A Combined SPM-ICA Approach to fMRI

Independent component analysis (ICA) and statistical parametric mapping (SPM) are two commonly used methods of analyzing fMRI measurements. Typically, these methods are applied separately to the measurements to produce brain maps indicating active brain regions in response to a stimulus or a performed task. However, ICA can also be used to develop a hemodynamic response model that can be used as a regressor in SPM of fMRI measurements. This may lead to a more accurate method of localizing brain activity that corresponds to performing a task or to various pathologies. In this study, BOLD fMRI data were acquired from a subject performing a finger flexion task in a block design paradigm. Both spatial and temporal ICA was performed on the subject´s BOLD fMRI measurements. Two hemodynamic response model signals were generated from ICA results to use as regressors in SPM of the subject data. IC maps and SPM-generated brain maps of the subject data using the canonical hemodynamic response model and the ICA-derived models were compared. In all cases, there was significant overlap in voxel activations