Peri-saccadic mislocalization in a rhesus macaque monkey depends on the visual appearance of the saccade target

Briefly presented high contrast stimuli are strongly mislocalized peri-saccadically. This phenomenon is likely a consequence, at least partially, of saccade vector information being internally relayed (via corollary discharge) to the visual system. Indeed, thalamic relaying of superior colliculus (SC) saccade-related motor bursts to the frontal cortex is necessary for peri-saccadic response field remapping in the latter (Sommer & Wurtz, 2006). Given recent evidence that SC motor bursts contain a visual sensory representation of the saccade target appearance (Baumann et al., 2022), we hypothesize that SC-sourced corollary discharge may transmit more than just vector information to the cortex. However, neuronally testing such a hypothesis requires a means to assess peri-saccadic visual mislocalization in non-human primates, and to ask whether such mislocalization depends on saccade target visual appearance. We trained a male monkey to repeatedly generate the same saccade within a session, but to two different possible targets (100% contrast grating of 1.5 deg radius and either 0.5 or 8 cycles/deg spatial frequency). After online saccade detection, we presented a flash (~10 ms; black; 0.45 by 0.45 deg) at one of three possible locations ahead of the saccade target center (10.6 deg; directly ahead or with +/- 45 deg deviation from saccade direction); flash onset occurred immediately or after 40 or 100 ms from saccade detection. After 500-800 ms from saccade end, we removed the grating, instructing the monkey to look towards where it previously saw the flash. Across 13735 trials from multiple sessions and different saccade directions, the monkey consistently exhibited robust mislocalization towards the saccade target location, like in classic human studies, and there was recovery with delayed post-saccadic flashes. Importantly, mislocalization strength depended on saccade target appearance. Thus, we established a suitable model for neuronally testing predictions regarding the functional role of sensory tuning in SC neuronal movement commands.