UCLA Students Try Brain Hacking, Part 1
Our beautiful OpenBCI equipment finally arrived in the mail, so over the past couple weeks, we had the chance to put together our first brain computer interface! Fortunately, with the help of several online guides and video tutorials, we were able to put the system together without too much difficulty.
While our future goals include building our own circuit board, we decided to start our adventures in BCIs by working with OpenBCI’s EEG headband kit (left, picture via OpenBCI) and Ganglion board (down to the right, picture via OpenBCI.)
The kit came with the black headband piece, 2 ear clip electrodes, and several wires that connected to either dry electrodes that snapped in place onto the band or dry comb electrodes. The snap electrodes are flat and work well in obtaining signals from direct contact areas like the forehead whereas the comb electrodes have several protrusions that allow them to push past hair and get closer to the scalp to get clearer signals.
Obviously, electrodes placed on different parts of the head will receive different signals since certain brain areas correspond with specific states of mind. Fortunately, we were able to figure out where to place the electrodes thanks to the premade holes in the headband and the standardized 10-20 international system of electrode placement.
Figure courtesy https://www.researchgate.net/figure/International-10-20-system-of-electrode-placement-7_fig2_256491418
Adapted with color coding
The Nz bubble at the top of the diagram represents a person’s nasion, or nose. The Iz bubble at the bottom of the diagram represents the participant’s inion (the small bump on the back of your head where the skull attaches to the neck), and the A1/A2 bubbles are on a person’s ears.
The bubbles that are highlighted in blue show the areas that the snap electrodes (left, picture via OpenBCI) can be placed while the bubbles highlighted in green shows where the comb electrodes (right, picture via OpenBCI) can be placed on the headband. In the future, if we choose to use a cap instead of a headband in or if we make alterations to the headband, we would be able to place electrodes in the other areas depicted on the diagram.
Next, in order to see the data that the electrodes collect, we had to first connect each electrode to OpenBCI’s Ganglion Board, which has 4 available channels.
We attached the two earclip electrodes, positions A1/2 on the diagram, to the D_G (driven ground) and REF (reference) pins to act as a base that the other electrodes’ voltages can be compared to. Earlobes don’t really have nerve endings or electrical activity, making them a suitable point of comparison.
For the four channels that would receive the input from the electrodes on the headband, it was important that we push the switches on each channel to their down position rather than keep them in their default up position. This way, the data from the four channels would be connected to the data from the REF pin so that we can see how the brain’s electrical activity is changing altogether from all four points rather than analyzing each channel’s data stream individually.
All in all, I think the part we probably had the most trouble with was simply getting OpenBCI’s GUI, the software that records and streams the data, to work. For some reason, the Ganglion board wasn’t immediately connecting over Bluetooth to our laptops. It probably also didn’t help that we were all very interested in the tech so we would barely let someone fiddle around with it for a minute or two before someone else declared they wanted a turn to look at it!
After refreshing the application a few times, ensuring all parts were properly connected and that bluetooth capabilities were, in fact, turned on, we were finally able to connect.
For our first test, we placed the electrodes somewhat randomly across the headband at the Fp1 and Af8 positions for the snap electrodes and the T7 and T8 positions for the spiky electrodes just to see what would happen. Sasha was the first to try on the headband, and after a few adjustments...
Tada - brain waves!!
At that point, we just played around with little to no understanding of the different applications on the GUI. With all the noise in our environment and how much everything was moving, the data we were looking at was filled with so many random jumps that it was essentially useless for any sort of application, but that didn’t dampen our excitement!
Our next challenge, acquiring clean data!