Manufacturing & Industrial Engineering Faculty Publications
Document Type
Article
Publication Date
9-11-2025
Abstract
Highlights
-
Novel hybrid architecture: Combined autoencoders with bidirectional LSTM networks for enhanced EEG signal classification, achieving 98% accuracy in distinguishing AD, FTD, and healthy controls.
-
Explainable AI integration: Implemented SHAP (SHapley Additive exPlanations) framework to enhance model transparency and identify entropy as the most influential feature for neurodegenerative disease detection.
-
Optimal temporal segmentation: Demonstrated that 5-s EEG windows with 50% overlap provide the best balance between classification accuracy and computational efficiency.
-
Comprehensive feature extraction: Utilized Power Spectral Density (PSD) analysis across standard frequency bands (Delta, Theta, Alpha, Beta, Gamma) following autoencoder-based dimensionality reduction.
-
Superior performance validation: Outperformed traditional machine learning methods (KNN: 38%, SVM: 40%) and unidirectional LSTM (84%) with the proposed Bi-LSTM approach achieving 98% accuracy.
-
Clinical applicability focus: Addressed interpretability challenges in deep learning for medical diagnosis, providing feature-level explanations essential for clinical trust and adoption.
Abstract
This study explores the use of deep learning techniques for classifying EEG signals in the context of Alzheimer’s disease (AD) and frontotemporal dementia (FTD). We propose a novel classification pipeline that combines autoencoders for feature extraction and bidirectional long short-term memory (Bi-LSTM) networks for analyzing patterns over time in EEG data. Given the complexity and high dimensionality of EEG signals, we employed an autoencoder to reduce data dimensionality while preserving key diagnostic features. The Bi-LSTM model effectively identified subtle temporal patterns in brain activity that are indicative of AD and FTD. To enhance interpretability, we applied SHapley Additive exPlanations (SHAP), providing insights into how individual features contribute to the model’s predictions. We evaluated our approach on a publicly available EEG dataset from OpenNeuro, which includes resting-state EEG recordings from 88 elderly participants—36 with AD, 23 with FTD, and 29 cognitively normal controls. EEG provides a non-invasive, cost-effective tool for brain monitoring, but presents challenges such as noise sensitivity and inter-subject variability. Despite these challenges, our approach achieved 98% accuracy while maintaining transparency, making it a promising tool for clinical applications in the diagnosis of neurodegenerative diseases.
Recommended Citation
Mouazen, Badr, Ahmed Bendaouia, Omaima Bellakhdar, Khaoula Laghdaf, Aya Ennair, El Hassan Abdelwahed, and Giovanni De Marco. "Transparent EEG analysis: leveraging Autoencoders, Bi-LSTMs, and SHAP for improved neurodegenerative diseases detection." Sensors 25, no. 18 (2025): 5690. https://doi.org/10.3390/s25185690
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.
Publication Title
Sensors
DOI
10.3390/s25185690
Included in
Industrial Engineering Commons, Manufacturing Commons, Medicine and Health Sciences Commons
Comments
© 2025 by the authors.
Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).