Abstract
How do primates localize flashed stimuli and why do eye-movements induce characteristic visual spatial mislocalization? Although of ultimate importance to everyday life, the answer to this question is still open. Physiological, behavioral and theoretical studies have suggested various neural mechanisms, from displacement of visual receptive fields during eye-movements to erroneously decoded eye-position. Here we propose a simple model for (mis-)localization during fixation and smooth pursuit eye-movements. Model predictions were verified in localization tasks in both human subjects and non-human primates. Our model assumes that localization is accomplished by a combination of two independent neural signals: a normally distributed position estimate, which is combined with an (erroneous) eye-position representation. To test our model, we employed localization of briefly presented stimuli during monocular fixation and smooth pursuit eye-movements in both, humans and monkeys. Human subjects reported perceived location by an established procedure (ruler), while monkeys indicated their percept by means of a saccade. Importantly, monocular viewing is challenged by a gap in retinal space, i.e. the blind spot. In contrast, our model predicts a gap-free, quasi continuous perceptual space with perceived locations covering also the area of the blind spot. According to our model, perceived space is shifted in the direction of smooth pursuit, as previously reported also for binocular eye-movements. Our experiments confirmed both model predictions in man and monkey. First, all subjects showed a continuous perceptual space with no obvious gap at the blind spot. Second, smooth pursuit induced an overall shift of the perceptual map with respect to fixation. Our findings imply a combination of two independent neural signals as the neural basis of flash localization: a visual map and an eye-position signal. Our model agrees well with previous studies suggesting the source of perceptual mislocalization during eye-movements to be an erroneous internal representation of eye-position.
Meeting abstract presented at VSS 2017