Linear Mixed-effects shape model to explore Huntington's Disease Data

From NAMIC
Jump to: navigation, search
Home < Linear Mixed-effects shape model to explore Huntington's Disease Data

Key Investigators

  • UIowa: Hans Johnson, Dave Welch
  • Utah: Manasi Datar, Josh Cates, Ross Whitaker

Objective

Make the linear mixed-effects shape model accessible for further exploration of Huntington's Disease Data

Approach, Plan

Meet with UIowa team to:

  • give an overview of results from the linear mixed-effects shape model
  • explain the ShapeWorks command line tools to optimize and analyze correspondences
  • discuss next steps toward making these tools accessible to UIowa team

Progress

preliminary results look promising

ITKv4 module for mixed-effects model available on GitHub

  • tested on Windows and Linux
  • application to run mixed-effects model on single component shapes (90%)

To Do

  • get working code for shape complexes (multiple structures per shape)
  • example parameter file for preprocessing shapes, initializing correspondences and applying mixed-effects model
  • extracting labels from input labelmap (various components in the shape) - clarify and possibly add a tool to the preprocessing stage


References

M Datar, P Muralidharan, A Kumar, S Gouttard, J Piven, G Gerig, RT Whitaker, PT Fletcher, Mixed-Effects Shape Models for Estimating Longitudinal Changes in Anatomy, STIA 2012