new kind of signal.
Simple and complex cells
The 1962 paper, "Receptive fields, binocular interaction and functional architecture in the cat's visual cortex" (Journal of Physiology 160:106–154), is the one most textbooks draw on. In it Hubel and Wiesel sorted orientation-selective cells into types.
Simple cells have receptive fields with separate, side-by-side excitatory and inhibitory zones arranged in parallel strips. A bright bar placed exactly on the excitatory strip, at the right angle, drives the cell; the same bar shifted onto the inhibitory strip suppresses it. Their responses could be largely predicted from a map of those zones.
Complex cells also prefer a particular orientation, but they do not have separate on and off zones. They respond to a correctly oriented line or edge anywhere within a larger receptive field, and many respond best to movement. In other words, a complex cell signals "there is an edge at this angle somewhere in this patch," with some tolerance for exact position.
In 1965, working on areas 18 and 19 of the cat, they described cells they called hypercomplex, which respond best to lines of a limited length and fire less when the line extends beyond the receptive field. Today these are usually called end-stopped cells, and they are known to occur in V1 as well.
Hubel and Wiesel proposed a wiring scheme to explain these properties. A simple cell, they suggested, could get its elongated field by summing input from several LGN cells whose round receptive fields fall along a straight line. A complex cell could then pool input from several simple cells with the same orientation but slightly different positions. This was the first concrete proposal for how the brain builds more abstract features out of simpler ones, step by step.
How well has the hierarchy held up?
The idea that LGN cells feed simple cells, which feed complex cells, has been influential and partly confirmed. Later work showed that the orientation selectivity of simple cells in cat layer 4 does depend substantially on aligned thalamic input, broadly as Hubel and Wiesel proposed. But the strict serial version proved too tidy. Recordings in the early 1970s showed that some complex cells receive direct input from the LGN rather than only from simple cells, so the two pathways run partly in parallel. Researchers also found that cortical circuits within V1, both excitatory and inhibitory, sharpen and modulate orientation tuning, and that responses are strongly affected by stimuli outside the classical receptive field. The simple/complex distinction itself is now often treated as two ends of a continuum rather than two sharply separate classes.
None of this overturns the main discovery. It refines the circuit explanation for it, which is the normal fate of a good first model.
Columns: the cortex is organized like a map
Because they marked their electrode tracks, Hubel and Wiesel could ask how cells with different properties were arranged in the cortex. When the electrode went in perpendicular to the cortical surface, cell after cell along the track tended to share the same preferred orientation. When the electrode ran at a slant, the preferred orientation shifted in small, regular steps as the electrode advanced.
They concluded that orientation is organized in columns, narrow regions running from the surface to the bottom of the cortex in which cells share an orientation preference. Vernon Mountcastle had already described columns in the somatosensory cortex in 1957; Hubel and Wiesel showed that the principle also governs vision, and that it applies to a feature the cortex itself creates.
They found a second, overlapping system for the two eyes. In the cat, as in primates, signals from the left and right eyes stay separate in the LGN and first combine in the cortex. Many cortical cells respond to both eyes, but most respond more strongly to one. Cells grouped by eye preference form ocular dominance columns. In the macaque monkey, these columns are especially crisp in the input layer 4C, where they appear as alternating stripes roughly half a millimeter wide. Hubel, Wiesel and Simon LeVay mapped them across the monkey's cortex in a 1977 paper using anatomical tracers, showing a pattern of interleaved bands that looks a little like a fingerprint.
From these observations they proposed the hypercolumn: a block of cortex about 1 mm across that contains a full set of orientation columns and a pair of ocular dominance columns (one per eye), all for roughly the same small patch of the visual field. Neighboring hypercolumns repeat the analysis for neighboring patches. Later optical imaging, from the 1980s onward, showed that orientation maps are arranged around "pinwheel" centers where all orientations meet, rather than the neat parallel slabs drawn in early diagrams, but the hypercolumn idea of a repeating module survived.
Beyond cats: monkeys, humans and rodents
In 1968 Hubel and Wiesel published a large study of the macaque monkey's striate cortex (Journal of Physiology 195:215–243). They found orientation selectivity, simple and complex cells and columns, along with features specific to primates such as cells that are not orientation-tuned in parts of layer 4C. Monkeys became the main model for the rest of their work, because the primate visual system is much closer to the human one.
Columns are not universal, though. Two-photon calcium imaging, which can record from hundreds of neurons at once, revealed in 2005 (Ohki and colleagues, Nature) that rat V1 neurons are sharply orientation-selective but are scattered in a salt-and-pepper arrangement, with neighboring cells preferring different orientations. Mice show the same pattern. So orientation selectivity looks like a general feature of mammalian visual cortex, while its orderly columnar layout appears in carnivores and primates but not in rodents. Why this is so, and whether columns provide any computational advantage, is still debated.
In humans, high-field fMRI at 7 tesla has resolved ocular dominance columns and, more controversially, signals related to orientation columns, but single-cell evidence comes mostly from animals.
Critical periods: experience wires the cortex
The second major strand of their work, begun in the early 1960s, asked whether these circuits are present at birth or shaped by experience. Their method was to suture one eyelid of a kitten closed for a period and then record from the cortex.
In 1963 Wiesel and Hubel reported in the Journal of Neurophysiology that monocular deprivation in young kittens had a dramatic effect: very few cortical cells could still be driven by the eye that had been closed, even though that eye and its retina were essentially normal. Most cells were now dominated by the open eye. The same closure in an adult cat had little effect.
In a 1970 paper in the Journal of Physiology they mapped the timing. Susceptibility to the effects of eye closure began suddenly near the start of the fourth week after birth, stayed high until somewhere between the sixth and eighth weeks, and then declined, disappearing around the end of the third month. This window is the classic example of a critical period.
A surprising result came from closing both eyes. If both eyes were deprived, many cells kept responding to both eyes, more than in single-eye closure. The lesson was that the two eyes compete for cortical territory. What matters is not only the amount of input but the balance between the eyes; an imbalance lets the active eye take over connections the inactive eye would have kept.
Anatomical studies, including work by Hubel, Wiesel and LeVay in monkeys, showed the physical basis: after monocular deprivation, the ocular dominance stripes in layer 4C belonging to the open eye widened while those of the closed eye shrank.
Other groups extended the approach. Colin Blakemore and Grahame Cooper reported in Nature in 1970 that kittens reared for their early months in a cylinder painted with only vertical or only horizontal stripes later behaved as if insensitive to the orientation they had not seen, and their cortical cells were biased toward the experienced orientation. The size of this effect has been debated, and some later attempts found weaker changes, but the general point that experience tunes orientation preferences is well supported.
From the lab to the clinic: amblyopia
The critical-period findings matter medically because they explain amblyopia, sometimes called lazy eye. In amblyopia, one eye sends a weaker or blurrier image to the brain in early childhood because of a squint (strabismus), a large difference in focus between the eyes, or something blocking the view, such as a congenital cataract. Even when the eye itself is later corrected, vision through it can remain poor because the cortex has wired itself around the other eye.
Hubel and Wiesel's work gave a biological rationale for treating these conditions early: removing congenital cataracts promptly and patching or blurring the stronger eye so the weaker one gets a fair share of cortical territory while the brain is still plastic. Later research has shown that the human visual system remains somewhat plastic beyond early childhood, so treatment in older children and even adults can produce gains, but early treatment generally works best.
Influence on artificial intelligence
The hierarchy of simple and complex cells inspired computer vision directly. In 1980 Kunihiko Fukushima published the neocognitron, a layered network with alternating "S-cells" and "C-cells" named after simple and complex cells. S-layers detected local features; C-layers pooled them to tolerate small shifts in position. This design is the direct ancestor of the convolutional neural networks that power modern image recognition, where convolution layers play the role of feature detectors and pooling layers provide position tolerance.
When trained CNNs are inspected, units in their first layers often end up as oriented edge and color-contrast detectors that resemble V1 receptive fields. That similarity is a useful analogy rather than proof that brains and CNNs compute the same way, and the differences are large: brains rely heavily on feedback connections and learn from far fewer labeled examples. For a non-technical walkthrough, see how image recognition works.
Common misconceptions
"Hubel and Wiesel found neurons that recognize objects." They did not. V1 cells respond to local features such as oriented edges. Cells selective for complex shapes and faces were found later in higher visual areas of the temporal lobe by other researchers.
"The cortex is a neat grid of identical columns." Early diagrams showed parallel slabs. Real orientation maps are curved and organized around pinwheel centers, and rodents have no orientation columns at all.
"Critical periods mean the brain stops changing after childhood." The specific sensitivity of ocular dominance to eye closure declines sharply after the critical period in cats, but adult brains retain many forms of plasticity, including the learning that underlies how memory and skills form.
"Visual cortex cells only care about orientation." V1 neurons are also tuned to spatial frequency, direction of motion, binocular disparity and, in primates, color, and their responses are modulated by context and attention.
Ethical context
The deprivation experiments involved suturing kittens' eyelids and later recording from their brains, and they attracted criticism from animal-welfare campaigners, especially from the late 1970s onward. Supporters point to their direct relevance to treating childhood cataracts and amblyopia. Today, comparable studies run under stricter animal-research oversight, and many questions are addressed with less invasive methods such as imaging in mice or human fMRI. Readers can weigh the trade-off for themselves; the scientific findings themselves are not in dispute.
Why the work still matters
Hubel and Wiesel did not just describe some neurons. They introduced a way of studying the brain: present controlled stimuli, record single cells, map what each cell "wants," and relate that to anatomy. They showed that the cortex transforms its input into new representations, that it is organized in repeating modules, and that experience during development shapes that organization. Each of these ideas opened a field of research, from brain mapping to developmental neuroscience to deep learning.
Their joint 1959 and 1962 papers remain among the most cited in neuroscience, and the experiments are collected, with their own commentary, in their 2005 book Brain and Visual Perception: The Story of a 25-Year Collaboration. For the basic biology of how neurons signal, see how the nervous system works.
FAQ
Q: What did Hubel and Wiesel discover? A: They found that most neurons in the primary visual cortex respond to lines or edges at a specific orientation rather than to spots of light, described simple and complex cells, showed that cells with similar orientation and eye preference are grouped into columns, and showed a critical period in which early visual experience shapes the cortex.
Q: How did they discover orientation selectivity? A: Partly by accident. In 1958, while they were trying to stimulate a cat cortical cell with spots on glass slides, the shadow of a slide's edge moving across the screen made the cell fire vigorously. Hubel later described the audio monitor going off "like a machine gun." They then showed that the cell responded to edges only at one orientation.
Q: Why did they use cats and monkeys? A: Cats were a standard model for visual physiology at the time, and Kuffler's lab had already worked out methods for recording from the cat retina. Cats also have good binocular vision. From the late 1960s they increasingly used macaque monkeys because the primate visual system is much closer to the human one.
Q: What is the difference between simple and complex cells? A: Simple cells have receptive fields with separate excitatory and inhibitory strips, so the exact position of a line matters. Complex cells prefer an orientation too but respond to a correctly oriented line anywhere within a larger receptive field, without separate on and off zones.
Q: What is a critical period? A: A window in development when experience has an outsized effect on brain wiring. In kittens, Hubel and Wiesel found that closing one eye affected the cortex strongly from about the fourth week to around the third month of life, but had little effect in adult cats.
Q: Did Hubel and Wiesel win a Nobel Prize? A: Yes. They shared one half of the 1981 Nobel Prize in Physiology or Medicine for discoveries concerning information processing in the visual system. Roger Sperry received the other half for his work on the specialization of the cerebral hemispheres.
Q: How did their work influence artificial intelligence? A: Kunihiko Fukushima's 1980 neocognitron explicitly copied their simple-cell and complex-cell hierarchy, and that design led to modern convolutional neural networks, which alternate feature-detecting and pooling layers.
Q: Do all mammals have orientation columns? A: No. Cats, ferrets and primates do. Rats and mice have orientation-selective neurons, but those neurons are mixed in a salt-and-pepper pattern rather than grouped into columns, as two-photon imaging showed in 2005.
Sources
- Hubel DH, Wiesel TN (1959). Receptive fields of single neurones in the cat's striate cortex. J Physiol 148:574–591 (PMC)
- Hubel DH, Wiesel TN (1962). Receptive fields, binocular interaction and functional architecture in the cat's visual cortex. J Physiol 160:106–154 (Wikidata record)
- Hubel DH, Wiesel TN (1968). Receptive fields and functional architecture of monkey striate cortex. J Physiol 195:215–243
- Hubel DH, Wiesel TN (1970). The period of susceptibility to the physiological effects of unilateral eye closure in kittens. J Physiol 206:419–436 (PMC)
- The Nobel Prize in Physiology or Medicine 1981 (NobelPrize.org)
- David Hubel, Nobel Lecture: Evolution of ideas on the primary visual cortex, 1955–1978 (PDF)
- Torsten Wiesel, Nobel Lecture (1981)
- Wurtz RH (2009). Recounting the impact of Hubel and Wiesel. J Physiol 587:2817–2823
- David Hubel (1926–2013): the man who developed our understanding of vision (PMC)
- Ohki K et al. (2005). Functional imaging with cellular resolution reveals precise micro-architecture in visual cortex. Nature 433:597–603
- Fukushima K (1980). Neocognitron. Biological Cybernetics 36:193–202 (PDF)
- The Scientist: A serendipitous shadow brought the brain's visual pathways to light
Read next:
Follow new articles
Add SciHub101 to your feed reader to see new science and technology stories.
About the Author
SciHub101 is an independent science and technology publication founded and edited by Buğra Sözeri, a software engineer and app developer based in Türkiye. AI tools assist with drafting and research. Source documentation in the existing archive is uneven, and articles are being reviewed and corrected. The Editorial Team byline identifies editorial responsibility; it does not certify specialist or independent human review. Where we compare products, anything built by our own team is labelled as such — see the editorial guidelines and About page for how corrections and disclosures work.
Related Articles

Why Do We Cringe? The Science of Vicarious Embarrassment
Vicarious embarrassment, known in German as Fremdscham, is the feeling of social pain on another person's behalf—you wince at their awkward moment even though they may feel nothing. This emotional contagion depends on empathy, theory of mind, and shared neural regions that also process physical pain. People higher in trait empathy cringe more intensely, and cringing can occur even when you know the other person does not care about the awkward situation. The anterior insula and anterior cingulate activate during vicarious embarrassment just as they do for empathy for others' pain, revealing that social mortification has neural roots.

How the Body Processes Alcohol: The Science of Metabolism and Intoxication
Alcohol is absorbed into the bloodstream within minutes of consumption, but the liver metabolizes it at a nearly constant rate of about one standard drink per hour, a rate that cannot be accelerated by food or hydration. The body treats ethanol as a toxin, converting it to acetaldehyde (more toxic than ethanol) and then to acetate through two enzyme-mediated reactions. Blood alcohol concentration (BAC) depends on body weight, sex, food intake, and genetics, making some individuals metabolically unable to process alcohol efficiently, as seen in alcohol flush syndrome. Once intoxicated, alcohol disrupts the balance of GABA and glutamate neurotransmitters in the brain, impairing judgment and coordination. Understanding alcohol metabolism reveals why hangovers occur, why some populations have lower tolerance, and why no trick can accelerate the clearance of alcohol from the body.

The Science of Sleep: What Happens in the Brain During REM Sleep
Sleep is not a passive shutdown but an active state of brain and body reorganization. During an 8-hour sleep cycle, humans alternate between non-REM (NREM) sleep—characterized by slow theta waves, reduced muscle tone, and energy conservation—and REM (rapid eye movement) sleep, when the brain shows activity patterns similar to wakefulness, the eyes move rapidly beneath closed eyelids, and skeletal muscles are paralyzed except for the diaphragm. REM sleep makes up approximately 20–25 percent of sleep in adults and is the stage where most vivid dreaming occurs. The neuroscience of REM reveals that memory consolidation (particularly emotional and procedural memories), creativity processing, and amygdala-driven emotion regulation all occur during this stage. The glymphatic system—a network of fluid pathways that clear metabolic waste from the brain during sleep—becomes most active during deep NREM sleep, not REM. This article explains the neuroscience of REM sleep, the discovery of the glymphatic system, and why sleep is essential to brain health.

What Happens to Your Body with No Sleep for 24 to 72 Hours
After 17 hours without sleep, your cognitive performance matches someone with a blood alcohol concentration of 0.05% — legally impaired to drive in most countries. After 24 hours, the equivalent is 0.10% — above the legal limit in all US states. After 72 hours, hallucinations are common, emotional regulation breaks down, and the immune system's ability to mount a response is measurably compromised. Sleep deprivation is not discomfort — it is a physiological insult to nearly every system in the body, and it unfolds on a precisely documented timeline.