• Yousef Mazaheri, Bharat B Biswal, B Douglas Ward, and James S Hyde.
• Department of Medical Physics and Radiology, Memorial Sloan Kettering Cancer Center, 1275 York Avenue, New York, NY 10021, USA. mazahery@mskcc.org
• Neuroimage. 2006 Aug 15; 32 (2): 603-15.

AbstractThe signal intensity during the dynamic approach to the equilibrium state of longitudinal magnetization is a function of sequence parameters, such as repetition time and flip angle, and depends on tissue characteristics, including longitudinal relaxation time of stationary tissue and the rate of blood inflow. A method is presented to extract information from data acquired during the transient state prior to T1 equilibrium using echo-planar acquisitions in T2*-weighted functional magnetic resonance imaging (fMRI) experiments. A voxel in a single slice acquisition is assumed to contain either stationary tissue or large vessels with flowing blood. Models are presented to characterize longitudinal magnetization relaxation of heterogeneous stationary tissue and blood inflow. The data were fitted to theoretical models for longitudinal relaxation of stationary tissue and inflowing blood assuming no residual signal prior to each RF excitation. Parameters were estimated at 3 T for each model using least squares estimation. A goodness-of-fit criterion was applied to exclude voxels that have transient data that does not fit the selected (best fit) model. Voxels that best fit the inflow model, measured at various TR and flip angles, were assumed to contain large draining veins and were excluded from functional maps. Histogram analysis of T1 distributions for activated voxels in a visual paradigm demonstrated the distributions are centered at T1 values of gray matter with tails at both sides of the center due to partial voluming of gray matter with white matter and CSF respectively. The mean gray matter volume fraction in activated voxels was about 0.9. The results indicate that transient data sets can provide additional information that is useful for both localization and characterization of the functionally relevant BOLD response.

36 keywords...

hide…