BRAINS Automated Pipeline

Objective:

Develop tools to eliminate the need for manual intervention for the analysis of brain morphology using MR images. We currently have a semiautomated pipeline that includes the following steps

• AC-PC Alignment of T1 Volume
• Co-registration of T2 weighted images to AC-PC Aligned T1
• Bias Field Correction
• Tissue Classification
• Talairach Atlas Segmentation
• Neural Network Segmentation
• Surface Generation
• Measurements of Volumetric and Surface Data

Proposed Pipeline:

• Alignment of T2 to raw T1 scan - Use BRAINSFit
• Brain Extraction - BRAINSMush
• Tissue Classification - KMeans + Bayesian
• Bias Field Correction - MRBiasCorrector
• AC-PC Detection T1 - ART
• Resample T1 Image - TBD
• AC-PC Detection T2 - ART
• Alignment of T2 to AC-PC aligned T1 - BrainsFit Initialized with ART results
• Brain Extraction - BRAINSMush + BRAINSCut
• Tissue Classification - TBD
• Talairach Brain Labeling - TBD
• Neural Network Labeling - BRAINSCut
• Measurements - TBD
• Surface Generation - BrainSurf
• Surface Labeling - TBD

Progress:

• A beta version of the automated workup exists within BRAINS2.
• BRAINS3 has been created using a TCL command line to connect components from BRAINS2, ITK, and VTK
  • Nearly all code is in ITK with a couple of exceptions: ROIs and Talairach Parameters
  • Interfaces have been created to seamlessly handle BRAINS and ITK images using ITK filters
  • Requires the WrapITK Interface
• Added a few new classes required for wrapping and utilization via the command line
• Added improved interpolation schemes including Sinc and B-Spline
  • Will become the standard for images used for tissue classification

To Do:

• Refine data types used in WrapITK to decrease the memory footprint
• Improve TCL scripts to verify that memory is being freed appropriately
• Look into loading only certain aspects of ITK when they are needed
• Implement Tissue classification in ITK
• Handle Talairach parameters using VTK structured grids
• Use new ROI format, used by BRAINSTracer, for working with ROIs
• Integrate some features through command line with BRAINSTracer

Key Investigators:

• University of Iowa: Vincent Magnotta, Hans Johnson, Greg Harris, Wen Li, Steven Dunn, Nancy Andreasen
• The MIND Institute: Jeremy Bockholt

Links:

• University of Iowa Department of Psychiatry
• University of Iowa MRI Center
• The MIND Institute

Papers:

1. Magnotta VA, Harris G, Andreasen NC, O'Leary DS, Yuh WT, Heckel D. Structural MR image processing using the BRAINS2 toolbox. Comput Med Imaging Graph. 26(4):251-64, 2002.
2. Harris G, Andreasen NC, Cizadlo T, Bailey JM, Bockholt HJ, Magnotta VA, Arndt S. Improving tissue classification in MRI: a three-dimensional multispectral discriminant analysis method with automated training class selection. J Comput Assist Tomogr. 23(1):144-54, 1999.
3. Magnotta VA, Andreasen NC, Schultz SK, Harris G, Cizadlo T, Heckel D, Nopoulos P, Flaum M. Quantitative in vivo measurement of gyrification in the human brain: changes associated with aging. Cereb Cortex. 9(2):151-60, 1999.
4. Powell S, Magnotta VA, Johnson H, Jammalamadaka VK, Pierson R, Andreasen NC. Registration and machine learning-based automated segmentation of subcortical and cerebellar brain structures. Neuroimage. 39(1):238-47, 2008.
5. Magnotta VA, Heckel D, Andreasen NC, Cizadlo T, Corson PW, Ehrhardt JC, Yuh WT. Measurement of brain structures with artificial neural networks: two- and three-dimensional applications. Radiology. 211(3):781-90, 1999.