Artifact removal and motor imagery classification in EEG using advanced algorithms and modified DNN

Area of Science:

• Biomedical Engineering
• Neuroscience
• Machine Learning

Background:

• Electroencephalography (EEG) based Brain-Computer Interfaces (BCIs) are challenged by artifacts and complex motor imagery tasks.
• Accurate EEG artifact removal and robust motor imagery classification are crucial for reliable BCI performance.

Purpose of the Study:

• To develop and evaluate an advanced approach for EEG artifact removal and four-class motor imagery classification.
• To enhance the accuracy and reliability of BCIs using novel signal processing and machine learning techniques.

Main Methods:

• Introduced Four Class Iterative Filtering (FCIF) for ocular artifact removal using iterative filtering and filter banks.
• Developed the FC-FBCSP algorithm, extending Filter Bank Common Spatial Pattern for four-class motor imagery.
• Employed a Modified Deep Neural Network (DNN) classifier to enhance feature discrimination.

Main Results:

• FCIF demonstrated effective mitigation of EEG artifacts, improving data quality.
• The FC-FBCSP algorithm combined with the Modified DNN classifier achieved high classification accuracy.
• The proposed method significantly outperformed baseline approaches on BCI Competition IV datasets.

Conclusions:

• The integrated approach of FCIF and FC-FBCSP with Modified DNN offers a superior solution for EEG artifact removal and motor imagery classification.
• This advancement holds potential for improving the performance and usability of BCI systems.