EEG Signal Spike Sorting Algorithm that uses Principle Component Analysis (PCA) and K-means machine learning algorithm for action ponentials. It uses Donoho's rule for the threshold standard deviation. I used a 5x STD threshold to detect spikes to avoid artifacts in windowing. After the spike detection from thresholding the spikes are overlapped in a 5ms window then sent into PCA. I used PC1 and PC2 for cluster in K-means which used 3 clusters for the detected spikes.
Threshold/Donoho's:
http://www.scholarpedia.org/article/Spike_sorting https://www.nature.com/articles/s41598-021-93088-w