The entire setup
What it does
EEG monitoring systems use electrodes in order to read magnetic signals created by the brain. These are caused when electricity is moved around the brain. Normally, healthy people's brains send signals down their spinal cords and through their nerves in order to control limbs, but in the case of quadra and paraplegics, the spinal cord has been severed, preventing electrical travel, and limb movement as a result. Through our hack, we bypass the nervous system and return control to patients allowing them to freely move their limbs with their minds.
How we built it
We used an EEG harness to create a ML model that we compared input to in order to determine whether or not to send a charge to an electrode attached to a person's ulnar nerve.
Challenges we ran into
Debugging low accuracy machine learning models was quite difficult.
Accomplishments that we're proud of
We created a highly accurate machine learning data aggregation model and created a way to enhance the lives of quadriplegics.
What we learned
We learned that creating machine learning models in noisy real world environments is a lot of work. This was also my (Alex) first time utilizing ML. Cognitive associations and biases.
What's next for Axon
In the future, former quadriplegics will go through a new kind of physical therapy where they train their minds along with their bodies. We foresee a radical medical procedure that utilizes ECoG (EEG nodes placed directly on the brain) or a neural lace to get more accurate and reliable readings. We also imagine replacing exterior electrodes with Pacemaker-like devices embedded directly onto peripheral nerves thus making Axon completely internal and providing a more precise range of motions.