Some of you may remember my last project with EEG-brain computer interfaces. It was fun, though the project was limited by budgetary constraints (which was one of its selling points - a fun EEG project for under $30). As such, the data was pretty noisy, giving me limited control.
Luckily, I now have the opportunity to explore further. Professor Abhishek’s lab has all of these OpenBCI boards and electrodes (which would cost thousands typically) and he was nice enough to give me essentially free reign on playing around with them! I’m pretty excited to explore, and I’ll document my journey here.
Firstly, I wanted to just get one of these headsets assembled and ready to use. The issue is that the state of the art requires you to put this gel on the electrodes to decrease impedance, which is sticky and gets everywhere and kind of gross. I didn’t want to do that, so I decided to try the headband set, which seemed the most convenient (other options: the headcap, and the Cortex Mark IV which I 3D printed at the CEID).
So on a Sunday, I got to work downloading the OpenBCI GUI and learning how to assemble the headset with the Cyton board (the lab also has the Daisy module! What luxury!). I assembled the headset and hooked it up to my computer. After playing around with the OpenBCI GUI a little, I wrote a small proof-of-concept script. I gathered my brain data from concentrating (doing mental math) and relaxing (thinking of landscapes) and trained a classifier to distinguish between the data, then I made a little game where concentrating made the player go up and relaxing made it go down. You can also listen to me explain it here:
Of course, since I only trained it on 20 5 second clips of each brain state, it had really low accuracy (65%), and the game is not the most exciting. There’s lots of ways to improve it.
Off-topic: I’ve met a few of the grad students at the lab and they’re super nice! There’s a gal who’s a masters student in computer architecture also in the Abhishek Bhattacharjee lab who’s doing research in the same room, and down the hall is a nice gal who’s a fourth-year PhD student in cloud security. It’s fun to see others around, though most of the time (or at least the times when I go in, which are usually weekends and afternoons) the rooms are empty.
TBC…