Fig. 6: Comparison of various cell-typing methods for MOs neurons.

a, Three connectivity-based clusters for MOs5 neurons (top row) along with the distribution of their somas (bottom left) and the overall projection patterns of MOs5 neurons (bottom right)³⁷. b, Key morphological features of the three connectivity-based MOs5 subtypes. The number of involved axonal morphometry samples is as follows: MOs5_1, 143; MOs5_2, 128; MOs5_3, 190. Similarly, the number of involved dendritic morphometry samples is as follows: MOs5_1, 143; MOs5_2, 128; MOs5_3, 190. c, Two metrics, neuron-beta and correlation coefficient, between single neurons and neuron populations in motor cortex, specifically MOp, MOs2/3, MOs5 and MOs6a subtypes. The number of single neurons and mesoscale projection experiments is indicated in brackets on the x axis. In b and c, the elements of box plots are consistent with those in Fig. 4l. d, Correlation of dendritic and axonal morphological features for MOs5 connectivity subtypes, along with examples of the first MOs5 cluster. Note that the clustered neurons in a might not have dendrite reconstructions; however, in this dendro-axonal correlation analysis only neurons in a but also with full dendrites and axons are counted. e, Transcriptomic profile-based single-neuron clustering of FRP-MOs neurons (n = 34,331) and more specific FRP-MOs layer 5 neurons (n = 9879), compared with the clustering based on connectivity and morphology features of FRP-MOs/FRP-MOs layer 5 neurons. Region labels in top left: CGE, caudal ganglionic eminence; L2/3 IT, intratelencephalic (IT) layers 2 and 3; L4/5/6 IT Car3, IT, Car3⁺, layers 4–6; MGE, medial ganglionic eminence; NP/CT/L6b, near-projecting(NP)/corticothalamic (CT)/layer 6b; PT, pyramidal tract.