Multi-dimensional characterizations of operator state: A validation of oculomotor metrics.

Abstract

Relationships between overt behavioral measures (such as reaction time (RT) and response accuracy) and psychophysiological indices of oculomotor, electroencephalographic (EEG), and cardiovascular activity were delineated within the context of a continuous performance task. Subjects maintained comparable mean performance levels across all task segments. However, variability in response speed and accuracy increased with time-on-task. The increased variability was associated with longer blink durations, decreased post-stimulus blink latencies, decreased anticipatory and reactive saccade velocities and amplitudes, and fewer and later reactive saccades. While blinks were inhibited prior to all stimuli, the post-stimulus period of inhibition was longest following imperative stimuli. Target stimuli were also associated with more efficient anticipatory eye-movements. In the absence of a blink, RTs were substantially delayed. When blinks were present, very short latency blinks were associated with more variable RTs and increased errors. If blink latencies were late, RTs were late as well. Trials containing especially long duration blinks were associated with decreases in performance accuracy. Target stimuli followed by reactive saccades were responded to more slowly and with less accuracy than when effective anticipatory eye-movements preceded stimulus onset. Furthermore, the larger the amplitude of the reactive saccade, the greater the increase in RT. Abstracting peripheral information (recheck saccades) also incurred a cost in terms of increased RT. These electrooculographic effects were accompanied by systematic changes in EEG and cardiovascular responses and exploratory multi-variate modeling indicated the degree to which both within, as well as between-subject performance variability, could be accounted for by various combinations of the psychophysiological measures.