Terence L. Tyson¹, Jorge Otero-Millan^1,2; ¹University of California, Berkeley, ²Johns Hopkins University

The gaze holding system can adapt when introduced to different input statistics. For example, gaze holding to an extreme eccentricity over a long duration biases the input to one side of the oculomotor range. This adaptation can manifest as rebound nystagmus, a directional drift that brings the eye back to the previously held gaze location followed by a corrective saccade in the opposite direction. From our experimental work, we showed that adaptation can change two aspects of the gaze holding system – the oculomotor setpoint and the leakiness of the oculomotor integrator. We define the setpoint as the eye position with zero velocity and the integrator leakiness as how quickly the eye drifts back to central gaze. Here, we attempt to model these adaptive mechanisms. The proposed model tracks the expected eye position and its variance, representing the setpoint and integrator leakiness, respectively. For the simulation, we provide artificial inputs to the model that mimics our gaze holding experiment and produces a curve that relates eye drift velocity to eye position. We then estimate the setpoint (x-intercept) and integrator leakiness (inverse slope) and compare these simulated results with our empirical findings. We show that our computational model can accurately reveal both changes in the setpoint and in the integrator leakiness given the biased input statistics introduced from prolonged eccentric gaze holding.

Acknowledgements: National Institute of Health (NIH) R00EY027846 research project grant