Fig. 2. Assessing quality of automated reconstructions.

a Automated and manual traces of example test neurons (left, inhibitory neurons; right, excitatory neurons—apical and basal dendrites are assigned for excitatory cells.) b Neuron reconstruction accuracy. Precision, recall, and f1-score values are calculated by comparing automated and manual trace nodes within a given distance (2, 5, and 10 μm). Box plots (n = 340 cells) are shown for axonal, dendritic and neurite (combined axonal and dendritic) nodes. The box extends from the first quartile (Q1) to the third quartile (Q3) of the data, with a line at the median. The whiskers extend from the box to the farthest data point lying within 1.5x the inter-quartile range (IQR) from the box. Mean  ± s.d. of the values over 340 cells are provided in Table S3. Scatter plots are shown in Fig, S17. c Generation of 2D axonal and dendritic ADRs. Scale bar, 100 μm. d Pearson’s r values (left) and average root-mean-squared error (right) between automatically vs. manually generated features (left and Fig. S16) and ADRs for each t-type (right and Fig. S18). ADRs are normalized to have unit norm. Source data are provided as a Source Data file.