A global coupling index of multivariate neural series with application to the evaluation of mild cognitive impairment

Task-state EEG signal classification for spatial cognitive evaluation based on multi-scale high-density convolutional neural network

In this study, a multi-scale high-density convolutional neural network (MHCNN) classification method for spatial cognitive ability assessment was proposed, aiming at achieving the binary classification of task-state EEG signals before and after spatial cognitive training. Besides, the multi-dimensional conditional mutual information method was used to extract the frequency band features of the EEG data. And the coupling features under the combination of multi-frequency bands were transformed into multi-spectral images. At the same time, the idea of Densenet was introduced to improve the multi-scale convolutional neural network. Firstly, according to the discreteness of multispectral EEG image features, two-scale convolution kernels were used to calculate and learn useful channel and frequency band feature information in multispectral image data. Secondly, to enhance feature propagation and reduce the number of parameters, the dense network was connected after the multi-scale convolutional network, and the learning rate change function of the stochastic gradient descent algorithm was optimized to objectively evaluate the training effect. The experimental results showed that compared with the classical convolution neural network (CNN) and multi-scale convolution neural network, the proposed MHCNN had better classification performance in the six frequency band combinations with the highest accuracy of 98%: Theta-Alpha2-Gamma, Alpha2-Beta2-Gamma, Beta1-Beta2-Gamma, Theta-Beta2-Gamma, Theta-Alpha1-Gamma, and Alpha1-Alpha2-Gamma. By comparing the classification results of six frequency band combinations, it was found that the combination of the Theta-Beta2-Gamma band had the best classification effect. The MHCNN classification method proposed in this research could be used as an effective biological indicator of spatial cognitive training effect and could be extended to other brain function evaluations.

Coupling feature extraction method of resting state EEG Signals from amnestic mild cognitive impairment with type 2 diabetes mellitus based on weight permutation conditional mutual information

This study aimed to find a good coupling feature extraction method to effectively analyze resting state EEG signals (rsEEG) of amnestic mild cognitive impairment(aMCI) with type 2 diabetes mellitus(T2DM) and normal control (NC) with T2DM. A method of EEG signal coupling feature extraction based on weight permutation conditional mutual information (WPCMI) was proposed in this research. With the WPCMI method, coupling feature strength of two time series in Alpha1, Alpha2, Beta1, Beta2 and Gamma bands for aMCI with T2DM and NC with T2DM could be extracted respectively. Then selected three frequency bands coupling feature matrix with the help of multi-spectral image transformation method to map it as spectral image characteristics. And finally classified these characteristics through the convolution neural network method(CNN). For aMCI with T2DM and NC with T2DM, the highest classification accuracy of 96%, 95%, 95% could be achieved respectively in the combination of three frequency bands (Alpha1, Alpha2, Gamma), (Beta1, Beta2 and Gamma) and (Alpha2, Beta1, Beta2). This WPCMI method highlighted the coupling dynamic characteristics of EEG signals, and its classification performance was better than all previous methods in aMCI with T2DM diagnosis field. WPCMI method could be used as an effective biomarker to distinguish EEG signals of aMCI with T2DM and NC with T2DM. The coupling feature extraction method used in this paper provided a new perspective for the EEG analysis of aMCI with T2DM.

Alpha Power and Functional Connectivity in Cognitive Decline: A Systematic Review and Meta-Analysis

Background:

Mild cognitive impairment (MCI) is a stage between expected age-related cognitive decline and dementia. Dementias have been associated with changes in neural oscillations across the frequency spectrum, including the alpha range. Alpha is the most prominent rhythm in human EEG and is best detected during awake resting state (RS). Though several studies measured alpha power and synchronization in MCI, findings have not yet been integrated.

Objective:

To consolidate findings on power and synchronization of alpha oscillations across stages of cognitive decline.

Methods:

We included studies published until January 2020 that compared power or functional connectivity between 1) people with MCI and cognitively healthy older adults (OA) or people with a neurodegenerative dementia, and 2) people with progressive and stable MCI. Random-effects meta-analyses were performed when enough data was available.

Results:

Sixty-eight studies were included in the review. Global RS alpha power was lower in AD than in MCI (ES = –0.30; 95% CI = –0.51, –0.10; k = 6), and in MCI than in OA (ES = –1.49; 95% CI = –2.69, –0.29; k = 5). However, the latter meta-analysis should be interpreted cautiously due to high heterogeneity. The review showed lower RS alpha power in progressive than in stable MCI, and lower task-related alpha reactivity in MCI than in OA. People with MCI had both lower and higher functional connectivity than OA. Publications lacked consistency in MCI diagnosis and EEG measures.

Conclusion:

Research indicates that RS alpha power decreases with increasing impairment, and could—combined with measures from other frequency bands—become a biomarker of early cognitive decline.

Feature Classification Method of Resting-State EEG Signals From Amnestic Mild Cognitive Impairment With Type 2 Diabetes Mellitus Based on Multi-View Convolutional Neural Network

The convolutional neural network (CNN) model is an active research topic in the field of EEG signals analysis. However, the classification effect of CNN on EEG signals of amnestic mild cognitive impairment (aMCI) with type 2 diabetes mellitus (T2DM) is not ideal. Even if EEG signals are transformed into multispectral images that are more closely matched with the model, the best classification performance can not be achieved. Therefore, to improve the performance of CNN toward EEG multispectral image classification, a multi-view convolutional neural network (MVCNN) classification model based on inceptionV1 is designed in this study. This model mainly improves and optimizes the convolutional layers and stochastic gradient descent (SGD) in the convolutional architecture model. Firstly, based on the discreteness of EEG multispectral image features, the multi-view convolutional layer structure was proposed. Then the learning rate change function of the SGD was optimized to increase the classification performance. The multi-view convolutional nerve was used in an EEG multispectral classification task involving 19 aMCI with T2DM and 20 normal controls. The results showed that compared with the traditional classification models, MVCNN had a better stability and accuracy. Therefore, MVCNN could be used as an effective feature classification method for aMCI with T2DM.

Resting-state EEG signal classification of amnestic mild cognitive impairment with type 2 diabetes mellitus based on multispectral image and convolutional neural network

OBJECTIVE

The purpose of this study is to judge whether this combination method of multispectral image and convolutional neural network (CNN) method can be used to distinguish amnestic mild cognitive impairment (aMCI) with Type 2 diabetes mellitus (T2DM) and normal controls (NC) with T2DM effectively.

APPROACH

In this study, the authors first combined EEG signals from aMCI patients with T2DM and NC with T2DM on five different frequency bands, including Theta, Alpha1, Alpha2, Beta1, and Beta2. Then, the authors converted these time series into a series of multispectral images. Finally, the images data were classified with the CNN method.

MAIN RESULTS

The classification effects of up to 89%, 91%, and 92% are obtained on the three combinations of frequency bands: Theta, Alpha1, and Alpha2; Alpha1, Alpha2, and Beta1; and Alpha2, Beta1, and Beta2. The spatial properties of EEG signals are highlighted, and its classification performance is found to be better than all the previous methods in the field of aMCI and T2DM diagnosis. The combination of multispectral images and CNN can be used as an effective biomarker for distinguishing the EEG signals in patients with aMCI and T2DM and in patients with NC with T2DM.

SIGNIFICANCE

The combined approach used in this paper provides a new perspective for the analysis of EEG signals in patients with aMCI and T2DM.

The feature extraction of resting-state EEG signal from amnestic mild cognitive impairment with type 2 diabetes mellitus based on feature-fusion multispectral image method

Recently, combining feature extraction and classification method of electroencephalogram (EEG) signals has been widely used in identifying mild cognitive impairment. However, it remains unclear which feature of EEG signals is best effective in assessing amnestic mild cognitive impairment (aMCI) with type 2 diabetes mellitus (T2DM) when combining one classifier. This study proposed a novel feature extraction method of EEG signals named feature-fusion multispectral image method (FMIM) for diagnosis of aMCI with T2DM. The FMIM was integrated with convolutional neural network (CNN) to classify the processed multispectral image data. The results showed that FMIM could effectively identify aMCI with T2DM from the control group compared to existing multispectral image method (MIM), with improvements including the type and quantity of feature extraction. Meanwhile, part of the invalid calculation could be avoided during the classification process. In addition, the classification evaluation indexes were best under the combination of Alpha2-Beta1-Beta2 frequency bands in data set based on FMIM-1, and were also best under the combination of the Theta-Alpha1-Alpha2-Beta1-Beta2 frequency bands in data set based on FMIM-2. Therefore, FMIM can be used as an effective feature extraction method of aMCI with T2DM, and as a valuable biomarker in clinical applications.

Estimating coupling strength between multivariate neural series with multivariate permutation conditional mutual information

Recently, coupling between groups of neurons or different brain regions has been widely studied to provide insights into underlying mechanisms of brain functions. To comprehensively understand the effect of such coupling, it is necessary to accurately extract the coupling strength information among multivariate neural signals from the whole brain. This study proposed a new method named multivariate permutation conditional mutual information (MPCMI) to quantitatively estimate the coupling strength of multivariate neural signals (MNS). The performance of the MPCMI method was validated on the simulated MNS generated by multi-channel neural mass model (MNMM). The coupling strength feature of simulated MNS extracted by MPCMI showed better performance compared with standard methods, such as permutation conditional mutual information (PCMI), multivariate Granger causality (MVGC), and Granger causality analysis (GCA). Furthermore, the MPCMI was applied to estimate the coupling strengths of two-channel resting-state electroencephalographic (rsEEG) signals from different brain regions of 19 patients with amnestic mild cognitive impairment (aMCI) with type 2 diabetes mellitus (T2DM) and 20 normal control (NC) with T2DM in Alpha1 and Alpha2 frequency bands. Empirical results showed that the MPCMI could effectively extract the coupling strength features that were significantly different between the aMCI and the NC. Hence, the proposed MPCMI method could be an effective estimate of coupling strengths of MNS, and might be a viable biomarker for clinical applications.

A new EEG synchronization strength analysis method: S-estimator based normalized weighted-permutation mutual information

Synchronization is an important mechanism for understanding information processing in normal or abnormal brains. In this paper, we propose a new method called normalized weighted-permutation mutual information (NWPMI) for double variable signal synchronization analysis and combine NWPMI with S-estimator measure to generate a new method named S-estimator based normalized weighted-permutation mutual information (SNWPMI) for analyzing multi-channel electroencephalographic (EEG) synchronization strength. The performances including the effects of time delay, embedding dimension, coupling coefficients, signal to noise ratios (SNRs) and data length of the NWPMI are evaluated by using Coupled Henon mapping model. The results show that the NWPMI is superior in describing the synchronization compared with the normalized permutation mutual information (NPMI). Furthermore, the proposed SNWPMI method is applied to analyze scalp EEG data from 26 amnestic mild cognitive impairment (aMCI) subjects and 20 age-matched controls with normal cognitive function, who both suffer from type 2 diabetes mellitus (T2DM). The proposed methods NWPMI and SNWPMI are suggested to be an effective index to estimate the synchronization strength.

Complex network analysis of resting state EEG in amnestic mild cognitive impairment patients with type 2 diabetes

Purpose: Diabetes is a great risk factor for dementia and mild cognitive impairment (MCI). This study investigates whether complex network-derived features in resting state EEG (rsEEG) could be applied as a biomarker to distinguish amnestic mild cognitive impairment (aMCI) from normal cognitive function in subjects with type 2 diabetes (T2D).

Method: In this study, EEG was recorded in 28 patients with T2D (16 aMCI patients and 12 controls) during a no-task eyes-closed resting state. Pair-wise synchronization of rsEEG signals were assessed in six frequency bands (delta, theta, lower alpha, upper alpha, beta, and gamma) using phase lag index (PLI) and grouped into long distance (intra- and inter-hemispheric) and short distance interactions. PLI-weighted connectivity networks were also constructed, and characterized by mean clustering coefficient and path length. The correlation of these features and Montreal Cognitive Assessment (MoCA) scores was assessed.

Results: Main findings of this study were as follows: (1) In comparison with controls, patients with aMCI had a significant decrease of global mean PLI in lower alpha, upper alpha, and beta bands. Lower functional connection at short and long intra-hemispheric distance mainly appeared on the left hemisphere. (2) In the lower alpha band, clustering coefficient was significantly lower in aMCI group, and the path length significantly increased. (3) Cognitive status measured by MoCA had a significant positive correlation with cluster coefficient and negative correlation with path length in lower alpha band.

Conclusions: The brain network of aMCI patients displayed a disconnection syndrome and a loss of small-world architecture. The correlation between cognitive states and network characteristics suggested that the more in deterioration of the diabetes patients' cognitive state, the less optimal the network organization become. Hence, the complex network-derived biomarkers based on EEG could be employed to track cognitive function of diabetic patients and provide a new diagnosis tool for aMCI.

Resting-state EEG coupling analysis of amnestic mild cognitive impairment with type 2 diabetes mellitus by using permutation conditional mutual information

OBJECTIVE

This study was meant to explore whether the coupling strength and direction of resting-state electroencephalogram (rsEEG) could be used as an indicator to distinguish the patients of type 2 diabetes mellitus (T2DM) with or without amnestic mild cognitive impairment (aMCI).

METHODS

Permutation conditional mutual information (PCMI) was used to calculate the coupling strength and direction of rsEEG signals between different brain areas of 19 aMCI and 20 normal control (NC) with T2DM on 7 frequency bands: Delta, Theta, Alpha1, Alpha2, Beta1, Beta2 and Gamma. The difference in coupling strength or direction of rsEEG between two groups was calculated. The correlation between coupling strength or direction of rsEEG and score of different neuropsychology scales were also calculated.

RESULTS

We have demonstrated that PCMI can calculate effectively the coupling strength and directionality of EEG signals between different brain regions. The significant difference in coupling strength and directionality of EEG signals was found between the patients of aMCI and NC with T2DM on different brain regions. There also existed significant correlation between sex or age and coupling strength or coupling directionality of EEG signals between a few different brain regions from all subjects.

CONCLUSIONS

The coupling strength or directionality of EEG signals calculated by PCMI are significantly different between aMCI and NC with T2DM.

SIGNIFICANCE

These results showed that the coupling strength or directionality of EEG signals calculated by PCMI might be used as a biomarker in distinguishing the aMCI from NC with T2DM.

A critical review: coupling and synchronization analysis methods of EEG signal with mild cognitive impairment

At present, the clinical diagnosis of mild cognitive impairment (MCI) patients becomes the important approach of evaluating early Alzheimer's disease. The methods of EEG signal coupling and synchronization act as a key role in evaluating and diagnosing MCI patients. Recently, these coupling and synchronization methods were used to analyze the EEG signals of MCI patients according to different angles, and many important discoveries have been achieved. However, considering that every method is single-faceted in solving problems, these methods have various deficiencies when analyzing EEG signals of MCI patients. This paper reviewed in detail the coupling and synchronization analysis methods, analyzed their advantages and disadvantages, and proposed a few research questions needed to solve in the future. Also, the principles and best performances of these methods were described. It is expected that the performance analysis of these methods can provide the theoretical basis for the method selection of analyzing EEG signals of MCI patients and the future research directions.