Detecting Cognitive Decline in Alzheimer's Disease using Brain Signals: An EEG Based Classification Approach

Alzheimer's disease is a long-term illness character-ized by a decline in abilities and the development of dementia. The main objective of this study is to explore the potential of using EEG-based classification for identifying and comprehending Alzheimer's disease in its stages. EEG, which measures brain ac-tivity, offers insights into the functioning of neurons. By applying machine learning techniques to EEG data, we have seen results in distinguishing between individuals with Alzheimer's disease and those without it. For our research, we used a dataset that included EEG recordings from individuals diagnosed with Alzheimer's disease, frontotemporal dementia, and healthy individuals for comparison. We extracted Mel coefficients (MFCC) as features and trained our models using decision trees and LightGBM classifiers. To evaluate their effectiveness, we assessed accuracy, sensitivity, specificity, and the area under the operating curve (AUC ROC). Our results showed accuracy in differentiating between Alzheimer's patients and healthy individuals, with an 89.3% accuracy rate along with AUC ROC values of 0.989 for the decision tree classifier and 98.64 % and 1.0 for the Igbm classifier, respectively. These findings highlight the potential of combining EEG-based classification with machine learning algorithms for diagnosis and personalized treatment planning for Alzheimer's disease.