EEG-based mental workload assessment
MSc thesis and research internship · g.tec medical engineering
Development and validation of EEG pipelines for N-back tasks manipulating memory load and presentation speed. The work covered human EEG acquisition, data-quality control, preprocessing, spectral and ERP analysis, feature extraction, classification and validation.
Scope
The project investigated whether EEG-derived markers and machine-learning models could distinguish mental-workload conditions across two N-back paradigms.
Work performed
- Participant setup, impedance control, acquisition monitoring and troubleshooting.
- Python/MNE-based preprocessing, quality control, spectral features and ERP analysis.
- Supervised classification with leakage-aware evaluation across separate recording sessions and participants.
- Analysis of observation-time requirements, temporal non-stationarity and cross-paradigm generalization.
Research output
The work contributed to the manuscript “EEG markers of mental workload in N-back tasks: temporal dynamics, classification, and generalization”, currently under review at Frontiers in Neuroergonomics, with shared first authorship.
Images from original thesis/research material.