fMRI activation detection by obtaining BOLD response of extracted balloon parameters with Particle Swarm Optimization

Abstract

Functional magnetic resonance imaging (fMRI) has immense worth in neuroimaging. The basic techniques used in fMRI data analysis were founded on a linear time-invariant system depicted with the impulse response function (HDR). However, there are evidences which accept nonlinear relations between the measured BOLD response and the task, especially in the case of rapid stimulation. Physiological models are replaced with previous impulse response function and use information based on physiological concept. In this paper, balloon model parameters as a physiological model were extracted using particle swarm optimization and then difference between obtained BOLD response from these parameters and measured signal from fMRI, were used to detect the active region.