Abstract
A growing number of physiological and psychophysical studies have reported that visual information is spatially remapped in a brief interval preceding the onset of an intended saccade. Researchers have interpreted these results as evidence of predictive remapping: an updating of visual or attentional receptive fields in anticipation of a saccade. Recently, this interpretation has been challenged by an alternative hypothesis that any apparent remapping is not predictive, but rather an epiphenomenon of saccadic target selection (Zirnsak & Moore, 2014). Last year (Michel & Parikh, 2015), we presented results from a reverse correlation study of transsaccadic perceptual discrimination designed to gather and evaluate evidence of remapping under either hypothesis. Observers were asked to report the average luminance polarity of a small, flickering Gaussian blob displayed peripherally for 500ms, whose luminance in each 10ms frame was selected randomly. We found that when irrelevant flickering blob stimuli were added to the display, laterally flanking the target at a distance equal to the magnitude of the instructed saccade, observers' reports were influenced by the luminance of the stimulus in a location corresponding to the prospective retinal location of the discrimination target (in a direction opposite the saccadic target). Importantly, information from this location only influenced the decision in the 100ms preceding the onset of the saccade, consistent with perisaccadic remapping. In the current study we modified the task by adding additional, intermediate, flankers to better characterize the spatial characteristics of the remapping. We found that the remapping is spatially precise. Of the 4 flankers we presented, only the one consistent with the prospective retinal location of the discrimination target influenced the discrimination in the perisaccadic interval. This result is consistent with remapping that is predictive and attention-based (e.g., Cavanagh, Hunt, Afraz, & Rolfs, 2010). We argue that it is inconsistent with remapping due to saccadic target selection.
Meeting abstract presented at VSS 2016