ANTsX neuroimaging-derived structural phenotypes of UK Biobank
Nicholas J Tustison1,2, Michael A Yassa3, Batool Rizvi3
1Department of Radiology and Medical Imaging, University of Virginia, Charlottesville, VA, USA. ntustison@virginia.edu.
Scientific Reports
|April 17, 2024
Summary
This study enhances UK Biobank neuroimaging data with new measurements for better brain health research. Different software tools show varied performance, highlighting the need for careful selection of imaging-derived phenotypes (IDPs).
Area of Science:
- Neuroscience
- Radiology
- Genetics
Background:
- UK Biobank is a major resource for studying lifestyle, environmental, and genetic factors in health.
- Neuroimaging provides imaging-derived phenotypes (IDPs) crucial for neuroscience research.
- Existing structural IDPs from FSL and FreeSurfer are being expanded.
Purpose of the Study:
- To augment UK Biobank neuroimaging structural IDPs with Advanced Normalization Tools (ANTs) measurements.
- To assess the clinical utility of these IDPs using predictive modeling.
- To compare the performance of different software libraries for IDP generation.
Main Methods:
- Integrated new cortical and subcortical measurements from ANTs, including hippocampal, extra-hippocampal, and cerebellar parcellations.
- Utilized predictive modeling with phenotypic correlates like age, fluid intelligence, and memory.
- Evaluated model accuracy using root-mean-squared-error and area-under-the-curve.
Main Results:
- New IDPs from ANTs, including medial temporal lobe and cerebellum measures, were added to the UK Biobank resource.
- Predictive modeling revealed varied performance between IDP sets from different software libraries (FSL, FreeSurfer, ANTs).
- Combined IDP sets showed different predictive accuracies, impacting clinical interpretability.
Conclusions:
- The study successfully expanded UK Biobank's neuroimaging IDP offerings with ANTs-derived measures.
- Performance differences between software packages necessitate careful consideration for IDP selection and application.
- Findings provide comparative insights into IDP utility and potential clinical relevance in neuroscience research.
