Abstract
Decades of research have shown that different stimulus dimensions engage different cortical areas along the traditional visual hierarchy. For example, neuronal populations in the primary visual cortex strongly respond to oriented gratings while more complex stimuli, such as dot motion kine-matograms or faces, engage higher level visual areas like area MT+/V5 and occipito-temporal face-sensitive areas, respectively. Extensive associative learning (e.g., fear conditioning), with exemplars taken from a specific stimulus dimension, systematically alters visuocortical responses to this dimension. Electrophysiological studies in humans, however, tend to conflate area-specific effects with stimulus onset effects, i.e., responses to low-level contrast or luminance changes associated with stimulus presentation over a blank screen. Thus, alternative hypotheses exist stating that associative learning (i) always involves changes in primary visual areas, irrespective of the feature dimension manipulated, versus (ii) selectively involves the tissue most sensitive to the feature dimension relevant for learning. In the present study, we used the steady-state potential technique and a classical aversive conditioning paradigm in which one exemplar from three stimulus dimensions along the visual hierarchy were selectively paired with an aversive outcome: sinusoidal gratings (Gabor patches), dot motion kinematograms, and faces. We minimized luminance changes across the trial by adding the stimuli to visual Brownian noise which was regenerated before each trial. We estimated the sources of the ssVEP using an L2 (minimum norm) inverse projection (n = 22) and examined the differences between conditioned stimuli paired (CS+) or not paired (CS−), with an aversive loud noise, using mass-univariate permutation-controlled F-tests. Results indicate that associative learning effects (measured as the difference between CS+ and CS−) involve separate and distinct brain regions: calcarine cortex for Gabor patches, mid-occipital gyrus for visual motion, and temporal cortex for faces. Thus, adaptive changes in visual cortex selectively occur in areas most sensitive to the critical stimulus dimension.