**WORKFLOW READ ME** 1. Use 3D Slicer module DicomToNrrdConverter to convert subject DICOM to NHDR. Use BMatrixGradientDirections. Use the .nhdr extension so that you will have easy access to the header file and information if editing is needed. 2. Create DTIPrep1.2 Protocols. -created default protocol from s3 -Edited Default protocol: -40% fail -NO to DiffusionbCheck -YES to ImagebCrop -YES to Slice, Interlace and Gradient bQuitonCheckFailure -YES to Vibration Checking (DominantDirection)--> V12 whole brain avg 6.598641975 std 0.062684789 -YES on DTI_bCompute -NO to saving frobenius. -Interlace Settings: (to fix B0s being dropped) -CorrelationDeviationBaseline: change from 2.5 to 4 -CorrelationDeviationGradient: change from 3 to 3.5 -Protocols located in 0_Processing IBIS_subj3_DTIPrep12_protocol_V12.xml 3. Assess the raw DWI -Get a feel for what the raw data looks like for each subject before quality control. -Is there motion? -Are there artifacts? If so, is there any pattern to the artifacts? -Note the gradients that contain artifacts or that have motion and make sure these are removed when doing the visual quality control. -Go through every slice in every direction for all gradients. 4. Run Auto QC using DTIPrep1.2 -Load individual subject nhdr and the study protocol, and hit "Run By Protocol". -QC outcomes are available in the IBIS_*_DWI_QCReport.txt. Note which gradients were removed and why. 5. Run Visual Quality Control with DTIPrep1.2 -Load IBIS_*_DWI_QCed.nrrd and IBIS_*_DWI_XMLQCResult.xml. -Assess if there is any motion between gradients, are there any artifacts, were the artifacts you saw in the raw data removed, were artifacts introduced? -Go through every slice in every direction for all gradients. -Exclude gradients that have artifacts. -Save visual checked DWI. 6. Convert Visual Quality Controlled DWI to DTI -DWI to DTI estimation module in Slicer 7. Visual Quality Control with Slicer4.2.2-1 -Load DTI and FA volumes. -Look at FA contrast to determine signal to noise ratio (SNR). -Using DTI volume, check colorFA of major tracts, glyph directions, and fiducial tractography of main tracts. Cingulum, genu, splenium, uncinate, corticospinal tracts bilaterally. -Fiducial tractography module settings (Tractography Interactive Seeding): Region Size 5mm Fiducial Seeding Step Size 1.0mm Max Number of Seeds 100 min 20mm Max 800mm FA Stopping Value 0.15 Stopping Track Curvature 0.7 Integration Step Length 0.5mm LINES 8. Re-computed brain mask using bet2 with the IBIS_*_DWI_VC_IDWI.nrrd as the input -bet2 -- we used the IDWI to compute the brain mask from because we found that this made a better brain mask than the B0 -You can also use AutoSeg to create a brain mask by using the B0 and IDWI 9. Quality Control the QCedVC IDWI bet2 brain mask -Slicer4.2.2-1 Editor- for removing the edge voxels -Load brain mask -Erode tool in the editor module - 4 neighbors- apply appropriately -save "_edit.nii.gz" -InsightSNAP_2.4 - for major edits -Load IBIS_*_DWI_VC_DTI_FA.nrrd as greyscale image -Load mask file: IBIS_*_DWI_VC_IDWI_bet2_mask_edit.nii.gz -increase contrast of greyscale image to better see brain -edit brain mask as needed and save edited mask with same name plus "_edited" before file extansion at end. -Did another round of erosion for the masks in Slicer and saved as "_edited2" 10. Create DTI images with edited bet2 mask applied -DWI to DTI estimation module in Slicer -Load the IBIS_*_DWI_VC.nrrd file -Load the final edited mask file, IBIS_*_DWI_VC_IDWI_bet2_mask_edited2.nii.gz -Once finished, loaded all subject's skull stripped DTI in Slicer to check skull strip results. 11. Atlas Building (DTIAtlasBuilder_1.3) -Add Safety Margin - 1- centerofHeadAlign -BSpline order 3, Log Euclidean, PPD -ANTS, SpatioTempDiffeo(SyN), Cross Correlation (CC), Region Radius-4 -Quality control the Atlases with MriWatcher 12. Label Map Creation in Slicer with Editor Module -Load FinalAtlasDTI.nrrd and FinalAtlasFA.nrrd -Make a separate label map file for each tract. OR if you make them in the same file, then use different label colors for each tract. -Genu- create label map in editor -Splenium- create label map in editor 13. Label Map Tractography - Tractography Label Map Seeding in Slicer -all default values except for what is below -FA stopping value: 0.15 -Integration step length: 0.5 14. FiberViewerLight -Used clustering algorithms in order until I felt that I wanted to keep all the remaining fiber classes left. Then I stopped editing the fiber bunlde and created the plane. -clean down to single fiber bundle. One that looks anatomically correct. -Plane creation: plane must only intersect the fiber bundle once, and it must be perpendicularly related to the fiber bundle from 2 views. It must be perpedicular to the flow of fibers at the point of intersection. -**When saving the plane, you MUST save it with the SAME NAME as the fiber bundle it is for with the addition "_plane.fvp" at the end. -copy final clean bundle and plane for each fiber tract to 3_Tractography/2_FinalFibers/ 15. DTIAtlasFiberAnalyzer -make directory with all VC_DTI_SkullStripped images -Copy Directory of Final Global Displacement Fields from Atlas Building directory to your DTIProfileAnalysis directory -Can use CSV from DTIAtlasBuilderResults.csv --> so long as you have not edited the path to this directory or any of the atlas input files. -Otherwise you can create a new csv. 1st column = subject name 2nd column = QCed_VC_SkullStripped_DTI 3rd Column = Global Displacement Fields from DTIAtlasBuilder -When plotting the profiles, create a text file to record all of the information on the plot window. Ideally this should be a file the tool saves and we will add this feature in the future. Also take a screen shot of this window so you have quick access to what the profiles looked like. -***Quality control occurs here as well. When plotting the diffusion profiles of the subjects against the diffusion profile of the atlas tract you are visualizing how well each subject is mapped into atlas space. If a subject does not map well into atlas space and that can not be fixed, the subject should be removed from analysis and the atlas rebuilt, and tractography redone. 16. FADTTS - requires Matlab. We used Matlab2011a. -create SubjectMatrices of covariates to be analyzed (max of 1 covariate/5 subjects). Subject x Covariate matrix (rows x columns) with or without a header row. -edit your FiberProfiles from DTIAtlasFiberAnalyzer --> remove Atlas column, remove header row, save as comma separated .txt -Edit your script as appropriate for your data -Make separate directories for each fiber and experiment 17. MergeStatWithFiber - last tab in DTIAtlasFiberAnalyzer -use corrected pvalue csv outputs from FADTTS -These need to be edited--> the header row needs to be added back in, and the first column must be arclength for that fiber. The other columns are the corrected p-values for each covariate in the order they appeared in the SubjectMatrix -Save output as vtk -Load this vtk in Slicer as a Model. In the Models module select View scalars. Must use the Slicer version (ex. Age_Slicer) to visualize the covariate p-values on the fiber.