Abstract
We are now in the era of millimeter-scale electron microscopy (EM) volumes collected at nanometer resolution (Shapson-Coe et al., 2021; Consortium et al., 2021). Dense reconstruction of cellular compartments in these EM volumes has been enabled by recent advances in Machine Learning (ML) (Lee et al., 2017; Wu et al., 2021; Lu et al., 2021; Macrina et al., 2021). Automated segmentation methods can now yield exceptionally accurate reconstructions of cells, but despite this accuracy, laborious post-hoc proofreading is still required to generate large connectomes free of merge and split errors. The elaborate 3-D meshes of neurons produced by these segmentations contain detailed morphological information, from the diameter, shape, and branching patterns of axons and dendrites, down to the fine-scale structure of dendritic spines. However, extracting information about these features can require substantial effort to piece together existing tools into custom workflows. Building on existing open-source software for mesh manipulation, here we present “NEURD”, a software package that decomposes each meshed neuron into a compact and extensively-annotated graph representation. With these feature-rich graphs, we implement workflows to automate a variety of tasks that would otherwise require extensive manual effort, such as state of the art automated post-hoc proofreading of merge errors, cell classification, spine detection, axon-dendritic proximities, and computation of other features. These features enable many downstream analyses of neural morphology and connectivity, making these new massive and complex datasets more accessible to neuroscience researchers focused on a variety of scientific questions.
Competing Interest Statement
XP is a co founder of Upload AI, LLC, a company in which he has financial interests. AST is co founder of Vathes Inc., and UploadAI LLC companies in which he has financial interests. JR is co founder of Vathes Inc., and UploadAI LLC companies in which he has financial interests.
Footnotes
↵* first author
Appendix tables added for cell type glossary, figure data size N specifics and NEURD feature documentation. Figure 1 updated to provide an overview of the paper. Figure 2 revised to include entire processing pipeline, and the feature set table was moved into the appendix table 2. Figure 28 was updates with reciprocal connection rates included and excitatory and inhibitory only subgraph experiment results for both the MICroNS and H01 dataset. Figures 29, 30 and 31 were added for clarity and to illustrate example neurons and their skeletonization. The main text and figure captions were updated for clarity and to include more specific evaluation metrics for certain statements made (for example, the test set accuracy for cell type classifcation is now included in the text and not just in the supplemental figure 19. More documentation and example ipynb notebooks for using NEURD were added to the GitHub repository, and then referenced in the main text.