Gaussian Mixture Models : spikes curves to classify

Hi Matlab Community I would like to use GMM to clster spikes signal issues from electrophysiological experiment. I'm using a hierarchical method coupled with a k-means. It works correctly, but I would like to compare wiht other methods. My spikes after being detected are stored in a matrix (sig) with n lines (n points for each spikes) and m columns (m spikes curves)

I tried some code for GMM, but I dont understand what i'm exactly doing. gm = gmdistribution.fit(sig,K); idx = cluster(gm,sig);

I would like to understand the method principle, how to use it and what do results mean. What is gm for example ?

Thanks a lot. Vincent.