feature detection

Meaning: Feature detection is the process by which specialized neurons in the visual cortex respond to particular elements of a stimulus, such as edges, line orientation, or direction of motion. Higher-level areas then combine these outputs into a representation of a whole object, so processing runs from the data upward: it is bottom-up. Hubel and Wiesel recorded from single cells in the visual cortex of cats and found that a cell fired most when a bar of light had one particular orientation. Later work described simple cells (oriented edges), complex cells (oriented edges, often moving), and hypercomplex cells (length or corners).

On the MCAT: Passages describe single-neuron recordings and ask which stimulus makes a cell fire most, or what happens when that feature is missing. Questions also ask you to label the process bottom-up rather than top-down, or to link it to parallel processing, in which color, shape, and motion are analyzed in separate pathways at the same time and then combined.

Common confusion: Feature detection does not explain all of perception. Top-down processing, which uses expectation and context, also shapes what we see, as when a smudged letter is read correctly from the surrounding word.