EEG-Based Brain-Computer Interface for Stroke Rehabilitation: Analyzing Motor Imagery and Neural Activation in a Public Dataset
Honors Capstone Project
1
Advisor(s)
Dr. Phillip R. Zoladz
Confirmation
1
Document Type
Paper
Location
ONU McIntosh Center; Wishing Well
Start Date
8-4-2025 4:00 PM
End Date
8-4-2025 4:15 PM
Abstract
This project investigates the use of a non-invasive brain-computer interface (BCI) to support motor function recovery following a stroke, with a focus on comparing relevant biomarkers of progress between BCI-assisted rehabilitation and standard stroke rehabilitation.
The research involves a proposal for implementing a BCI protocol in which stroke survivors utilize motor imagery tasks to control a virtual or robotic exoskeleton arm. To enhance the analytical component of this study, publicly available electroencephalography (EEG) datasets on stroke rehabilitation and BCI will be analyzed. This approach strengthens the foundation of the research proposal by effectively “pilot testing” the proposed methods with existing data.
EEG datasets are sourced from BNCI Horizon 2020 and PhysioNet, which provide extensive records of stroke patients performing motor tasks using an exoskeleton robotic arm. These datasets serve as benchmarks for comparing cortical activation patterns in stroke patients undergoing BCI-assisted rehabilitation versus those following standard rehabilitation protocols. EEG signals will be analyzed using Python-MNE, with preprocessing steps including artifact removal, signal filtering, and epoch segmentation. Feature extraction will focus on event-related desynchronization and event-related synchronization within the mu and beta frequency bands, as these are key indicators of motor-related brain activity.
Following signal processing, biomarkers will be correlated with existing clinical scores to assess how neurophysiological changes correspond with functional motor recovery. This analysis aims to provide insights into the effectiveness of BCI-based interventions in stroke rehabilitation compared to traditional rehabilitation approaches.
Recommended Citation
Al-Olimat, Aleece K., "EEG-Based Brain-Computer Interface for Stroke Rehabilitation: Analyzing Motor Imagery and Neural Activation in a Public Dataset" (2025). ONU Student Research Colloquium. 19.
https://digitalcommons.onu.edu/student_research_colloquium/2025/Papers/19
Restricted
Available to ONU community via local IP address and ONU login.
EEG-Based Brain-Computer Interface for Stroke Rehabilitation: Analyzing Motor Imagery and Neural Activation in a Public Dataset
ONU McIntosh Center; Wishing Well
This project investigates the use of a non-invasive brain-computer interface (BCI) to support motor function recovery following a stroke, with a focus on comparing relevant biomarkers of progress between BCI-assisted rehabilitation and standard stroke rehabilitation.
The research involves a proposal for implementing a BCI protocol in which stroke survivors utilize motor imagery tasks to control a virtual or robotic exoskeleton arm. To enhance the analytical component of this study, publicly available electroencephalography (EEG) datasets on stroke rehabilitation and BCI will be analyzed. This approach strengthens the foundation of the research proposal by effectively “pilot testing” the proposed methods with existing data.
EEG datasets are sourced from BNCI Horizon 2020 and PhysioNet, which provide extensive records of stroke patients performing motor tasks using an exoskeleton robotic arm. These datasets serve as benchmarks for comparing cortical activation patterns in stroke patients undergoing BCI-assisted rehabilitation versus those following standard rehabilitation protocols. EEG signals will be analyzed using Python-MNE, with preprocessing steps including artifact removal, signal filtering, and epoch segmentation. Feature extraction will focus on event-related desynchronization and event-related synchronization within the mu and beta frequency bands, as these are key indicators of motor-related brain activity.
Following signal processing, biomarkers will be correlated with existing clinical scores to assess how neurophysiological changes correspond with functional motor recovery. This analysis aims to provide insights into the effectiveness of BCI-based interventions in stroke rehabilitation compared to traditional rehabilitation approaches.