Transsaccadic prediction of visual features: Weighting, scope and generalization

点击翻译按钮获取中文摘要

We move our gaze several times each second to bring the high-resolution foveal region to interesting objects for detailed visual sampling. As a consequence, with each new fixation, objects change their retinal position and, owing to the visual system's nonhomogeneity, their spatial resolution as well. The main goal of the project is to investigate how the visual system uses predictions to deal with the different visual inputs sampled at the periphery and fovea. In the first round we have conducted several learning experiments in which we swapped objects across saccadic eye movements unnoticed by most participants. Afterwards, we tested how these altered visual experiences affect predictions of peripheral and foveal input. We showed that peripheral perception was biased toward previously associated foveal input. Moreover, visual search was biased toward previously associated peripheral input. The overarching goals of the second period are (1) to characterize the weighting function of predicted foveal and actual peripheral input, (2) to determine the scope of transsaccadic prediction, and (3) to investigate the specificity, generalization, and persistence of transsaccadic prediction.