Large enhancement of perfusion contribution on fMRI signal

Department of Radiology, Center for Magnetic Resonance Research, Medical School, University of Minnesota, Minneapolis, MN 55455, USA.
Journal of cerebral blood flow and metabolism: official journal of the International Society of Cerebral Blood Flow and Metabolism (Impact Factor: 5.41). 03/2012; 32(5):907-18. DOI: 10.1038/jcbfm.2012.26
Source: PubMed


The perfusion contribution to the total functional magnetic resonance imaging (fMRI) signal was investigated using a rat model with mild hypercapnia at 9.4 T, and human subjects with visual stimulation at 4 T. It was found that the total fMRI signal change could be approximated as a linear superposition of 'true' blood oxygenation level-dependent (BOLD; T(2)/T(2)(*)) effect and the blood flow-related (T(1)) effect. The latter effect was significantly enhanced by using short repetition time and large radiofrequency pulse flip angle and became comparable to the 'true' BOLD signal in response to a mild hypercapnia in the rat brain, resulting in an improved contrast-to-noise ratio (CNR). Bipolar diffusion gradients suppressed the intravascular signals but had no significant effect on the flow-related signal. Similar results of enhanced fMRI signal were observed in the human study. The overall results suggest that the observed flow-related signal enhancement is likely originated from perfusion, and this enhancement can improve CNR and the spatial specificity for mapping brain activity and physiology changes. The nature of mixed BOLD and perfusion-related contributions in the total fMRI signal also has implication on BOLD quantification, in particular, the BOLD calibration model commonly used to estimate the change of cerebral metabolic rate of oxygen.

  • [Show abstract] [Hide abstract]
    ABSTRACT: The study of brain activation in small animals is of high interest for neurological research. In this study, we proposed a protocol to monitor brain activation in rats following whisker stimulation using the short half-life PET tracer [(15)O]H2O as a marker for cerebral blood flow. This technique enables the study of baseline and activation conditions in fast succession within the same scanning session. Furthermore, we compared the results obtained from PET imaging with additional BOLD-fMRI data acquired in the same animals within the same anesthetic session in immediate succession. Although the maximum relative signal changes during brain activity observed with PET were substantially higher compared to the BOLD-fMRI results, statistical analyses showed that the number of activated voxels in PET was lower compared to the fMRI measurements. Furthermore, there was a difference in the activation centers in both the shape and location between PET and fMRI. The discrepancy in the number of activated voxels could be attributed to a lower overall contrast-to-noise ratio of the PET images compared to BOLD-fMRI, whereas the difference in the spatial location indicates a more fundamental process, involving the different physiological origins of the PET and BOLD-fMRI response. This study clearly demonstrates that [(15)O]H2O-PET activation studies may be performed in small laboratory animals, and shows the complimentary nature of studying brain activation using [(15)O]H2O-PET and fMRI.
    No preview · Article · Dec 2013 · NeuroImage
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: A neuroimaging technique based on the saturation-recovery (SR)-T1 MRI method was applied for simultaneously imaging blood oxygenation level dependence (BOLD) contrast and cerebral blood flow change (ΔCBF), which is determined by CBF-sensitive T1 relaxation rate change (ΔR1CBF). This technique was validated by quantitatively examining the relationships among ΔR1CBF, ΔCBF, BOLD and relative CBF change (rCBF), which was simultaneously measured by laser Doppler flowmetry under global ischemia and hypercapnia conditions, respectively, in the rat brain. It was found that during ischemia, BOLD decreased 23.1±2.8% in the cortical area; ΔR1CBF decreased 0.020±0.004s-1 corresponding to a ΔCBF decrease of 1.07±0.24 ml/g/min and 89.5±1.8% CBF reduction (n=5), resulting in a baseline CBF value (=1.18 ml/g/min) consistent with the literature reports. The CBF change quantification based on temperature corrected ΔR1CBF had a better accuracy than apparent R1 change (ΔR1app); nevertheless, ΔR1app without temperature correction still provides a good approximation for quantifying CBF change since perfusion dominates the evolution of the longitudinal relaxation rate (R1app). In contrast to the excellent consistency between ΔCBF and rCBF measured during and after ischemia, the BOLD change during the post-ischemia period was temporally disassociated with ΔCBF, indicating distinct CBF and BOLD responses. Similar results were also observed for the hypercapnia study. The overall results demonstrate that the SR-T1 MRI method is effective for noninvasive and quantitative imaging of both ΔCBF and BOLD associated with physiological and/or pathological changes.
    Full-text · Article · Apr 2015 · PLoS ONE