Nonlinear analysis of spontaneous MEG activity in alzhemer's disease

Chapter in book


Research Areas

Publication Details

Author list: Fernández A, Gómez, C., García, M., Hornero, R.
Publisher: Nova Science Publishers, Inc.
Publication year: 2013
Start page: 103
End page: 118
Number of pages: 16
ISBN: 978-162417350-9
Languages: English-Great Britain (EN-GB)


Alzheimer's disease (AD) is one of the most common disorders among
elderly population and it is considered the main cause of dementia in
western countries. This progressive, degenerative brain disorder is
characterized by neural loss and the appearance of neurofibrillary
tangles and senile plaques. Although a definite diagnosis is only
possible by necropsy, a differential diagnosis with other types of
dementia and with major depression should be attempted. The differential
diagnosis includes medical history studies, physical and neurological
evaluation, mental status tests, and neuroimaging techniques. Nowadays,
magnetoencephalography (MEG) recordings are not used in AD clinical
diagnosis, despite its potential as aid diagnostic tool. The aim of this
chapter is to show recent analyses of the spontaneous MEG activity in
AD. For this purpose, five minutes of recording were acquired with a
148-channel whole-head magnetometer in 36 patients with probable AD and
26 control subjects. MEG data were analyzed by means of three nonlinear
measures: Lempel-Ziv complexity (LZC), sample entropy (SampEn) and
detrended fluctuation analysis (DFA). LZC measures the number of
different substrings and the rate of their recurrence along the original
time series. SampEn is an embbeding entropy that quantifies the signal
regularity. Finally, DFA is a method designed to quantify correlations
in noisy and non-stationary time series. LZC and SampEn showed that MEG
recordings are less complex and more regular in AD patients than in
control subjects. DFA revealed changes in the fluctuations of MEG
signals. Significant differences between AD patients and elderly
controls were found. using the three nonlinear methods (p-values <
0.05, Welch t-test). We used receiver operating characteristic (ROC)
curves with a leave-one-out cross-validation procedure to assess the
ability of the methods to classify AD patients and control subjects. The
highest area under the ROC curve was achieved with DFA (0.7959) whereas
the highest accuracy was obtained with LZC algorithm (72.58%). We
conclude that nonlinear analyses from spontaneous MEG activity could be
useful to increase our insight into AD. This chapter shows that MEG
recordings reflect the alterations on AD patients' brains. These
alterations may be associated with deficiencies in information
processing. © 2013 by Nova Science Publishers, Inc. All rights reserved.


No matching items found.


No matching items found.

Last updated on 2019-10-08 at 00:15