Across two datasets featuring sustained attention tasks, we confirm that increased random respiratory variability during task performance is associated with increased sigh frequency, and that sighs reset this variability. We further identify novel influences on sigh frequency: First, the degree of respiratory phase-locking to task stimuli positively correlates with both sigh frequency and increased respiratory random variability. Second, sigh frequency was higher with fixed stimulus intervals and lower with variable intervals, suggesting that dynamic and more engaging task conditions reduce sighing. Third, paced breathing reduces sigh frequency compared to spontaneous breathing, despite similar inspiratory volume variability – challenging prior claims that restrictive breathing increases sighing. This reduction may stem from increased task engagement during paced breathing, reducing the need for variability-resetting sighs. Fourth, variability analysis before and after sighs confirms that sighs reset respiratory variability, supporting the resetter hypothesis of sigh function, but we found no improvement in subjective focus or task performance, suggesting that sigh-induced attentional benefits may be context-dependent. Fifth, we demonstrate for the first time in humans that sighs co-occur with changes in pupil diameter, implicating noradrenaline-related arousal mechanisms. Sigh-related pupil dynamics differed between groups, suggesting interactions with both task structure and breathing patterns. Overall, our findings highlight the sensitivity of sigh frequency to respiratory phase-locking and task demands. The link between sighs and pupil responses suggests a broader role for the respiratory system in modulating arousal. Future research should further explore how respiration interacts with cognitive engagement, and noradrenaline-driven attentional shifts.

Sighs are spontaneous deep breaths thought to play a homeostatic role in respiratory control. Their relationship to ongoing respiratory variability has been repeatedly demonstrated. How sighs are related to task engagement, performance, structure and to arousal has remained unclear. Presently, we investigated sigh behaviour across two sustained attention tasks using respiratory belt recordings. Participants completed either a Gradual Contrast Change Detection task (dataset Grad) or a Paced Auditory Cue Entrainment task (dataset PACE), with subgroups performing the latter under spontaneous (NIB) or slow-paced (IB) breathing conditions. Sighs were identified as breaths at least twice the mean inspiratory volume (Vi). We analysed the total variability (coefficient of variation; CV) and structured variability (lag-1 autocorrelation; AR) of respiratory rate (RR) and Vi, their changes over the task and around sigh events. In groups with spontaneous breathing (Grad, NIB), sigh frequency was positively related to CV in both RR and Vi suggesting a relationship to the overall variability. Sigh frequency and CV increased over the task duration, while post-sigh dynamics showed decreased CV and increased Vi-AR, supporting a role for sighs in resetting the temporal structure in respiratory variability. In the IB group, sigh frequency was drastically reduced, and no pre-post sigh changes were observed. Sighs were also associated with changes in pupil diameter, implicating involvement of the noradrenaline-mediated arousal system. Sighs were not related to any proceeding alterations in task performance or subjective engagement. Finally, respiratory phase-locking to task timing predicted increased sigh frequency and random variability, highlighting a novel linkage between this adaptive cognitive phenomenon, yet maladaptive physiological consequences. The lack of task performance differences may need to be clarified using more demanding tasks which could draw out variability. These findings do support a functional role of sighs in regulating respiratory variability, phase-locking behaviour, and pupil-linked arousal during prolonged cognitive tasks.