SS-RIME: A Scale-Stabilized Approach to EEG Cognitive Workload Classification
Sensors · 25.04.2026 · 26 (9) · 2679
DOI: 10.3390/s26092679 ↗Kurz gefasst
Fachartikel in der Zeitschrift Sensors, online erschienen am 25.04.2026. Verfasst von Kais Khaldi und 4 weiteren Autor:innen. Beteiligte Einrichtung: Department of Computer Science, College of Computer and Information Sciences, Jouf University, Sakaka 72388, Saudi Arabia. Hauptthemen laut MEDLINE: Electroencephalography, Cognition, Workload. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 53 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: This research was funded by the Deanship of Graduate Studies and Scientific Research at Jouf University, Jouf University.
Abstract (Originalsprache)
Accurate and interpretable assessment of cognitive workload from EEG remains a central challenge in neuroergonomics and real-time human–machine interaction. To address the limitations of existing Empirical Mode Decomposition (EMD) and Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) approaches, particularly their instability, limited neuroscientific grounding, and sensitivity to amplitude fluctuations, this paper introduces Scale-Stabilized Relative Intrinsic Mode Energy (SS-RIME), a theoretically motivated and physiologically informed feature extraction framework. SS-RIME integrates instantaneous frequency stabilization to enforce a consistent oscillatory hierarchy across subjects, delta (1–4 Hz) and theta (4–7.5 Hz) spectral weighting based on established frontal-midline activity, and cross-IMF energy normalization to reduce amplitude-driven variability. Applied to 64-channel EEG recorded during N-back tasks, the proposed framework achieved high performance, outperforming both classical machine-learning baselines and deep learning models such as EEGNet, DeepConvNet, and ShallowConvNet. SS-RIME yielded accuracies of 99.12±0.41% (0 vs. 2-back), 97.84±0.63% (0 vs. 3-back), and 92.31±1.12% (2 vs. 3-back), demonstrating strong cross-subject generalization. Theta-dominant IMFs over frontal midline regions emerged as the most discriminative components, supporting the neuroscientific validity of the stabilized and spectrally weighted Hilbert–Huang representation. With an inference time below 20 ms per epoch, SS-RIME is computationally efficient and suitable for real-time neuroergonomics applications, providing a robust, explainable, and physiologically grounded solution for EEG-based cognitive workload decoding while addressing key methodological gaps in prior EMD/CEEMDAN and deep learning approaches.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Computer Science, College of Computer and Information Sciences, Jouf University, Sakaka 72388, Saudi Arabia.
- Department of Information System, College of Computer and Information Sciences, Jouf University, Sakaka 72388, Saudi Arabia.
- Department of Mathematics, College of Science, Jouf University, Sakaka 72388, Saudi Arabia.
ORCID-Profile: Kais Khaldi, Afrah Alanazi, Inam Alanazi, Sahar Almenwer, Anis Mohamed
Förderung
- This research was funded by the Deanship of Graduate Studies and Scientific Research at Jouf University
- Jouf University
Zitieren
APA
Khaldi, K., Alanazi, A., Alanazi, I., Almenwer, S., & Mohamed, A. (2026). SS-RIME: A Scale-Stabilized Approach to EEG Cognitive Workload Classification. Sensors, 26(9), 2679. https://doi.org/10.3390/s26092679
Vancouver
Khaldi K, Alanazi A, Alanazi I, Almenwer S, Mohamed A. SS-RIME: A Scale-Stabilized Approach to EEG Cognitive Workload Classification. Sensors. 2026;26(9):2679. doi:10.3390/s26092679
BibTeX
@article{Khaldi2026,
author = {Khaldi, Kais and Alanazi, Afrah and Alanazi, Inam and Almenwer, Sahar and Mohamed, Anis},
title = {SS-RIME: A Scale-Stabilized Approach to EEG Cognitive Workload Classification},
journal = {Sensors},
year = {2026},
volume = {26},
number = {9},
pages = {2679},
doi = {10.3390/s26092679}
}
EUVIMED