Quntification of Brain Morphometry and Anatomical Change in NeonatesAnders M. Dale, PhDUniversity of California, San Diego

Outline Adaptation and extension of FreeSurfer methods for neonatal brain scans Build neonatal brain atlasTracking of longitudinal anatomical change Automated identification of white matter tracts

Improvements to MRI acquisition methods Real-time motion correction Correction of B0 distortions in EPI/DTI data

FreeSurfer Anatomical SegmentationSubcortical SegmentationCortical Parcellation

Quantitative Morphometric Reporting

FreeSurfer Segmentation of 3 Year OldJosh Kuperman, UCSD

Pediatric Brain Development: Change in T1-contrast27 day old3 month oldLinda Chang & Thomas Ernst, U of Hawaii

Full-Brain Segmentation in 2-Week Old BabyCourtesy of Ellen Grant, MGH & Josh Kuperman, UCSD

Anatomical Change over 6 Months in MCI:Preliminary Results from the Alzheimers Disease Neuroimaging Initiative (ADNI)Holland et al., 2008

Change over 6 MonthsChange over 12 MonthsAnatomical SegmentationQuantification of Longitudinal ChangeIndividual ADNI Subject (70 y.o. Female, Diagnosed with MCI)MMIL / UCSD

Tracking of Longitudinal Anatomical ChangeAffine Registration (TP1)Josh Kuperman & Dominic Holland, UCSD

Tracking of Longitudinal Anatomical ChangeAffine Registration (TP2)Josh Kuperman & Dominic Holland, UCSD

Tracking of Longitudinal Anatomical ChangeNon-Linear Registration (TP1)Josh Kuperman & Dominic Holland, UCSD

Tracking of Longitudinal Anatomical ChangeNon-Linear Registration (TP2)Josh Kuperman & Dominic Holland, UCSD

Tracking of Longitudinal Anatomical ChangeNon-Linear Registration (Jacobian)Josh Kuperman & Dominic Holland, UCSD

Tract Identification Using DTI StudioDonald Hagler & Josh Kuperman, UCSD

Fully Automated DTI TractographyCombining Atlas Priors and Diffusion DirectionDonald Hagler & Josh Kuperman, UCSD

Outline Adaptation of FreeSurfer segmentation methods for neonatal brain scans Build neonatal brain atlas Cortical surface reconstruction and parcellation Segmentation of subcortical structuresTracking of longitudinal anatomical change Automated identification of white matter tracts

Improvements to MRI acquisition methods Real-time motion correction Correction of B0 distortions in EPI/DTI data

* /GE Title or job number /*

Host Process

Nav reconPROMO motion estimation6 rigid body motion paramstx, ty, tz, rx, ry, rzNOTE: Navs placed during intrinsic dead-time of 3D sequences for zero impact on scan timeNavdataPROspective MOtion Correction (PROMO)SegmentNavDeadTimeNavDeadTimeNavDeadTimeNavDeadTimePSDm = || 1 0 ||01Sorted rescanlistm > mt123

* /GE Title or job number /*

Nav=3-plane spirals3-Plane Navigators3 single-shot spirals in 38 msflip = 8FOV = 32 cm10 mm x 10 mm x 10 mmDeadTime=Minimum 40 ms for nav recon and motion estimation

* /GE Title or job number /*

3-Plane Spiral MasksAcquired spiralsMaskMasked spirals

PROMO OffPROMO Real-Time Motion Correction T110 Year Old Boy

Tim Brown, Josh Kuperman, J. Cooper Roddey, UCSD; Eric Han, GEPROMO OnCurrently in routine use for research studies at UCSD

Clinical trial in Peds planned at UCSD, UCSF, Mayo & Stanford

Real-Time Motion Correction XETA T2Nathan White & J. Cooper Roddey, UCSD; Eric Han, GEPROMO onPROMO off

Spatial Distortion in EPI at 3T Due to B0 VariationPE direction: A>P Before CorrectionDominic Holland, UCSD

PE direction: P>A Before CorrectionSpatial Distortion in EPI at 3T Due to B0 VariationDominic Holland, UCSD

PE direction: A>P After CorrectionSpatial Distortion in EPI at 3T Due to B0 VariationDominic Holland, UCSD

PE direction: P>A After CorrectionSpatial Distortion in EPI at 3T Due to B0 VariationDominic Holland, UCSD

Registration of DTI EPI and 3D Structural T1 DataDominic Holland, UCSD

Integration of Imaging and Genomics:Mapping the Genetic Basis of Brain Development

Heritability of Cortical ThicknessRimol et al., 2008

Genome-Wide Association Studies (GWAS)SNP rs327067SNP rs756928SNP rs115970SNP rs192646TOP Study

Tracking of Longitudinal Anatomical ChangeAffine Registration (TP1)Josh Kuperman & Dominic Holland, UCSD

Tracking of Longitudinal Anatomical ChangeAffine Registration (TP2)Josh Kuperman & Dominic Holland, UCSD

Area Patterning of the Mammalian CortexOLeary et al., Neuron, 2007

Next Steps Adaptation and extension of FreeSurfer methods for neonatal brain scans Build neonatal brain atlas need additional scans, with optimized MRI sequence work with MGH & UH for manual segmentationTracking of longitudinal anatomical change extend method to model tissue contrast change Automated identification of white matter tracts need neonatal scans with appropriate protocol manual identification of tracts

Improvements to MRI acquisition methods Port methods from GE to Siemens platform Incorporate genetic information(?)

Pediatric Brain Development: Change in T1-contrast27 day old3 month oldLinda Chang & Thomas Ernst, U of Hawaii

White Mater Development in Neonatal MonkeysJulia Hamstra & David Amaral, UC Davis; Dominic Holland, UCSD1 week4 weeks12 weeks8 weeks

Genetic Correlation: Primary Visual (V1)Rimol et al.,2008

Genetic Correlation: aSTS/TPRimol et al.,2008

Genetic Correlation of Cortical ThicknessRimol et al.,2008

Subcortical Segmentation in 2-Week Old BabyCourtesy of Ellen Grant, MGH & Josh Kuperman, UCSD

FreeSurfer Segmentations of 3-4 Year Olds: The Good, the Bad, and the UglyJosh Kuperman, UCSD

Image Quality Issue?

*****************Show spiral moviesSplit into 3 slides Slide as is (but generalized to include XETA) minus bullet 1 and 3. Add justification for MP-RAGE-like contrast. Refer to as T1-weighted 3D imaging as opposed to IR-SPGR. Include time. Have spiral movies and info from first bullet. Include spiral parameters. Also show that there is no burn-in on the 3D images (use no motion / no application case vs no play out at all). Framework and 3rd bulletShow spiral moviesSplit into 3 slides Slide as is (but generalized to include XETA) minus bullet 1 and 3. Add justification for MP-RAGE-like contrast. Refer to as T1-weighted 3D imaging as opposed to IR-SPGR. Include time. Have spiral movies and info from first bullet. Include spiral parameters. Also show that there is no burn-in on the 3D images (use no motion / no application case vs no play out at all). Framework and 3rd bulletShow spiral moviesSplit into 3 slides Slide as is (but generalized to include XETA) minus bullet 1 and 3. Add justification for MP-RAGE-like contrast. Refer to as T1-weighted 3D imaging as opposed to IR-SPGR. Include time. Have spiral movies and info from first bullet. Include spiral parameters. Also show that there is no burn-in on the 3D images (use no motion / no application case vs no play out at all). Framework and 3rd bullet**********************