Frequency characteristics of the smooth pursuit component in tracking eye movements

Abstract

Frequency responses of the human eye tracking system vary remarkably according to the motion of the visual target. In general, periodic (predictable) target motion yields a higher gain and a smaller phase lag in comparison with random (unpredictable) target motion. This paper deals with the question of which of the saccadic and smooth pursuit systems causes this variability.

Two kinds of target motions, sine (predictable) and Gaussian random (unpredictable) waves, were presented to the subject with either of two instructions: to fixate on the target or to fixate on the target and do mental arithmetic simultaneously. Recorded eye movements were separated into their saccadic and smooth components, and the respective frequency characteristics were analysed.

The result shows that the large phase lag in non-predictive tracking is mainly due to the saccadic system. At low frequencies, the smooth component reveals a rather large phase lead in non-predictive tracking but not in predictive tracking.

We also describe a mathematical model of the pursuit system to explain the experimental result.