Eyes wide open: regulation of arousal by temporal expectations

To perform a task continuously over an extended period of time, it is necessary to maintain an adequate level of arousal. In cognitive research, traditional studies have used repetitive, monotonous tasks to learn about the dynamics of arousal in tasks that require sustained vigilance, such as driving or monitoring a factory line. To date, studies have rarely considered whether observers use task-embedded regularities in such continuous contexts to anticipate events and regulate arousal accordingly. In the current study, we explored whether arousal is an adaptive process that reacts to temporal stimulus predictability. Across two experiments, we used pupillometry as a proxy measure of arousal in human observers performing continuous tasks. Within the tasks, we manipulated the temporal predictability of relevant events by presenting stimuli either in a fixed rhythm or at varied intervals. Temporal predictability led to the lowering of tonic levels of arousal. Trial-wise temporal structures led to short-scale changes in pupil size related to temporal anticipation. Accordingly, we suggest that arousal is sensitive to varying levels of temporal predictability and dynamically adapts to task demands to bring about performance benefits as well as energetic efficiencies overall. Statement of RelevancePeople often have to sustain focus and high levels of performance on extended and non-stimulating tasks (e.g., driving, sowing, monitoring data acquisition). A critical factor to the success (or failure) in sustained performance is arousal - the energetic state of the cognitive system. Here we used pupil dilation as a proxy to study levels of arousal during sustained performance on monotonous tasks. We reveal that arousal is dynamically regulated to support performance according to an important fundamental property of any task: its temporal structure. When the timing of task-relevant events is predictable, arousal levels fluctuate accordingly, saving energy overall while also optimally guiding performance. Our study and findings add ecological validity to the study of temporal expectations, by moving investigations beyond typical trial-by-trial designs. They also carry significant implications for clinical studies relying on sustained performance tasks.