A Collaborative Pilot Study with Algoma University for Word Detection Using EEG Auditory Evoked Potentials

---

Project Lifespan: 2016 - Present

The inconsistency of phoneme generation may have been too large with respect to inter subject variability, thus, in this pilot study I opted to use whole words instead. I tried to extend work done by Aushivan et.al, where they found a way to isolate signals with 99% accuracy using imagined words using images of EEG’s rather than the time varying multichannel signal. I used images generated by 3D EEG source localization techniques along with machine learning to perform the same task but in real-time.