Classification of Electroencephalogram (EEG) Signals Using Linear Discriminant Analysis

---

Abstract

The cognitive behavior of brain can be analyzed using EEG signals. Nowadays, EEG signals are widely used to study brain related activities and various disorders. Artificial intelligence tools are widely used to analyze these signals which are captured using network of electrodes fixed on the human scalp and transferred on mobile device for further analysis. In the present study, EEG signal analysis is performed on the online data set containing motor imagery information. Various parameters of EEG signals are pre-processed before analyzing. EEG signal pre-processing is done using Independent Component Analysis (ICA). For filtering purpose, FIR filter is used, as it shows linear phase response. Eigenvalues as features are selected. Morlet wavelet transform is used to perform time-frequency analysis and to compute average power present in signals. Principal Component Analysis (PCA) is done for dimensionality reduction. Further, EEG signals are classified using Linear Discriminant Analysis (LDA).

---

Keywords

EEG signals; Pre-processing; ICA; Artifacts; Down sampling; Filtering; FIR; ERP; Feature extraction; Eigenvalues; TFR; Morlet wavelet; Average Power; PCA; Classification by LDA