Sleep and Sustained Attention in Adults

Background

In adults, sleep duration has an impact on sustained attention: the ability to stay focused and continue processing stimuli during a task, particularly during repetitive, mundane tasks that would typically result in habituation and distraction (Lim & Dinges, 2010; Robertson et al., 1997). When adults do not receive enough sleep their performance on sustained attention tasks is impaired: they are slower responding to the stimuli and very inconsistent in their response speed, they have more responses that are slower than 500ms (considered a lapse in attention), and they make more omission errors, when you don’t respond to the target, and commission errors, when you respond to a no-go or absent target. While this is a well-documented effect in adults, research on children has been inconclusive. Two meta-analyses failed to find an association, and this is concerning because sustained attention is important in supporting children’s learning.

Most studies measuring sustained attention rely on vigilance tasks, where the participant responds to the appearance of a rare target. The novelty of the target appearing may increase alerting to an extent where over-arousal hinders the accurate measurement of sustained attention (Johnson et al., 2007). The fixed sustained attention to response task (SART) instead requires the participant to inhibit their response to a predictable, cyclically occurring target, so there is no arousal induced by a rare target. The unengaging nature of the task likely increases the demand on the sustained attention system relative to other vigilance tasks, providing a more accurate measure of this ability (Lewis et al., 2017).

This study is part of the Spotlight Project, a longitudinal investigation of attentional control in primary school aged children. The primary aim of the current study is to investigate if short sleep duration predicts a previously documented haphazard response profile on a sustained attention task for each of the three years of the study, and examine the relations between sleep and sustained attention longitudinally over the three years. Shorter sleep in children predicted poorer sustained attention (more omission errors, more slow responses, more behavioural inattention), but also faster responding and less variability in responding. This haphazard response profile is intriguing, and this adult study will investigate if poor sleep in young adults completing the fixed SART will predict a similar response haphazard profile, or align with previous adult findings. This will provide us with a broader developmental picture of the relations between sleep and sustained attention.

Research Questions / Hypotheses

Does short sleep in young adults produce a haphazard response profile similar to that observed in the children, or will their performance echo previous adult findings?

Shorter sleep duration and poorer sleep quality were hypothesised to predict deficits on the fixed SART similar to previous adult findings: longer response times, increased variability, more omission and commission errors, and more very long responses.

Participants

286 REP participants completed the study. Of these students, 4 were excluded from analyses for improperly completing the questionnaire, 2 did not disclose their sex, 26 did not complete the fixed SART, and 144 did not complete the fixed SART adequately, as they made too many omission errors or frequently responded in less than 100ms.

Methods

Participants completed a sleep habits questionnaire, a 21-item survey about their sleep duration, sleep behaviour (e.g. snoring, napping, restlessness), and sleep habits around bedtimes, waketimes, waking during the night. Participants reported their typical sleep habits for the past week on a 5-point scale ranging from never (0) to always (7).

Participants completed the fixed sustained attention to response task (fixed SART) online via the platform Pavlovia. This is an experimental task measuring sustained attention.

Generalised and linear regression models predicted performance on the SART from sleep quality and duration as separate predictors.

Results

The findings were largely inconclusive. Where sleep did predict SART performance, analyses of model fit found that these models were unlikely to be a better fit to the data than null models without the sleep predictor.

The number of participants excluded for responding too rapidly, and also the distribution of the included participant response time data may indicate a ceiling effect, where participants were not waiting for the response cue or not meaningfully attending to the stimuli.

Implications

The pattern of responding on the SART did not tell us much about the relations with sleep, but it did highlight the importance of completing tasks like the fixed SART under supervision where researchers can confirm participants understand the task and participants are motivated to attend for the duration of the task. This research will be presented in an Honours thesis in 2021.