Abstract
Purpose: To measure the neuronal basis of orientation-specific contrast adaptation in humans. Methods: To induce and maintain adaptation, subjects continuously viewed a moving, vertical, high-contrast (75%) grating. After an initial 30-s adaptation period, subjects performed a sequence of two-interval forced choice discrimination judgements. Each trial, which lasted 2.5 s, included two 0.5-s intervals with slightly different contrasts. Subjects pressed a button to indicate which interval had the higher contrast. The stimulus was maintained at the adapting level (∼75% contrast) during most of these trials. Test trials, once every 15 s, were presented at various lower contrast levels. The task sequence consisted of 5 high-contrast trials followed by a lower contrast test trial, and this sequence was repeated 12 times in the course of each 3.5-minute fMRI scan. To measure the effects of orientation-selective adaptation, we compared the fMRI responses when the test gratings were either parallel or orthogonal to the adapter. Scans with parallel and orthogonal test patterns were interleaved. fMRI data were collected using a fast, spiral pulse sequence (3T scanner, custom surface coil, acquisition rate 1 Hz, 8 oblique slices perpendicular to the calcarine sulcus). Results: The amplitude of the fMRI responses were smaller for parallel than for orthogonal test gratings. This effect was statistically significant in all visual areas measured: V1, V2, V3, and V4v. Conclusion: Orientation-selective contrast adaptation reduces the neuronal responses to test gratings with the same orientation as the adapter, while perpendicular test gratings produce larger responses, presumably in an unadapted pool of neurons.