Retinal image motion distracts visual motion memory even when generated by eye movement

Interference between visual short-term memory (VSTM) and task-irrelevant sensory distractors is a well-documented phenomenon across a wide range of visual features, which is believed to stem from neural interactions between mnemonic and sensory processing. An overlooked question in the ongoing debate is whether this VSTM distraction is linked to the original sensory information or the ultimate perceptual experience. Here we addressed this issue by leveraging the perceptual invariance during ocular tracking of an object (i.e., smooth pursuit), where retinal image motion induced by smooth pursuit is encoded in early visual areas involved in motion processing, including the middle temporal (MT) and the medial superior temporal (MST) areas, but is perceptually suppressed. Our results showed that retinal image motion during the VSTM maintenance (delay period) attracted the memorized motion speed, consistent with previous studies. Importantly, the impact was comparable whether the retinal image motion was due to physical displacement of objects in world coordinates or to apparent motion induced by smooth pursuit without actual motion in world coordinates. In fact, observers’ responses in the VSTM task were partially predicted by retinal image motion during the delay period by a cross-condition classifier, where the distraction effect induced by the smooth pursuit-induced apparent motion was replicable using a classifier trained on data from the condition with the world-coordinate motion, and vice versa. These findings provide behavioral evidence that sensory inputs can distract VSTM without conscious perception and also suggest that the VSTM system shares neural substrates with sensory processing, but not with perception.