Back
Comparison of the BOLD hemodynamic response function at 3T and 9.4T
{Ultra-high-field (UHF) functional magnetic resonance imaging (fMRI) has shown tremendous potential for advancing non-invasive neuroscience in the human brain. To enable analysis of fMRI data at UHF, we need to measure the hemodynamic response function (HRF) at UHF and compare it to that obtained at conventional field strengths. Methods: 2-s duration visual stimulation consisted of round regions of flickering colored dots that appear for 0.67 s, followed by a second differently colored region of dots at a second location, then a third. Subjects (N \textequals 6) had to push a button with the color matching the dot display, a fast-paced and moderately demanding task. Each stimulus was followed by a 28-s period during which subjects performed a slow-paced color-detection task at fixation. 16 stimuli were presented per run, and 5 runs per session. fMRI data were collected using a point-spread-function-corrected EPI sequence that obtained 1.5-mm voxels at 3T, 1-mm at 9.4T, both at 1.25-s TR. We averaged the BOLD response across the 80 stimulus presentations to measure the HRF in visual cortex (V1), superior colliculus (SC), and lateral geniculate nucleus (LGN). Results: Good quality HRFs were measured in all regions at both field strengths (Figure, upper row). Cortical HRFs always had significant undershoots, but these were not observed in subcortical regions. No significant \textquotedblleftinitial dips\textquotedblright were observed. In V1, time-to-peak for 9.4T HRFs was significantly longer, 6.77$\pm$0.28 s, then was observed at 3T, 5.51$\pm$0.25 s (Figure, lower row). Subcortical HRFs were significantly faster than in cortex, but again 9.4T HRFs were slower (LGN 5.18$\pm$0.16 s, SC 4.77$\pm$0.22 s) than those measured at 3T (LGN 4.54$\pm$0.19 s, SC 4.08$\pm$0.19 s). Conclusions: HRFs recorded at UHF exhibit significantly slower temporal dynamics than at 3T. Also, sub-cortical HRFs are significantly faster than those observed in cortex, with minimal undershoot. Analysis of UHF and sub-cortical fMRI needs to take these differences into account to permit accurate linear analysis of experimental results.}
@misc{item_3005973, title = {{Comparison of the BOLD hemodynamic response function at 3T and 9.4T}}, booktitle = {{48th Annual Meeting of the Society for Neuroscience (Neuroscience 2018)}}, abstract = {{Ultra-high-field (UHF) functional magnetic resonance imaging (fMRI) has shown tremendous potential for advancing non-invasive neuroscience in the human brain. To enable analysis of fMRI data at UHF, we need to measure the hemodynamic response function (HRF) at UHF and compare it to that obtained at conventional field strengths. Methods: 2-s duration visual stimulation consisted of round regions of flickering colored dots that appear for 0.67 s, followed by a second differently colored region of dots at a second location, then a third. Subjects (N \textequals 6) had to push a button with the color matching the dot display, a fast-paced and moderately demanding task. Each stimulus was followed by a 28-s period during which subjects performed a slow-paced color-detection task at fixation. 16 stimuli were presented per run, and 5 runs per session. fMRI data were collected using a point-spread-function-corrected EPI sequence that obtained 1.5-mm voxels at 3T, 1-mm at 9.4T, both at 1.25-s TR. We averaged the BOLD response across the 80 stimulus presentations to measure the HRF in visual cortex (V1), superior colliculus (SC), and lateral geniculate nucleus (LGN). Results: Good quality HRFs were measured in all regions at both field strengths (Figure, upper row). Cortical HRFs always had significant undershoots, but these were not observed in subcortical regions. No significant \textquotedblleftinitial dips\textquotedblright were observed. In V1, time-to-peak for 9.4T HRFs was significantly longer, 6.77$\pm$0.28 s, then was observed at 3T, 5.51$\pm$0.25 s (Figure, lower row). Subcortical HRFs were significantly faster than in cortex, but again 9.4T HRFs were slower (LGN 5.18$\pm$0.16 s, SC 4.77$\pm$0.22 s) than those measured at 3T (LGN 4.54$\pm$0.19 s, SC 4.08$\pm$0.19 s). Conclusions: HRFs recorded at UHF exhibit significantly slower temporal dynamics than at 3T. Also, sub-cortical HRFs are significantly faster than those observed in cortex, with minimal undershoot. Analysis of UHF and sub-cortical fMRI needs to take these differences into account to permit accurate linear analysis of experimental results.}}, year = {2018}, slug = {item_3005973}, author = {Ress, D and Kim, J and Taylor, AJ and Scheffler, K and Hagberg, G and Himmelbach, M} }