When Eyes Wander Around: Mind-Wandering as Revealed by Eye Movement Analysis with Hidden Markov Models

Area of Science:

• Cognitive Psychology
• Neuroscience
• Educational Technology

Background:

• Mind-wandering significantly impacts learning efficiency in today's digitally saturated and distracting environments.
• Effective detection of mind-wandering is essential for optimizing educational strategies and interventions.

Purpose of the Study:

• To investigate if eye movement patterns can distinguish between focused attention and mind-wandering states.
• To introduce and validate a novel approach using hidden Markov models (EMHMM) for attention state detection.

Main Methods:

• Utilized a wearable eye-tracker to record eye movements during a sustained attention to response task.
• Applied eye movement analysis with hidden Markov models (EMHMM) to analyze spatial and temporal eye-movement data.
• Employed clustering techniques to identify distinct eye movement patterns associated with attentional states.

Main Results:

• Identified two primary eye movement patterns: centralized and distributed, through EMHMM clustering.
• Participants exhibiting the centralized pattern demonstrated superior target detection performance.
• Individuals with the centralized pattern reported higher levels of self-perceived focus compared to those with the distributed pattern.

Conclusions:

• Distinct eye movement patterns correlate with different attentional states, specifically focused attention versus mind-wandering.
• EMHMM provides a novel and effective method for analyzing attention dynamics through eye movement.
• This research offers a non-intrusive method for monitoring mind-wandering in classroom settings, potentially enhancing learning outcomes.