Computational methods of EEG signals analysis for Alzheimer's disease classification

Abstract

Computational analysis of electroencephalographic (EEG) signals have shown promising results in detecting brain disorders, such as Alzheimer's disease (AD). AD is a progressive neurological illness that causes neuron cells degeneration, resulting in cognitive impairment. While there is no cure for AD, early diagnosis is critical to improving the quality of life of affected individuals. Here, we apply six computational time-series analysis methods (wavelet coherence, fractal dimension, quadratic entropy, wavelet energy, quantile graphs and visibility graphs) to EEG records from 160 AD patients and 24 healthy controls. Results from raw and wavelet-filtered (alpha, beta, theta and delta bands) EEG signals show that some of the time-series analysis methods tested here, such as wavelet coherence and quantile graphs, can robustly discriminate between AD patients from elderly healthy subjects. They represent a promising non-invasive and low-cost approach to the AD detection in elderly patients.

Figure 1

The location on the scalp of the 19 EEG original signal channels (groups A vs. C), represented by circles and colored according to the p value for $\mathcal{C}$, $\mathcal{F}$, $\mathcal{Q}$, $\mathcal{E}$, $\mathrm{\Delta }$, and $\mathcal{I}$, respectively.

Figure 2

The location on the scalp of the 19 EEG beta band signals (groups A vs. C), represented by circles and colored according to the p value for $\mathcal{C}$, $\mathcal{F}$, $\mathcal{Q}$, $\mathcal{E}$, $\mathrm{\Delta }$, and $\mathcal{I}$, respectively.

Figure 3

The location on the scalp of the 19 EEG alpha band signals (groups A vs. C), represented by circles and colored according to the p value for $\mathcal{C}$, $\mathcal{F}$, $\mathcal{Q}$, $\mathcal{E}$, $\mathrm{\Delta }$, and $\mathcal{I}$, respectively.

Figure 4

The location on the scalp of the 19 EEG theta band signals (groups A vs. C), represented by circles and colored according to the p value for $\mathcal{C}$, $\mathcal{F}$, $\mathcal{Q}$, $\mathcal{E}$, $\mathrm{\Delta }$, and $\mathcal{I}$, respectively.

Figure 5

The location on the scalp of the 19 EEG delta band signals (groups A vs. C), represented by circles and colored according to the p value for $\mathcal{C}$, $\mathcal{F}$, $\mathcal{Q}$, $\mathcal{E}$, $\mathrm{\Delta }$, and $\mathcal{I}$, respectively.

Figure 6

The location on the scalp of the 19 EEG original signal channels (groups B vs. D), represented by circles and colored according to the p value for $\mathcal{C}$, $\mathcal{F}$, $\mathcal{Q}$, $\mathcal{E}$, $\mathrm{\Delta }$, and $\mathcal{I}$, respectively.

Figure 7

The location on the scalp of the 19 EEG beta band signals (groups B vs. D), represented by circles and colored according to the p value for $\mathcal{C}$, $\mathcal{F}$, $\mathcal{Q}$, $\mathcal{E}$, $\mathrm{\Delta }$, and $\mathcal{I}$, respectively.

Figure 8

The location on the scalp of the 19 EEG alpha band signals (groups B vs. D), represented by circles and colored according to the p value for $\mathcal{C}$, $\mathcal{F}$, $\mathcal{Q}$, $\mathcal{E}$, $\mathrm{\Delta }$, and $\mathcal{I}$, respectively.

Figure 9

The location on the scalp of the 19 EEG theta band signals (groups B vs. D), represented by circles and colored according to the p value for $\mathcal{C}$, $\mathcal{F}$, $\mathcal{Q}$, $\mathcal{E}$, $\mathrm{\Delta }$, and $\mathcal{I}$, respectively.

Figure 10

The location on the scalp of the 19 EEG delta band signals (groups B vs. D), represented by circles and colored according to the p value for $\mathcal{C}$, $\mathcal{F}$, $\mathcal{Q}$, $\mathcal{E}$, $\mathrm{\Delta }$, and $\mathcal{I}$, respectively.

Figure 11

Boxplots for the best electrodes and frequency bands for $\mathcal{C}$, $\mathcal{F}$, $\mathcal{Q}$, $\mathcal{E}$, $\mathrm{\Delta }$, and $\mathcal{I}$, respectively.

Figure 12

The computational cost for the methods under consideration to computing the measures $\mathcal{C}$ (WC), $\mathcal{F}$ (FD), $\mathcal{Q}$ (QE), $\mathcal{E}$ (WE), $\mathrm{\Delta }$ (QG), and $\mathcal{I}$ (VG) as a function of a random time series (white noite) of length T.