Correlation (Pearson correlation coefficient) of the synaptic organizations for the vnoNa (light blue) and vNa architectures (dark red) with the somatic distances (x axis) and dendritic clustering of synapses (y axis) for the three cluster arrangements. Each architecture is characterized by its more consistent synaptic dissimilarity matrix (cf. C): synaptic weight based for vnoNa, and subunit assignment based for vNa. Clustering-based correlations (y axis) are computed only for clustered synapses (top left block of corresponding dissimilarity matrices, see also B). Boxplots in (C) show median (horizontal line), 25th and 75th percentiles (box), 1.5 interquartile ranges (whiskers), and outliers (circles) across five separate training datasets (C, purple) or 10 training dataset pairs (C, blue and red). Crosshairs in (D) show mean 1 SD along each axis across five training datasets. See alsoTable S1.
