DTI SubprojectProcessing & Analysis
Norbert SchuffVA Medical Center & University of
C lif i S iCalifornia, San Francisco
ObjectivesObjectives
D t i if DTI • Determine if DTI can:– provide a robust marker of PD progression
serve as an adjunct to clinical assessment and – serve as an adjunct to clinical assessment and PET
Scope Of WorkScope Of Work• DTI computations:
– Rigorous quality control g q y– Accurate alignment to structural MRI– Calculation of:
• standard DTI measures (FA, MD)standard DTI measures (FA, MD)• advanced DTI measures (geodesic FA, generalized
FA)• eigenvectors (tractography)g ( g p y)
• Exploratory Group Analyses – Region-of-interest basedRegion of interest based– Regionally unbiased– Joint analyses of DTI with PET
MRI AcquisitionsMRI Acquisitions
T1ww “Proton density” T2wStructuralMRI
Bow Diffusion weightedDTI Bow Diffusion weightedDTI
Automated Quality ControlFrame-to-Frame Similarity
Good Quality Motion Instability
ection
s Brain Signal
Slices
Diffusion Dire
Slices
Noise
DTI Quality Control: First 8 ScansDTI Quality Control: First 8 Scans
50
Baylor Northwestern JohnH
30
40
SNR
DW
I
3102 3105 3301 3556 3102 3105 3301 3556 3102 3105 3301 3556
Subject ID
20
Baylor Northwestern JohnH
8
10
12
oise
Sim
ilari
ty
3102 3105 3301 3556 3102 3105 3301 3556 3102 3105 3301 3556
4
6
Whi
te N
o
3102 3105 3301 3556 3102 3105 3301 3556 3102 3105 3301 3556
Subject ID
DTI-MRI Co-registrationDTI MRI Co registrationRigid alignment Raw DWI
Nonlinear alignment(in collaboration with Dr. Tom Fletcher, U Utah)
DTI InterpretationDTI Interpretation
High Sensitivity – Low Specificity DTI Intensity Electron MicrographsHigh Sensitivity – Low Specificity y g p
Axon Model & DTI Components
Summary:1. Mean displacement distance of tissue
water that is capture on DTI using a
(A) Non myelinated olfactory nerve of the garfish
clinical MRI scanner ~ 10‐20 um
2. Structural features of axons other than myelin are sufficient to give rise to DTI contrast
(A) Non‐myelinated olfactory nerve of the garfish(B) Myelinated optic nerve of the garfish
C. Beaulieu, NMR in Biomedicine 2002; 15; 435‐55
DTI ComputationsDTI ComputationssMRI Fractional Anisotropy Kullbeck‐Leibler
Anisotropy
SummarySummary• Accomplishments
MRI sites were qualified– MRI sites were qualified– DTI of first 8 subjects QC’ed– MRI quality is generally very good– Standard and advanced DTI measures implemented
What needs to be done• What needs to be done– Format for uploading derived data to LONI– Processing of DTI as data arriveg– Interim group analyses (n=20 per group?)