Comparing the spatial and temporal reproducibility of brain activation using three fMRI techniques: BOLD, FAIR, and VASO

Abstract

In order to conduct the longitudinal fMRI study, measuring of the reproducibility of functional imaging is important. Our immediate study, we compared the spatial activation map reproducibility of three fMRI techniques with 8Hz checkerboard visual stimuli in three experiment sessions. The interval from experiment session one session two is one hour, and the interval form experiment session one to session three is three months. In addition to traditional blood oxygenation level dependent (BOLD) 11,121 fMRI, CBF and CBV fMRI measured by flow-sensitive alternating inversion recovery (FAIR) [3][4][5] and vascular space occupancy (VASO) [6] techniques were also used. The latter two developed techniques are more suitable in comparing the longitudinal variation of single subject or among subjects since the quantification capability of CBF and CBV. However, signal change of FAIR and VASO is smaller than BOLD (1/2 and 1/3, respectively) and may reduce the reproducibility of fMRI activation. An equivalent ROI of three techniques was selected in which contained all the activation areas located around primary visual cortex Percentage of activation overlapping was then calculated from three experiments. The mean overlapping percentage of 7 participants between experiment session one and two was 67.0%, 63.7%, and 65.2%, for BOLD, FAIR, and VASO respectively. There is no significant difference in reproducibility among these techniques. Although FAIR showed slightly better overlapping percentage (61.9%) than the percentage of BOLD and VASO (54.0% and 56.7%) in experiment session one and three, there is still no significant difference between three fMRI techniques. This result indicated that there is no significant difference of reproducibility between three fMRI techniques, although FAIR and VASO show smaller signal change than BOLD. © 2007 IEEE.