Skip to content
This repository was archived by the owner on Nov 15, 2019. It is now read-only.

m2g - reliable and scalable connectome estimation

Latest

Choose a tag to compare

@gkiar gkiar released this 25 Nov 16:43
· 12 commits to master since this release

Overview

This release contains the newest version of m2g, and has been shown to perform as a reliable, large-scale, and robust pipeline for structural MR connectome generation, with all previous features and functionality as well as some additions. Reliability remains paralleled to that of the previous release, though accessibility of the code as well as versatility for computing graphs with downstream inference in mind has been greatly increased. The documentation has been expanded upon to allow users to more easily navigate using and benefiting from m2g. First steps towards integration of downstream derivatives into a spatial database and lims system have been taken, though will only be flushed out in the next release.

Functionality

  • Integrated multi-atlas small graph generation
  • Integrated list-files
  • Improved performance on large clusters by resolving unintentional multithreading
  • Standalone evaluation of pipeline performance
  • Compliance of derivatives and pipeline output with defined data specification
  • Big graph generation integrated with compression
  • Data converters created for fibers into SWC skeleton format

Organization

  • Removed unnecessarily workflows/dev pipelines

Documentation

  • Rearranged tutorials and docs pages
  • Provided thorough explanation of how to install, run, and analyze pipeline (either all together or each independently)