Abstract
An event-related fMRI paradigm was used to study orientation-specific pattern adaptation within neuronal responses in the human visual cortex. When a pair of brief visual stimuli are presented in succession, the fMRI response to the second stimulus can be calculated by subtracting out the response to the first stimulus when presented alone. However, if the two stimuli are similar, the fMRI response to the second stimulus is usually suppressed compared to the response to the first stimulus. This suppression is commonly attributed to nonlinearities in the fMRI signal, but may instead be the result of neuronal adaptation to the first stimulus. This adaptation hypothesis was tested by measuring the suppression of the second stimulus, while varying its orientation and SOA compared to the first stimulus. Subjects viewed successive pairs of 1-second full-contrast 0.25 c/deg, 8Hz counterphase modulated gratings that subtended 30 x 30 degrees of visual angle. Pairs of stimuli either had the same orientation (vertical or horizontal) or were orthogonal. SOAs varied from 1 to 4 seconds. In area V1, the suppression effect was greater for same-orientation pairs than for orthogonal pairs, and decreased with increasing SOA. Area V2 showed similar results. In comparison, in higher visual areas, suppression effects were similar in magnitude for same-orientation and orthogonal pairs. These results are consistent with suppression being the result of orientation-selective adaptation of neuronal responses rather than nonlinearities in the fMRI signal. This adaptation effect is a useful experimental tool. For example, our data provides a measurement of the time-course of recovery to adaptation that is consistent with previous psychophysical estimates. In addition, measuring suppression effects as a function of orientation can provide a measure of the orientation selectivity of neurons across visual areas.