Motor Imagery EEG Classification using Time Series features Extraction based on Scalable Hypothesis

Abstract

The development of brain-computer interfaces (BCIs) as a tool for communication and control among people with disabilities has been encouraging. The creation of successful BCI systems depends on the classification accuracy of EEG signals. The FRESH algorithm is used in this study to propose a method for classifying EEG signals based on time series features. Conventional signal processing methods may not capture the necessary information for classification as EEG signals are inherently time varying. On the other hand, time series features capture the temporal properties of the EEG signal and can be used to extract discriminative data for classification. We applied different classification models such as KNN, SVM, and LDA to the extracted features obtained using the FRESH algorithm, to classify the EEG signals of healthy subjects. Our proposed method achieved high classification accuracies of 94% to 97% for 5 types of upper-limb complex pre-movements, performed in 50 trials. This represents a significant improvement compared to previous methods, indicating the effectiveness of our approach. The proposed method, incorporating the FRESH algorithm, has significant implications for the development of BCI systems as accurate classification of EEG signals is essential for efficient communication and control. This study highlights the importance of time series features in EEG signal classification for BCI systems. By capturing the temporal characteristics of the EEG signal using the FRESH algorithm, our proposed method provides more discriminative information for classification than traditional signal processing techniques. The high classification accuracies achieved by our method demonstrate the effectiveness of the FRESH algorithm in improving the accuracy of BCI systems. Further research can investigate the potential of time series features extracted using the FRESH algorithm in other areas of BCI development, including the control of robotic limbs and the detection of neurological disorders.