Comparison of preprocessing techniques for effective cognitive analysis using electroencephalography (EEG)

Electroencephalography (EEG) serves as a significant technique to analyze the cognition. The purpose of this study is to compare EEG preprocessing techniques and identify the most suitable pipeline for reliable analysis of cognitive processing in sports and non-sports group. The study attempts the preprocessing of the EEG signal in active states from both groups using two methods. • The first method (Method A) called TRIM-EEG (Tri State Method) is an amalgamation of techniques that include filtering using notch, ocular artifact removal using Independent Component Analysis (ICA), and frequency selection using bandpass filtering. • The second method (Method B) uses the Variational Decomposition Mode (VMD) technique, which decomposes the signal and its parameters were selected using Shuffled Frog Leaping Algorithm (SfLA) with spectral kurtosis. Error metrics were used to analyze the performance and TRIM-EEG had better values. The resultant preprocessed EEG signal yielded by the TRIM-EEG method was subjected to a broad feature analysis to differentiate between sports and non-sports groups using five different classifiers. Support Vector Machine (SVM) and Random Forest (RF) yielded higher accuracy 93.33 %. The performance of classifiers and the comparative analysis of EEG preprocessing demonstrates that TRIM-EEG method can be used as an effective technique in cognitive analysis studies.