Da Ma1
1Wake Forest University School of Medicine, Winston-Salem, NC, United States
Synopsis
This
talk will present several different approaches and best practice to achieve effective
quantify and visualize brain morphology. First, we will demonstrate qualitative
and quantitative methods to perform, visualize, and evaluate the effect of
multi-site neuroimage data harmonization for removing unwanted confounding
factors. Second, we will illustrate visualization methods for various brain
morphology metrics, such as structural volume, cortical thickness, that can
reveal individualized brain morphology patterns. Lastly, we will introduce the parametric
statistical analysis and visualization approach for presenting the groupwise
brain morphology variation across large population.
Neuroimage data harmonization
When
processing and analyzing large multicenter databases, the effects of multiple
confounding covariates may increase the heterogeneity within the data. Such
uncontrolled variation may hinder the capability of big data analysis to detect
changes due to the actual effect of interest, for example, changes due to
disease. Data harmonization is therefore important to control the effect of confounding
covariates and improve the effectiveness of the downstream tasks such as
statistical analysis and machine learning methods. In addition, efficient ways of
evaluation and assessment methods are needed to evaluate the effectiveness of the
data harmonization methods. Visualization and computational methods will be
introduced to evaluate the goodness of harmonization, both quantitative and
qualitative (1).Brain morphology metrics visualization
Furthermore,
when performing computational analysis of big neuroimage data, various brain
morphologies metrics are often derived such as brain volume, cortical
thickness, cortical surface area, curvature, etc. It is therefore important to visualize
and project these brain morphology metrics onto the individual brain space. Effective
projection of harmonized data onto the brain space might provide powerful visualizations
about the individualized disease patterns.Parametric statistical analysis of groupwise brain morphology variation
Finally,
voxel-wise or cortical-vertex-wise statistical analyses are powerful tools to
reveal groupwise brain morphology differences across distinctive populational
groups. For example, different subtypes of neurodegenerative diseases might reveal
distinctive atrophy patterns. Therefore, it is also important to perform
effective quantification and visualization of the voxel-/vertex-wise statistical
analysis to reflect morphological variations in the localized brain regions in
the populational groupwise space (2). Acknowledgements
No acknowledgement found.References
1. Ma D, Popuri K,
Bhalla M, Sangha O, Lu D, Cao J, et al. Quantitative assessment of field
strength, total intracranial volume, sex, and age effects on the goodness of
harmonization for volumetric analysis on the ADNI database. Hum Brain Mapp.
2019 Apr 1;40(5):1507–27.
2. Ma
D, Cardoso MJ, Zuluaga MA, Modat M, Powell NM, Wiseman FK, et al. Substantially
thinner internal granular layer and reduced molecular layer surface in the
cerebellar cortex of the Tc1 mouse model of down syndrome – a comprehensive
morphometric analysis with active staining contrast-enhanced MRI. NeuroImage.
2020 Dec;223:117271.