In infancy, it's the extremes of arousal that are 'sticky': Naturalistic data challenge purely homeostatic approaches to studying self-regulation.

Most theoretical models of arousal/regulatory function emphasise the maintenance of homeostasis; consistent with this, most previous research into arousal has concentrated on examining individuals' recovery following the administration of experimentally administered stressors. Here, we take a different approach: we recorded day-long spontaneous fluctuations in autonomic arousal (indexed via electrocardiogram, heart rate variability and actigraphy) in a cohort of 82 typically developing 12-month-old infants while they were at home and awake. Based on the aforementioned models, we hypothesised that extreme high or low arousal states might be more short-lived than intermediate arousal states. Our results suggested that, contrary to this, both low- and high-arousal states were more persistent than intermediate arousal states. The same pattern was present when the data were viewed over multiple epoch sizes from 1 s to 5 min; over 10-15-minute time-scales, high-arousal states were more persistent than low- and intermediate states. One possible explanation for these findings is that extreme arousal states have intrinsically greater hysteresis; another is that, through 'metastatic' processes, small initial increases and decreases in arousal can become progressively amplified over time. Rather than exclusively using experimental paradigms to study recovery, we argue that future research should also use naturalistic data to study the mechanisms through which states can be maintained or amplified over time.