Automated Delineation, Parcellation and Analysis of the Cerebellum from MR Images

Period of Performance: 06/01/2013 - 05/31/2014


Phase 1 SBIR

Recipient Firm

Neurobehavioral Research, Inc.
Honolulu, HI 96814
Principal Investigator


DESCRIPTION (provided by applicant): Cerebellar involvement has been implicated in a large spectrum of disorders. Reduced cerebellar volumes have been reported in alcoholism, autism spectrum disorders, schizophrenia, mood disorders, and essential head tremors. As an example of the work demonstrating the more widespread role of the cerebellum in brain function, Sullivan and colleagues have demonstrated impairment in static upright balance with excessive sway in individuals with pathology of the anterior superior vermis, and have shown that alterations in nodes of the frontocerebellar circuitry can predict executive dysfunction in prefrontal regions. There has been an explosion of studies in the last decade examining functional brain networks using task-activated fMRI, and resting state functional connectivity MRI . For the most part, cerebellar morphometric studies and studies of the cerebellar components of brain networks have been ignored, largely as a consequence of lack of tools to support such investigations. The time is ripe for reliable and valid tools to support such studies. We believe that a cerebellar analysis toolkit (CATK) not only has a waiting audience and market, but its availability will act as a catalyst for the incorporation of the cerebellum in studies of brain structure and function, and thus will be of great significance. ! The goal of this SBIR project is to develop an integrated cerebellar segmentation / parcellation and analysis suite. The core component will be MR image delineation of the cerebellum and its subregions. The suite will include a comprehensive set of image and surface analysis tools, including the capability for shape analysis of cerebellar structures, and for visualization and surface editing. We will develop an integrated analysis environment with high performance tools implementation on GPUs using CUDA. CATK will have three major components: 1) cerebellar delineation, 2) A CUDA-based surface and image processing toolbox, and 3) A CUDA/openGL based rendering/surface editing tool. The product will provide a low cost means for extracting valuable information from existing MR image databases as well as from new data.