Robust Bayesian Detection of Brain Activated Regions with Functional Mri

The functional Magnetic Resonance imaging (fMRI) is currently the most prominent method used for functional brain imaging, and it is a big step forward in the process of answering the main question asked to all the functional imaging methods: What are the brain regions involved in mediating a specific brain function? And thought the fMRI’s obvious qualities have allowed for its fast acceptance and development, its limitations are far from letting this question to become a "closed problem". The fMRI experiments usually look for the change in blood oxygenation and blood volume resulting from altered neural activity. This signal, called blood-oxygenation-leveldependent (BOLD), results from the endogenous paramagnetic contrast property of the deoxygenated hemoglobin. Hence, increased blood flow reduces the local concentration of deoxygenated hemoglobin causing an increase in the MR signal on a T2*-weighted image [1]. It is commonly accepted, and has been empirically proved, that there is a strong correlation between neural activity and the vascular response that leads to the consequent increase of this BOLD signal citation. This relation, know as the hemodynamic response function (HRF), is at the core of fMRI data analysis. Inferences about which brain regions are involved in the particular stimulated cognitive and sensorimotor functions are based on how well the BOLD signal correlates with this stimulation, mediated by the HRF. Contrary to the impression one might get in a brief review of the literature, there are not many ways to analyze fMRI time-series with a diversity of statistical and conceptual approaches. In fact, with very few exceptions, every analysis is a variant of the general linear model (GLM), that expresses the observed response variable in