Abstract
Natural vision with eye movements affords a presaccadic extrafoveal preview of soon-to-be-fixated objects which facilitates their subsequent foveal processing. Recently, several studies have started to investigate the effects of the validity of the presaccadic preview on electrophysiological correlates of postsaccadic object recognition by combining simultaneous EEG/eye-tracking with intra-saccadic display changes. A systematic analysis of these neural preview effects and their spatiotemporal features across studies may help to elucidate mechanisms underlying trans-saccadic predictions and integration and may also enable comparisons with more established EEG phenomena (such as the visual MMN). In this work, we aggregated the data of 13 published EEG/eye-tracking studies (20 experiments, n=373) that have investigated extrafoveal-to-foveal preview effects for heterogenous stimulus categories (words, characters, faces, objects, Gabors) both with and without an intervening saccade. Evoked potentials from all studies were jointly analyzed in one linear mixed model (“mega analysis”). Results reveal a remarkably consistent electrophysiological signature of extrafoveal preview (in)validity on subsequent foveal brain responses across studies, with strongest effects observed on the late parts of the occipitotemporal N1/N170 component (more negative after invalid previews, peak effect across studies at 226 ms) with an opposite hemispheric lateralization for words vs. faces/objects. However, results also show robust earlier influences on foveal processing beginning around 90 ms and peaking at ~130 ms (falling flank of P1 more positive after invalid previews). We highlight several factors modulating these neural preview effects, in particular stimulus category, active saccade execution (stronger effects after saccades) and display change awareness. We also show that more delayed occipitotemporal preview effects (peaking at 260 ms) are robust against a strong change in lower-level features across the saccade (a change in letter case), indicating that postsaccadic EEG measures can distinguish extrafoveal processing/predictions at different levels of the visual hierarchy (e.g., visual features vs. more abstract letter representations).