Abstract
Several studies have shown that the variability of smooth pursuit eye speed can match perceptual speed discrimination thresholds. While Kowler & McKee (Vision Research, 1987) and Gegenfurtner et al (Journal of Vision, 2003) have shown this to hold during steady state pursuit, Osborne et al (Nature, 2005) claimed that the eye speed variability during the initiation phase of pursuit is also due to sensory noise alone. They measured pursuit variability using an innovative new method, but they did not determine perceptual variability of their monkey observers. We therefore measured pursuit and perceptual variability using the same stimuli.
Subjects had to pursue a Gaussian dot in a step-ramp paradigm. The dot moved at one of 5 different speeds (5.5 to 16.4 deg/s) for 1 s. Individual traces were fitted against their average to estimate their variability in latency and amplitude. The analysis was performed for short intervals containing the initiation period only and for longer intervals including steady-state pursuit. In a different set of trials, subjects had to compare the speed of the stimuli to a memorized standard (11.3 deg/s) for presentation durations ranging from 150 ms to 1s. In a third set of trials, subjects performed both tasks, eye movements and psychophysical judgments, for stimuli presented for 1 s.
In agreement with published studies, we found Weber fractions of less than 10% in the psychophysical speed discrimination task, even for the shortest presentation duration of 150 ms. Pursuit variability was about four times as high for the shortest analysis interval (250 ms) containing the initiation phase only. Only for pursuit analysis intervals of 400–500 ms, pursuit variability approached perceptual variability.
Our results show that the motor system contributes significantly to the variability of smooth pursuit eye movements during the initiation phase.