VERT (Virtual Environment Rehabilitation Tool) is a proof-of-concept implementation of our team’s vision for the future of brain injury rehab. VERT includes three main components.
Firstly, EEG data is streamed live to a C# script that uses Brainflow’s concentration model. After extensive experimentation, we settled on a linear discriminant analysis machine learning model fed EEG band power averages from the last 10 seconds. An OpenBCI Ganglion reads the EEG signals with 4 electrodes placed on the F7, Fp1, Fp2, and F8 locations (based on the 10-20 model).
The second component of VERT is the exercise library. We chose to implement a version of ‘The Tower of Hanoi’ puzzle as it is regularly used in studying cognition. The exercises are designed in Unity to be played on the Oculus Quest. The exercises use Oculus controllers and Ganglion EEG data for a combination of motor and cognitive control. In our Tower Of Hanoi implementation, concentration above a set threshold will cause the selected disc (determined by the user’s gaze) to hover. Once the disc is hovering, the user can use their hand position to move it left or right until concentration is lost. We induce concentration by focusing hard and imagining ourselves moving the disc.
Lastly, users can view their historical performance in the Evaluation Portal. Currently, the portal is a locally hosted node.js web page using the “SB Admin 2” bootstrap template. Future work would include making a published website where both customers and their healthcare professionals can evaluate rehabilitation progress.