The Computational Anatomy project has attempted to do for shape and form what Chomsky's theory of computational linguistics did for natural language understanding. We will discuss the notions of brain parsing and "geodesic positioning" central to Computational Anatomy. Parsing is a core technology in Computational Anatomy allowing for the understanding of high resolution MR brain imagery at 1mm scale in terms of a neuro-ontology of semantic labels.
Brain parsing supports high throughput indexing allowing us to build with the Mori laboratory several BrainClouds in which brains can be searched, retrieved and "machine learned".
We will examine several applications in Medical Imaging, including our pediatric and geriatric brain clouds.
We will discuss some results on one of the nation studies called BIOCARD for characterizing the onset and trajectory of preclinical Alzheimer's disease in temporal lobe structures - amygdala, hippocampus, and entorhinal cortex - as well as the study of Huntingdons disease as part of the US's PREDICT project studying high dimensional imaging markers as a function of genetic loading.