Model-based clustering representation and interpretation.

<p><b>(A)</b> Graph showing the resulting BIC values depending on the number of clusters. Results are shown in a range from two to ten clusters. The model that achieved the highest BIC value had six clusters. <b>(B)</b> Representative examples of dendritic spines with a <i>p</i>* = 1 (highest membership probability) from the six different clusters. <b>(C)</b> 2D projection of the 6D probability distributions representing the membership probability of each spine to each cluster according to classical multidimensional scaling. Spines were colored combining cluster colors according to their probabilities of membership to each cluster. <b>(D)</b> The absolute value of the logarithm of the total variance for each cluster, i.e., |log₁₀det(<b>Σ</b>_<i>i</i>)|, where <b>Σ</b>_<i>i</i> is the variance-covariance matrix of cluster <i>i</i>. It is a value that summarizes the heterogeneity of morphologies within a cluster.