Could a thought become a reliable interface? This proof of concept uses EEG signals to control a virtual light switch and thermostat, making an abstract neurotechnology question visible and testable.
Users can create an account, connect an EEG device, and control a digitally animated light switch and thermostat with their brain's neuroelectrical signals.
MVP / proof-of-concept journal update — March 13, 2020.
Milecia recorded our EEG data with her four-electrode OpenBCI device: 100 one-second bursts. Ryan used that data to train a predictive model with 97% accuracy. The back-end Python API returns a number from 0 to 5, each corresponding to one of six commands the user may be thinking: no, yes, up, down, left, or right.
Ellen and Roenz built the React JS interface and the back-end database connectivity for securely storing user data. When someone clicks the Record EEG button—the pink brain with rotating blue-and-white circles pictured below—the front end calls the Python API and runs this sequence:
The front end maps that returned command to a specific response. It is a little like recording a voice command for Siri or Google Assistant, except the one-second recording captures neuroelectrical signals instead of speech.
The next step is continuous, real-time prediction: filter out unrelated thoughts, recognize the six specific commands—no, yes, up, down, left, and right—and expand the available actions.
For a proof of concept, those six commands are enough to show the interaction end to end.
