Figure 7. Morphometric analysis of electron microscopic images reveals three types of neurons.

(A) Based on results of the multivariate statistical tool principal component analysis (PCA; see Supplementary Figure 11) we built a partial least square discriminant analysis (PLS-DA) model to find classify the groups found by PCA. PLS-DA identified three principal components which together accounted for ~80% of the variability of data. (B) The PLS-DA score plot reveals the presence of well-separated three groups of neurons. (C) The loading plot shows the contribution of the variables to the separation of the three groups. (D) The variable importance on projection plot shows that all variables contributed significantly to the separation of the groups. The contribution of TRPV1-CB1 average distance (TR-CB dist), TRPV1 density (TR dens) and CB1 receptor density (CB dens) is highly significant. TR-CB crit: number of TRPV1-CB receptor critical distances; TR-TR crit: number of TRPV1-TRPV1 critical distances; TR-TR dist: TRPV1-TRPV1 average distance; CB-TR dist: CB1 receptor-TRPV1 average distance. (E,F) Bar charts showing average values of morphometric data of neurons belonging to the three groups shown in (B). Statistical analysis (ANOVA followed by Fischer’s post-hoc test) shows that neurons is Group 1 (yellow) are different from neurons both in Group 2 (red) and 3 (purple) in CB1 receptor density (CB dens), TRPV1 – CB1 receptor average distances (TR-CB dist) and the number of TRPV1 – CB1 receptor critical distances (TR-CB crit). Neurons in Group 2 (red) are different from neurons in Group 3 (purple) mainly in the distribution of TRPV1 (TRPV1 density (TR dens); TRPV1-TRPV1 average distance (TR-TR dist) and the number of TRPV-TRPV1 critical distance (TR-TR crit)). CB1 receptor-TRPV1 average distances (CB-TR dist) are also different between neurons in Group 2 and 3. distribution. The number of CB1 receptor TRPV1 critical distances (CB-TR crit) is not different between the 3 groups.