EEG Dataset Collection for Mental Workload Predictions in Flight-Deck Environment

Hernández-Sabaté, Aura; Yauri, José; Folch, Pau; Álvarez, Daniel; Gil, Debora · 2024 · Crossref

DOI: 10.3390/s24041174

archive: archived pipeline: cataloged verified

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

Summary

This paper addresses the critical safety issue of high mental workload (MW) in aviation, which can lead to pilot error and catastrophic accidents. While deep learning methods offer potential for detecting cognitive states via physiological data, they require large, annotated datasets that are currently scarce, particularly for flight-deck scenarios. To bridge this gap, the authors present a new, publicly available EEG dataset specifically designed for MW recognition and transfer learning to flight environments. The dataset comprises EEG recordings collected from three distinct experimental scenarios using an Emotiv Epoc X headset (14 electrodes, 128 Hz sampling). The first scenario involved 16 participants performing variants of the N-back test to induce low, medium, and high MW through memory and arithmetic tasks. The second scenario featured 17 participants playing "Heat-the-Chair," a serious game designed to simulate multitasking and interruptions akin to air traffic control interactions. The third scenario involved two professional pilots operating an Airbus A320 simulator across five missions with varying complexities, including standard flights and emergency situations like engine failure or wind shear. Each experiment included baseline periods, task phases, and recovery periods, with data annotated using theoretical difficulty levels, self-perceived workload via NASA-TLX questionnaires, and performance metrics. The authors validated the dataset on three levels. First, they confirmed the validity of the theoretical MW complexity by correlating it with subjects' self-perceived difficulty and performance scores. Second, they demonstrated significant differences in EEG temporal patterns across the different theoretical difficulty levels, validating the physiological signal quality for MW assessment. Third, they assessed the dataset's utility for artificial intelligence by training and evaluating deep learning models for MW recognition. The results indicated that the dataset effectively supports the development of AI systems capable of distinguishing between cognitive states. The significance of this work lies in providing a comprehensive, open-access resource that combines controlled cognitive tasks with realistic flight simulation data. By including both laboratory-induced workload and real-world pilot scenarios, the dataset facilitates the training of models that can transfer knowledge from controlled environments to complex flight decks. This resource aims to advance research in human-machine interaction and pilot monitoring systems, ultimately contributing to improved aviation safety through better detection of cognitive overload.

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, self report data