David Hunter Hubel (1926-2013) was a Canadian-American neuroscientist who, working with Torsten Wiesel at Johns Hopkins and later Harvard Medical School, uncovered how the brain’s visual cortex processes what the eye sees. Beginning in the late 1950s, the pair recorded from single neurons in the visual cortex of anaesthetized cats while showing them patterns of light, and found that individual cells responded most strongly to edges and bars at particular orientations - the basic feature detectors of vision.
Their experiments revealed a layered, hierarchical organization. “Simple” cells responded to edges at a specific position and orientation, while “complex” cells responded to an edge at a given orientation across a range of positions, building tolerance to where exactly the stimulus fell. They also mapped how the cortex is arranged into orientation columns and ocular dominance columns, an orderly functional architecture rather than a tangle of wiring.
For these discoveries about information processing in the visual system, Hubel and Wiesel shared half of the 1981 Nobel Prize in Physiology or Medicine, with the other half going to Roger Sperry. Their picture of vision as a hierarchy of increasingly complex feature detectors directly inspired later artificial vision systems - Kunihiko Fukushima cited it when designing the neocognitron, the architectural ancestor of the convolutional neural networks that now dominate computer vision.