Yunshan Cai¹, Jude F. Mitchell^1,2,3, Sara S. Patterson^1,2,3,4; ¹Neuroscience Graduate Program, University of Rochester, ²Center for Visual Science, University of Rochester, ³Brain and Cognitive Sciences, University of Rochester, ⁴Flaum Eye Institute, University of Rochester

Optokinetic nystagmus (OKN) is a reflexive eye movement for gaze stabilization with slow-phase tracking followed by a resetting saccade. OKN is highly conserved across vertebrates, with differing patterns of monocular temporal-nasal symmetry depending on eye placement and cortical involvement. In lateral-eyed species, horizontal OKN (hOKN) is primarily driven by nasally-tuned ON direction-selective retinal ganglion cells (ON-dsRGCs) projecting to the nucleus of the optic tract (NOT). In contrast, forward-facing species show abundant cortical projections to the NOT that support binocular integration, improving tracking gain at higher speeds and providing monocular hOKN symmetry. Cortical inputs are thought to be the sole driver of adult primate hOKN. Failed maturation of corticofugal (cortex-to-NOT) projections leads to pathological nystagmus and permanent hOKN asymmetry, consistent with a direct contribution from retinal direction signals. The existence of ON-dsRGCs in the primate retina was only recently confirmed; however, their contributions to hOKN remain an open question. We systematically quantified hOKN in two marmosets. We measured slow-phase gain and temporal-nasal asymmetries across stimulus velocities under binocular and monocular viewing using full-field drifting gratings and a digital dual-Purkinje eye tracker (Wu et al., 2023). Under monocular viewing, both animals exhibited stronger nasal than temporal gain. Under binocular viewing, gain increased and greater nasal-temporal symmetry was observed. An unexpected multi-modal distribution of slow-phase velocities occurred in 17% of the trials (374/2243). A Gaussian mixture model captured a clear bimodal distribution within these trials. The most prominent bimodal distributions were found at 40°/s, with two stable gain modes (monocular: 20.8±8°/s, 8.1±5.6°/s; binocular: 26.3±5.7°/s, 11.9±5.3°/s). Such bimodality represents a previously undescribed feature of hOKN behavior that could reflect differential cortical and retinal contributions. Overall, the common marmoset hOKN behavior is highly consistent with previous measurements in other non-human primates and humans.

Acknowledgements: YC and SSP from NIH R00-EY035323 and Research to Prevent Blindness, JFM from NIH R01- EY030998, and P30-EY001319