Optimized EEG-BCI Techniques to determine the level of concentration using CSP with KNN

Abstract

The increasing prominence of Digital Learning Methodologies, particularly self-learning through video tutorials, highlights a critical need to assess listener engagement and instructional efficacy without direct feedback. This paper presents a novel approach to determine the concentration level of a listener while watching video tutorials using Electroencephalography (EEG) brain wave signals, and consequently, to estimate the performance of the trainer. The methodology involves a multi-stage process: initially, EEG brain wave datasets undergo rigorous data preprocessing to ensure signal quality. Subsequently, relevant features are extracted using the Common Spatial Pattern (CSP) algorithm, optimizing discriminative information. Finally, the K-Nearest Neighbors (KNN) algorithm is employed to accurately classify the listener's concentration level and infer the efficiency of the tutorial video. Experimental results demonstrate a robust performance evaluation, achieving an accuracy of approximately 93%. This research primarily aims to provide an objective method for predicting learning concentration from EEG brain waves, offering valuable insights for enhancing digital educational content.