Here, we tested that healthy elderly (Nold) and Alzheimer’s disease (AD) individuals can be discriminated with a moderate accuracy using resting state eyes-closed electroencephalographic (rsEEG) markers. Eyes-closed rsEEG data were collected in 100 Nold and 120 AD subjects. eLORETA freeware estimated the source current density (SCD) and functional connectivity (lagged linear connectivity, LLC) in frontal, central, parietal, occipital, temporal, and limbic regions. Delta (2–4 Hz), theta (4–8 Hz), alpha 1 (8–10.5 Hz), and alpha 2 (10.5–13 Hz) were the frequency bands of interest. Univariate (i.e., single rsEEG marker with receiver operating characteristic, ROC, curve) and multivariate (i.e., multiple rsEEG markers with artificial neural networks, ANNs) classifiers were used. The best accuracy was of 76% with univariate classifiers and 77% with multiple classifiers. The present results suggest that both univariate and multivariate rsEEG classifiers allowed a moderate classification rate between Nold and AD individuals. Interestingly, the accuracy based on multiple rsEEG markers as inputs to ANNs was similar to that obtained with a single rsEEG marker, unveiling their information redundancy for classification purposes. In future AD studies, multiple rsEEG markers should also include other classes of independent linear (i.e. directed transfer function) and nonlinear (i.e. entropy) variables to improve the classification.

Classification of Healthy Subjects and Alzheimer’s Disease Patients with Dementia from Cortical Sources of Resting State EEG Rhythms: Comparing Different Approaches

Brunetti, A.;Ferri, R.;Nobili, F.;
2019-01-01

Abstract

Here, we tested that healthy elderly (Nold) and Alzheimer’s disease (AD) individuals can be discriminated with a moderate accuracy using resting state eyes-closed electroencephalographic (rsEEG) markers. Eyes-closed rsEEG data were collected in 100 Nold and 120 AD subjects. eLORETA freeware estimated the source current density (SCD) and functional connectivity (lagged linear connectivity, LLC) in frontal, central, parietal, occipital, temporal, and limbic regions. Delta (2–4 Hz), theta (4–8 Hz), alpha 1 (8–10.5 Hz), and alpha 2 (10.5–13 Hz) were the frequency bands of interest. Univariate (i.e., single rsEEG marker with receiver operating characteristic, ROC, curve) and multivariate (i.e., multiple rsEEG markers with artificial neural networks, ANNs) classifiers were used. The best accuracy was of 76% with univariate classifiers and 77% with multiple classifiers. The present results suggest that both univariate and multivariate rsEEG classifiers allowed a moderate classification rate between Nold and AD individuals. Interestingly, the accuracy based on multiple rsEEG markers as inputs to ANNs was similar to that obtained with a single rsEEG marker, unveiling their information redundancy for classification purposes. In future AD studies, multiple rsEEG markers should also include other classes of independent linear (i.e. directed transfer function) and nonlinear (i.e. entropy) variables to improve the classification.
2019
978-3-030-01844-3
978-3-030-01845-0
File in questo prodotto:
File Dimensione Formato  
DelPercio et al. EEG-AD classification chapter Springler Verlag 2018.pdf

accesso chiuso

Tipologia: Documento in versione editoriale
Dimensione 1.36 MB
Formato Adobe PDF
1.36 MB Adobe PDF   Visualizza/Apri   Richiedi una copia

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11567/943103
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 1
  • ???jsp.display-item.citation.isi??? 1
social impact