Brain Computer Interfaces(BCIs) have emerged as a transformative technology, enabling communication between the brain and external computing devices. Through the translation of brain activity to control signals, BCIs offer significant potential in domains across neurorehabilitation, assistive devices and human-computer interactions. Among the various BCI paradigms, Motor Imagery (MI)-based BCIs leverage electroencephalography (EEG) signals to detect imagined movements, eliminating the requirement for physical motion.
The accurate classification of MI EEG signals continues to be a crucial challenge in BCI research, as it directly influences the system’s responsiveness and usability. Traditional MI classification research so far focuses on distinguishing between broad movement categories, such as left-hand and right-hand imagery. However, recent studies suggest that targeting single-joint movements (e.g., wrist, knee, elbow) might lead to more precise and natural control mechanisms.
Accordingly, this study focuses on a single-joint, specifically the wrist, based MI task classification. The introduction of a novel model capable of distinguishing distinct wrist movements and investigating model performance across distinct joint movements, is the main aim of this study. Thus, contributing to improved precision in BCI applications.
Classifying EEG signals for single-joint movements (wrist pronation / supination) is challenging due to low signal-to-noise ratios and overlapping spatial features. Existing MI-based BCI systems primarily focus on broad limb movements, limiting precision and hindering the development of applications requiring fine motor control, such as prosthetics and neurorehabilitation.
- Develop a robust classification model for wrist pronation/supination MI tasks using EEG data.
- Evaluate different preprocessing, feature extraction and machine learning approaches.
- Address inter-subject variability in single-joint MI-based EEG classification.
The primary objective is to classify MI tasks for single joint movements, focusing on the wrist, with high accuracy to enhance the usability of Brain-Computer Interface(BCI) applications in neurorehabilitation and assistive technology.
The higher overview of steps involved in classifying MI tasks is shown below.
EEG signals of 25 healthy subjects performing wrist movements, pronation and supination, are obtained through the publicly available dataset on GigaDB(https://gigadb.org/dataset/100788).
- 60 Hz notch filter will be applied to raw EEG data to reduce the effect of external electrical noises such as DC noise of power supply and the scan rate of the monitor.
- Band-pass filter will be applied to remove both low-frequency and high-frequency components that are irrelevant to the Motor Imagery (MI) task.
- Independent Component Analysis(ICA) technique will be applied to remove artifacts such as eye-blinking effects, muscle movement, and other non-neural interferences.
The model’s performance will be assessed using standard classification metrics, including accuracy, precision, recall, F1-score, and Matthews Correlation Coefficient (MCC). Additionally, Area Under the ROC Curve (AUC-ROC) will provide insights into the model’s discriminative capabilities. Statistical significance tests, such as paired t-tests, will be conducted to validate performance improvements.
- Better Neurorehabilitation outcomes: Enables more targeted therapy for patients with motor impairments.
- Enhanced BCI control precision: Improved usability for prosthetic devices and assistive technology.
- Contribution to future research in fine motor control using single-joint EEG-based studies.