Sustained attention is an essential behavior in life, but often leads to performance decrements with time. Computational accounts of sustained attention suggest this is due to brief disruptions in goal-directed processing, or microlapses. Decreases in gamma spectral power are a potential candidate for indexing microlapses and discriminating between low and high performers in sustained attention tasks, while increases in beta, alpha, and theta power are expected to exhibit compensatory effort to offset fatigue. The current study tests these hypotheses in a 10-minute Psychomotor Vigilance Test, a context that eliminates confounds with measuring gamma frequencies. 34 participants ( Mage = 22.60; SDage = 4.08) volunteered in the study. Results suggested frontal gamma power declined with time-on-task, indicating reduction in central cognition. Beta power increased with time-on-task, suggesting compensatory effort; however, alpha and theta power did not increase. Additionally, gamma power discriminated between low and high performers, potentially suggesting motivational differences between the groups.

Cognitive fatigue emerges in wide-ranging tasks and domains, but traditional vigilance tasks provide a well-studied context in which to explore the mechanisms underlying it. Though a variety of experimental methodologies have been used to investigate cognitive fatigue in vigilance, relatively little research has utilized electroencephalography (EEG), specifically event-related potentials (ERPs), to explore the nature of cognitive fatigue, also known as the vigilance decrement. Moreover, much of the research that has been done on vigilance and ERPs uses non-traditional vigilance paradigms, limiting generalizability to the established body of behavioral results and corresponding theories. In this study, we address concerns with prior research by (1) investigating the vigilance decrement using a well-established visual vigilance task, (2) utilizing a task designed to attenuate possible confounding ERP components present within a vigilance paradigm, and (3) informing our interpretations with recent findings from ERP research. We averaged data across electrodes located over the frontal, central, and parietal scalp. Then, we generated waveforms locked to the onset of critical low-frequency or non-critical high-frequency events during a 40 min task that was segregated into time blocks for data analysis. There were three primary findings from the analyses of these data. First, mean amplitude of N1 was greater during later blocks for both low-frequency and high-frequency events, a contradictory finding compared to past visual vigilance studies that is further discussed with respect to current interpretations of the N1 in visual attention tasks. Second, P3b mean amplitude following low-frequency events was reduced during later blocks, with a later onset latency. Third and finally, the decrease in P3b amplitude correlated with individual differences in the magnitude of the vigilance decrement, assessed using d′. The results provide evidence for degradations of cognitive processing efficiency brought on by extended time on task, leading to delayed processing and decreased discriminability of critical stimuli from non-critical stimuli. These conclusions are discussed in the context of the vigilance decrement and corresponding theoretical accounts.

As task environments become more automated and operators become more passive monitors, it is important to understand the underlying mechanisms of the vigilance decrement to help inform task development and interventions. Some vigilance studies have reported an end-spurt effect, where performance increases at the end of the task. This is commonly purported to be the result of increased resources from motivation or arousal; however, self-regulation of attentional resources throughout the vigil has not been addressed. To provide insight into the mechanisms underlying the vigilance decrement, we utilize event-related potentials to examine the vigilance decrement beyond the end-spurt. Thirty-two individuals aged 18 – 36 ( M = 22.60; SD = 4.08) completed a vigilance task with an unknown end time while EEG data was collected. Results revealed non-monotonic trends in N1 and P3 components, suggesting that individuals might self-regulate resources during the vigil, adapting to the constraints of the task environment.

No abstract

No abstract