Dynamic multilayer networks reveal mind wandering

Xu, Zhongming; Tang, Shaohua; Di, Zengru; Li, Zheng · 2024 · Crossref

DOI: 10.3389/fnins.2024.1421498

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study addresses the lack of objective, neural-based methods for detecting mind-wandering (MW) during learning tasks. While MW negatively impacts educational outcomes, its neural mechanisms remain poorly understood, particularly regarding the dynamic fluctuations in functional connectivity. Previous approaches often relied on static network analyses or single-frequency band metrics, overlooking the complex, multi-scale dynamics of brain activity. The authors propose a dynamic multilayer network analysis framework using electroencephalography (EEG) to capture these temporal changes and distinguish MW from focused learning. The researchers collected EEG data from 14 healthy participants engaged in a video-learning task under two conditions: focused learning (high-interest videos) and mind-wandering (low-interest videos with future planning cues). They constructed weighted multiplex networks for each EEG segment, integrating two layers of functional connectivity: amplitude envelope correlation (AEC) and imaginary phase-locking value (IPLV) across delta, theta, alpha, beta, and gamma frequency bands. To avoid arbitrary windowing, they employed an extended distance measure/closeness centrality (EDMCC) algorithm to segment the data based on structural changes. Network states were defined by clustering overlapping node closeness centrality vectors using the Louvain algorithm, resulting in four recurring network motifs. Dynamic characteristics, including state frequency, duration, coverage, and transition probabilities, were then analyzed. Finally, a hidden Markov model (HMM) was trained on these state sequences to classify mental states. The results revealed that multilayer network states were consistent across frequency bands, with states A and B showing high cross-frequency similarity. Transition patterns between network states were non-random, indicating structured dynamics. Significant differences in dynamic metrics were observed between MW and focused learning conditions, particularly in the delta, alpha, and beta bands for AEC, and the theta band for IPLV. The HMM classifier, using state sequences as input, achieved a mean area under the receiver operating characteristic curve (AUC) of 0.888 for within-participant detection of mind-wandering. This work demonstrates that dynamic multilayer network analysis provides a robust framework for characterizing the temporal evolution of brain connectivity during mind-wandering. By integrating multiple connectivity measures and frequency bands, the method captures nuanced neural signatures that static analyses miss. The high classification accuracy suggests potential for real-time, non-invasive MW detection in educational settings, offering a pathway to develop interventions that mitigate the negative effects of mind-wandering on learning outcomes.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• neuro workload indices