Abstract
A principal goal of attention research is to develop tasks with clear behavioral signatures of attentional fluctuations. Measures that index attentional states often fall under two broad umbrellas: decision tasks, in which participants make responses based on the changing requirements of each trial, and rhythm tasks, in which participants respond rhythmically to a uniform stimulus (e.g., a metronome tone). In the former, response speeding typically precedes errors (indicative of attention failures). In the latter, increased response variability precedes subjective reports of off-task states. We developed and validated the rhythmic visual response task (RVRT); a rhythm task incorporating trial-unique scene stimuli. The RVRT incorporates two important advances from both task categories: (1) it is free from the influence that differential decision-making has on fluctuations in attentional states, and (2) trial-unique stimuli enable later cognitive judgments to be mapped to specific moments in the task. These features allow a relatively unobtrusive measure of mind wandering that facilitates the downstream assessment of its consequences. Participants completed 900 trials of the RVRT, interrupted periodically by thought probes that assessed their attentional state. We found that both response time variance and speed predicted depth of mind wandering. Encouraged by these findings, we used the same analysis approach on archival data to demonstrate that the combination of variance and speed best predicted attentional states in several rhythm and decision task datasets. We discuss the implications of these findings and suggest future research that uses the RVRT to investigate the impact of spontaneous mind wandering on memory, decision-making, and perception.