Predicting lapses of attention with sleep-like slow waves

The human brain continuously generates conscious experiences, and attention is crucial for directing cognitive resources. However, attention can shift inward, leading to mind wandering (MW), where focus drifts to task-unrelated thoughts, or mind blanking (MB), a complete cessation of conscious thought. These attentional lapses, characterized by shifts away from external demands, can manifest as sluggish or impulsive behaviors. It's unclear whether these diverse lapses stem from distinct physiological causes or share a common origin. Prior models have suggested separate neurophysiological states for MW and MB. However, the increase in both sluggish and impulsive responses following sleep deprivation and in attention-deficit disorders implies a potential shared mechanism. Sleepiness has been linked to both MW and MB, with experiences resembling both observed during hypnagogia (the transition to sleep). This suggests a potential role for sleep pressure, which increases with time awake and is only relieved by sleep. Extended wakefulness can lead to local sleep, characterized by electroencephalographic (EEG) signatures of non-rapid eye movement (NREM) sleep, namely high-amplitude slow waves, in specific brain regions while the individual remains awake. These slow waves, during both global and local sleep transitions, are linked to reduced neuronal firing, causing behavioral errors. While previous research has associated wake slow waves with behavioral lapses, their impact on conscious experience (MW and MB) is unclear. This study hypothesizes that local sleep, manifested as slow waves during wakefulness, underlies both the behavioral and phenomenological aspects of attentional lapses, proposing a unified physiological explanation.

Existing research on attentional lapses, encompassing mind wandering (MW) and mind blanking (MB), has explored their behavioral and subjective experiences, but the underlying neural mechanisms remain incompletely understood. Some studies suggest distinct neurophysiological mechanisms for MW and MB, while others propose shared underlying processes. The impact of sleep deprivation and sleepiness on both MW and MB highlights the need for investigating the role of sleep pressure in these lapses. Prior work has established a link between local sleep, characterized by the emergence of sleep-like slow waves in the awake brain, and behavioral errors. However, the connection between these slow waves and subjective experiences like MW and MB remains to be fully elucidated.

Twenty-six healthy adults (ten females, age: 29.8 ± 4.1 years) participated. Participants performed modified Sustained Attention to Response Tasks (SARTs) with faces and digits as stimuli. Continuous presentation of stimuli was interrupted every 30–70 seconds, prompting participants to report their mental state (task-focused, MW, MB) and vigilance level. High-density EEG and pupil size were recorded continuously. Behavioral data included accuracy (misses, false alarms) and reaction times (RTs). Slow waves were operationally defined as high-amplitude waves in the delta (1–4 Hz) band, detected using an established algorithm. Analyses included linear mixed-effects models (LMEs) to examine relationships between mental states, slow waves, and behavioral measures, as well as hierarchical Bayesian Drift-Diffusion Modeling (HDDM) to analyze decision-making processes in detail. Model comparisons assessed the impact of mental states and slow waves on behavior and EEG parameters.

Participants reported being task-focused only about 48% of the time, with the remaining time spent in MW or MB. Both MW and MB were associated with decreased task performance, but in distinct ways. MB showed more misses and slower RTs than MW, suggesting sluggishness, while MW demonstrated more false alarms and faster RTs, suggesting impulsivity. Both MW and MB were associated with reduced subjective vigilance ratings and smaller pupil size, suggesting lower arousal. Importantly, slow waves in the EEG preceded and co-occurred with both behavioral and subjective markers of attentional lapses. Globally, increased slow wave density, amplitude, and slopes were negatively correlated with vigilance ratings and pupil size. Locally, frontal slow waves predicted MW, characterized by an increase in density, amplitude, and slopes. Centro-parietal slow waves were associated with MB, along with increased slopes in those areas. At the single-trial level, frontal slow waves were linked to faster RTs and more false alarms (impulsivity), while posterior slow waves were associated with slower RTs and more misses (sluggishness). HDDM analysis further revealed that slow waves reduced the decision threshold (α), leading to impulsivity, and increased non-decision time (τ), indicating slower processing. Posterior slow waves specifically slowed evidence accumulation for Go responses (VGo), while frontal slow waves slowed evidence accumulation for NoGo responses (VNoGo).

This study provides compelling evidence that attentional lapses, encompassing both MW and MB, are preceded and accompanied by local sleep-like slow waves, even in well-rested individuals. The location of these slow waves predicts the type and behavioral profile of the lapse. The results support the concept of local sleep as a mechanism underlying attentional lapses, with the spatial distribution of slow waves explaining the variability in subjective experiences and behavioral outcomes. The findings integrate behavioral, phenomenological, and physiological levels of explanation, offering a unified account of attentional lapses. The observed relationship between frontal slow waves and MW aligns with the role of the frontal cortex in executive functions and response inhibition. The association between posterior slow waves and MB is consistent with the role of parietal cortices in sensorimotor integration and suggests a disruption of awareness when these regions temporarily go offline.

This study demonstrates that sleep-like slow waves, even in well-rested individuals, precede and co-occur with attentional lapses, including mind wandering and mind blanking. The location of these slow waves predicts the specific behavioral and phenomenological characteristics of these lapses. This provides strong evidence for local sleep intrusions as a proximate cause for attentional lapses. Future research could explore the generalizability of these findings to more demanding tasks and naturalistic settings, investigate the interplay between slow waves and other neural mechanisms influencing attention, and explore potential applications in educational and professional settings using brain-machine interfaces.

The study utilized relatively undemanding tasks, which might have increased the likelihood of sleepiness and local sleep compared to more engaging tasks. Generalizing these findings to more complex or naturalistic settings requires further investigation. The study primarily relied on correlational data; future studies incorporating intracranial recordings or sleep deprivation could strengthen the causal link between slow waves and attentional lapses. The continuous visual stimuli presented in the task limited the ability to fully disentangle the effects of slow waves from task-related events.