Paul Jolly^1,2 (), Michele A. Cox^1,2, Ashley M. Clark^1,2, Sanjana Kapisthalam^1,2, Yuanhao H. Li^1,2, Ruitao Lin^1,2, T. Scott Murdison³, Alina Neverodska^1,2, Jie Z. Wang^1,2, Bin Yang^1,2, Zhetuo Zhao^1,2, Michele Rucci^1,2; ¹University of Rochester, ²Center for Visual Science, University of Rochester, NY, USA, ³Reality Labs, USA

It is well established that humans tune their saccades according to visual demands. Recent research has shown that while centering objects of interest onto the high-acuity fovea, saccades also yield luminance transients that facilitate neural encoding (Mostofi et al, 2020) and enhance visual sensitivity (Boi et al, 2018) in a low range of spatial frequencies. These previous studies were conducted with the head of the observer strictly immobilized. Under natural viewing conditions, most saccadic gaze shifts involve coordinated movements of the eye and head, and it remains unclear whether the luminance transients resulting on the retina from these joint movements differ from those measured under head immobilization. In this study, we simultaneously measured head and eye movements using a custom apparatus that enables arcminute-level resolution. This device is composed of a motion capture system (Optitrack) integrated with the oscillating field monitor (Eibenberger et al, 2016), a magnetic-induction eye-tracker. Subjects (N=17) wore scleral eye coils and a tightly-fitting helmet while performing four real-world tasks with varying acuity demands: reading a Snellen eye-chart, threading a needle, sorting beads, and searching targets in natural scenes. As expected, saccade amplitude distributions varied considerably across tasks. However, all distributions peaked for very small saccades, ranging from approximately 10’ in Snellen to 1.5 deg in sorting. On the retina, head and eye movements yielded traces highly similar to those given by saccades alone under head immobilization, so that the power spectra of the resulting luminance modulations were also similar to those previously reported for head-fixed saccades. Because of the differences in amplitude distributions, these modulations emphasized high spatial frequencies in the Snellen task and extended to lower spatial frequencies in the other tasks. These results suggest that the luminance transients from head-eye saccades contribute to task-relevant spatial representations, suggesting an additional function of head-eye coordination.

Acknowledgements: Research supported by Reality Labs and NIH grants EY018363 and P30 EY001319