EEGDecoder-x: an explainable deep learning framework for cross-subject EEG-based detection of Alzheimer's and Creutzfeldt–Jakob disease - Report - MDSpire
Advertisement
EEGDecoder-x: A Transparent Deep Learning Approach for Cross-Subject Detection of Alzheimer's Disease and Creutzfeldt–Jakob Disease Using EEG Data
Clinical Report: EEGDecoder-x for Detecting Alzheimer's and CJD Using EEG
Overview
The EEGDecoder-x framework demonstrates high accuracy in distinguishing between Alzheimer's disease (AD), Creutzfeldt–Jakob disease (CJD), and healthy controls using EEG data, achieving 97.22% classification accuracy.
Background
Neurodegenerative diseases like Alzheimer's and Creutzfeldt–Jakob disease pose significant diagnostic challenges due to overlapping early-stage symptoms. Current diagnostic methods often rely on expensive and invasive procedures.
Data Highlights
Metric
Value
Classification Accuracy
97.22%
Subjects with AD
12
Subjects with CJD
12
Healthy Controls
12
Key Findings
EEGDecoder-x achieved 97.22% classification accuracy in distinguishing AD, CJD, and healthy controls.
The framework includes a hybrid attention network for effective spatio-temporal feature extraction.
Utilizing a Leave-One-Subject-Out evaluation paradigm addresses issues of data leakage in model validation.
Early-stage EEG abnormalities in CJD can be non-specific, complicating differential diagnosis.
Clinical Implications
The EEGDecoder-x framework offers a promising tool for the early detection of neurodegenerative diseases, emphasizing the need for interpretable models in clinical settings. Its high accuracy and explainability may facilitate better diagnostic processes for conditions like AD and CJD.
Conclusion
EEGDecoder-x represents a significant advancement in EEG-based diagnostics for neurodegenerative diseases, combining high accuracy with essential interpretability for clinical application.
