NeuroImage 50 (2010) 1036–1043
Contents lists available at ScienceDirect
NeuroImage j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y n i m g
Technical Note
Functional MR angiography using phase contrast imaging technique at 3T MRI Chang-Ki Kang a, Sang-Hoon Kim a, Hyon Lee a, Chan-A Park a, Young-Bo Kim a, Zang-Hee Cho a,b,⁎ a b
Neuroscience Research Institute, Gachon University of Medicine and Science, Incheon, 405-760, South Korea Department of Radiological Sciences, University of California, Irvine, CA 92697, USA
a r t i c l e
i n f o
Article history: Received 23 October 2009 Revised 14 December 2009 Accepted 10 January 2010 Available online 18 January 2010 Keywords: Functional angiography Phase contrast angiography Vascular response Velocity fMRI
a b s t r a c t Purpose: Phase contrast magnetic resonance angiography (PC MRA) is an important, non-invasive method for obtaining quantitative information on blood flow. The purpose of the present study was to investigate the dynamic cerebral arterial response in humans to visual stimulation by quantitative velocity analysis using PC MRA. Methods: Functional PC MRA (fPCA) images at 3 Tesla MRI were acquired with a total acquisition time of 10 m 18 s for 2 rest and 1 visual stimulation sessions using a flashing checker board. Twelve healthy subjects participated in this study. Circular regions of interest (ROIs) consisting of 21 pixels were analyzed to determine the velocity changes due to visual stimulation in selected vessel segments supplying the visual cortex. Time-of-flight (TOF) MRA and PC MRA reference volume images were then compared to measure the signal intensity and quantitative blood velocity changes in the cerebral vessels. Results: The mean velocity changes measured by PC MRA between rest and stimulation were 0.64 and 0.42 cm/s within the peak ROI and distal ROIs of the selected blood vessels, which corresponds to a 42.9% and 30.1% increase, respectively. Both TOF MRA and PC MRA reference volume images also exhibited an increase in the signal intensity of the target vessels during stimulation. Conclusion: This study demonstrated that fPCA could play a role in the quantitative analysis of the functional cerebral vascular response by measuring dynamic vascular changes in intensity and velocity in small arteries. © 2010 Elsevier Inc. All rights reserved.
Introduction Since Roy and Sherrington observed the relationship between cerebral blood flow and neuronal activity in 1890 (Roy and Sherrington, 1890), many physiological studies have been conducted to understand the mechanisms of the cerebral vascular response to neural stimulation (Iadecola, 2004). This neurovascular coupling or functional hyperemia has become the basis of several modalities in modern functional brain imaging, including functional magnetic resonance imaging (fMRI), positron emission tomography (PET), and transcranial doppler (TCD) (Attwell and Iadecola, 2002; Seitz and Roland, 1992). In the early 1990s, the introduction of blood oxygenation level dependent (BOLD) fMRI revolutionized functional brain mapping (Ogawa et al., 1990). However, BOLD fMRI only provides an indirect means to measure neural activity due to the complexity of its signal sources and factors that influence them, such as oxygen metabolism, cerebral blood flow, and blood volume (D'Esposito et al., 2003). Furthermore, the spatial specificity of BOLD contrast for the site of neural activity could be attenuated, since it is based on the detection of changes in deoxyhemoglobin concentration in vascular networks ⁎ Corresponding author. Neuroscience Research Institute, Gachon University of Medicine and Science, Incheon, South Korea, 1198 Kuwol-dong, Namdong-gu, Incheon, 405-760, South Korea. Fax: +82 32 460 8230. E-mail address:
[email protected] (Z.-H. Cho). 1053-8119/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2010.01.038
ranging from small capillaries to large draining veins (Duong et al., 2000; Menon et al., 1993). Recently several alternative approaches for the localization of neural activation, such as cerebral blood flow (CBF) fMRI based on arterial spin labeling (ASL), have also been proposed (Kim, 1995; Kwong et al., 1995). The ability to quantitatively measure CBF has made this an important technique for functional imaging, as the magnitude of CBF change has been shown to correlate with metabolic changes and the extent of neural activation (Duong et al., 2001; Seitz and Roland, 1992). However, this technique is hindered by a relatively low signal to noise ratio (SNR) and consequently low functional sensitivity. PET, a standard method in the assessment of cerebral perfusion, can also provide quantitative values for CBF, cerebral blood volume (CBV), oxygen extraction fraction (OEF) and the cerebral metabolic rate for oxygen consumption (CMRO2) (Heiss, 2000). However, the acquisition time of PET is often too long for the study of the hemodynamic response of functional brain activity. Furthermore, its limited spatial resolution prevents it from detecting the individual response of small vessels. TCD was shown to detect the instantaneous velocity of blood flowing through the basal cerebral arteries with excellent temporal resolution (Aaslid et al., 1982). It has been used to measure cerebral vascular dynamics and provide a quantitative index of cerebral blood flow velocity that correlates highly with regional CBF (Bishop et al., 1986). However, its spatial resolution is confined to measurements of
C.-K. Kang et al. / NeuroImage 50 (2010) 1036–1043
1037
the major cerebral arteries and can only elucidate lobular or hemispheric organization of neural activity (Deppe et al., 1997). More recently, the dynamic response of the small arteries supplying the motor cortex to finger tapping tasks was measured by functional magnetic resonance angiography (fMRA) using a 7 Tesla (T) MRI scanner (Cho et al., 2008). This study suggested the potential to measure the functional arterial response using time-of-flight (TOF) MRA sequences at 7T. However, 7T MRI is not widely available yet, and TOF MRA signals can only provide limited physiological information because its signal depends on several factors, including vascular directionality and the saturation effect of moving nuclear spins. In contrast to fMRA, quantitative velocity analysis of the vascular response to neural stimulation could be determined using phase contrast MRA (PC MRA). PC MRA employs dedicated motion sensitive bipolar gradients to quantitatively measure blood flow and phase shifts induced by moving nuclear spins, which are then used to calculate velocity (Moran, 1982; Moran et al., 1985). Consequently, in the arena of clinical and basic science research, PC MRA has become one of the most prevalent and noninvasive methods for velocity measurements in the cerebral vessels, and extensive validation in 0.5 T, 1.5 T (Bakker et al., 1999; Evans et al., 1993; Maier et al., 1989; Zananiri et al., 1991) and 3 T MRI systems (Harloff et al., 2009; Lotz et al., 2005) has been made. For functional studies, phase contrast MRI has been previously used to calculate the blood oxygen saturation response of pial veins near the motor cortex using the blood flow velocity (Haacke et al., 1997; Hoogenraad et al., 1998). In the present study, we applied the concept of phase contrast imaging to investigate the dynamic arterial response of the blood vessels supplying the occipital cortex in order to determine the potential of PC MRA as a functional technique. We also conducted 3 T fMRA imaging using TOF MRA in order to ascertain the previous fMRA study conducted at 7T MRI and compared these results with reference volume PC MRA data. Materials and methods Subjects and MR system Twelve healthy volunteers (ten males and two females, age range 21–31 years) from local universities participated in the study after signing a form of informed consent. The experiments were approved by the institutional review board (IRB). MR imaging was performed using a 3 T MRI scanner (Allegra, Siemens, Germany). All MR images were acquired using a custom-built quadrature transmit/receive (Tx/ Rx) surface coil (7 × 11 cm2 of a loop) covering the visual cortical area, which was designed for 3T MRI and tuned to 128 MHz in the presence of a human head (Fig. 1a). Subjects were positioned supine inside the bore of the magnet, instructed to pay attention to the checker board stimulation, and requested to avoid head movement during entire experiment. Functional brain imaging (PC MRA, fMRI and TOF MRA) PC MRA imaging consisted of one reference volume scan and 3 velocity encoded directional scans. PCA sequence parameters were optimized for velocity encoding as follows: repetition time (TR)/echo time (TE)/flip angle (FA) = 43.6 ms/6.7 ms/30°, field of view (FOV) = 113 × 172 mm, matrix size = 168 × 256, number of slices = 36, slice thickness = 1 mm, and 20 cm/s velocity encoding (VENC) for all three directions. Functional PC MRA (fPCA) images consisted of one stimulation scan between two rest scans, where each scan took 3 min 26 s. Functional MRI (fMRI) was also conducted to investigate the correlation of the BOLD activation area with PC MRA related vascular changes. An echo planar image (EPI) sequence was used for fMRI with TR/TE/FA = 3000/35/90°, FOV = 113 × 172, and a 2-mm isotropic
Fig. 1. Radiofrequency (RF) coil and scout image. A quadrature transmit/receive (Tx/Rx) surface coil used in this study is shown in panel (a) which was tuned to 128 MHz for 3T MRI. A scout image obtained with the surface coil is shown in panel (b). The white box indicates the ROI used for functional angiographic imaging.
voxel resolution. fMRI data was processed with SPM2 (Wellcome Department of Cognitive Neurology, London, UK). Active pixels above a family-wise error (FWE) corrected P-value threshold of 0.05 (cluster sizes = 20) were overlaid onto T1 anatomic images. In order to determine whether fMRA imaging using a 3T MRI scanner was consistent with observation published at 7T (Cho et al., 2008), a three dimensional (3D) fast low-angle shot (FLASH) gradient-echo sequence TOF MRA sequence was acquired with a TR/TE/FA = 20/3.1/30°, FOV = 113 × 172, matrix size = 168 × 256, number of slices = 52, and slice thickness = 0.67 mm with a coverage of about 35 mm in the visual cortex (Fig. 1b). A partial Fourier was applied in the phase-encoding and slice-encoding directions to further improve the temporal resolution. The same stimulation protocol was used for the fMRA scan as fPCA. The acquisition time was 1 min 40 s per scan. The main target vessels were the posterior cerebral arteries (PCA) and their branches supplying blood to visual cortex. Since PC MRA reference volume scans share similar imaging parameters with TOF MRA, the PC MRA reference images were also compared to the fMRA results at 3T. TOF and PC MRA reference images were evaluated with maximum intensity projections (MIP). The average signal intensities of the vessels of interest (VOI) during the visual task were compared to those at rest.
1038
C.-K. Kang et al. / NeuroImage 50 (2010) 1036–1043
Velocity analysis PC MRA measures three directional (x, y, z) flow velocities depending on the VENC gradients and one reference without a VENC gradient. The images acquired with flow encoding gradients were processed to extract the 3 directional components of flow, which were calculated by phase subtraction of the flow encoding images from the reference image. In this study, the VENC gradient was set to 20 cm/s for all three directions, since a low VENC has been shown to improve the visualization of small vascular structures (Conturo and Smith, 1990), and in our preliminary trials using a custom-built surface coil, the maximum velocity measured in the branches of the PCA was less than 20 cm/s (data not shown). For further analysis, we utilized a complex difference reconstruction technique, which has been shown to reduce the partial volume effect (Bernstein and Ikezaki, 1991) and to detect the functional response using magnitude and phase changes (Rowe, 2005). Since the selected VENC of 20 cm/s was sufficiently greater than the measured velocity within the distal target PCA branches, we assumed a linear relationship between the complex difference signal and flow velocity. PCA branches, the vessels supplying blood to the visual cortex, were selected for quantitative velocity analysis as shown in Fig. 2a. Background and tissue intensity values, collected from outside and inside the brain, respectively, were converted into velocities; the means of which were 0.95 ± 0.03 (MEAN ± SEM) and 1.09 ± 0.06 cm/s, respectively. Pixels exceeding 1.5 cm/s within the selected ROI were defined as vessels (Fig. 2b and c). After subtracting the background and tissue signals, quantitative velocity analysis was conducted on the remaining vessels by counting the number of pixels and by assigning the pixels to different groups based on their velocities, where Velocity Group #2 ranged from above 1.5 to 2.5 cm/s, Group #3 from above 2.5 to 3.5 cm/s, and so on (Fig. 2d). Each velocity group was expressed as a
ratio between the number of pixels in the group and the total number of pixels which exceed 1.5 cm/s during the rest scan. Vessels that were located close to the activation site established by fMRI were selected for the further analysis. MIP images between rest and visual tasks were acquired first. Circular ROIs containing 21-pixels, which correspond to 3.3 mm in diameter, were manually selected to segment the target vessels (Fig. 3a). The average velocity and its percent change were calculated (Fig. 3b). For each subject, we defined the ROI having the maximum velocity change as the peak ROI (the arrow in Fig. 3b) and all the distal vessel segments including the peak ROI as the distal ROIs (Groups IV to VII). The mean change in vessel diameter upon stimulation was also measured within the peak ROI segment as previously described (Cho et al., 2008). Statistical analysis The difference between rest and visual stimulation sessions within each subject was compared using the paired t-test. The number of pixels and their calculated velocity changes during stimulation were also analyzed using the paired t-test. Statistical significance was defined as P b 0.05. All statistical analysis was carried out using SPSS 15.0 (SPSS Inc., Chicago, IL). Results fMRA using TOF MRA at 3T showed signal intensity changes in the arterial vessels during visual stimulation as shown in Fig. 4a. Additionally, we directly compared TOF MRA images with PC MRA reference images as shown in Fig. 4b. Signal intensity changes within the selected vessel during stimulation were 11.6% and 9.0% for TOF MRA and PC MRA reference images, respectively (Fig. 4c).
Fig. 2. Description of the methodology used for velocity analysis. Panel (a) shows the MIP image obtained from a representative magnitude summed image and the manually selected ROI drawn to include the branches of the PCA (within the black square). A velocity histogram from a representative subject is displayed in panel (b). Pixels which exceed 1.5 cm/s (the vertical dashed line) were further analyzed as shown in the inset graph. For convenience, the plot only shows data up to 4 cm/s. Panel (c) shows background tissue and vessel images corresponding to values under (left panel) and over (right panel) the threshold value of 1.5 cm/s, respectively. The graph in panel (d) was plotted as the ratio of the number of pixels in each velocity group to the total number of pixels over the threshold value. Velocity Group #2 ranges from 1.5 to 2.5 cm/s, Group #3 ranges from 2.5 to 3.5 cm/s, and Group #8+ indicates a velocity of 7.5 cm/s and above.
C.-K. Kang et al. / NeuroImage 50 (2010) 1036–1043
1039
rest were 1.47 ± 0.08 and 1.36 ± 0.06 cm/s within the peak and distal ROIs, respectively. During stimulation the peak and distal ROI velocities increased to 2.11 ± 0.15 and 1.78 ± 0.08 cm/s, respectively (Fig. 7a; P b 0.001), which corresponded to an increase of 42.9 ± 4.12% and 30.1 ± 2.53% with respect to rest, respectively (Fig. 7b). The diameters in the peak ROI segments were 1.06 ± 0.24 mm and 1.72 ± 0.18 mm during rest and visual tasks, respectively, corresponding to a 62.3% diameter change (Fig. 7c; P b 0.001). Discussion
Fig. 3. Selection of peak and distal ROIs and the average velocity within each ROI. The schematic in panel (a) shows the circular ROIs selected for the analysis of vessels with large signal changes during stimulation. Each circle consisted of 21 pixels, corresponding to a diameter of 3.3 mm. Panel (b) shows a graph of the mean velocity (empty and filled circles) and its percent change between stimulation and rest (diamonds). The ROI with the greatest intensity change was defined as the peak ROI (arrow). In this case, ROI IV was defined as the peak ROI and ROIs IV to VII were defined as distal ROIs.
Conventional fMRI image exhibiting a typical activation map in the visual cortex during checker-board stimulation is shown in Fig. 5a. Velocity encoded PC MRA images clearly show the vessels supplying blood to the activated sites (Fig. 5b). At a lower imaging resolution of 2 mm per slice, vascular changes in vessels closely oriented to the area of activation by visual stimulation were visible (Fig. 5c). Imaging at a slice thickness of 1 mm resulted in greater visualization of the vascular changes in the small vessels close to the activation site (Fig. 5d). We calculated the overall average velocity change in the PCA and its branches during visual stimulation of 12 subjects. We found that 8.2% more pixels responded during visual stimulation with respect to rest (Fig. 6a; P b 0.001) and the average velocity in the pixels was 3.53 ± 0.07 cm/s, while the velocity at rest was 3.43 ± 0.06 cm/s. This corresponded to a 2.9% velocity increase during stimulation (Fig. 6b; P = 0.014). Pixels from the PC MRA images of the vessels during rest and stimulation were counted and stratified into velocity groups, and the pixel percentage change between rest and stimulation was calculated and plotted as shown in Fig. 6c. The results revealed that the greatest changes were observed in the pixels in Groups #7 and #8+ (17.26% and 17.67%, respectively), i.e. those vessels with the highest blood flow velocity as shown in Fig. 6d. Further analysis was performed on target vessels with high functional response using a 21-pixel circular ROI. Mean velocities at
In the present study, we have demonstrated that PC MRA imaging could be used to measure the dynamic arterial response of small arteries supplying blood to the occipital cortex during visual stimulation. Furthermore, the results showed that the presented method could provide quantitative information on dynamic blood flow velocity changes in small cerebral vessels during visual stimulation. Taking temporal resolution as well as SNR into account, we selected a slice thickness of 1 mm, at which level the small arteries could be well depicted within a reasonable time as shown in Fig. 5. However, the acquisition time increased from 1 min 40 s at a slice thickness of 2 mm to 3 min 26 s at 1 mm. We also conducted a number of trials at higher isotropic resolution (0.67 mm per slice); however, the acquisition time took approximately 5 min (data not shown). Further studies will be needed to optimize the temporal resolution as well vessel contrast for this technique. The mean velocity through the PCA and all its branches at rest (3.43 cm/s) and the change during visual stimulation (2.9%) in PC MRA were different from studies using TCD, which were 41.2 cm/s and 4.4%, respectively (Lisak et al., 2005). This difference may arise from the different imaging territories and the different imaging equipment used. While TCD only determines the flow through the main trunk of the PCA, the main contributors to the mean velocity determined by PC MRA were the small vessels closest to the surface coil, because of the increased sensitivity of the coil for vessels near the visual cortex. In preliminary tests using a volume coil, which better reflects the velocity in the main trunk of the PCA and its proximal branches, the average velocity in the main trunk as well as all the branching arteries of the PCA were 37.8 ± 4.9 and 11.95 ± 0.78 cm/s, respectively. In contrast, the calculated velocity of the main trunk (16.77 cm/s) and all the branches of the PCA (3.43 cm/s) was significantly lower when using the surface coil due to the low sensitivity of the coil for the center of the imaging field. The velocity values for the PCA using the volume coil were similar to previous reports that measured PCA velocity using transcranial Doppler (Lisak et al., 2005); however, the signal intensity was greatly diminished in the periphery making the visualization and accurate velocity analysis of the small arteries supplying the visual cortex difficult. In the present study, the maximum pixel changes were seen in the vessels with high blood flow like Velocity Group #7. In contrast, Doppler studies in animals have shown that, in general, vessel segments proximal to the site of neuronal activity exhibit the highest signal change during stimulation, which may correspond to the Group #2 (Iadecola et al., 1997). This might be due to the partial volume effect which is influenced by the imaging resolution, even though we used a complex difference reconstruction method to reduce this effect. This difference can also be explained by the inherent limitation of MIP for z-directional information (inferior-superior direction) leading to a diminished signal from the vasculature. In addition, the neurovascular response occurs within a few seconds following neural activation. However, the acquisition time of 3 min 26 s utilized for the fPCA scans in the present study only provides information on the average vascular dynamic effect. In order to investigate the details of functional hyperemia, temporal resolution and the aforementioned limitations need to be overcome.
1040
C.-K. Kang et al. / NeuroImage 50 (2010) 1036–1043
Fig. 4. fMRA response demonstrated by TOF and PC MRA reference images at 3T. Panel (a) shows representative MIP images obtained by TOF MRA scans during rest and visual stimulation followed by another rest session. The original MIP image (left) is displayed with expanded images showing the PCA and its branches during each session (right). Dashed ROIs were manually selected, and the averaged signal intensity change was calculated to quantify the vascular signal intensity change. (b) shows analogous MIP images obtained from reference volume PC MRA data. (c) The average signal intensity changes within the ROIs were plotted. The signal increased 11.6% during visual stimulation for TOF MRA and 9.0% for PC MRA.
However, the velocity changes in the vessels proximal to the neural activity sites (Fig. 7) are consistent with previous observations using Doppler, which showed a maximum change of 40% in the velocity through the vessels supplying the somatosensory cortex of rats (Ances et al., 1998, 1999). Previous studies also showed that the velocity in the pial veins of the motor cortex increased by 43.9% at motor activity (2.71 ± 1.44 cm/s at rest and 3.90 ± 1.91 cm/s at stimulation) (Haacke et al., 1997), which was similar to the 42.9% velocity change in the small arteries observed in the present study. The diameter change within the peak velocity ROI segment of the selected vessel upon stimulation was 62.3%, which was similar to previous TOF MRA-based studies (Cho et al., 2008). The change in the
number of voxels which was defined as a vessel was reflecting CBV changes, corresponding to 8.2% (Fig. 6a). Further studies aimed at improving in the spatial resolution and protocols dedicated for the CBF and/or CBV contributions could provide a more accurate distinction between them. Furthermore, we found that reference volume PC MRA data could supply data similar to that of fMRA data acquired with TOF MRA performed at 3 T. fMRA images showed the arterial changes during stimulation were consistent with previous studies conducted using a 7 T MRI scanner (Cho et al., 2008). The result revealed that the vascular response to visual stimulation correlated well using both sequences; however, the overall vessel visualization by TOF MRA
C.-K. Kang et al. / NeuroImage 50 (2010) 1036–1043
1041
Fig. 5. Activation maps of fMRI and MIP images of PC MRA during visual stimulation. A representative activation map acquired with EPI fMRI is shown in panel (a). The crossing point of the green lines indicates the area of peak activation. Panel (b) shows a representative MIP image from the magnitude summed image acquired by PC MRA. The green cross indicates the point in the plane corresponding to the peak activation determined by fMRI. Panels (c) and (d) show expanded images of the ROI vessels during rest and stimulation sessions at a slice thickness of 2 mm and 1 mm, respectively. Arrows indicate small vessels showing the significant responses to the stimulation.
Fig. 6. Changes in the overall number of pixels and velocity during stimulation. Panel (a) The average change in the number of pixels at rest and stimulation of 12 subjects increased by 8.2% during visual stimulation (⁎P b 0.001). The pixels over a velocity of 1.5 cm/s during the resting state were counted and normalized to one. Panel (b) Velocities were converted from the pixel intensities obtained during rest and stimulation. The average velocities were 3.43 and 3.53 cm/s during rest and visual stimulation, respectively (⁎P = 0.014). Note that the velocity was calculated using the following formula: VENC × √3 × Pixel signal intensity (Psi) / Max. Intensity = 20 × √3 × Psi / 4096. The histogram in panel (c) shows the number of pixel expressed as a ratio between rest and stimulation (white and black bars, respectively) for each velocity group and their percent change (grey filled diamonds). Panel (d) shows pixels with a velocity of 6.5 cm/s and over, corresponding to Velocity Groups #7 and #8+ (refer to the right image of Fig. 2c displaying pixels with a velocity of 1.5 cm/s and over).
1042
C.-K. Kang et al. / NeuroImage 50 (2010) 1036–1043
Fig. 7. Changes in velocity within selected ROIs during stimulation. Average velocities of the peak and distal ROIs of the PCA branches during rest and stimulation for 12 subjects are shown in panel (a). At the peak ROIs, the average velocities were 1.47 and 2.11 cm/s during the rest and visual stimulation, respectively. At distal ROIs, the velocities were 1.36 and 1.78 cm/s, respectively (⁎P b 0.001; ⁎⁎ P b 0.001). Panel (b) Average velocity changes of the peak and distal ROIs were 42.9% and 30.1% at the peak and distal ROIs, respectively. Panel (c) The average diameters within the peak ROI segments were 1.06 ± 0.24 mm and 1.72 ± 0.18 mm during the rest and stimulation states, respectively, which corresponds to a 62.3% increase (⁎P b 0.001).
was enhanced compared to the reference PC MRA data. This result might be due to different MR parameters used, especially the acquisition time of 1 min 40 s for TOF MRA compared to the approximately 51 s for the PC MRA reference scans. The percent signal change in vessels was consistent to previously published reports using 7T MRI (Cho et al., 2008), although there were methodological differences between this study and the 7T study, including ROI selection, coil sensitivity, and different tasks and cortical areas examined. These results demonstrate that fMRA is feasible at the more clinically available 3T MRI and can be used for the study of the arterial vascular response to neural activity. In addition, PC MRA reference images, which showed analogous change in the cerebral vasculature with TOF MRA, could also be used to elucidate functional arterial changes induce by neural stimulation. In summary, we demonstrated that quantitative changes in blood flow velocity in the branches of the PCA during checker-board stimulation could be measured using the noninvasive technique, PC MRA. We also demonstrated that fMRA imaging was feasible at 3T and showed that PC MRA reference images could also be used to detect dynamic vascular changes during visual stimulation in parallel with fMRA. Even with its limited spatial and temporal resolution, our data suggest that fPCA can be a promising quantitative technique for the investigation of neurovascular coupling in the human brain in vivo. Conflict of interest statement The authors declare no conflict of interest.
Acknowledgments This work was supported by a grant (A085136) of the Korea Healthcare Technology R&D Project from the Ministry for Health, Welfare & Family Affairs and a grant (2009K001285) from Brain Research Center of the 21st Century Frontier Research Program funded by the Ministry of Education, Science and Technology, of the Republic of Korea.
References Aaslid, R., Markwalder, T.M., Nornes, H., 1982. Noninvasive transcranial Doppler ultrasound recording of flow velocity in basal cerebral arteries. J. Neurosurg. 57, 769–774. Ances, B.M., Detre, J.A., Takahashi, K., Greenberg, J.H., 1998. Transcranial laser Doppler mapping of activation flow coupling of the rat somatosensory cortex. Neurosci. Lett. 257, 25–28. Ances, B.M., Greenberg, J.H., Detre, J.A., 1999. Laser Doppler imaging of activation-flow coupling in the rat somatosensory cortex. NeuroImage 10, 716–723. Attwell, D., Iadecola, C., 2002. The neural basis of functional brain imaging signals. Trends Neurosci. 25, 621–625. Bakker, C.J., Hoogeveen, R.M., Viergever, M.A., 1999. Construction of a protocol for measuring blood flow by two-dimensional phase-contrast MRA. J. Magn. Reson. Imaging 9, 119–127. Bernstein, M.A., Ikezaki, Y., 1991. Comparison of phase-difference and complexdifference processing in phase-contrast MR angiography. J. Magn. Reson. Imaging 1, 725–729. Bishop, C.C., Powell, S., Rutt, D., Browse, N.L., 1986. Transcranial Doppler measurement of middle cerebral artery blood flow velocity: a validation study. Stroke 17, 913–915. Cho, Z.H., Kang, C.K., Han, J.Y., Kim, S.H., Park, C.A., Kim, K.N., Hong, S.M., Park, C.W., Kim, Y.B., 2008. Functional MR angiography with 7.0 T is direct observation of arterial response during neural activity possible? NeuroImage 42, 70–75. Conturo, T.E., Smith, G.D., 1990. Signal-to-noise in phase angle reconstruction: dynamic range extension using phase reference offsets. Magn. Reson. Med. 15, 420–437. D'Esposito, M., Deouell, L.Y., Gazzaley, A., 2003. Alterations in the BOLD fMRI signal with ageing and disease: a challenge for neuroimaging. Nat. Rev. Neurosci. 4, 863–872. Deppe, M., Knecht, S., Henningsen, H., Ringelstein, E.B., 1997. AVERAGE: a Windows program for automated analysis of event related cerebral blood flow. J. Neurosci. Methods 75, 147–154. Duong, T.Q., Silva, A.C., Lee, S.P., Kim, S.G., 2000. Functional MRI of calcium-dependent synaptic activity: cross correlation with CBF and BOLD measurements. Magn. Reson. Med. 43, 383–392. Duong, T.Q., Kim, D.S., Ugurbil, K., Kim, S.G., 2001. Localized cerebral blood flow response at submillimeter columnar resolution. Proc. Natl. Acad. Sci. U. S. A. 98, 10904–10909. Evans, A.J., Iwai, F., Grist, T.A., Sostman, H.D., Hedlund, L.W., Spritzer, C.E., Negro-Vilar, R., Beam, C.A., Pelc, N.J., 1993. Magnetic resonance imaging of blood flow with a phase subtraction technique. In vitro and in vivo validation. Invest. Radiol. 28, 109–115. Haacke, E.M., Lai, S., Reichenbach, J.R., Kuppusamy, K., Hoogenraad, F.G., Takeichi, H., Lin, W., 1997. In vivo measurement of blood oxygen saturation using magnetic resonance imaging: a direct validation of the blood oxygen level-dependent concept in functional brain imaging. Hum. Brain Mapp. 5, 341–346. Harloff, A., Albrecht, F., Spreer, J., Stalder, A.F., Bock, J., Frydrychowicz, A., Schollhorn, J., Hetzel, A., Schumacher, M., Hennig, J., Markl, M., 2009. 3D blood flow characteristics
C.-K. Kang et al. / NeuroImage 50 (2010) 1036–1043 in the carotid artery bifurcation assessed by flow-sensitive 4D MRI at 3T. Magn. Reson. Med. 61, 65–74. Heiss, W.D., 2000. Ischemic penumbra: evidence from functional imaging in man. J. Cereb. Blood Flow Metab. 20, 1276–1293. Hoogenraad, F.G., Reichenbach, J.R., Haacke, E.M., Lai, S., Kuppusamy, K., Sprenger, M., 1998. In vivo measurement of changes in venous blood-oxygenation with high resolution functional MRI at 0.95 tesla by measuring changes in susceptibility and velocity. Magn. Reson. Med. 39, 97–107. Iadecola, C., 2004. Neurovascular regulation in the normal brain and in Alzheimer's disease. Nat. Rev. Neurosci. 5, 347–360. Iadecola, C., Yang, G., Ebner, T.J., Chen, G., 1997. Local and propagated vascular responses evoked by focal synaptic activity in cerebellar cortex. J. Neurophysiol. 78, 651–659. Kim, S.G., 1995. Quantification of relative cerebral blood flow change by flow-sensitive alternating inversion recovery (FAIR) technique: application to functional mapping. Magn. Reson. Med. 34, 293–301. Kwong, K.K., Chesler, D.A., Weisskoff, R.M., Donahue, K.M., Davis, T.L., Ostergaard, L., Campbell, T.A., Rosen, B.R., 1995. MR perfusion studies with T1-weighted echo planar imaging. Magn. Reson. Med. 34, 878–887. Lisak, M., Trkanjec, Z., Mikula, I., Demarin, V., 2005. Mean blood flow velocities in posterior cerebral arteries during visual stimulation. Mt. Sinai J. Med. 72, 346–350. Lotz, J., Doker, R., Noeske, R., Schuttert, M., Felix, R., Galanski, M., Gutberlet, M., Meyer, G.P., 2005. In vitro validation of phase-contrast flow measurements at 3 T in comparison to 1.5 T: precision, accuracy, and signal-to-noise ratios. J. Magn. Reson. Imaging 21, 604–610.
1043
Maier, S.E., Meier, D., Boesiger, P., Moser, U.T., Vieli, A., 1989. Human abdominal aorta: comparative measurements of blood flow with MR imaging and multigated Doppler US. Radiology 171, 487–492. Menon, R.S., Ogawa, S., Tank, D.W., Ugurbil, K., 1993. Tesla gradient recalled echo characteristics of photic stimulation-induced signal changes in the human primary visual cortex. Magn. Reson. Med. 30, 380–386. Moran, P.R., 1982. A flow velocity zeugmatographic interlace for NMR imaging in humans. Magn. Reson. Imaging 1, 197–203. Moran, P.R., Moran, R.A., Karstaedt, N., 1985. Verification and evaluation of internal flow and motion. True magnetic resonance imaging by the phase gradient modulation method. Radiology 154, 433–441. Ogawa, S., Lee, T.M., Kay, A.R., Tank, D.W., 1990. Brain magnetic resonance imaging with contrast dependent on blood oxygenation. Proc. Natl. Acad. Sci. U. S. A. 87, 9868–9872. Rowe, D.B., 2005. Modeling both the magnitude and phase of complex-valued fMRI data. NeuroImage 25, 1310–1324. Roy, C.S., Sherrington, C.S., 1890. On the regulation of the blood-supply of the brain. J. Physiol. 11, 85–158 117. Seitz, R.J., Roland, P.E., 1992. Vibratory stimulation increases and decreases the regional cerebral blood flow and oxidative metabolism: a positron emission tomography (PET) study. Acta Neurol. Scand. 86, 60–67. Zananiri, F.V., Jackson, P.C., Goddard, P.R., Davies, E.R., Wells, P.N., 1991. An evaluation of the accuracy of flow measurements using magnetic resonance imaging (MRI). J. Med. Eng. Technol. 15, 170–176.