Shengwei Zhang1, Konstantinos Arfanakis1
1Department of Biomedical Engineering, Illinois Institute of Technology, Chicago, IL, United States
This work demonstrated that diffusion imaging studies employing automated ROI selection by means of conventional atlas-based segmentation suffer from misregistration. We adopted the main principles of TBSS and proposed an alternative automated approach, termed here skeletonized atlas-based segmentation, which is relatively immune to misregistration. Furthermore, a new skeletonized atlas was developed in ICBM152 space based on the IIT2 DTI template. The combination of the whole brain DTI template and skeletonized WM atlas has the potential to increase the accuracy and sensitivity of both voxel-wise and ROI analyses in DTI as well as high-angular resolution diffusion imaging studies.