Mind Wandering Influences EEG Signal in Complex Multimodal Environments.

Area of Science:

• Cognitive Neuroscience
• Human-Computer Interaction
• Neuroscience

Background:

• Mind wandering (MW) significantly affects vigilance and performance, particularly in human operators monitoring automated systems, potentially leading to out-of-the-loop (OOTL) situations.
• Electroencephalogram (EEG) systems show promise in quantifying and mitigating MW by analyzing its impact on event-related potentials (ERPs) and brain oscillations.
• Understanding the factors influencing EEG markers of MW in complex, real-world scenarios remains crucial for operator performance and safety.

Purpose of the Study:

• To investigate the gradual emergence of attentional decoupling during mind wandering.
• To differentiate the effects of sensory modality (visual vs. auditory) on EEG markers of MW.
• To explore the relationship between task-related and task-unrelated MW episodes and their impact on cognitive states.

Main Methods:

• Utilized EEG to measure event-related potentials (ERPs) and alpha-wave activity in 18 participants.
• Participants supervised an automated drone (visual task) and responded to auditory beeps (auditory task).
• Incorporated steady-state auditory response (ASSR) using 40-Hz amplitude-modulated brown noise and categorized self-reported MW episodes.

Main Results:

• Task-unrelated MW was associated with lower N1 ERP amplitude and higher alpha-wave activity, supporting attentional decoupling.
• Task-related MW showed higher P3 ERP amplitude, indicating different cognitive processes compared to task-unrelated MW.
• No significant influence of attentional states on ASSR amplitude was observed, suggesting modality-specific effects.

Conclusions:

• EEG effectively tracks MW in ecologically relevant laboratory tasks, demonstrating its utility in studying operator cognitive states.
• Results highlight the complex influence of perceptual decoupling on operator behavior and EEG measures, with variations based on MW type.
• The study underscores the need to consider different types of MW and their distinct neural correlates for effective human-automation interaction.