The Development of an Electroencephalography (EEG)-Derived, 3D-Printed Brain-Computer Interface (BCI) NeuroProsthesis Utilizing Machine Learning for Chronic Impairment
23 Pages Posted: 21 Nov 2018
Date Written: February 9th, 2018
In the United States alone, there are 20 million people suffering from limb loss and 50 million with paralysis. Artificial prostheses have been manufactured to restore aesthetic appearance and locomotory function of impaired extremities. However, most prosthetic appendages are passive, incapable of restoring complete function. Hence, Brain-Computer Interfaces (BCIs) utilizing electroencephalograms (EEGs) have been explored for their uninterrupted communication between the brain and external device-output. Nonetheless, a fully functional BCI product has not been available for biomedical application thus far. Our goal is to develop a supervised machine learning algorithm, written in Python, to predict patterns in EEG signals customized for the impaired. Emotiv EPOC, a cost-effective BCI, will be used to detect motor imagery. The output signals will be integrated with Raspberry Pi 3 to control a 3D-printed transradial prosthetic. This novel, cost-effective, and facile approach will bring a comprehensive and non-invasive resolution to global impairment worldwide.
Keywords: Neurology, Biomedical Engineering, Computer Science, Machine Learning, AI, Neuroscience
JEL Classification: I12
Suggested Citation: Suggested Citation