Using independent component analysis to detect active regions in brain fMRI for tactile stimulation

Functional magnetic resonance imaging (fMRI) provides a powerful, non-invasive way to investigate brain activity. Blood oxygen level dependent (BOLD) fMRI measures the decrease in deoxyhemoglobin that corresponds to the gradually increased flow of blood to the active regions of the brain. Therefore, one important task in fMRI analysis is to correctly identify the active neural regions. In this research, a new algorithm is proposed to detect brain regions engaged by tactile stimulation as measured by fMRI. Independent component analysis (ICA) is a technique that attempts to separate functional image data into spatially independent non-Gaussian components. The proposed ICA-based method is a two-staged process for selecting the spatially independent component that best matches with the time course of tactile stimulation. Compared to the commonly used Statistical Parametric Mapping (SPM) algorithms, the proposed method is more sensitive in detecting some of the active regions and requires fewer input parameters in processing the fMRI data.