Project: TR&D 3 (Intrinsic Shape Analysis)

University Of Southern California, Los Angeles CA

Investigators

Linked publications & trials

Abstract

Once relevant features have been extracted from imaging datasets as detailed in Projects 1 and 2, users of the LONI Resource need to interpret the resulting data to address biologically relevant questions. At the data interpretation stage, considerable flexibility and interactivity is needed to allow users to tailor analysis methods to the specific biological questions of interest. Data interpretation potentially involves a number of processes: i) data dimensionality reduction; ii) integration with biological measures derived from non-imaging sources; iii) statistical model selection and parameter estimation; and iv) model validation and hypothesis testing. Application of these processes is not necessarily sequential, and the final interpretation of the data may involve iterative use of these processes to explore a number of possibilities. Visualization tools play an important role in data interpretation and error checking, allowing users to inspect raw data and statistical results in an anatomic context. Due to the interactive nature of the data interpretation stage of analysis, issues of performance and usability are particularly important.