Utilization of a combined EEG/NIRS system to predict driver drowsiness

Nguyen, Thien; Ahn, Sangtae; Jang, Hyojung; Jun, Sung Chan; Kim, Jae Gwan · 2017 · Crossref

DOI: 10.1038/srep43933

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 critical safety issue of driver drowsiness, which contributes significantly to fatal automobile accidents. While various detection methods exist, including vehicle performance monitoring and behavioral recording, these are often susceptible to external environmental factors. Physiological signal measurement offers a more robust alternative, yet single-modality approaches often lack sufficient accuracy. The authors propose a combined Electroencephalography (EEG) and Near-Infrared Spectroscopy (NIRS) system to simultaneously monitor neuronal electrical activity and cerebral hemodynamics, aiming to identify the most informative parameters for accurate, real-time drowsiness prediction. The experimental design involved nine healthy subjects performing a simulated driving task using a racing wheel and screen setup. Subjects underwent both awake and drowsy states, with drowsiness defined by eye closures exceeding two seconds. The researchers recorded EEG (64 channels), NIRS (prefrontal cortex), Electrooculography (EOG), and Electrocardiography (ECG) signals. Data preprocessing included filtering and Independent Component Analysis to remove artifacts. The study employed frequency domain analysis to compare EEG band powers, statistical t-tests to identify significant differences in NIRS hemoglobin concentrations, and Fisher’s Linear Discriminant Analysis (FLDA) for state classification. Additionally, time series analysis was utilized to detect the transition from awake to drowsy states. The results demonstrated that EEG beta band power and NIRS oxy-hemoglobin (HbO) concentration changes were the most discriminative parameters. During the drowsy state, EEG showed increased power in lower frequency bands (delta, theta, alpha) and decreased power in higher bands (beta, gamma), with the frontal beta band showing the strongest correlation with eye closure. NIRS data revealed a significant difference in HbO changes between states, with a sharp increase in HbO concentration preceding the first eye closure. Classification accuracy improved when combining modalities: the mean accuracy for combined EEG/NIRS was 79.2%, compared to 70.5% for EEG alone and 73.7% for NIRS alone. Crucially, the simultaneous sharp decrease in beta band power and increase in HbO change occurred several seconds before eye closure. The authors developed a Drowsiness Detection Index (DDI) based on these two parameters, setting thresholds for HbO variation (>0.05 mM/DPF) and beta band percentage change (>20%). This combined index achieved 100% true positive detection with 0% false negatives, predicting drowsiness an average of 3.6 seconds before the first eye closure. In contrast, using either parameter alone resulted in significant false negatives. The study concludes that a combined EEG/NIRS system, specifically targeting frontal beta power and HbO changes, offers a superior method for early drowsiness detection compared to single-modality approaches, providing a valuable lead time for intervention to prevent accidents.

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.

StageOutcomeToolModelPromptAttemptsCompleted
discover success Crossref 1 2026-08-09
archive success canonical_url 1 2026-08-09
extract success pdftotext 4 2026-08-10
clean success clean 2 2026-08-10
chunk success chunk 2 2026-08-10
embed success embed Qwen/Qwen3-Embedding-8B 2 2026-08-10
promote success 1 2026-08-09
summarize success llm qwen3.6-27b-nvidia summ-v5 2 2026-08-10
tag success vector_similarity 17 2026-08-11
verify partial 2 2026-08-10

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

Topics

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

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).