Difference between revisions of "DBP2:Harvard:Brain Segmentation Roadmap"

From NAMIC Wiki
Jump to: navigation, search
Line 60: Line 60:
  
 
:* Algorythm was used to trace and analyze anterior limb of the internal capsule on 1.5T data. It generated reacher representation of frontal fiber projections, it also turned out to be more sensitive to group differences in white matter integrity that conventional deterministic tractography (insert poster here).   
 
:* Algorythm was used to trace and analyze anterior limb of the internal capsule on 1.5T data. It generated reacher representation of frontal fiber projections, it also turned out to be more sensitive to group differences in white matter integrity that conventional deterministic tractography (insert poster here).   
:* Algorythm was also used to trace smaller white matter fiber tracts, such as Cingulum, Fornix, Uncinate Fasciculsu, Arcuate Fasciculus on Santa Fee dataset (link to Santa Fee conference page)   
+
:* Algorythm was also used to trace smaller white matter fiber tracts, such as Cingulum, Fornix, Uncinate Fasciculsu, Arcuate Fasciculus on 3T "Santa Fee" dataset (link to Santa Fee conference page)   
  
 
; C - References
 
; C - References
Line 80: Line 80:
 
:** Whole brain segmentation, and automatic extraction of regions interconnected by Arcuate Fasciculus (Inferior frontal and Superior Temporal Gyri), as well as another ROI that would guide the tract ("waypoint" ROI). This step has been accomplished for the entire dataset of 23 schizophrenia subjects and 23 controls.  
 
:** Whole brain segmentation, and automatic extraction of regions interconnected by Arcuate Fasciculus (Inferior frontal and Superior Temporal Gyri), as well as another ROI that would guide the tract ("waypoint" ROI). This step has been accomplished for the entire dataset of 23 schizophrenia subjects and 23 controls.  
 
:** White matter segmentation, in order to prevent algorithm from traveling through the ventricles, where diffusivity is high. This has been done also for the entire dataset now.  
 
:** White matter segmentation, in order to prevent algorithm from traveling through the ventricles, where diffusivity is high. This has been done also for the entire dataset now.  
:** Linear registration of labelmaps to the DTI space. This step has been done for all subjects, however we were not satisfied with the results of registration, especially in the frontal areas, when using tools available through slicer. We have redone tis step, using nonlinear registration and demon's algorythm,New registration protocol, b-spline, has been developed and made available to us few weeks ago (May 2008), and results are much more promising. We are currently running the registration for all our subjects. This step should be completed within the next few weeks.   
+
:** Linear registration of labelmaps to the DTI space. This step has been done for all subjects, however we were not satisfied with the results of registration, especially in the frontal areas, when using tools available through slicer. We have redone this step, using nonlinear registration and demon's algorythm (slicer), and results are much more precise. This step has been also done now on all subjects. 
:** Applying stochastic tractography algorithm, to find the path connecting two ROIs. This has been also done for the subset of subjects (19 in total), however we will redo it after the new registration is completed.  
+
:** Seeding tracts. We have piloted in on ten cases using 5000 seeds per voxel, but since it is quite time consuming running it on even most powerfull computers in the lab, we have experimented with smaller number of seeds per voxel. We tested 1000 seeds, which gave virtually identical results. We are now applying this stem to all our cases. Currently, all steps take about 6-8 hours per case to complete. (should be done by the end of November).   
 +
Applying stochastic tractography algorithm, to find the path connecting two ROIs. This has been also done for the subset of subjects (19 in total), however we will redo it after the new registration is completed.  
 
:** Extracting path of interest, and calculating FA along the path for group comparison. Presentation of previous results for 7 schizophrenics and 12 control subjects, can be found here: [[Media:NAMIC_AHM_Arcuate.ppt|Progress Report Presentation]]. Once new registration is completed, and tracts extracted, we intend to use another NAMIC tool for tract parametrization (Mahnaz will make it available for us within few weeks), to look at the diffusion properties along the tracts, which will make group comparison much more precise.  
 
:** Extracting path of interest, and calculating FA along the path for group comparison. Presentation of previous results for 7 schizophrenics and 12 control subjects, can be found here: [[Media:NAMIC_AHM_Arcuate.ppt|Progress Report Presentation]]. Once new registration is completed, and tracts extracted, we intend to use another NAMIC tool for tract parametrization (Mahnaz will make it available for us within few weeks), to look at the diffusion properties along the tracts, which will make group comparison much more precise.  
 
; D - New developments (As of August 15th 2008) :
 
; D - New developments (As of August 15th 2008) :

Revision as of 19:16, 3 November 2008

Home < DBP2:Harvard:Brain Segmentation Roadmap

Back to NA-MIC Collaborations, Harvard DBP 2

Stochastic Tractography for VCFS

Roadmap

The main goal of this application is to characterize anatomical connectivity abnormalities in the 
brain of patients with velocardiofacial syndrome (VCFS), and to link this information with deficits in schizophrenia. 
The arcuate fasciculus including seed, midpoint and target ROI's.

This page describes the technology roadmap for stochastic tractography, using newly acquired 3T data, NAMIC tools and slicer 3.

A - Algorythm 


Algorythm

A - Descrition
  • Most tractography methods estimate fibers by tracing the maximum direction of diffusion. A limitation of this approach is that, in practice, several factors introduce uncertainty in the tracking procedure, including, noise, splitting and crossing fibers, head motion and image artifacts. To address this uncertainty, stochastic tractography methods have been developed to quantify the uncertainty associated with estimated fibers (Bjornemo et al., 2002). Method uses a propagation model based on stochastics and regularization, which allows paths originating at one point to branch and return a probability distribution of possible paths. The method utilizes principles of a statistical Monte Carlo method called Sequential Importance Sampling and Resampling (SISR). Based on probability functions, using a sequential importance sampling technique (Bjornemo et al., 2002 insert link to the paper), one can generate thousands of fibers starting in the same point by sequentially drawing random step directions. This gives a very rich model of the fiber distribution, as contrasted with single fibers produced by conventional tractography methods. Moreover, from a large number of sampled paths, probability maps can be generated, providing better estimates of connectivity between several anatomical locations.
B - Possible Applications
  • Since diffusion direction uncertainty within the gray matter is quite significant; principal diffusion direction approaches usually do not work for tracking between two gray matter regions. Thus if one requires finding connections between a priori selected anatomical gray matter regions, defined either by anatomical segmentations (in case of using structural ROI data), or functional activations (in case of megring DTI with fMRI), stochastic tractography seems to be the method of choice. Insert an example of application to the anatomical data (Figure 1), and to to fMRI data (Figure 2).
  • Stochastic Tractography is also comparable, if not better, in defining large white matter fiber bundles, espetially those traveling through white matter regions characterized by increased diffusion uncertainty (fiber crossings). Example of such application to internal capsule(Figure 3)
C - References
  • Björnemo M, Brun A, Kikinis R, Westin CF. Regularized stochastic white matter tractography using diffusion tensor MRI. In Fifth International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI'02). Tokyo, Japan, 2002;435-442. (insert paper here)


Module

Fill it in once its ready. Write exactly, step by step what is its functionality, adding snapshots:

Loading DTI in Nrrd format
eddy current distortion correction
movement correction
smoothing
tensor estimation
creating WM mask
loading rois
creating probability map
visual inspection (looking at areas where mask does not exist, but FA exceeds 0.3)
thresholding
visualizing path, tracts
getting numbers (thresholded FA, FA weighted by probability)


Work accomplished

A - Optimization and testing of stochastic tractography algorythm 
  • Original methodological paper, as well as our first atempts to use the algorythm (CC+ and matlab scripts) have been done on old "NAMIC" 1.5T LSDI data (available here).
  • Tri worked hard on making sure algorythm works on new high resolution 3T data (available here).
  • Tests have been done also o the spiral diffusion phantom, to make sure diffusion directions and scanner coordinates are handled properly by the algorythm (insert figure).
B - Clincal applications
  • Algorythm was used to trace and analyze anterior limb of the internal capsule on 1.5T data. It generated reacher representation of frontal fiber projections, it also turned out to be more sensitive to group differences in white matter integrity that conventional deterministic tractography (insert poster here).
  • Algorythm was also used to trace smaller white matter fiber tracts, such as Cingulum, Fornix, Uncinate Fasciculsu, Arcuate Fasciculus on 3T "Santa Fee" dataset (link to Santa Fee conference page)
C - References
  • Add our IC poster for clinical reference


Work in progress

A - Optimization and Testing of stochastic tractography module 
  • Julien is testing now separate components of work flow, including the masks, and impact of their precision on stochastic output, as well as impact of number of seeding points on tractography results.
  • At the same time, we are testing the module on Max Plank dataset, as well as running it on tractography comparison project dataset (link here)
  • We are discussing possibility of adding project specific functionality to module, such as third ROI to guide tractography, nonlinear registration button for merging fMRI and DTI data.
B - Clincal applications
  • We have started the project of investigating Arcuate Fasciculus using Stochastic Tractography. This structure is especially important in both VCFS and schizophrenia, as it connects language related areas (Brocka and Wernicke's), and is involved in language processing quite disturbed in schizophrenia patients. It also can not be reliably traced using deterministic tractography.
    • Whole brain segmentation, and automatic extraction of regions interconnected by Arcuate Fasciculus (Inferior frontal and Superior Temporal Gyri), as well as another ROI that would guide the tract ("waypoint" ROI). This step has been accomplished for the entire dataset of 23 schizophrenia subjects and 23 controls.
    • White matter segmentation, in order to prevent algorithm from traveling through the ventricles, where diffusivity is high. This has been done also for the entire dataset now.
    • Linear registration of labelmaps to the DTI space. This step has been done for all subjects, however we were not satisfied with the results of registration, especially in the frontal areas, when using tools available through slicer. We have redone this step, using nonlinear registration and demon's algorythm (slicer), and results are much more precise. This step has been also done now on all subjects.
    • Seeding tracts. We have piloted in on ten cases using 5000 seeds per voxel, but since it is quite time consuming running it on even most powerfull computers in the lab, we have experimented with smaller number of seeds per voxel. We tested 1000 seeds, which gave virtually identical results. We are now applying this stem to all our cases. Currently, all steps take about 6-8 hours per case to complete. (should be done by the end of November).

Applying stochastic tractography algorithm, to find the path connecting two ROIs. This has been also done for the subset of subjects (19 in total), however we will redo it after the new registration is completed.

    • Extracting path of interest, and calculating FA along the path for group comparison. Presentation of previous results for 7 schizophrenics and 12 control subjects, can be found here: Progress Report Presentation. Once new registration is completed, and tracts extracted, we intend to use another NAMIC tool for tract parametrization (Mahnaz will make it available for us within few weeks), to look at the diffusion properties along the tracts, which will make group comparison much more precise.
D - New developments (As of August 15th 2008) 



D - Subject comparison 
  • Since the DTI scanning protocol was established and tested on schizophrenia subjects, and thus data collection is much more advanced there, and since the ultimate goal is to compare anatomical connectivity abnormalities in VCFS with these in schizophrenia, this project has two benefits- it gives schizophrenia comparison data, as well as leads to establishing the tractography protocol that will be easily applicable to VCFS, once more data is collected.
  • We are looking into ways to compare stochastic tractography results between groups. Tract based and volume based measures are considered (both included as parts of the NA-MIC Kit)
    • In addition, Ipek is developing local analysis tools and may have a tool available in Fall 2008.

Updates/Progress

A - Optimization of stochastic tractography algorithm 
  • Stochastic Tractography algorythm has been used to analyze anterior limb of the internal capsule using 1.5T data, data was presented at the 2007 ACNP symposium, and we are now working on the manuscript.
  • We have optimized our algorithm to work with high resolution diffusion data acquired on 3T magnet, and tested it on multiple white matter fiber tracts (cingulum bundle, arcuate fasciculus, fornix). We are working on methods paper.
B – Slicer 3 Stochastic Tractography module and testing plus documentation 
  • Stochastic Tractography module has been completed, and presented at the AHM in SLC. Its now part of the slicer3. It works fine with GE data format, needs to be optimized for other data types. Martin is looking into it, we hope to have Julien, our new software engineer, to take over the module maintenance as soon as he is on board (July 2008)
C - Analysis of small anatomical structures 
  • We have started the project of investigating Arcuate Fasciculus using Stochastic Tractography. The pipeline for the analysis includes
    • Whole brain segmentation, and automatic extraction of regions interconnected by Arcuate Fasciculus (Inferior frontal and Superior Temporal Gyri), as well as another ROI that would guide the tract ("waypoint" ROI). This step has been accomplished for the entire dataset of 23 schizophrenia subjects and 23 controls.
    • White matter segmentation, in order to prevent algorithm from traveling through the ventricles, where diffusivity is high. This has been done also for the entire dataset now.
    • Linear registration of labelmaps to the DTI space. This step has been done for all subjects, however we were not satisfied with the results of registration, especially in the frontal areas, when using tools available through slicer. New registration protocol, b-spline, has been developed and made available to us few weeks ago (May 2008), and results are much more promising. We are currently running the registration for all our subjects. This step should be completed within the next few weeks.
    • Applying stochastic tractography algorithm, to find the path connecting two ROIs. This has been also done for the subset of subjects (19 in total), however we will redo it after the new registration is completed.
    • Extracting path of interest, and calculating FA along the path for group comparison. Presentation of previous results for 7 schizophrenics and 12 control subjects, can be found here: Progress Report Presentation. Once new registration is completed, and tracts extracted, we intend to use another NAMIC tool for tract parametrization (Mahnaz will make it available for us within few weeks), to look at the diffusion properties along the tracts, which will make group comparison much more precise.
D - New developments (As of August 15th 2008) 
  • Doug left the lab for the PhD program at NYU, but will continue working on the project part time. All registrations are now completed, tracts have been generated for half of the subjects included in the analysis. We hope to submit an abstract for ICOS schizophrenia conference due September 15th.
  • Julien, our new software engineer, has joined the lab, and is getting acquainted with stochastic tractography software. He will be responsible for improving the STM (Stochastic Tractography) module, and making sure software works with STM compliant datasets. Progress

Staffing Plan

  • Sylvain, Yogesh and Doug are the DBP resources charged with adapting the tools in the NA-MIC Kit to the DBP needs
  • Julien is our new NAMIC software engineer
  • Polina is the algorithm core contact
  • Brad is the engineering core contact

Schedule

  • 10/2007 - Optimization of Stochastic Tractography algorythm for 3T data. DONE
  • 10/2007 - Algorythm testing on Santa Fe data set and diffusion phantom. DONE
  • 01/2008-AHM - Prototype Stochastic Tractography module in Slicer 3. DONE
  • 01/2008-AHM - Working on the ways of extracting and measuring diffusion properties within the Arcuate Fasciculus using Slicer 3 module. DISCUSSED AT AHM, STILL WORK IN PROGRESS
  • 03/2008 - Start of the module application to group data. NOT STARTED YET
  • 07/2008 - BatchMake workflow.
  • 10/2008 - Data analysis and paper write up.
  • 01/2009-AHM - Groupwise tract based and volume based analysis of the multiple white matter tracts as a NA-MIC Workflow.

Team and Institute

  • PI: Marek Kubicki (kubicki at bwh.harvard.edu)
  • DBP2 Investigators: Sylvain Bouix, Yogesh Rathi, Julien de Siebenthal
  • NA-MIC Engineering Contact: Brad Davis, Kitware
  • NA-MIC Algorithms Contact: Polina Gollard, MIT

Publications

In print