Human Systems Integration Division @ NASA Ames

The smooth pursuit system causes the eyes to move to cancel "retinal slip" of a target and presumably uses the results of visual motion computations to control motor ouput. The overall latency of this system can be thought of as having two components: an input-processing component that depends on stimulus properties, and a motor output component that does not.

This poster presents a method for assessing pursuit latency that we call the eye movement correlogram. The subject is instructed to maintain fixation of a target that moves on a pseudo-random trajectory, designed to make prediction impossible.1 The trajectories are computed by integrating a white noise velocity profile. Some low-pass filtering may be applied to the velocity signal before integration to smooth the trajectory. The subjects' eye movements are recorded, and the signals are differentiated to produce eye velocity. Saccades are detected using a velocity criterion, and values of the smooth velocity are interpolated in the neighborhoods of saccades. The resulting smooth velocity signal is cross correlated with the stimulus velocity. When averaged across a number of trials (each having a different random trajectory), an impulse response-like function is revealed, which is the correlogram. The time at which the peak of this function occurs can be interpreted as the latency of the pursuit system.