Abstract:Researchers are currently exploring methods to detect early-stage Alzheimer's Disease (AD) and other forms of dementia such as frontotemporal dementia (FTD), especially through non-invasive biomarkers, i.e., measurements that reflect bio- logical processes. This paper utilizes a dataset of electroencephalogram (EEG) recordings, a noninvasive biomarker, to distinguish individuals with Alzheimer's or frontotemporal dementia from healthy control subjects. This paper explores the usage of machine learning methods to more accurately predict the cognitive status of patients from these non-invasive EEG Biomarkers. We found that AD patients could be easily separated from healthy controls based on their EEG features using simple linear classifiers with an accuracy of 77% and the AD, FTD, and healthy controls with an accuracy of around 57% (randomly selecting the right class is about one third or 33% in the 3-way classification).

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to distinguish patients with Alzheimer's disease (AD), frontotemporal dementia (FTD) from healthy control groups through non - invasive electroencephalogram (EEG) biomarkers. Specifically, the study aims to use features extracted from EEG recordings by machine - learning methods to predict patients' cognitive states more accurately, in order to achieve early detection of these diseases. ### Research Background - **Alzheimer's disease (AD)**: A progressive brain disease that leads to the destruction of brain cells and synapses, and eventually results in cognitive decline and the development of dementia. - **Frontotemporal dementia (FTD)**: The second - leading cause of early - onset dementia, mainly affecting the anterior regions of the brain. - **Biomarker**: Currently, there is an increasing interest in the early detection of these neurodegenerative diseases, especially non - invasive biomarkers such as EEG. ### Research Objectives - **Utilize EEG recordings**: As non - invasive biomarkers, use machine - learning methods to distinguish AD, FTD patients and healthy control groups. - **Improve classification accuracy**: Improve the accuracy of classification through multiple classifiers (such as linear discriminant analysis (LDA), support vector machine (SVM)). ### Main Findings - **Feature extraction**: The relative power spectral density (PSD) of five frequency bands (δ, θ, α, β, γ) and the α/θ and β/θ power ratios were extracted from EEG recordings. - **Statistical test**: Using the Mann - Whitney U test, it was found that there were significant differences in δ, θ, α, β, α/θ and β/θ features between AD and healthy control groups; there were significant differences in θ, β, γ, α/θ and β/θ features between CN and FTD groups. - **Classification performance**: - **Binary classification (AD vs. healthy control)**: The accuracy of the LDA classifier was 77.32%, the recall rate was 69.20%, the precision rate was 80.32%, the F1 - score was 72.16%, and the balanced accuracy was 76.61%. - **Ternary classification (AD, FTD, healthy control)**: The accuracy of the LDA classifier was 55.69%, the recall rate was 55.69%, the precision rate was 55.69%, the F1 - score was 55.69%, and the balanced accuracy was 52.00%. ### Discussion - **Research significance**: This study shows the potential of non - invasive and low - cost computational models based on EEG biomarkers in the early detection of AD and FTD. - **Limitations**: The data set is small, which may lead to over - fitting; biomarker changes in AD and FTD patients at different stages may affect the accuracy of the model. - **Future directions**: Expand the database to include more patients at different stages, and combine other features (such as MEG) to improve classification performance. ### Conclusion This study successfully achieved the preliminary classification of AD and FTD patients through EEG recordings and machine - learning methods, showing the potential application value in the early diagnosis of these diseases.

Classifying Alzheimers Disease and Dementia Patients Using Non-invasive EEG Biomarkers

Alzheimer’s Disease and Frontotemporal Dementia: A Robust Classification Method of EEG Signals and a Comparison of Validation Methods

Identification of Alzheimer's disease from central lobe EEG signals utilizing machine learning and residual neural network

Combining EEG signal processing with supervised methods for Alzheimer’s patients classification

Using Shallow Neural Networks with Functional Connectivity from EEG signals for Early Diagnosis of Alzheimer's and Frontotemporal Dementia

Classification of Alzheimer's Disease and Frontotemporal Dementia Using Electroencephalography to Quantify Communication between Electrode Pairs

Deep Learning-based Classification of Dementia using Image Representation of Subcortical Signals

Fully Automated Discrimination of Alzheimer's Disease Using Resting-State Electroencephalography Signals.

A Comparative Study on Feature Extraction Techniques for the Discrimination of Frontotemporal Dementia and Alzheimer's Disease with Electroencephalography in Resting-State Adults

Neural Biomarker Diagnosis and Prediction to Mild Cognitive Impairment and Alzheimer’s Disease Using EEG Technology

Fusing convolutional learning and attention-based Bi-LSTM networks for early Alzheimer's diagnosis from EEG signals towards IoMT

EEG Window Length Evaluation for the Detection of Alzheimer’s Disease over Different Brain Regions

Comparative analysis of machine learning algorithms for Alzheimer's disease classification using EEG signals and genetic information

A Self-driven Approach For Multi-class Discrimination In Alzheimer's Disease Based On Wearable EEG.

Machine Learning Classification of Alzheimer's Disease Stages Using Cerebrospinal Fluid Biomarkers Alone

Identifying Alzheimer Disease Dementia Levels Using Machine Learning Methods

Diagnosis of Alzheimer's disease via resting-state EEG: integration of spectrum, complexity, and synchronization signal features

Mild Cognitive Impairment detection based on EEG and HRV data

EEGAlzheimer'sNet: Development of transformer-based attention long short term memory network for detecting Alzheimer disease using EEG signal

Working Memory Classification Enhancement of EEG Activity in Dementia: A Comparative Study

An Automated Approach for the Detection of Alzheimer's Disease From Resting State Electroencephalography