Extended Data Fig. 20: related to Fig. 5. Single neuron reconstruction data.
a, Digital images (left panel) show two coronal planes (2401 and 2501) through the viral injection site in one representative case (195409). The sections were counterstained with propidium iodide-stained (PI) cellular nuclei to reveal cytoarchitectonic background and facilitate identification of soma locations of labeled neurons. Four L2/3 IT neurons (#03, 06, 23, 24), one L5 ET (#01) and one L6 (#22) neurons were selected for reconstruction. Scale bars, 500 µm. b, c, Analyses of projection target patterns for MOp neurons from the MouseLight dataset and schematic of cell type specific networks. b, Schematic depiction of the major targets contacted by three MOp cells (identified by their MouseLight name) and the pairwise comparisons to quantify the differences in regions targeted (Δ values to the right). c, Histogram of pairwise differences in regions targeted by MOp neurons (“real”) compared to those with randomized targeted regions while normalizing the number of regions invaded by each neuron and the number of neurons invading each region (“shuffled”). The real distribution is broader than the shuffled distribution (CVshuffled=0.431, half-height-widthshuffled=25; CVreal=0.479, half-height-widthreal=31; half-height widths pairs of horizontal arrows) and the statistical difference is highly significant (1-tailed Levene’s test of variances: p < 10−25). The left tail reflects similarity between neurons likely in the same projection class; the right to differences between neurons from different classes. d, Top views of the CCF show the brain-wide reconstructions rendered in 3D from example Cre line tracer experiments (colored by layer and projection class key) or individual single cells (red) assigned to each cluster.
