Properties of human eye-head gaze shifts in an anti-gaze shift task.

We investigated the metrics and kinematics of human eye-head gaze shifts using the anti-gaze shift task. Surprisingly, no systematic difference was found between peak gaze velocities of large pro- and anti-gaze shifts. In a follow-up experiment that equated perceived stimulus luminance across multiple eccentricities, pro-gaze shifts were consistently faster than anti-gaze shifts. In both experiments, we did not observe any head-only errors where initial head motion dissociates from gaze direction, even though many subjects generated such movements in other paradigms. These experiments confirm the influence of stimulus luminance on comparative movement velocity, and demonstrate that the behavioural set assumed in this task discourages head-only errors.