Lapses in sustained attention predicted by changes in visually-guided movements

Previous research has linked keypress response times (RTs) to gradual shifts in sustained attention. Specifically, RTs are typically shorter and/or more variable prior to commission errors on no-go trials in the gradual-onset continuous performance task (Rosenberg et al., 2013), indicating a link between keypress response latency and fluctuations in sustained attention over time. We expanded on this research in a novel task where we asked participants to make repetitive visually-guided movements. Participants were required to reach out and press a display every time a continuously moving ball collided with a stationary box and to withhold their response every time the ball narrowly missed the box. The latter no-go trials occurred only 10% of the time, so participants were frequently reaching to the target. We measured data from both hand and eye movements. We found that movement initiation latency (but not movement time) was shorter preceding commission errors on no-go trials relative to trials where the participant correctly withheld their response (correct rejections). This measure proved quite sensitive, with significant differences observed even when only examining the single trial that preceded the no-go trial. In other words, when participants initiate their movements quickly, they are more likely to make subsequent errors that indicate lapses of sustained attention. Pupil size was correlated at the subject level with initiation latency, but did not predict commission errors. Together these data suggest that studying complex behavior may be valuable in uncovering sensitive measures that indicate when a person is losing focus during a sustained attention task. Further, by isolating specific components of motor output, we may be able to link drifts in sustained attention to specific sub-processes of attention and cognitive control (e.g., Erb et al., 2016).