Probabilistic Shape Statistics

Alzheimer’s disease is characterized pathologically by the accumulation of amyloid plaques and neurofibrillary tangles. These pathological processes are thought to precede the macroscopic changes in the form of brain tissue loss and a subsequent decline in cognitive abilities. The changes, commonly known as atrophy, can be observed with imaging modalities such as MRI. Determining the evolution of regional shape changes may provide insights into distinct patterns in the progression of Alzheimer’s disease, and allow for earlier diagnosis and prognosis in the furture.

Statistical models have been proposed to disentangle, in a data-driven fashion, the distinct subtypes of Alzheimer’s disease. However, none of the models have taken into account regional shape due to its nonlinear nature. In the project, we develop statistical shape analysis methodology that will allow analysis of shape data and thereby understanding of the evolution of the human brain shape in Alzheimer’s disease.