Abstract
Grouping separate regions with similar feature components is critical for forming coherent percepts of whole objects and complete scenes. When these components are ambiguous, they all may be perceived as identical in all regions despite different neural representations for each region (Kovács et al., PNAS 1996; Slezak & Shevell, JOSA A 2018). In general, perceptual ambiguity may be resolved by monocularly-driven or binocularly-driven neural mechanisms. This study presented two discs, each one designed to induce ambiguity at a different level of neural representation, to determine if a grouping process can act on different kinds of competing neural representations. Alternatively, the difference in neural level of representation may inhibit grouping. METHOD Two equiluminant chromatic discs were presented above and below a fixation cross, one using standard binocular rivalry (SBR) and one using interocular-switch rivalry (ISR; Christiansen et al., JOV 2017). Observers reported when both discs appeared the same color. This “mixed-method” measurement was tested in conjunction with conditions presenting both discs using only one method (either both ISR or both SBR, the latter with separate conventional and patchwork trials (as in Kovács et al., PNAS 1996)). To assess the null hypothesis of no grouping, independence predictions were determined by measuring percepts of the top or bottom disc presented alone. RESULTS/CONCLUSIONS All five observers showed significant grouping for discs presented in the same method, no matter whether SBR or ISR (14 of 15 contrasts significant, each at p< 0.05). On the other hand, four of the five observers showed no significant difference between the independence prediction and the “mixed-method” measurements (p>0.05). These results support the view that grouping consistently occurs when two rivalrous discs are resolved at the same neural level, but less so or not at all when the discs are resolved by neural processes at different levels of the visual system.
Acknowledgement: Supported by NIH EY-026618