Optimal number of clusters identified using the GMM clustering algorithm.

The proportion of cluster agreement represents a measure of consensus or stability amongst the number of clusters across the 1000 repeated runs. The total number of clusters was identified by fitting the GMM 1000 times with each run initialised by a different seed to obtain a unique set of clustering solutions for that seed. The Adjusted Rand Index (ARI) assesses the stability of the clustering solution by measuring the proportion of neuronal pairs consistently classified in both clustering solutions to the total number of neuronal pairs across 100 other solutions.