Computational mechanisms underlying eye-head coordination

Abstract

Gaze or eye-head coordination is an ethologically and physiologically important process that directs the visual system to objects of interest. Due to its relative simplicity, it is an ideal system to study multi-effector coordination. Whereas in the past, gaze control models have been developed on experimental data that focused on the mean responses, I used variability as a tool to study and model temporal coordination, spatial coordination and tested stochastic models of gaze control to test the predictions of the independent, common and interactive models of gaze control. I also used McLaughlin’s task paradigm of gaze adaptation to study gaze coordination under more dynamic conditions and studied its implication for models of gaze control. In the context of temporal coordination, the results suggest a common movement preparation phase underlying eye and head movement initiation. In the case of spatial coordination, I found evidence for the presence of a global gaze error feedback. However, the gaze learning results, which showed effector-specific learning suggested the presence of separate controllers that contribute unequally to gaze adaptation. Taken together, my results collectively suggest that an interactive model, that has a separate internal saccadic feedback controller and a global gaze feedback controller maybe the best model that explains my results.