Analysis of MEG recordings from Alzheimer's disease patients with sample and multiscale entropies

Abstract

Abstract-Alzheimer's disease (AD) is one of the most prominent neurodegenerative disorders. The aim of this study is to analyze the magnetoencephalogram (MEG) background activity in AD patients using sample entropy (SampEn) and multiscale entropy (MSE). The former quantifies the signal regularity, while the latter is a complexity measure. These concepts, irregularity and complexity, are linked although the relationship is not straightforward. Five minutes of recording were acquired with a 148-channel whole-head magnetometer in 20 patients with probable AD and 21 control subjects. Our results show that MEG recordings are less complex and more regular in AD patients than in control subjects. Significant differences between both groups were found in some MEG channels with both methods (p < 0.01, Student's t-test with Bonferroni's correction). Using receiver operating characteristic curves, accuracies of 75.6% with SampEn and of 87.8% with MSE were reached. Our findings show the usefulness of these entropy measures to increase our insight into AD.

Use of magnetoencephalography (MEG) to study functional brain networks in neurodegenerative disorders

“…The loss of spectral entropy was found in all brain regions [51]. A lower sample entropy and a lower multiscale entropy in AD were demonstrated by Gomez et al [52] confirming the notion of a loss of complexity. In another study Gomez et al [53] showed a slower decline rate of auto-mutal information of MEG in AD, reflecting an abnormal type of brain dynamics in this disease.…”

Complexity Analysis of Spontaneous Brain Activity in Attention-Deficit/Hyperactivity Disorder: Diagnostic Implications

“…Several complexity estimates have been applied to EEG and MEG: Correlation Dimension, First Lyapunov Component, Auto-mutual Information, Lempel-Ziv Complexity (LZC), etc. (17,18).…”

Characterization of early partial seizure onset: Frequency, complexity and entropy

“…The sample entropy SampEn (Richman and Moorman, 2000) is derived from the approximate entropy (Pincus et al, 1991) a method closely related to the Kolmogorov entropy and characterizing the rate of generation of new information. SampEn is used for Alzheimer patients assessment (Gomez et al, 2007) and was included in the features used for seizure detection algorithm (Aarabi et al, 2009). The permutation entropy permH (Bandt and Pompe, 2002) was developed as a fast and robust measure of complexity which, in known chaotic systems, behaves similarly to the Lyapunov exponents.…”