During summer 2022, after my first year in college,
I worked in the
Yale Computer Systems Lab.
A brain-computer interface (BCI) was thrust into my hands,
and I was esentially told to mess around with it. I made spellers
(programs where you can type characters with just your brain and attention),
and used my machine learning knowledge to try
a new technique with
CNNs
Convolutional Neural Networks
and
LSTMs
Long Short-Term Memory
to make
EEG
electroencephalography
data coherent,
which wasn't super effective, but it was an
interesting approach and a very cool way to spend a summer.
Given the short timeframe, my lack of neuroscientific abiltiy,
the disorganization of the project, and the nascency of the technology,
I wasn't really able to do much novel or useful work with the BCI.
I did discover that my brain is not very easily readable with
EEG — I apparently suffer from "BCI Illiteracy",
as do many people —
and that BCI technology is not very effective if it's not
surgically implanted into your brain.