Functional brain imaging at 1.5 T: Feasibility study and comparison between conventional gradient echosequences and echo-planar imaging

Introduction: Functional brain MRI localizes the regions of the brain activated during sensorimotor or cognitive tasks using contrast based on local variations in blood flow and oxygenation. During neuronal activity, venous deoxyhaemoglobin concentration decreases leading to a signal increase due to a longer T2 which is better seen on long TE gradient echo (GE) sequences. Echo-planar imaging has an excellent temporal resolution, but is still technically difficult to perform and has a limited spatial resolution. Conventional imaging sequences are widely available and allow higher spatial resolution. We have compared a conventional GE sequence (SPGR), with echo-planar imaging (GE-EPI) in brain activation studies on a clinical 1.5 T scanner. Material and methods: Motor and visual stimulations were used in 5 healthy volunteers. The voxel size was 32 mm3 in EP1 and 4 or 16 mm3 in SPGR. The temporal resolution was 3 s in EPI and 9 s in SPGR. With each technique, 48 images were continuously acquired at the same location during alternating periods of stimulation or rest. Activation maps were calculated and the time course of the signal was analyzed. Results: Both sequences gave comparable sensitivity to activation in conditions of high signal to noise ratio. Higher resolution with SPGR was accompanied by higher signal changes during activation but also by larger baseline fluctuations. Activated areas were smaller and better delineated with SPGR than with EPI. Discussion: The high resolution of conventional sequences improves the anatomical localization of small activated areas. However, those sequences are sensitive to motion artifacts and to inflow effects and allow only single slice acquisitions. EPI remains the method of choice when temporal resolution is a priority (3D or multislice acquisition).