Enhance motor imagery EEG classification using DWT and chirplet transform

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Brain Computer Interfaces (BCIs) enable individuals with motor impairments to interact with their environment.
• Effective BCI operation depends on accurate classification of electroencephalography (EEG) signals, particularly those from Motor Imagery (MI).
• Continuous improvement in EEG signal classification methods is crucial for advancing BCI technology.

Purpose of the Study:

• To enhance the performance of BCI systems by improving EEG signal classification.
• To demonstrate the importance of time-related data in MI activity classification.
• To introduce and evaluate novel signal processing techniques for EEG data.

Main Methods:

• Implementation of Discrete Wavelet Transform (DWT) and Chriplet Transform for EEG signal processing.
• Utilizing Visual Studio Code Python for the implementation and testing of the proposed methods.
• Evaluation of the proposed methods on the CBCIC and BCI Competition IV datasets.

Main Results:

• The proposed method achieved 91% efficiency.
• An accuracy of 94.8% was obtained for the CBCIC dataset.
• An accuracy of 93.72% was achieved for the BCI Competition IV dataset with a response time of 1.03 seconds.

Conclusions:

• Discrete Wavelet Transform and Chriplet Transform significantly enhance EEG signal classification for BCIs.
• The proposed approach demonstrates the critical role of time-domain features in accurately classifying MI tasks.
• The method offers a dependable and efficient solution for improving BCI performance, particularly for individuals with movement limitations.