Sensory tuning in neuronal movement commands: potential perceptual consequences

We normally perceive a stable visual environment despite repetitive eye movements. To achieve such stability, visual processing integrates information across saccades, and laboratory hallmarks of such integration are robustly observed by presenting brief peri-movement probes. In one phenomenon, peri-saccadic suppression, visual sensitivity to such probes is strongly suppressed, and in another, peri-saccadic mislocalization, perceived probe locations are grossly erroneous. Both phenomena are believed to depend, at least in part, on corollary discharge associated with saccade-related neuronal movement commands. However, we found in a companion study (VSS 2022) that superior colliculus (SC) motor bursts, a known source of corollary discharge (Sommer & Wurtz, 2004), paradoxically exhibit strong sensory tuning in them. Therefore, here we investigated whether both peri-saccadic suppression and mislocalization depend on saccade-target visual properties. We asked eight humans to generate saccades to either high (5 cycles/deg) or low (0.5 cycles/deg) spatial frequency gratings. We always placed a high contrast target spot at grating center, to ensure matched saccades across image types. In one experiment (suppression), peri-saccadic probes (~12 ms duration) appeared at one of four fixed locations, which was indicated by subjects via button press (4AFC paradigm), relative to the saccade target. We varied probe contrasts across trials to calculate perceptual thresholds. In a second experiment (mislocalization), we used high contrast probes, and the subjects pointed (via mouse cursor) at their perceived locations. In both experiments, we observed significantly stronger peri-saccadic suppression and mislocalization for low spatial frequency saccade targets, consistent with our neuronal results (VSS 2022). This was despite matched saccade metrics and kinematics to trials with high spatial frequency targets. To the extent that peri-saccadic suppression and mislocalization depend on corollary discharge, our results suggest that such discharge relays more than just saccade vector properties to the visual system; saccade-target visual features can also be peri-saccadically processed.