System evaluation for in vivo imaging of Amyloid Beta plaques in a mouse brain using statistical decision theory

We have developed a method that can be used to assess high resolution imaging systems for relevant molecular imaging applications that require a good combination of spatial resolution and sensitivity. The in-vivo imaging of Amyloid Beta plaques, which are characteristic in the neuropathology of Alzheimer disease, in the brain of a mouse has proven to be a major challenge in current research due to their heterogeneous structure of micro-scale range (<= 100 mu m). Detecting these plaques or changes in their size and distribution by an ideal observer of the data from an imaging system can be employed as a figure of merit to optimize the hardware configuration of that imaging system for this important application. This method can be implemented on the raw data and does not require image reconstruction studies. We have derived a test statistic for the binary detection task that employs a stochastic object model to describe these plaques. The parameters of the object model were obtained from in-vitro plaque images.