Classification of target and non-target stimuli based on EEG signal during n-back working memory task

The proposed study explores the use of EEG signal analysis combined with machine learning to assess working memory functionality in healthy aged individuals performing n-back tasks. EEG data, sampled at 500 Hz, was pre-processed to remove artifacts, and features were extracted from time domain and frequency domain, with a focus on energy in the alpha, beta, and theta band, which are known to be associated with cognitive functions. Synthetic Minority Over-Sampling Technique (SMOTE) was used to address class imbalance issues, enhancing the performance of classification models. Support Vector Machine (SVM) and K-Nearest Neighbors (KNN) algorithms were used for classification, with SVM achieving the highest accuracy of 63% based on energy in alpha band and statistical features. The results highlight the significance of specific frequency bands and feature combinations in distinguishing working memory states, providing an edge for the design of robust cognitive assessment tools.