Abstract
Whether writing a grant application, or driving a car, research suggests we spend nearly half our waking life thinking about something other than what we are currently doing (Killingsworth & Gilbert, 2010). Instead, we are mind-wandering (MW), which is the process of attention drifting from task focus to the processing of task-unrelated information. Importantly, our current understanding of MW is limited by two methodological pitfalls. First, behavioural tasks used to measure MW provide very coarse temporal resolution (several seconds to minutes). Second, these tasks rely on subjective reports to determine the occurrence of MW. The current project presents a novel, objective measure of MW that reliably predicts the onset of MW with a temporal resolution of 100ms. In this task, participants track an object with their eyes and report self-caught MW with a keypress. To assess MW objectively we correlated eye and target position over time (using a 100ms moving window). We observed that large deviations in this correlation reliably preceded self-reported MW. Taking trials where such deviations occurred (83%) and treating the deviations as indicators of MW onset, histograms of MW duration were created for 9 different MW phenotypes. These phenotypes represent three classes of self-reported distraction (external, internal task-related, internal task-unrelated) and three levels of self-reported task focus. When mostly paying attention to the task, distraction type did not affect MW duration; however, when only some attention was paid to the task, MW duration increased as distraction became more internal and less task-related. Interestingly, despite mean MW durations of 1500-3000ms, the duration distributions displayed a strong positive skew with some episodes lasting over 15 seconds. These results strongly challenge the use of arbitrary MW duration windows currently employed in MW task analysis, and further justify the development of an objective online measure of MW.
Meeting abstract presented at VSS 2018