Differences between revisions 2 and 189 (spanning 187 versions)
⇤ ← Revision 2 as of 2010-11-01 14:06:49 →
Size: 1756
Editor: AllisonStevens
Comment:
← Revision 189 as of 2017-03-16 14:11:24 → ⇥
Size: 0
Editor: AllisonMoreau
Comment: multimodal diffusion integration and diffusion processing tutorials were merged into one. deleting original diff page that didn't include mulitmodal
Deletions are marked like this. Additions are marked like this.
Line 1: Line 1:
=== Diffusion Use Case: ===
10 MIND subjects (can't use schizophrenia data)

1. Creation of DTI maps with dt_recon (FA, diffusivity, IVC).
 * We should talk about potentially adding more metrics to this.
 * After #1, you need to make sure to point out where the registration file is that represents the correspondence between str and diff spaces. So when you want to do ROI analysis using wmparc or any other FS recon files, people will know how to transform them to the right space.
 *run dtk on data to look at tracts?

2. ROI analysis of #1 with wmparc.mgz
 * need to look a little more closely at what this is doing, and how we might suggest doing this.

3. Comparing FA point by point along the tract (using tracula; need interesting question to test)
 * (I think this technically uses the FSL tools, but I guess various things we do use them.)

4. Recreate importing freesurfer aparc as ROI

5. Ways of registering across subjects - using CVS

6. voxel based statistics with mri_glmfit
 * There might be some things to discuss with regard to procedures- e.g. do we use an FA threshold or something like that, or just compare all voxels, etc.

Study questions:
 *effects of age on FA of a specific tract (These are the ages: 47 57 36 50 46 41 52 63 38 30. Not sure how much age effect we'll be able to see.)
 *gender study



=== Definition of dt_recon output: ===
ADC: apparent diffusion coefficients (ln(S0/S1)/(b1-b0))

RA: relative anisotropy (sqrt(var(lambda))/mean(lambda))

radialdiff: radial diffusivity (lambda2+lambda3)/2

IVC: inter-voxel correlation

where lambda = (lambda1, lambda2, lambda3) the eigenvalues
S0 = signal intensity without the diffusion weighting
S1 = signal with the gradient