Diagnostic Multi-modal Image Processing Via Complexity Measures

Large volumes of multi-modal data are typically produced in current imaging studies of brain structure and function. While a number of useful data mining techniques for extracting relevant information from these large volumes of data have been developed, summarizing the spatial patterns in the various images in a way that is interpretable for clinicians would greatly enhance the utility of such information extraction. This project proposes the use of information theoretic, complexity measures, for the generation of interpretable, spatial (and spatio-temporal in the case of functional imaging) summary information from multi-modal imaging data for diagnostic purposes. These methods are currently being applied, e.g., to generate predictors for early indications of cognitive decline.