Developing Phenotypic-Specific Brain Templates for DTI-Based Analysis
L LeMerise1*, J Guerrero2, S Hartley3, A Alexander4, B Christian5, (1) University of Wisconsin-Madison, Madison, WI, (2) University of Wisconsin Madison, Madison, WI, (3) University Of Wisconsin-Madison, Madison, WI,(4) University of Wisconsin - ADCL, Madison, WI, (5) University of Wisconsin, Madison, WI
Presentations
SU-H300-IePD-F9-2 (Sunday, 7/10/2022) 3:00 PM - 3:30 PM [Eastern Time (GMT-4)]
Exhibit Hall | Forum 9
Purpose: Parameter maps from diffusion magnetic resonance imaging (dMRI) enable analyses of tissue brain microstructure that offer the potential to inform RX treatment planning or therapy response. For clinical trial applications, dMRI studies may include voxel-based analysis, atlas-based labeling, and group average fiber tracking. These methods require spatial normalization of the quantitative maps. Studies have shown that population specific templates offer advantages in group-level analyses compared to using standard targets like the MNI atlas, by not imposing a specific type of anatomy on non-typically anatomical brains. This is particularly relevant when studying heterogeneous conditions like brain cancer or populations that are challenging to image due to excessive motion such as autistic or down syndrome groups. This work’s goal is to implement the method known as T1-w-imaging-diffusion (Ti-Di) fusion for robust estimation of DTI-based templates using a phenotypic-specific Down Syndrome (DS) cohort with unique brain structure.
Methods: Images from 38 DS participants ages 25.5 to 55.6 years were used to create a brain template. The framework methods consisted of co-registering individual dMRI to T1-weighted (T1-w) images, non-linear estimation of a T1-w template, spatial normalization of the T1-w images to the template, applying the transformations to the DTI maps, and finally averaging the DTI maps to generate DTI templates.
Results: Visual inspection of the registered images indicates greater similarities between individual images and the DS-specific template compared to co-registration to the MNI atlas, particularly with respect to the preservation of enlarged ventricles and smaller cerebellum, hallmark features in DS. This suggests potential improvements in spatial alignment of parameter maps across individuals, which is the subject of ongoing and future work.
Conclusion: Improved intersubject alignment of dMRI data will improve power in detecting differences or associations in analyses of diffusion in neuroimaging studies of this and other neurological conditions such as brain cancer.
Keywords
MRI, Image Processing, Image Fusion
Taxonomy
Contact Email