Cesar Caballero Gaudes1,
2, Fikret Isik Karahanoglu3, Franois Lazeyras2,
Dimitri Van de Ville, 23
1Basque
Center on Cognition, Brain and Language, Donostia-San Sebastian, Guipuzcoa,
Spain; 2Department of Radiology and Medical Informatics,
University of Geneva, Geneva, Switzerland; 3Medical Image
Processing Lab, Ecole Polytechnique Federale de Lausanne, Lausanne, Vaud,
Switzerland
Paradigm-free mapping (PFM) enables to map the BOLD response in space and time without prior knowledge of the timing of the events using sparsity-promoting estimators, such as the Dantzig Selector or LASSO. Here, we extend PFM by using hierarchical structured sparsity-promoting estimators (Group-LASSO, Weighted-Fusion and Group-Weighted-Fusion) along with an informed basis set in order to deal with BOLD response variability, and use the monotone fast iterative shrinkage thresholding algorithm to solve this deconvolution problem. Experimental results demonstrate that the new approach has superior abilities to characterize single-trial BOLD responses with no prior timing information at 3T.