Predicting attentional lapses using response time speed in continuous performance tasks

Abstract

Failures of sustained attention, including lapses and mind-wandering, have serious consequences on cognitive task performance. In recent years, real-time triggering methods have been used to isolate periods of optimal and suboptimal attention based on patterns of response times in monotonous continuous performance tasks. In a closed-loop fashion, these triggering designs reduce the need for retrospective processing to identify periods of poor attention by using simple intrasubject response time boundaries to trigger events based on inferred attentional state. In the current review, we first discuss studies that used principal component analysis to identify response patterns that precede both task errors and phenomenological reports of mind-wandering. Then, we review designs that used real-time triggering conditions to reinforce the relationship between lapsing and memory encoding. Finally, we describe important next steps to generalize the utility of the triggering procedure across populations, validate lapse countermeasures, and shine light on the limited human capacity to maintain vigilance.