کنفرانس مهندسی برق ایران

صفحه اصلی / سی و چهارمین کنفرانس بین المللی مهندسی برق

EEG-Based Signal Processing and Machine Learning Approach for Alzheimer’s Disease Detection Using FFT

نویسندگان :

Mahdiyeh Tofighi Milani¹

1- دانشگاه صنعتی سهند

کلمات کلیدی :

Alzheimer’s Disease،Electroencephalogram،Machine Learning،Fast Fourier Transform

چکیده :

Alzheimer’s disease (AD) is one of the most common and complex neuropsychological disorders worldwide. It is indeed a very difficult and sensitive kind of problem to figure out the nature of the illness especially within the early stages. To a large extent, these are investigative tools like MRI and PET scans, cognitive and psychological tests, biological markers, and genetic and molecular examinations that pave the way to a diagnosis for the disease; however, these are expensive and can be invasive in some cases. Because of this, the study presents a solution that is quite far from being invasive and yet very affordable; this solution is based on electroencephalogram (EEG) signals and machine learning algorithms to differentiate patients with Alzheimer’s disease from healthy controls. First, in the suggested approach, the EEG signals are decomposed into various frequency sub-bands through the Fast Fourier Transform (FFT). Afterward, the ReliefF algorithm selects statistical and nonlinear features, i.e., mean, variance, kurtosis, skewness, Shannon entropy, mobility, and complexity, and these features are input to different classifiers such as k-Nearest Neighbors (KNN), Support Vector Machine with a linear kernel (L-SVM), Decision Tree (DT), Linear Discriminant Analysis (LDA), Random Forest (RF), and Naïve Bayes (NB). Better accuracy and speed of the model were achieved as a result of the identification of up to 60 features with the help of the ReliefF algorithm. The assessment of the proposed method performance with EEG data from Alzheimer's and healthy subjects demonstrated that the RF classifier was the most effective one, delivering an average accuracy of 95.36%.