A Computational Framework for Ultrastructural Mapping of Neural Circuitry

• DOI: 10.1371/journal.pbio.1000074
• Publication Date: 2009-03-30T14:18:23Z
• Authors: James R Anderson; Bryan W Jones; Jia-Hui Yang; Marguerite V Shaw; Carl B Watt; Pavel Koshevoy; Joel Spaltenstein; Elizabeth Jurrus; Kannan UV; Ross T Whitaker; David Mastronarde; Tolga Tasdizen; Robert E Marc
• Keywords: Connectomics; Terabyte; Profiling (computer programming); Ground truth; Brain atlas

Circuitry mapping of metazoan neural systems is difficult because canonical regions (regions containing one or more copies all components) are large, regional borders uncertain, neuronal diversity high, and potential network topologies so numerous that only anatomical ground truth can resolve them. Complete a specific requires synaptic resolution, region coverage, robust classification. Though transmission electron microscopy (TEM) remains the optimal tool for mapping, process building large serial section TEM (ssTEM) image volumes rendered by need to precisely mosaic distorted tiles register mosaics. Moreover, most molecular class markers poorly compatible with imaging. Our objective was build complete framework ultrastructural circuitry mapping. This combines strong TEM-compliant small molecule profiling automated tile mosaicking, slice-to-slice registration, gigabyte-scale browsing volume annotation. Specifically we show how ultrathin datasets their resultant classification maps be embedded into ssTEM scripted acquisition tools (SerialEM), mosaicking registration (ir-tools), slice viewers (MosaicBuilder, Viking) used manage terabyte-scale volumes. These methods enable large-scale connectivity analyses new legacy data. In well-posed tasks (e.g., in retina), previously would require decades assembly now completed months. Perhaps importantly, fusion profiling, SerialEM, ir-tools assembly, data viewers/annotators also allow as prospective discovery nonneural practical screening methodology neurogenetics. Finally, this provides mechanism parallelization imaging, analysis across an international user base, enhancing productivity cohort microscopists.