NJIT Applied Mathematics Colloquium

Friday, February 17, 2012, 11:30am

Cullimore Lecture Hall II
New Jersey Institute of Technology

Sparse Paradigm Free Mapping: detection of activations and resting state networks in fMRI. 

University of South Carolina 

The ability to detect single trial responses in functional Magnetic Resonance Imaging (fMRI) studies of the brain is important,
yet traditionally a known experimental paradigm is used where the timing of activations has to be specified in the model. The main task in our work is to detect signals in space and time in very high-dimensional datasets, taking into account the physical properties inherent in the data collection, as well as the noise.

Paradigm Free Mapping (PFM) is a method that detects single trial responses without specifying prior information on the
timing of the events. The PFM method is based on the deconvolution of the fMRI signal using a hemodynamic response
function, and involves a ridge regression estimator for signal deconvolution and a baseline signal period for statistical
inference. A sparse version of PFM uses the Dantzig Selector, and this method obtains high detection rates of activation, comparable to a model-based analysis, but requiring no information on the timing of the events or baseline period. The practical operation of this sparse PFM method was assessed with single-trial fMRI data acquired at 7 Tesla, where it automatically detected all task-related events. Further the methodology is used to detect resting state networks, where a spatio-temporal network of activations is observed while the volunteers are apparently at rest.

The work is joint with Cesar Caballero Gaudes, Natalia Petridou, Susan Francis and Penny Gowland.