Limits...
Cerebral blood flow measurement using fMRI and PET: a cross-validation study.

Chen JJ, Wieckowska M, Meyer E, Pike GB - Int J Biomed Imaging (2008)

Bottom Line: The CBF changes were compared using 3 types of region-of-interest (ROI) masks.FAIR measurements of CBF changes were found to be slightly lower than those measured with PET (average DeltaCBF of 21.5 +/- 8.2% for FAIR versus 28.2 +/- 12.8% for PET at maximum stimulation intensity).Finally, a t-test comparison of the slopes of the linear fits of PET versus ASL DeltaCBF for all 3 ROI types indicated no significant difference from unity (P > .05).

View Article: PubMed Central - PubMed

Affiliation: McConnell Brain Imaging Centre, Montreal Neurological Institute, McGill University, Montreal, PQ, Canada H3A 2B4. jean.chen@mail.mcgill.ca

ABSTRACT
An important aspect of functional magnetic resonance imaging (fMRI) is the study of brain hemodynamics, and MR arterial spin labeling (ASL) perfusion imaging has gained wide acceptance as a robust and noninvasive technique. However, the cerebral blood flow (CBF) measurements obtained with ASL fMRI have not been fully validated, particularly during global CBF modulations. We present a comparison of cerebral blood flow changes (DeltaCBF) measured using a flow-sensitive alternating inversion recovery (FAIR) ASL perfusion method to those obtained using H(2) (15)O PET, which is the current gold standard for in vivo imaging of CBF. To study regional and global CBF changes, a group of 10 healthy volunteers were imaged under identical experimental conditions during presentation of 5 levels of visual stimulation and one level of hypercapnia. The CBF changes were compared using 3 types of region-of-interest (ROI) masks. FAIR measurements of CBF changes were found to be slightly lower than those measured with PET (average DeltaCBF of 21.5 +/- 8.2% for FAIR versus 28.2 +/- 12.8% for PET at maximum stimulation intensity). Nonetheless, there was a strong correlation between measurements of the two modalities. Finally, a t-test comparison of the slopes of the linear fits of PET versus ASL DeltaCBF for all 3 ROI types indicated no significant difference from unity (P > .05).

No MeSH data available.


Related in: MedlinePlus

Correlation plots of group-average fMRI CBFpercent changes with respect to PET CBF changes in the V1, t-map, and GM ROIs. The dots represent results from baseline and 4levels of visual stimulation, and the triangles represent the results for the5% hypercapnia condition. The equations resulted from the linear fitting of thedata, represented by the dotted lines. The hypercapnic ΔCBF measured using PET and MR are similar. Inall 3 cases, the χ2 values have probabilities wellwithin the range of acceptance, and t-testresults indicate the slope of the linear fit is not significantly differentfrom unity.
© Copyright Policy - open-access
Related In: Results  -  Collection


getmorefigures.php?uid=PMC2553188&req=5

fig6: Correlation plots of group-average fMRI CBFpercent changes with respect to PET CBF changes in the V1, t-map, and GM ROIs. The dots represent results from baseline and 4levels of visual stimulation, and the triangles represent the results for the5% hypercapnia condition. The equations resulted from the linear fitting of thedata, represented by the dotted lines. The hypercapnic ΔCBF measured using PET and MR are similar. Inall 3 cases, the χ2 values have probabilities wellwithin the range of acceptance, and t-testresults indicate the slope of the linear fit is not significantly differentfrom unity.

Mentions: The ΔCBF values measured with FAIR and PET in allthe ROIs during baseline and 4 graded levels of visual stimulation arepresented in Figure 5. PET data is shown as having much higher standarddeviation than FAIR measurements, even when the GM ROIs were used. The group-averagedresults seem to provide a better indication of the CBF changes. Table 2 summarizesthe FAIR and PET ΔCBF group averages. The FAIR baseline signalvalue was obtained from the 6-minute baseline data, whereas in PET, since onlyone baseline volume was obtained, the baseline percent change was fixed at 0%.The regions of activation-induced ΔCBF in both the V1 and t-map ROIs were localized to the expected site of activation.However, as seen in Figure 6, the ΔCBF in the GM ROIs are lower, since many less thanmaximally activated voxels are likely to have been included in the mask. Inaddition, a postblurring data resolution of 14 × 14 mm2 implies thatboth GM and WM contribution can be expected in the same voxel; the GM signalintensity thus diminished in both PET and FAIR [3]. We further observed that forall ROIs, increases of FAIR ΔCBF appear to correspond well with increases invisual stimulation intensity, in agreement with previous observations [8]. This was not the case for PETdata. Finally, for the hypercapnic condition, ΔCBF in all three ROIs were similar for FAIR andPET.


Cerebral blood flow measurement using fMRI and PET: a cross-validation study.

Chen JJ, Wieckowska M, Meyer E, Pike GB - Int J Biomed Imaging (2008)

Correlation plots of group-average fMRI CBFpercent changes with respect to PET CBF changes in the V1, t-map, and GM ROIs. The dots represent results from baseline and 4levels of visual stimulation, and the triangles represent the results for the5% hypercapnia condition. The equations resulted from the linear fitting of thedata, represented by the dotted lines. The hypercapnic ΔCBF measured using PET and MR are similar. Inall 3 cases, the χ2 values have probabilities wellwithin the range of acceptance, and t-testresults indicate the slope of the linear fit is not significantly differentfrom unity.
© Copyright Policy - open-access
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC2553188&req=5

fig6: Correlation plots of group-average fMRI CBFpercent changes with respect to PET CBF changes in the V1, t-map, and GM ROIs. The dots represent results from baseline and 4levels of visual stimulation, and the triangles represent the results for the5% hypercapnia condition. The equations resulted from the linear fitting of thedata, represented by the dotted lines. The hypercapnic ΔCBF measured using PET and MR are similar. Inall 3 cases, the χ2 values have probabilities wellwithin the range of acceptance, and t-testresults indicate the slope of the linear fit is not significantly differentfrom unity.
Mentions: The ΔCBF values measured with FAIR and PET in allthe ROIs during baseline and 4 graded levels of visual stimulation arepresented in Figure 5. PET data is shown as having much higher standarddeviation than FAIR measurements, even when the GM ROIs were used. The group-averagedresults seem to provide a better indication of the CBF changes. Table 2 summarizesthe FAIR and PET ΔCBF group averages. The FAIR baseline signalvalue was obtained from the 6-minute baseline data, whereas in PET, since onlyone baseline volume was obtained, the baseline percent change was fixed at 0%.The regions of activation-induced ΔCBF in both the V1 and t-map ROIs were localized to the expected site of activation.However, as seen in Figure 6, the ΔCBF in the GM ROIs are lower, since many less thanmaximally activated voxels are likely to have been included in the mask. Inaddition, a postblurring data resolution of 14 × 14 mm2 implies thatboth GM and WM contribution can be expected in the same voxel; the GM signalintensity thus diminished in both PET and FAIR [3]. We further observed that forall ROIs, increases of FAIR ΔCBF appear to correspond well with increases invisual stimulation intensity, in agreement with previous observations [8]. This was not the case for PETdata. Finally, for the hypercapnic condition, ΔCBF in all three ROIs were similar for FAIR andPET.

Bottom Line: The CBF changes were compared using 3 types of region-of-interest (ROI) masks.FAIR measurements of CBF changes were found to be slightly lower than those measured with PET (average DeltaCBF of 21.5 +/- 8.2% for FAIR versus 28.2 +/- 12.8% for PET at maximum stimulation intensity).Finally, a t-test comparison of the slopes of the linear fits of PET versus ASL DeltaCBF for all 3 ROI types indicated no significant difference from unity (P > .05).

View Article: PubMed Central - PubMed

Affiliation: McConnell Brain Imaging Centre, Montreal Neurological Institute, McGill University, Montreal, PQ, Canada H3A 2B4. jean.chen@mail.mcgill.ca

ABSTRACT
An important aspect of functional magnetic resonance imaging (fMRI) is the study of brain hemodynamics, and MR arterial spin labeling (ASL) perfusion imaging has gained wide acceptance as a robust and noninvasive technique. However, the cerebral blood flow (CBF) measurements obtained with ASL fMRI have not been fully validated, particularly during global CBF modulations. We present a comparison of cerebral blood flow changes (DeltaCBF) measured using a flow-sensitive alternating inversion recovery (FAIR) ASL perfusion method to those obtained using H(2) (15)O PET, which is the current gold standard for in vivo imaging of CBF. To study regional and global CBF changes, a group of 10 healthy volunteers were imaged under identical experimental conditions during presentation of 5 levels of visual stimulation and one level of hypercapnia. The CBF changes were compared using 3 types of region-of-interest (ROI) masks. FAIR measurements of CBF changes were found to be slightly lower than those measured with PET (average DeltaCBF of 21.5 +/- 8.2% for FAIR versus 28.2 +/- 12.8% for PET at maximum stimulation intensity). Nonetheless, there was a strong correlation between measurements of the two modalities. Finally, a t-test comparison of the slopes of the linear fits of PET versus ASL DeltaCBF for all 3 ROI types indicated no significant difference from unity (P > .05).

No MeSH data available.


Related in: MedlinePlus