FMRI TASK DESIGN AND STATISTICAL ANALYSIS SOFTWARE

DESCRIPTION (provided by applicant): The overall goal of the project is to develop a new generation of functional MRI design and analysis software. The proposed software is designed to lead to an order of magnitude improvement over current methods in the estimation and detection of fMRI activation. This phase II project has three main aims: 1. to extend the utilization of a optimal spatial filtering technique to increase the SNR in the fMRI signal; 2. to develop a computer-aided design tool to permit the creation of novel fMRI experiments; and 3. integration of this software into an integrated suite providing a wide variety of spatiotemporal measurements of human brain function from structural MRI, fMRI, EEG, and MEG. The software will also incorporate methods to greatly increase the statistical power and reliability for detecting task-related physiological responses with fMRI, including the use of correlated noise compensation, and optimal spatial filtering. These methods will be applicable to both blocked and event-related paradigms. The software will generate optimal rapid-presentation event-related (RPER) fMRI designs. Such designs allow for new classes of functional imaging experiments and permit comparisons between fMRI and traditional electrophysiological or behavioral studies. The software will be designed to be utilized within our suite of MRI software tools. In particular, the results will be able to be presented on flattened and inflated cortical surfaces for superior visualization. The area of multi-modal brain imaging will be greatly advanced by supporting event-related designs and cortical surface formats within a single computational platform. PROPOSED COMMERCIAL APPLICATIONS: The proposed would be of interest to researchers using functional MRI, i.e., well over 1000 worldwide and growing rapidly. The proposed software will increase the functionality of fMRI by at least 20% but we propose to sell it for only about 0.1% of the fMRI equipment cost. Put another way, the proposed software could have the number offMRI sessions needed to achieve a given level of significance, and thus would play for itself after a single study.