Commutator filter: A novel technique for the identification of structures producing significant susceptibility inhomogeneities and its application to functional MRI

A robust postprocessing method to detect regions of magnetic field change is presented. High resolution gradient echo data is acquired to create a complex baseline image. From this image, two low resolution images are obtained by applying a low pass filter on the magnitude image and the complex image, respectively. Subtraction of the magnitude of the filtered complex image from the filtered magnitude image gives a new image called a commutator image in which regions of susceptibility-induced field variation are greatly enhanced. Results on phantoms and applications to visualize paramagnetic venous blood vessels in the human brain are presented. The use of this technique on a 3D gradient echo data set represents a new approach to obtain an MR venogram. The application of this method to blood oxygenation level dependent (BOLD) functional brain MR imaging is demonstrated and the important role susceptibility plays in this model is verified.