Human oculomotor function: reliability and diurnal variation.

To provide information on test-retest reliability for seven oculomotor paradigms currently used in studies of schizophrenia and other neuropsychiatric conditions, we tested eight controls at four weekly intervals, twice in the morning (8-10 AM) and twice in the afternoon (3-5 PM). Intraclass correlation coefficients were significant (p < .05) for both AM and PM pairs of measures as well as for mean AM and PM pairs for closed-loop pursuit gain, open-loop pursuit gain (using velocity as the measure), saccadic frequency during pursuit and fixation, visually and nonvisually guided saccadic latency and velocity, antisaccadic latency, and premature reflexive saccades during the memory-guided saccade task. Acceleration as a measure of open-loop gain (for slower targets) and accuracy of saccades to a moving target were only reliable at PM testing time. Nonvisually guided saccadic accuracy and inappropriate reflexive saccades during the antisaccade task were not reliable, possibly due to the narrow range of values for these measures. Except for approximately 10% fewer saccades during pursuit and fixation in the morning, there were no consistent diurnal differences. These findings suggest that, in a small sample of subjects, most measures of oculomotor function are stable across time and may reflect underlying neurophysiologic traits.