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.

Recently, the gamma band ( γ; 70-100 Hz) has been implicated in sustained attention decay across a vigil consistent with computational models of fatigue. Frontal γ indexing centrally controlled sustained attention and parietal γ linked to gated sensory processes declined across a 10-minute vigilance task, a pattern observed for faster but not slower performers. The anatomical distribution of γ activity indicates neural communication, or connectivity, within the fronto-parietal network. We used Granger Prediction to evaluate fatigue effects on network γ connectivity. Results showed stronger directional connectivity for frontal→parietal versus parietal→frontal over time, indicating that top-down control of attention largely remained intact. However, parietal→frontal early γ connectivity increased with time, suggesting a network shift to enhanced sensory-directed processes after only 8 minutes. This pattern of connectivity was mirrored by fast but not slower performers. Our findings provide new directions for computational accounts of fatigue mechanisms and highlight the importance of individual differences.

No abstract