Abstract
When the two eyes receive different inputs, the visual system either fuses the images into one coherent percept, or engages in binocular rivalry. What neural computations decide between rivalry and fusion? If these computations compare opposing neural signals for fusion and rivalry, and adjusts their weights based on stimulus history, then adapting to either fusion or rivalry might bias the subsequent perception of an ambiguous, partially fusible stimulus towards the other state. Stimuli were constructed using 0.5 contrast orthogonal square-wave gratings (±45°; 0.7 cpd). There were two types of adapters: rivaling orthogonal gratings presented dichoptically, and fusible plaids (the sum of the orthogonal gratings) presented binocularly. The ambiguous test patterns were plaids where a different component grating in each eye was presented with reduced contrast (0.3). During each trial, subjects viewed the adaptor for 6 s followed by the ambiguous test for 8 s. Subjects pressed one of four buttons to report percepts of: +45 grating, –45 grating, piecemeal mixed, or fused plaid. For the rivalrous adapter, the test pattern was always initially perceived as a fused plaid and then began to rival after 6.24±0.44 s (SE; n=6). For the fused adapter, the test pattern began to rival much sooner, either immediately upon presentation, or shortly thereafter (3.25±0.27 s). Similar results were obtained regardless of whether the orientations of the rivalrous adapting gratings matched that of the higher or lower contrast component of each eye’s test plaid. These results strongly suggest that adaptation affected an opponent neural mechanism that determines whether a stimulus fuses or rivals. This mechanism may depend upon the ocular opponency units proposed by a recent computational model of rivalry (Said & Heeger 2013).
Meeting abstract presented at VSS 2015