David Melcher¹, Michele Deodato²; ¹New York University Abu Dhabi

During active vision, the extrafoveal preview of a saccade target facilitates recognition of the postsaccadic stimulus when it is valid, but impairs it when invalid, a phenomenon known as the preview effect. It is unclear whether this effect reflects genuine facilitation from valid previews or whether it is inflated by the perceptual cost of invalid ones. Moreover, it remains unknown whether preview effects for object recognition scale with the amount of time that parafoveal information is available before a saccade, as predicted by continuous evidence-accumulation accounts. Across two experiments (behavioral and EEG), we used a gaze-contingent paradigm in which participants executed rapid saccades toward peripheral stimuli that served as valid, invalid, or absent (placeholder) previews. After the saccade, a target stimulus appeared and participants reported whether it depicted a face or a house. Behaviorally, we found that reaction times reflected clear perceptual benefits from valid previews and costs from invalid ones, demonstrating that preview effects arise from both facilitation and interference, similar to findings in reading research. Benefits and costs increased as a function of preview duration, indexed by saccadic reaction times. Interestingly, even absent previews produced time-dependent shifts in performance, suggesting the presence of a temporal weighting of extrafoveal evidence. Overall, this indicates that preview mechanisms are not “one-shot” but instead reflect continuous temporal integration of peripheral information. EEG responses supported this view: early postsaccadic neural responses were modulated by preview validity and preview time, revealing distinct neural signatures for facilitation and interference. Together, these findings show that preview effects in object recognition arise from a time-dependent weighting of extrafoveal evidence. This supports a framework in which the visual system continuously accumulates and updates sensory evidence across saccades to guide rapid object recognition.

Acknowledgements: NYUAD Center for Brain and Health, funded by Tamkeen under NYU Abu Dhabi Research Institute grant CG012