Influences of Information Processing Modality on Mental Workload Recognition Performance Based on EEG Feature Extraction
DOI: 10.54941/ahfe1004750
archive: archived pipeline: cataloged
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This study investigates whether the modality of information processing (visual-only versus audio-visual) significantly impacts the accuracy of mental workload recognition using electroencephalogram (EEG) signals. The research is motivated by the gap between previous EEG-based workload studies, which predominantly utilized single-visual or dual-visual tasks, and real-world operational environments, such as aircraft cockpits, where operators routinely process simultaneous auditory and visual information. Accurate, real-time workload recognition is critical for optimizing human-machine interaction and preventing performance degradation caused by overload or underload. The experimental design involved 20 subjects performing simulated flight tasks under two scenarios: a visual single-modality task and an audio-visual dual-modality task. In the dual-modality scenario, subjects processed auditory alarms at a rate of three per minute in addition to the visual flight task. Within each scenario, two levels of mental workload were induced by varying task complexity: a low-workload airfield traffic pattern task and a high-workload ground-attack task involving complex avionics and weapon system operations. Experimental sequences followed a Latin-square design to mitigate practice and fatigue effects. Data collection included 30-channel EEG signals and NASA-TLX subjective workload ratings. EEG data were preprocessed using band-pass filtering (1–30 Hz) and independent component analysis to remove ocular artifacts. Two feature extraction methods were employed: Power Spectral Density (PSD), which yielded a 120-dimensional feature vector based on delta, theta, alpha, and beta band energies, and Common Spatial Pattern (CSP), which produced a 4-dimensional feature vector from specific central electrodes. Mental workload levels were classified using a Support Vector Machine with a Radial Basis Function kernel, optimized via grid search and 4-fold cross-validation. Results confirmed that the experimental design successfully induced distinct workload levels, as evidenced by significantly higher NASA-TLX scores in high-workload conditions for both modality types. Regarding recognition performance, PSD features yielded higher accuracy than CSP features in both scenarios (visual: 0.8914 vs. 0.8630; audio-visual: 0.8752 vs. 0.8330). However, statistical tests revealed no significant difference in recognition accuracy between the visual single-modality and audio-visual dual-modality scenarios for either feature extraction method. While PSD features showed a marginal trend toward better performance in the dual-modality condition compared to CSP, this difference was not statistically significant. The authors attribute the lack of modality effect to the brief and discrete nature of the auditory alarms, which did not substantially alter EEG patterns associated with workload. The study concludes that information processing modality does not significantly influence EEG-based mental workload recognition performance. This finding suggests that EEG-based workload monitoring systems developed for visual tasks may be generalizable to multi-modal operational environments without substantial loss in accuracy. Furthermore, the study highlights that while PSD features offer slightly superior classification accuracy, CSP features provide a more computationally efficient alternative with minimal performance trade-off, making them suitable for real-time applications where processing speed is a constraint.
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.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | Crossref | — | — | 1 | 2026-08-09 |
| archive | success | canonical_url | — | — | 1 | 2026-08-09 |
| extract | success | cached | — | — | 4 | 2026-08-23 |
| clean | success | clean | — | — | 1 | 2026-08-09 |
| chunk | success | chunk | — | — | 1 | 2026-08-09 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-08-09 |
| promote | success | — | — | — | 1 | 2026-08-09 |
| summarize | success | llm | qwen3.8-27b-gittensor | summ-v5 | 3 | 2026-08-23 |
| tag | success | vector_similarity | — | — | 11 | 2026-08-11 |
| verify | success | — | — | — | 2 | 2026-08-09 |
Summary generated by qwen3.8-27b-gittensor on 2026-08-23; verification: pending re-verification.
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).
- Empirical Findings: physiological data