Changes in pupil size track self-control failure.

People are more likely to perform poorly on a self-control task following a previous task requiring self-control (ego-depletion), but the mechanism for this effect remains unclear. We used pupillometry to test the role of attentional effort in ego-depletion. We hypothesized that an elevated pupil diameter (PD)-a common physiological measure of effort-during an initial task requiring self-control should be negatively associated with performance on a subsequent control task. To test this hypothesis, participants were first assigned to either a high- or low-demand attention task (manipulation; a standard ego-depletion paradigm), after which all participants completed the same Stroop task. We then separately extracted both sustained (low-frequency) and phasic (high-frequency) changes in PD from both tasks to evaluate possible associations with lapses of cognitive control on the Stroop task. We first show that in the initial task, sustained PD was larger among participants who were assigned to the demanding attention condition. Furthermore, ego-depletion effects were serially mediated by PD: an elevated PD response emerged rapidly among the experimental group during the manipulation, persisted as an elevated baseline response during the Stroop task, and predicted worse accuracy on incongruent trials, revealing a potential indirect pathway to ego-depletion via sustained attention. Secondary analyses revealed another, independent and direct pathway via high levels of transient attentional control: participants who exhibited large phasic responses during the manipulation tended to perform worse on the subsequent Stroop task. We conclude by exploring the neuroscientific implications of these results within the context of current theories of self-control.

Life is uncertain, eat dessert first: Uncertainty causes controlled and unemotional eaters to consume more sweets.

Sometimes even dieters with the best self-control overindulge. Uncertain situations may undermine the self-control of even well-controlled eaters. This study was designed to test the hypothesis that uncertainty increases unhealthy snacking. Participants were either told that they would be giving a speech, that they would be listening to a speech, or that they would find out later whether they were to give a speech or not. Among participants who typically reported good control over their eating or scored low on emotional eating, participants who were uncertain about whether they would be giving a speech ate more candy than participants who expected to not have to give a speech and even those who expected to have to give a speech. Participants who reported poor control over their eating or scored high on emotional eating did not eat significantly more when uncertain. These findings suggest that, for people who are typically able to control their eating, uncertainty increases snacking more than certainty of a negative outcome.