Optimized electroencephalogram and functional near-infrared spectroscopy-based mental workload detection method for practical applications

Chu, Hongzuo; Cao, Yong; Jiang, Jin; Yang, Jiehong; Huang, Mengyin; Li, Qijie; Jiang, Changhua; Jiao, Xuejun · 2022 · Crossref

DOI: 10.1186/s12938-022-00980-1

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 accurately detecting mental workload in complex human–machine systems, such as military operations and driving, where excessive cognitive load can lead to performance degradation and safety risks. While multimodal detection using electroencephalogram (EEG) and functional near-infrared spectroscopy (fNIRS) signals offers superior accuracy compared to single-modal approaches, existing methods often rely on complex, high-channel configurations that hinder practical application. The authors aimed to optimize the signal acquisition configuration by reducing the number of EEG channels while maintaining or improving detection accuracy, thereby creating a more portable and efficient monitoring system. The researchers conducted an experiment with 20 volunteers performing a Multi-Task Attribute Battery (MATB) task designed to induce four distinct levels of mental workload. Data collected included subjective workload ratings, 64-channel EEG signals, and two-channel fNIRS signals measuring oxygenated and deoxygenated hemoglobin in the prefrontal cortex. To optimize the EEG configuration, the authors analyzed feature importance across all channels and incrementally added channels to a support vector machine (SVM) model to determine the point of diminishing returns in classification accuracy. They then compared the performance of three feature sets—EEG only, fNIRS only, and fused EEG–fNIRS—using three classifiers: support vector machine (SVM), random forest (RF), and decision tree (DT). The results demonstrated that classification accuracy increased with the number of EEG channels but plateaued after 26 channels, leading the authors to select this optimized configuration. Physiological analysis revealed that theta band power in the prefrontal region and oxygenated hemoglobin concentrations increased with task difficulty, while alpha band power in the occipital region decreased. Notably, the study identified a significant increase in beta1 and beta2 band power in the occipital lobe with higher workload, a finding reported for the first time. In terms of classification, the fused EEG–fNIRS feature set outperformed single-modal sets. The random forest classifier achieved the highest four-level mental workload detection accuracy of 76.25 ± 5.21% using the optimized 26-channel EEG and two-channel fNIRS configuration. This multimodal approach also exhibited lower standard deviation in accuracy, indicating greater model stability and robustness compared to unimodal methods. The study concludes that optimizing channel configuration to 26 EEG channels and two frontal fNIRS channels provides a balance between high detection accuracy and practical feasibility. The proposed method achieves higher accuracy than previously reported results for four-level workload classification while significantly reducing equipment complexity. These findings suggest that the optimized multimodal detection system is well-suited for real-time monitoring in high-stakes environments, potentially enhancing operator safety and performance in military, aviation, and driving applications.

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
• workload measurement

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