Human brain activity was evoked by a dynamic random-dot display in whi
ch a square-wave grating appeared and disappeared at regular intervals
. Grating visibility was determined by one of four different contrasts
: texture, stereo disparity, luminance, or color. Scalp fields measure
d with 31 electrodes were used to estimate epicortical potential field
s. The estimation procedure required detailed anatomical data for each
subject. These were obtained from magnetic resonance images. A three-
dimensional digitizer and a stereotactic headgear were used to accurat
ely merge the frame of reference of the magnetic resonance image with
that of the evoked potential. Epicortical potential fields provided a
better indicator of where brain activity is evoked than did scalp fiel
ds. These procedures also corrected for anatomical variations between
scalp and brain from subject to subject. In two right-handed female su
bjects, evoked activity was observed in the left posterior parietal an
d the right occipital, parieto-occipital and posterior temporal cortic
es. Evoked activity was observed in the left parietal cortex for lumin
ance processing, in the right parietal cortex for texture processing a
nd in the right temporal cortex for color processing, which was select
ive for the particular contrast.