This project explores various machine learning algorithms and their ability to forecast Local Field Potentials (LFPs). These ML techniques range from classical shallow models to modern deep learning networks. While it is intially designed to work on LFPs, they can be used beyond brainwaves.
Currently we do not have a strong dataset splitting apart a deterministic signal from the full extracellular potential recordings. It is believed an unsupervised learning method would potentially be able to extract information that can be used to isolate the deterministic signal. Specifically, we intend to try:
Variational AutoEncoders
K-Means Clustering of Wavelets
Linearly independent dimensional reduction of signals (PCA)
Currently we do not have a strong dataset splitting apart a deterministic signal from the full extracellular potential recordings. It is believed an unsupervised learning method would potentially be able to extract information that can be used to isolate the deterministic signal. Specifically, we intend to try: