Abstract
When a static stimulus appears at two locations in succession, one can perceive the illusion of apparent motion (AM) across the two locations. Previous research using fMRI and a forward encoding model shows that intermediate visual features of a transforming object, which are absent in the retinal input but are spatiotemporally interpolated during AM, are reconstructed in population-level feature-selective tuning responses in the region of V1 corresponding to the AM path (Familiar, Chong, & Shim, 2014 VSS). However, little is known about how this interpolated feature information is reconstructed in non-stimulated regions of early visual cortex via top-down processing. Here using fMRI and a novel method of feature channel-based functional connectivity we examined functional interactions between the region of V1 that is retinotopically mapped to the AM path (V1-AM path) and high-level visual areas that are implicated in motion and shape processing (MT+ and LOC) during the perception of AM. Participants viewed rotational AM between two oriented gratings in the periphery and the population-level orientation channel responses in the V1-AM path, MT+ and LOC ROIs were reconstructed. We used the response at each orientation channel as input to compute channel-based inter-area correlations between V1 and each of the higher-level areas. The results showed that correlations between V1-AM path and MT+, as well as V1-AM path and LOC, were higher at orientation channels corresponding to the interpolated orientations on the AM path (0° or 90° depending on AM condition) when AM was perceived compared to when AM was abolished by simultaneously presenting the two gratings. This channel-selective increase in functional connectivity was not evident at other orientation channels. These results suggest that signals from high-level motion and shape processing areas, such as MT+ and LOC, may provide feedback information that mediates dynamic feature reconstruction in early retinotopic cortex during AM.
Meeting abstract presented at VSS 2015