Abstract
Binocular rivalry occurs when sensory inputs from two eyes are incongruent. Our visual system resolves the ambiguity by preferring one of the stimuli at a particular moment, creating alternating percepts. It is yet unclear how neural representations of low-level visual features are modulated by perceptual alternations in human visual areas. In two experiments, we used fMRI-based encoding models to reconstruct channel responses to suppressed/dominant orientations and motion directions during binocular rivalry. For orientation rivalry, participants first underwent an fMRI session and viewed contrast-reversing sinusoidal gratings of six possible orientations with both eyes. BOLD response patterns were used to train an encoding model of six orientation-selective channels. In binocular orientation rivalry sessions, we presented two orthogonally oriented gratings to the two eyes, and each binocular rivalry run was followed by a replay run to simulate the perceptual effect in the immediately preceding run. Relative to the reconstructed channel responses from the replay runs, we found that, in V1, the channel response to the suppressed orientation was higher during binocular rivalry, while the channel response to the dominant orientation was weaker. These effects were also found in V2 and V3, albeit less pronounced. In V4, the channel response profiles in the rivalry and replay conditions became indistinguishable. For motion direction rivalry, we used RDKs moving in opposite directions with 100% motion coherence as rivalry stimuli. Compared with the reconstructed channel responses from the replay runs, we found elevated responses to the suppressed direction and lowered responses to the dominant direction in V1-V3, V3A, and MT+. Similar to the orientation rivalry experiment, no significant difference was found between the rivalry and replay conditions in V4. Taken together, our findings suggest that during the binocular rivalry of low-level visual features, the suppressed information is substantially represented in early and intermediate visual areas.
Meeting abstract presented at VSS 2018