Neural basis of contrast appearance measured with fMRI

Purpose: To study the neural basis of contrast appearance by comparing fMRI responses with psychophysical observations. Background: We assume that fMRI response is proportional to local average neural activity. And we hypothesize that there is a distinct brain area that mediates perceived contrast. If such a brain area exists, then its fMRI responses to two different spatial patterns should be the same if and only if their perceived contrasts are identical. Psychophysical Methods: Test stimuli were peripheral, 8 Hz, counterphase gratings of various spatial frequencies and contrasts. Subjects adjusted the contrast of a comparison stimulus, in a two alternative temporal forced-choice double-staircase, to match the contrast appearance of each test stimulus. fMRI Methods: Gratings were alternated every 40 seconds with uniform fields. fMRI response (using a T2* weighted spiral acquisition) was quantified as the amplitude of the (40 sec period) sinusoid that best fit each pixel's time-series. These responses were then averaged over each of several regions of interest. We determined the contrast of the comparison stimulus that produced the same fMRI response as each test stimulus. Results: Both fMRI response and perceived contrast depend strongly on spatial frequency. Psychophysically determined contrast matches across spatial frequencies correspond reasonably well with matches determined from fMRI responses in V1. The correspondence is more variable in extrastriate regions. Conclusions: fMRI coupled with a psychophysical matching paradigm can be used to study the neural correlate of contrast appearance. Contrast appearance may be mediated by neurons in V1.