DBP:Harvard:Collaboration:MITDTI

Back to Harvard DBP 1

Current Projects

• Diffusion measures along fiber tracts of the cingulum bundle
  • We are trying to obtain diffusion measures (FA, trace, etc.) along fiber tracts of the cingulum bundle, comparing schizophrenic patients with control subjects.
  • Participants:
    • MIT: CF Westin, Lauren O'Donnell, Raul San Jose
    • PNL: Marc Neithammer, Sylvain Bouix, Marek Kubicki, Mark Dreusicke
• Clustering of anatomically distinct fiber tracts
  • Lauren O'Donnell (LMI) has developed a method to cluster fiber tracts based on their location and the regions they connect
  • Participants:
    • MIT: Lauren O'Donnell, CF Westin
    • PNL: Marek Kubicki, Mark Dreusicke, Martha Shenton
• FA/trace measurements of the corpus callosum and anterior commissure
  • Using ROI labelmaps drawn on sagittally acquired DTI data (higher resolution in sagittal plane), we are calculating diffusion measures in the corpus callosum and anterior commissure of schizophrenic patients.
  • This project has involved the coregistration of sagittal and coronal tensor data.
  • Fiber tracking (and clustering) is currently being applied to these cases
  • Participants:
    • MIT: Lauren O'Donnell, CF Westin, Scott Hoge
    • PNL: Marek Kubicki, Mark Dreusicke

Progress

• Software Testing: Scalar measure outputs from the DTMRI module of Slicer (FA and trace) are being compared with standard protocol measurements.
• Fiber tracking using multiple ROIs: We are testing the use of multiple ROIs to identify fiber tracts of interest.
• Fiber clustering: A unique method has been developed to cluster fiber tracts based on their anatomical connectivity.
• High resolution DTI acquisition: We are experimenting with the fiber tracking of higher resolution DTI scans.
• Coregistration of sagittal/coronal DTI: To use ROI labelmaps drawn on sagittally acquired data, we have developed a method to coregister the tensors of sagittal and coronal scans, so that labelmaps can be applied to full-brain coronal DTI acquisitions.
• DICOM files: We have ensured that DICOM DTI images are able to read in the Slicer.
• Masking FA/trace maps: Because of file type incompatability, scalar volumes from the DTMRI module could not be masked with ROI labelmaps. The problem has since been solved.

Publications

• Lauren O'Donnell, Carl-Fredrik Westin. White Matter Tract Clustering and Correspondence in Populations. MICCAI 2005
• Lauren O'Donnell, Marek Kubicki, Martha E. Shenton, W. Eric L. Grimson, Carl-Fredrik Westin. A Method for Clustering White Matter Fiber Tracts. AJNR (In Press)

Links

• Fiber Clustering in Slicer