Related Experiment Video
Updated: Sep 16, 2025
11:01
Examining the Characteristics of Episodic Memory using Event-related Potentials in Patients with Alzheimer's Disease
Published on: August 30, 2011
13.7K
Graph Spectral Analysis Using Electroencephalography in Alzheimer Disease and Frontotemporal Dementia Patients
María Paula Bonomini1,2,3, Eduardo Ghiglioni1,4, Noelia Belén Ríos1,4
1Instituto Argentino de Matemática "Alberto P. Calderón", Saavedra 15, Bs. As., Argentina.
International Journal of Neural Systems
|July 8, 2025
Summary
Graph Spectral Analysis (GSA) using electroencephalography (EEG) shows promise for diagnosing Alzheimer's disease. Regional GSA metrics, particularly in the Theta band, demonstrated high classification performance in distinguishing AD patients from controls.
Area of Science:
- Neuroscience
- Network Science
- Medical Imaging
Background:
- Graph theory aids in understanding brain dysfunction via MagnetoEncephaloGraphy (MEG) and fMRI.
- Limited research exists on graph theory with reduced electrode sets like 10-20 Electroencephalography (EEG).
Purpose of the Study:
- To apply Graph Spectral Analysis (GSA) to EEG signals for Alzheimer's disease (AD) diagnosis.
- To evaluate the utility of global and regional GSA metrics in classifying AD patients versus controls.
- To identify key parameters and brain rhythms valuable for AD diagnosis.
Main Methods:
- Calculated global GSA metrics from adjacency or Laplacian matrices of EEG signals.
- Computed regional GSA metrics, assessing centrality in five cortical regions (frontal, central, parietal, temporal, occipital).
- Utilized these metrics in a binary AD/control classification using k-fold cross-validation.
Main Results:
- The Theta band exhibited the highest connectivity and synchronizability across all groups.
- Theta band showed preserved connections and shortest average distances among temporal electrodes.
- Regional GSA parameters, especially in the Theta band, yielded the highest classification performance (mean accuracy, sensitivity, and F1-score).
Conclusions:
- Regional GSA parameters generally outperformed global metrics for most rhythms in AD classification.
- The Theta band and regional analysis show significant potential for AD diagnosis using EEG.
- Further investigation into refined GSA parameters is encouraged for improved diagnostic specificity.
