Decoding the cognitive states of attention and distraction in a real-life setting using EEG

Kaushik, Pallavi; Moye, Amir; Vugt, Marieke van; Roy, Partha Pratim · 2022 · Crossref

DOI: 10.1038/s41598-022-24417-w

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 challenge of tracking attention and distraction in real-life settings, moving beyond the highly controlled laboratory environments that dominate existing research. The authors argue that previous studies often rely on artificial tasks, restricted participant movement, and biased self-reports, limiting the generalizability of findings to daily scenarios where attention lapses can have serious consequences, such as in driving or industrial work. To overcome these limitations, the researchers investigated whether electroencephalogram (EEG) data collected during a naturalistic, complex social interaction could be used to decode cognitive states using machine learning. The experiment involved 24 male Tibetan monks from Sera Jey monastic university in India, who engaged in 46 monastic debates. This setting was chosen for its inherent variability in attention states and its naturalistic context, which imposes no constraints on speech or movement. EEG signals were recorded simultaneously from pairs of participants using a 32-channel Biosemi system. Crucially, attention and distraction labels were not derived from self-reports but were annotated by three senior monks observing video recordings. These expert raters identified instances of intense attention or serious distraction based on behavioral cues such as eye gaze, facial expressions, and the relevance of utterances. Data preprocessing included downsampling, bandpass filtering, and Independent Component Analysis to remove artifacts. The study also examined the impact of debate experience by comparing participants with 3–10 years of training against those with 14–25 years. The results demonstrated significant neural differences between attention and distraction states. Attention was associated with increased left frontal alpha power, increased left parietal theta power, and decreased central delta power compared to distraction. Statistical analysis using False Discovery Rate correction confirmed significant differences in delta, theta, and alpha bands across frontal and parietal regions. Furthermore, less experienced participants exhibited more frequent distraction instances than their more experienced counterparts. In terms of classification, the study employed Long Short-Term Memory (LSTM) models to predict cognitive states on a single-trial basis. The LSTM model achieved a maximum classification accuracy of 95.86% using delta waves and 95.4% using theta waves, significantly outperforming traditional machine learning approaches reported in prior laboratory-based studies. The significance of this work lies in its demonstration that high-accuracy decoding of attention and distraction is feasible using EEG data collected in uncontrolled, real-life environments. By utilizing second-person observer annotations instead of self-reports, the study provides a less intrusive and potentially more reliable method for labeling cognitive states. These findings support the development of practical Brain-Computer Interfaces capable of tracking attention in real-time during daily activities, thereby enhancing safety and performance monitoring in high-stakes environments without the need for artificial laboratory constraints.

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
• distraction detection algorithms

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Empirical Findings: physiological data