Predictive velocity estimation in the pursuit reflex response to pseudo‐random and step displacement stimuli in man.

1. Eye movements have been recorded in man in response to various forms of continuous and discontinuous target motions in the horizontal plane in an attempt to establish the mechanisms of prediction in the pursuit reflex. 2. In an initial experiment the target motion was composed of four sinusoids, each of peak velocity +/‐ 3.3 deg/s. The three lowest frequencies (0.11, 0.24 and 0.37 Hz) remained constant whereas the highest frequency (F4) was varied from 0.39 to 2.08 Hz. When F4 was 0.39 Hz, all frequency components had a high level of eye velocity gain (mean 0.92) but as F4 was increased there was a significant (P less than 0.001) decline in gain for all three low frequencies which reached a minimum (mean 0.53) when F4 was 1.56 Hz. However, the gain for F4 always remained at a high level, comparable to that evoked by a discrete frequency sinusoid of the same frequency. 3. When the highest‐frequency sinusoid was replaced by a square wave of identical amplitude a similar decline in gain for the low frequencies was observed. Eye velocity exhibited a quasi‐sinusoidal modulation at the frequency of the square wave even though the rapid steps did not constitute a suitable stimulus to the visual velocity feed‐back mechanisms. 4. When only two sinusoids were mixed to form the pursuit stimulus a similar break‐down in gain of the lower‐frequency component was observed which reached a minimum (mean gain 0.58) when F2 was between 1 and 2 Hz. This implies that the predictability of stimulus motion is dependent, not on the complexity of the stimulus, but on its highest‐frequency component. 5. Presentation of square‐wave target displacements alone confirmed that smooth eye movements could be evoked by such a stimulus. Eye velocity was at a maximum between 1.0 and 1.5 Hz and was predictive of ensuing target displacement. Responses to staircase step sequences of varying duration indicated that prediction was based on an assessment of the duration of the preceding sequence. 6. Tachistoscopic presentation of targets during low‐frequency sinusoidal oscillation indicated that illumination of the target for very short periods (10‐320 ms) as few as two times per cycle during minimum velocity phases enhanced the perception of continuous movement. A predictive eye movement was evoked with a high level of peak velocity which then decayed until the subsequent exposure of the target.(ABSTRACT TRUNCATED AT 400 WORDS)