Paul N. Russell

Working memory load and the vigilance decrement

In this study, we examined the impact of concurrent verbal and spatial working memory demands on performance on an alpha-numeric successive target detection task. Seven hundred and forty-five participants performed a target detection task while simultaneously performing either a spatial or a verbal working memory task or they performed matched no-memory control tasks. The vigilance decrement, both an increase in target detection response times and a decrease in perceptual sensitivity A′ to target stimuli over time, was exacerbated by concurrent working memory load. The spatial and verbal working memory loads both impacted vigilance performance suggesting utilization of common executive resources. Overall, these results support the view that the vigilance decrement results from high cognitive resource demands (e.g., hard work), not from cognitive under-load (e.g., boredom or mindlessness).

Search asymmetry, sustained attention, and response inhibition

In the present experiment, we used search asymmetry to test whether the sustained attention to response task is a better measure of response inhibition or sustained attention. Participants performed feature present and feature absent target detection tasks using either a sustained attention to response task (SART; high Go low No-Go) or a traditionally formatted task (TFT; high No-Go low Go) response format. In addition to performance, we employed functional near infrared spectroscopy (fNIRS) to measure lateral cerebral oxygenation levels and self-reports of Tense Arousal, Energetic Arousal, task related and unrelated thoughts occurring during the tasks. Detections were lower and reaction times longer in the feature absent search than the feature present search regardless of response format. Detections were lower, but reaction times shorter in the SART than the TFT regardless of feature search. Greater right than left frontal hemisphere activation occurred in the SART than the TFT. In addition, the SART was more fatiguing based on self-reports than the TFT, but there were no differences in Task-Unrelated Thoughts across task conditions. Overall, the results suggest the SART places high response inhibition, not necessarily sustained attention, demands on participants.

Brief mental breaks and content-free cues may not keep you focused

The dominant theory amongst vigilance researchers is resource theory. From this theoretical perspective the increase in lapses with time on task, the vigilance decrement, is due to the depletion of the cognitive resources necessary for the maintenance of performance. Alternative under-load theories have recently been suggested: mindlessness and goal habituation. Advocates of mindlessness theory suggest the vigilance decrement results from conscious disengagement due to task monotony. From this theoretical perspective, the inclusion of content unrelated cues should draw the participants out of their mindless state. An alternative proposal is that vigilance decrements are a result of goal habituation. From this perspective a momentary deactivation and reactivation of task goal by inserting a brief task switch in a vigilance task should eliminate the vigilance decrement. In order to examine these alternative theories, we had 498 participants perform a visual vigilance task with either the inclusion of brief task switches, content-free cues, or in comparative control conditions during the vigilance task. All experimental groups had an equivalent robust vigilance decrement measured by both a decline in perceptual sensitivity over time and an increase in response latency over time. There was, moreover, no difference in self-reported mental workload or task-unrelated thoughts across the experimental groups. Bayesian analyses resulted in substantial evidence in favour of the null hypothesis, in agreement with the expectations based on resource theory and contrary to the expectations based on either the mindlessness or goal-habituation theories.

Rest is best: The role of rest and task interruptions on vigilance

We examined the impact task interruptions have on visuospatial vigilance in two experiments. In the first experiment participants were randomly assigned to one of three interruptions: participants were given a complete rest (rest), participants completed an alphanumeric vigilance task (letter), or participants performed the primary vigilance task (continuous). In the second experiment participants were randomly assigned to one of the conditions from the first experiment or to two further conditions, in which participants (spatial memory) performed a spatial match to sample task, or participants (verbal memory) performed a letter match to sample task. Vigilance performance post-interruption was best for rest, worst for continuous, and varied for the other interruption tasks. Overall, the results suggest the vigilance decrement is due to the repeated use of particular executive resources, but there may, in addition be domain specific interference when the primary task and activities during a break make use of the same resources.

Visuospatial and verbal working memory load: effects on visuospatial vigilance.

In this study, we examined the impact of concurrent verbal and visuospatial working memory demands on performance of a visuospatial successive target detection task. Three hundred and four participants performed a visuospatial vigilance task while simultaneously performing either a spatial or verbal working memory task that either required a memory load during the vigil or did not require a memory load during the vigil. Perceptual sensitivity A′ to vigilance target stimuli was reduced by concurrent memory load, both verbal and visuospatial. The decline in perceptual sensitivity to vigilance targets, the vigilance decrement, was steeper for a visuospatial memory task than a verbal memory task, regardless of concurrent memory load. Memory performance after vigilance detection trials was much lower for visuospatial than verbal items, even though memory performance before vigilance detection trials was higher for visuospatial than verbal items. Together, this indicates increased interference when a visuospatial vigilance task is paired with a visuospatial memory task, than when paired with a verbal memory task. Overall, the visuospatial and verbal working memory loads both impacted vigilance target detection, suggesting utilization of common executive resources. There may, however, be domain specific interference, and this may be exacerbated for two visuospatial tasks.

The Effects of Arousing Negative and Neutral Picture Stimuli on Target Detection in a Vigilance Task

Objectives: The present study was designed to explore whether target detection in a vigilance task is influenced by task-irrelevant negative emotional and neutral picture stimuli and to test predictions derived from the boredom-mindlessness versus resource depletion accounts of vigilance performance. Background: Previous research indicates that emotional stimuli can capture spatial attention. Research on the effect of negative emotional and neutral visual stimuli on temporal aspects of attention has not, however, been researched in detail. Method: For this study, 51 participants (15 men and 36 women) were assigned at random to one of three vigilance conditions: a visual vigil with task-irrelevant negative-arousing pictures, a visual vigil with task-irrelevant neutral pictures, or a no-picture visual vigil control. Vigilance performance was assessed in all conditions. Results: Overall performance efficiency was negatively influenced by the negative-arousing pictures and was interpreted to favor resource depletion to boredom-mindlessness accounts of vigilance performance. Conclusion: Task-unrelated negative emotional stimuli appear to impair absolute levels of target detections in a vigilance task. Application: In monitoring settings where negative emotional stimuli are present, the intrusion of negative emotional stimuli should be mitigated via alterations in the system design, or if this is implausible, the monitors may need additional stress coping and emotional resilience training.

Reliable- and unreliable-warning cues in the Sustained Attention to Response Task

The Sustained Attention to Response Task (SART) is a Go-No-Go signal detection task developed to measure lapses of attention. In this study, we examined the impact that warning signals, reliable and unreliable, have on SART performance. Eighteen participants performed a no-warning, reliable-warning, or unreliable-warning SART. Response times were faster, errors of commission lower, but errors of omission higher in the reliable-warning SART in comparison with the no-warning or unreliable-warning SART. There was a significant negative correlation between participants’ errors of commission rate and their response times in the unreliable-warning and no-warning SART. This correlation was reduced in the reliable-warning SART. Making the task perceptually easier reduces the errors of commission, in contradiction to the mindlessness perspective, and reduces the speed-accuracy trade-off. These results, overall, support the view that the SART is primarily a measure of response strategy, not sustained attention per se.

Target predictability, sustained attention, and response inhibition.

We examined whether the sustained attention to response task is a better measure of response inhibition or sustained attention. Participants performed a number detection task for 37.3min using either a Sustained Attention to Response Task (SART; high Go low No-Go) or a more traditionally formatted vigilance task (TFT; high No-Go low Go) response format. Participants performed these tasks using either a regular fixed ordered stimuli set (1-9, sequentially repeated), in which the target number appeared predictably, or a random order (1-9, random presentation), in which the target number appeared at random. We utilized functional near infrared spectroscopy (fNIRS) to measure cerebral oxygenation levels in the right and left frontal areas. We also used post-task participant reports of arousal, and conscious thoughts occurring during the tasks. Performance differed for the both response format and target predictability. Greater right than left frontal hemisphere activation occurred in the TFT than the SART with time-on-task. In addition, the SART response format resulted in elevated self-reports of task-related thoughts than the TFT response format. The results suggest the SART, random or fixed ordered, places high response inhibition, not necessarily sustained attention, demands on participants. Elevated levels of task-related thoughts during the SART format condition in comparison to the TFT condition does not appear to be in accord with the claim that the SART induces mindlessness.

Effects of breaks and goal switches on the vigilance decrement.

Abstract In this report, we examined the impact of rest breaks and goal switches on the vigilance decrement—the change in performance with time-on-task—and subjective stress and workload. In study 1, we examined the decrement in a visual line-length discrimination task employed in previous research. The task had a linear decrement in perceptual sensitivity with time-on-task based on the results of a trend analysis (orthogonal polynomial contrasts). In addition, the task had a profile of self-reported stress and workload similar to other moderately demanding vigilance tasks. In study 2, we examined the decrement in the same task with the factorial inclusion of rest breaks and goal switches late in the task. The inclusion of rest breaks resulted in a significant cubic trend in perceptual sensitivity. After the first rest break, there was a reversal of the linear downward decrement in perceptual sensitivity, but the second rest break did not elicit this change. The inclusion of goal switches did not cause any significant deviation from the linear trend and did not attenuate the decrement. In addition, the inclusion of rest breaks resulted in reduced self-reported effort and temporal demand. The results of the present study cohere with the expectations of a resource depletion account of the vigilance decrement.

Feature absence–presence and two theories of lapses of sustained attention

We conducted two experiments using search asymmetry to test two theories of sustained attention lapses currently debated in the literature: the mindlessness theory and the resource theory. In the first experiment, participants performed a feature present and a feature absent sustained attention to response task (SART). The mindlessness theory predicts more commission errors in a feature present SART than a feature absent SART, while resource theory predicts the reverse, i.e. more commission errors in a feature absent SART than feature present SART. The latter occurred. In the second study, participants performed a naturalistic scene vigilance task with targets more analogous to a feature present search or more analogous to a feature absent search. The vigilance decrement occurred, with an increase in reaction time to targets over the course of the vigil. The decrement was more pronounced with the targets more analogous to a feature absent search. Overall, the results of both studies support a resource theory of sustained attention lapses, not the mindlessness theory.