[Task Classifcation of Right-hand and Foot Motion Imagery Based on Wavelet Packet Transform]

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Brain-computer interfaces (BCIs) offer communication solutions for individuals with lost motor function.
• Electroencephalography (EEG) signal recognition is crucial for BCI development.
• Event-related desynchronization (ERD) analysis aids in understanding neural activity during motor tasks.

Purpose of the Study:

• To classify right-hand and foot motion imagery tasks using EEG signals.
• To evaluate the effectiveness of Wavelet Packet Transform (WPT) and Transfer Learning (TL) for BCI.
• To assess the potential for online BCI applications.

Main Methods:

• EEG signals were decomposed using Wavelet Packet Transform (WPT).
• Relevant channels and frequency bands related to ERD were analyzed.
• Wavelet packet energy was calculated from selected nodes.
• Transfer Learning (TL) was applied to classify BCI Competition III data (IVa).

Main Results:

• The proposed method achieved ideal classification results on the BCI Competition III dataset.
• The combination of WPT and TL demonstrated high accuracy in distinguishing motor imagery tasks.
• The approach proved to be simple and effective for EEG-based BCI.

Conclusions:

• The developed WPT and TL method is a viable and effective technique for EEG-based BCI.
• This approach holds significant value for the online application of brain-computer interfaces.
• The findings support the advancement of communication aids for individuals with motor impairments.