Functional MRI (fMRI) has become a major tool for physicians, psychologists, and neuroscientists in their work of trying to understand brain function and architecture. Because many investigators use fMRI and because new study paradigms are being developed to assess different parts of the brain, ways to evaluate the post-processing tools for these datasets are needed. Many different algorithms have been incorporated into packages that permit corrections for motion artifacts in the datasets and provide the means of applying statistical analysis in order to extract information from the studies. How well a particular analysis approach works on a particular dataset is often difficult to understand because activation is not precisely repeatable and may change between data collections. Therefore, using real data to assess how well a particular processing approach works is usually not appropriate. For this reason, a realistic computer-generated phantom can provide a reasonable way of testing a particular processing tool or algorithm so that the user has confidence in the results that could be quite useful as a way of testing new software during the development phase or in teaching new users how to employ a particular tool for processing fMRI data. The overall goal of this application is two-fold: 1) to extend, refine, and evaluate a previously developed computer-generated phantom designed to evaluate the capabilities of available algorithms and processing packages for functional magnetic resonance imaging (fMRI) datasets, and 2) to make this phantom available to the user community. ? ?
