Abstract
As we move forward in the environment we experience a radial expansion of retinal images, of which the center corresponds to the instantaneous direction of self-motion (heading). Humans and non-human primates can precisely perceive their heading directions even when they are making simultaneous pursuit eye movements, which shifted the center of the expansion pattern on retina. Previous controversial studies have shown that both retinal and extra-retinal strategy can account for accurate heading perception during pursuit. Here we propose that the visual system flexibly combine the retinal and the extra-retinal signals in a Bayesian-like statistically optimal fashion. To test predictions derived from this Bayesian-like framework, we devised a pair of experiments in which we independently manipulated the reliability of each input signals in a two alternative heading direction discrimination task. Two aspects of visual signal reliability were tested separately, the simulated heading speed and element motion coherency. Our results showed that the contribution of extra-retinal signals to heading judgments increased with increasing pursuit eye movement speed, and decreased with increasing simulated heading speed. Similarly, observers were relying on more of retinal signals instead of extra-retinal signals to support their heading judgments when we increased the coherency of element motion in retinal flow. We hope our findings can be helpful to unify previous differing observations in a rather sample framework.
Meeting abstract presented at VSS 2014