Comparison of two different approaches for brain activity detection in fMRI: SPM-MAP and SPM-GLM

The functional MRI {Magnetic Resonance Imaging), fMRI, is today a widespread tool to study and evaluate the brain from a functional point of view. The blood-oxygenation-level-dependent (BOLD) signal is currently used to detect the activation of brain regions with a stimulus application, e.g., visual or auditive. In a block design approach the stimuli (called paradigm in the fMRI scope) are designed to detect activated and non activated brain regions with maximized certainty. However, corrupting noise in MRI volumes acquisition, patient motion and the normal brain activity interference makes this detection a difficult task. The most used activation detection fMRI algorithm, here called SPM-GLM [1] uses a conventional statistical inference methodology based on the t-statistics In this paper we propose a new Bayesian approach, by modeling the data acquisition noise as additive white Gaussian noise (AWGN) and the activation indicators as binary unknowns that must be estimated. Monte Carlo tests using both methods have shown that the Bayesian method, here called SPM-MAP, outperforms the traditional one, here called SPM-GLM, for almost all conditions of noise and number of paradigm epochs tested.