Pupil-based states of brain integration across cognitive states

Arousal is a potent mechanism that provides the brain with functional flexibility and adaptability to external conditions. Within the wake state, arousal levels driven by activity in the neuromodulatory systems are related to specific signatures of neural activation and brain synchrony. However, direct evidence is still lacking on the varying effects of arousal on macroscopic brain characteristics and across a variety of cognitive states in humans. Using a concurrent fMRI-pupillometry approach, we used pupil size as a proxy for arousal and obtained patterns of brain integration associated with increasing arousal levels. We carried out this analysis on resting-state data and data from two attentional tasks implicating different cognitive processes. We found that an increasing level of arousal was related to a non-linear pattern of brain integration, with increasing brain integration from intermediate to larger arousal levels. This effect was prominent in the salience network in all tasks, while other regions showed task-specificity. Furthermore, task performance was also related to arousal level, with accuracy being highest at intermediate levels of arousal across tasks. Taken together, our study provides evidence in humans for pupil size as an index of brain network state, and supports the role of arousal as a switch that drives brain coordination in specific brain regions according to the cognitive state.