Purchase this article with an account.
Melchi Michel, Umang Parikh; The timecourse of spatial information integration across saccades. Journal of Vision 2015;15(12):1305. doi: https://doi.org/10.1167/15.12.1305.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
While most real-world tasks involve multiple fixations, visual performance thresholds are typically measured for single-fixation tasks. How might visual information integration during otherwise equivalent individual fixations differ between single-fixation and multiple-fixation tasks? In particular, how does an intervening saccade affect the characteristics of spatial and temporal integration used in a visual judgment? We investigated this question using reverse correlation in a transsaccadic perceptual discrimination task. 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. The Gaussian blob target was presented across a 10° saccadic eye movement cued to the right or left of the fixation location. Successful performance of this task required integrating information temporally across the intervening saccade. In the main experimental condition, two additional flickering blobs were added to the display, displaced laterally 10° to the left and right of the target. These flanking blobs were added to determine how observers use information from two locations—one corresponding to the prospective retinal location of the target and the other corresponding to the prospective location of ‘remapped’ receptive field(s) associated with the presaccadic target location—both of which have been reported to influence perceptual judgments. Observers were explicitly directed to ignore these flankers. We averaged the time-varying blob luminances aligned to the onset of the saccade across trials to determine the spatiotemporal pattern of visual information integration. The result of this analysis is a psychophysical kernel that indicates the relative weight given to each of the three spatial locations as a function of time. The computed kernel indicates that observers begin integrating information from the postsaccadic target location about 100ms before the onset of a saccade. We discuss this as evidence of perisaccadic remapping in the direction opposite the saccade.
Meeting abstract presented at VSS 2015
This PDF is available to Subscribers Only