Skip to main content
NCBI home page
Journal of Cerebral Blood Flow & Metabolism logoLink to Journal of Cerebral Blood Flow & Metabolism
. 2010 Jan 13;30(5):913–922. doi: 10.1038/jcbfm.2009.284

Quantitative functional magnetic resonance imaging of brain activity using bolus-tracking arterial spin labeling

Michael E Kelly 1,2, Christoph W Blau 1, Karen M Griffin 3, Oliviero L Gobbo 4, James FX Jones 3, Christian M Kerskens 1,2,*
PMCID: PMC2949184  PMID: 20068581

Abstract

Blood oxygen level dependent (BOLD) functional magnetic resonance imaging (fMRI) is the most widely used method for mapping neural activity in the brain. The interpretation of altered BOLD signals is problematic when cerebral blood flow (CBF) or cerebral blood volume change because of aging and/or neurodegenerative diseases. In this study, a recently developed quantitative arterial spin labeling (ASL) approach, bolus-tracking ASL (btASL), was applied to an fMRI experiment in the rat brain. The mean transit time (MTT), capillary transit time (CTT), relative cerebral blood volume of labeled water (rCBVlw), relative cerebral blood flow (rCBF), and perfusion coefficient in the forelimb region of the somatosensory cortex were quantified during neuronal activation and in the resting state. The average MTT and CTT were 1.939±0.175 and 1.606±0.106 secs, respectively, in the resting state. Both times decreased significantly to 1.616±0.207 and 1.305±0.201 secs, respectively, during activation. The rCBVlw, rCBF, and perfusion coefficient increased on average by a factor of 1.123±0.006, 1.353±0.078, and 1.479±0.148, respectively, during activation. In contrast to BOLD techniques, btASL yields physiologically relevant indices of the functional hyperemia that accompanies neuronal activation.

Keywords: aging, animal studies, functional MRI (fMRI), neurovascular coupling, perfusion weighted MRI

Introduction

Neurovascular coupling, defined as the close spatial and temporal relationship between neural activity and cerebral blood flow (CBF), forms the basis for neuroimaging techniques that map regional changes in the brain activity. The increase in local CBF that follows neural activity (Leniger-Follert and Hossmann, 1979) is mediated by complex interactions between neurons, glial cells, and vascular smooth muscle cells (D'Esposito et al, 2003) and can be used to provide contrast between an active brain region and surrounding resting tissue. Functional magnetic resonance imaging (fMRI) has emerged as the foremost method for visualizing these vascular correlates of brain activation.

The first fMRI maps of brain activation were generated by measuring the first passage of an exogenous intravascular contrast agent, termed as dynamic susceptibility contrast magnetic resonance imaging (Belliveau et al, 1991). The original technique provided regional cerebral blood volume (rCBV) maps of the brain during photic stimulation of the human visual cortex. Dynamic susceptibility contrast magnetic resonance imaging has since become a useful method for perfusion quantification that enables the quantification of mean transit time (MTT), CBF, and CBV, and changes in these parameters during focal increases in neural activity have been reported in both animal and human studies (Mandeville et al, 1998; Li et al, 2000). However, because contrast administration is required for each measurement, the use of dynamic susceptibility contrast magnetic resonance imaging for fMRI applications has declined.

Most present-day fMRI studies apply the blood oxygenation level dependent (BOLD) technique (Ogawa et al, 1990). The BOLD signal depends on the change in concentration of deoxyhemoglobin in the microvasculature when a neural event occurs. The increase in local blood flow in the activated region results in a decrease in the concentration of deoxyhemoglobin and this in turn leads to an increase in the BOLD signal, which can be imaged by a suitable magnetic resonance imaging protocol (Bandettini et al, 1992).

The physiologic basis of the BOLD contrast mechanism is complex and the quantification of the changes in parameters such as CBF, CBV, and the metabolic rate of oxygen consumption (CMRO2) using this technique has proven difficult. A number of quantitative approaches for the BOLD technique have been proposed (Buxton and Frank, 1997; Davis et al, 1998; Hyder et al, 1998), but these approaches typically rely on assumptions about a number of parameters, which cannot be determined directly with noninvasive MRI techniques (Li et al, 2000).

Alterations in the BOLD signal with aging and disease provide a considerable challenge for the interpretation of BOLD fMRI studies (D'Esposito et al, 2003). For example, it is well known that Alzheimer's disease is associated with a decrease in cerebral microvessel density, flattening of endothelial cells, and smooth muscle cell degeneration. Each of these effects has a function in reducing both resting CBF and the increase in CBF produced by activation (Hock et al, 1997). It follows that if the BOLD fMRI technique is not sensitive to alterations in brain pathologic study, it may not be the optimum technique to assess neural activity in cohorts of aged and or diseased subjects.

To improve the quantitative information from BOLD studies, a calibration technique and corresponding deoxyhemoglobin dilution model that quantifies the CMRO2 has been proposed (Davis et al, 1998). This approach estimates the maximum possible BOLD fMRI signal change by inducing either hypercapnia or hyperoxia. These states cause an increase in CBF without increasing the CMRO2 and thereby calibrate the BOLD signal for quantification of CMRO2. A recent review of these techniques (Bulte et al, 2009) suggests that inspiration of a CO2/O2 mixture should not be used for calibration, as the stimuli can produce BOLD signal changes that cannot be accounted for by a model based purely on CMRO2 consumption. This, coupled with the invasiveness of the technique, particularly in aged or diseased subjects, is a significant disadvantage of the quantitative BOLD approach. More recently, noncapnic and nonhypoxic challenges have also been applied for calibration of the BOLD signal. For example, BOLD measurements have been combined with electrophysiologic recordings during electrical stimulation of the rat forepaw to establish the relationship between hyperemic components and neural activity (Sanganahalli et al, 2009). This approach has identified a component of the BOLD signal that can be attributed to changes in CMRO2.

Arterial spin labeling (ASL) (Detre et al, 1992) uses tagged arterial blood as an endogenous contrast agent as opposed to the change in deoxyhemoglobin concentration used by the BOLD technique and has been applied to fMRI studies in the rat brain (Kerskens et al, 1996). Activation imaging using ASL has several potential advantages over the BOLD fMRI technique. Arterial spin labeling is insensitive to magnetic susceptibility effects that can degrade BOLD studies. By interleaving labeled images with control images, long-term instabilities in the image intensity can be alleviated (Siewert et al, 1996). This makes ASL more suitable for longitudinal studies in which changes in the perfusion response to neural activity are tracked over time. There have also been indications that the ASL signal is better localized at the site of neural activity than the BOLD signal (Wang et al, 2003). However, the primary advantage of ASL fMRI over BOLD fMRI is the fact that the signal is directly related to blood flow, not blood oxygenation, and, therefore, contains information that can be extracted to quantify the perfusion of active brain regions. Changes in physiologic parameters such as CBF during neural activity can be accurately quantified and tracked over time.

The quantification of CBF (mL/g/min) during neuronal activation is typically achieved with ASL by applying the modified Bloch equation approach (Kwong et al, 1992). This formulation assumes that the exchange of water molecules between the capillaries and the extravascular water is rapid. However, it has been well established through the use of diffusion sensitive gradients in combination with ASL that the fraction of labeled arterial water exchanging with tissue water decreases from 0.9 at normal flow rates (1.0 mL/g/min) to as low as 0.6 at elevated flow rates (4.0 mL/g/min) (Silva et al, 1997). As a result, the assumption of near perfect capillary permeability to water leads to errors in perfusion quantification when CBF is elevated because of neural activity.

A recently developed technique, bolus-tracking ASL (btASL) (Kelly et al, 2009a), is applied in this study to ASL data acquired during neuronal activation in the left primary somatosensory cortex forelimb (S1FL) region of the rat brain. This noncompartmental approach to ASL perfusion quantification is based on a Fokker–Planck equation for the variation in the distribution of labeled spins within the brain over time. The MTT, capillary transit time (CTT), relative cerebral blood volume of labeled water (rCBVlw), relative cerebral blood flow (rCBF), and perfusion coefficient, P, are quantified without the use of constants such as the blood–brain partition coefficient, λ, or the permeability-surface area product on which compartmental models depend (Parkes and Tofts, 2002). Variations in these parameters during neuronal activation can provide a novel quantitative assessment of the changes in blood flow dynamics in an activated brain region.

Materials and methods

Theory

A theoretical model to facilitate the quantification of cerebral perfusion with ASL has been described in detail earlier (Kelly et al, 2009a). The model is based on the following Fokker–Planck equation for the concentration of labeled spins, c, within the entire volume, V, from the labeling plane to the image plane:

graphic file with name jcbfm2009284e1.jpg

This equation describes the variation of c within V over time using three terms. The first term incorporates transport of labeled spins because of bulk flow, F. The second term, which includes the perfusion coefficient, P, takes into account random effects such as pseudo-diffusion within the microvasculature (Kim and Kim, 2006) and the exchange of labeled spins between the capillary bed and extravascular water. These effects result in the dispersion of the labeled spins on traveling from the labeling plane to the image plane. The third term represents the T1 relaxation of the labeled spins.

The btASL (described below) can be described by a rectangular input function or bolus, c0(t), defined by c0(t)=C0(Θ(t)−Θ(t−τ)) where C0 is the initial concentration of inflowing spins at the labeling plane and Θ(t) is a rectangular input function defined as Θ(t) for t⩾0 and Θ(t)=0 for t<0. The solution to Equation (1) for a rectangular input function has been found (Kelly et al, 2009a):

graphic file with name jcbfm2009284e2.jpg

From this it can be seen that the solution has the conventional structure of a residue detection experiment with bolus dispersion (Petersen et al, 2006) and is equivalent to the convolution of an arterial input function (AIF) with a relaxation function, m(t), and residue function, r(V,t). The MTT and CTT can be calculated from the first and second moments of r(V,t), respectively. The MTT is calculated from the first moment of r(V,t) as

graphic file with name jcbfm2009284e3.jpg

which is defined as the average time it takes a particle to traverse the vasculature (Meier and Zierler, 1954). The CTT is calculated from the second moment of r(V,t) as

graphic file with name jcbfm2009284e4.jpg

which has been interpreted as the time taken for labeled arterial water to be distributed at the region of interest (ROI) (Kelly et al, 2009a).

Anesthesia and Animal Preparation

Female Wistar rats (n=5, 250 to 350 g body mass) aged between 3 and 5 months were used for all experiments in accordance with the protocols approved by the Trinity College Dublin and University College Dublin ethics committees and the Irish Department of Health and Children. The rats were initially anesthetized using isoflurane gas in oxygen administered through a facemask (3% at 1 L/min of oxygen) and a 22-gauge subcutaneous plastic cannula (Introcan. B. Braun, Sligo, Ireland) was placed between the shoulder blades. Isoflurane was discontinued after 15 to 30 mins and the rats were switched to sedation with a continuous subcutaneous infusion of medetomidine (loading dose bolus of 0.035 mg/kg followed by continuous infusion at 0.1 mg/kg/hr), which has been shown to provide suitable conditions for fMRI studies in rats (Weber et al, 2006) and is used for the first time with ASL in this study.

Temperature was monitored with a rectal thermometer and maintained using a warming surface controlled by a water pump-driven temperature regulator. The respiration signal was continuously monitored using custom hardware and software (SA Instruments Inc., Stony Brook, NY, USA). The electrical stimulation resulted in neuronal activation when the respiration rate was in the range of 60 to 80 breaths per minute. The partial pressure of carbon dioxide (pCO2) for each animal was measured using a calibrated transcutaneous blood gas analyzer (TCM4, Radiometer Copenhagen, Willich, Germany) before commencing the magnetic resonance imaging scanning. The mean pCO2 for the 60 to 80 breaths per minute respiration rate range was 49.1±6.7 mm Hg.

Anatomic Scan

All imaging was carried out on a 7T, preclinical magnetic resonance imaging system (Biospec 70/30 USR, Bruker Biospin, Ettlingen, Germany). The imaging slice with maximal S1FL coverage was located by comparing images acquired using a rapid acquisition with relaxation enhancement high resolution anatomic scan with the rat brain atlas. The following parameters were used for the rapid acquisition with relaxation enhancement acquisitions: slice thickness=2 mm, echo time, TE=12 ms, repetition time, TR=3.134 secs, FOV=3.0 × 3.0 cm, image matrix=256 × 256. This imaging slice was then used for the subsequent ASL sequence.

Arterial Spin Labeling Sequence

The btASL sequence described earlier (Kelly et al, 2009a) was used to provide signal–time curves of the passage of a 3-secs bolus through the S1FL region. The sequence consisted of a 5-secs preparation interval containing the inversion pulse, followed by snapshot fast low angle shot image acquisition (Kerskens et al, 1996). A 7.2-cm diameter bird-cage linear resonator, optimally tuned to 300.3 MHz, coil was used for transmission of the ASL and fast low angle shot excitation pulses. Signal detection was performed using an actively decoupled, circularly polarized, 20 mm diameter surface coil.

During the labeling phase of the ASL measurement, the principle of flow-induced fast adiabatic passage (Dixon et al, 1986) was used to supply inverted arterial spins to the imaging location. A rectangular pulse with maximum B1 amplitude of 120 mG was approximated by 10 shorter pulses with durations defined by a duty cycle of 80% to reduce demands on the RF amplifier. The RF power of the inversion pulse was set to achieve inversion at the desired location. The pulse had a bandwidth of 2.8 kHz and the gradient strength was set to 14 mT/m to provide an inversion region thickness of 4.7 mm. The pulse frequency was offset by −12 kHz, resulting in a tagging location 2 cm proximal to the imaging slice. A control image with the offset frequency reversed (+12 kHz) was also acquired, in which inflowing spins were left undisturbed. Corresponding pairs of labeled and control images were subtracted to provide perfusion-weighted maps (Figure 1). The preparation interval contained the inversion pulse and two variable delays. The delays and the inversion pulse duration were varied (Kelly et al, 2009a) to allow the full signal–time curve to be plotted (Figure 2).

Figure 1.

Figure 1

Open in a new tab

(A) ASL perfusion map acquired during electrical stimulation of the forepaw of animal 1. The corresponding increase in cerebral perfusion in the right somatosensory cortex can be clearly identified. (B) ASL perfusion map acquired in resting state (control). ASL perfusion maps for animals 2–5 can be found in Supplementary Figure 1.

Figure 2.

Figure 2

Open in a new tab

Least-squares fit of ASL data (yellow and red asterisks for control and activation experiments, respectively) to theoretical model (blue and green solid lines for control and activation experiments, respectively) for the left S1FL of animal 1. LSF results for animals 2–5 can be found in Supplementary Figure 2.

The snapshot fast low angle shot imaging was acquired with centric phase encoding to enhance sensitivity to contrast provided by inflowing labeled magnetization (Holsinger and Riederer, 1990). The following parameters were used: slice thickness=2 mm, TR=6.66 ms, TE=2.99 ms, RF flip angle=30°, field of view=3.0 × 3.0 cm, image matrix=128 × 64, receiver bandwidth=50 kHz. The total time for the ASL preparation and image acquisition for a single measurement point was 5.426 secs.

Stimulation Protocol

A square pulse nerve and muscle stimulator (Grass Technologies Inc., West Warwick, RI, USA) was used to electrically stimulate the right forepaw (3 V, 5 Hz, 5 ms duration pulses). This resulted in neuronal activation in the S1FL region, at +0.2 mm Bregma (Figure 1). Stimulation was started 10 secs before the beginning of the btASL sequence and was maintained throughout the labeling phase of the experiment, which was 59.7 secs in duration. This was followed by a rest period of 5 mins and was repeated six times for signal averaging.

Before the acquisition of the btASL signal–time curves (Figure 2), the ASL sequence was used to ensure that the electrical stimulus was providing a consistent hemodynamic response in the activated S1FL region. The inversion pulse was fixed at the position within the preparation interval that produces the maximum perfusion-weighted signal in the btASL signal–time curves (Figure 2). Eight perfusion-weighted images were acquired during 60 secs of electrical stimulation to investigate the variations in the ASL signal within the activated S1FL region during stimulation, with the results for each animal shown in Figure 3.

Figure 3.

Figure 3

Open in a new tab

Variation in ASL signal in the activated left S1FL region during 60 secs stimulation of the right forepaw animal 1 to 5.

Curve Fitting and Calculation of Transit Times

The activated left S1FL region was manually selected as the ROI in each dataset using the ImageJ ROI tool (Rasband W.S., Bethesda, MD, USA). Concentration–time curves for this ROI were created by calculating the mean signal intensity within the ROI and plotting the change in signal versus time. The resultant curves from the left S1FL region during electrical stimulation (red asterisks) and in the resting state (yellow asterisks) are shown in Figure 2 for one animal. To fit the theoretical model to these curves, the solution in Equation (2) was rewritten as follows (Kelly et al, 2009a):

graphic file with name jcbfm2009284e5.jpg

where erfc is the complementary error function and τ is the bolus duration and T1 is the longitudinal relaxation time in the S1FL region, as measured by the rapid acquisition with relaxation enhancement sequence with variable repetition time. An average T1 value of 1.74 secs was measured for the S1FL region (n=5) and used in the fitting procedure.

To facilitate quantification of the transit times from the curve-fitting procedure, Equation (5) was parameterized by three fitting parameters, A[0], A[1], and A[2] as follows: A[0]=c0/2, A[2]=MTT/(2.CTT), and A[2]=CTT/4. The curve-fitting routine in Mathematica (Wolfram Research Inc., Version 5.1, Champaign, IL, USA) was used to find the least-squares fit of Equation (6) to the experimental concentration–time curves by the Levenberg–Marquardt method.

Estimation Regional Cerebral Blood Volume of Labeled Water

The fitting parameter A[0]=c0/2, which represents the amplitude of the fitted curves in Figure 2, is directly proportional to the area under the curve or zeroth moment of the curve defined by Equation (2). A[0] was, therefore, used to estimate the rCBVlw in the left S1FL during neuronal activation and in the resting state. It should be noted that the subscript lw is used here to signify that the ASL signal is composed of contributions from both the intra- and extravascular spaces (Silva et al, 1997), and as a result, the volume estimated from the area under the ASL signal–time curve is not purely intravascular in nature.

A[0] was normalized by the degree of inversion (DOI), α, for each experiment to provide the final measure of rCBVlw (Table 2). The left and right middle cerebral arteries were clearly identifiable from the ASL perfusion maps (Figure 1). For both arteries, the pixel with the maximum signal intensity was selected and the DOI was estimated from the ratio of the signal intensity obtained in the labeling phase of the ASL experiment to the signal intensity obtained in the same arteries in the control phase. The final values for α, shown in Table 2, were calculated from the average of the values obtained for the left and right middle cerebral arteries and were in the range of 84% to 96%, which is typical for the flow-induced adiabatic inversion technique (Zhang et al, 1993).

Regional Extent of Neuronal Activation

To establish the regional extent of the neuronal activation in terms of the MTT, CTT, and rCBVlw, pixel-by-pixel maps of the perfusion parameters were created. The maps were formed by creating a signal–time curve for each pixel within the brain and performing the least-squares fit of Equation (5) to these signal–time curves on a pixel-by-pixel basis. Figures 4A, C, and E show the MTT, CTT, and rCBVlw maps obtained during electrical stimulation of the right forepaw of animal 1. The decrease in both the MTT and CTT and the increase in the rCBVlw in the activated left S1FL region can be clearly identified. Figures 4B, D, and F show the MTT, CTT, and rCBVlw maps obtained in the resting state (control).

Figure 4.

Figure 4

Open in a new tab

Pixelwise maps of the MTT, CTT, and rCBVlw of animal 1 during neuronal activation (A, C, E) and in the resting state (B, D, F). The color bars are in units of time (s) for the MTT and CTT maps and in arbitrary units (a.u.) for the rCBVlw maps. The decrease in both the MTT and CTT and the increase in rCBVlw can be clearly identified in the activated S1FL region.

Measured Change in Regional Cerebral Blood Flow and Perfusion Coefficient

The change in rCBF and P in the left S1FL during neuronal activation can be calculated from the measured changes in MTT, CTT, and rCBVlw. First, from the definition of the MTT in Equation (3), the ratio of the rCBF during the activation experiment, rCBFact, to the rCBF during the control experiment, rCBFctrl, is given by

graphic file with name jcbfm2009284e6.jpg

Similarly, from the definition of CTT in Equation (4), the ratio of the perfusion coefficient during the activation experiment, Pact, to the perfusion coefficient during the control experiment, Pctrl, is given by

graphic file with name jcbfm2009284e7.jpg

Results

Variation in Transit Times During Neuronal Activation

Table 1 shows the individual MTT and CTT values obtained from the fitting parameters of the least-squares fit. The error in the individual transit times was calculated from the standard error of the curve-fitting procedure. Both transit times were found to decrease in all subjects during neuronal activation. The MTT decreased on average from 1.939±0.175 secs (control) to 1.616±0.207 secs (activation) and the CTT decreased on average from 1.606±0.106 secs (control) to 1.305±0.201 secs (activation) (errors represent one standard deviation from the mean). These results are illustrated in Figures 5A and B, in which a two-tailed paired t-test was used to test for statistical significance in the magnitude of the decrease in both transit times. The level for statistical significance (P-value) was set to 0.01. A significant decrease in both transit times was discovered, with P=0.0012 for the MTT and P=0.0082 for the CTT.

Table 1. MTT and CTT measured during resting state (control) and neuronal activation.

Animal Control Activation
  MTT (secs) CTT (secs) MTT (secs) CTT (secs)
1 1.976±0.122 1.527±0.091 1.641±0.056 1.048±0.035
2 1.949±0.140 1.476±0.103 1.518±0.102 1.171±0.075
3 2.202±0.121 1.709±0.092 1.965±0.125 1.555±0.096
4 1.737±0.182 1.606±0.161 1.506±0.107 1.428±0.096
5 1.833±0.128 1.711±0.115 1.450±0.132 1.321±0.167
Open in a new tab

Errors in individual transit times are calculated from the standard error of the curve fitting by applying Gaussian error propagation.

Figure 5.

Figure 5

Open in a new tab

(A) Change in MTT, (B) CTT, and (C) rCBVlw during neuronal activation. A statistically significant decrease in both transit times and a statistically significant increase in rCBVlw was measured in the activated S1FL when compared with the same S1FL region in the control experiment (two-tailed paired t-test: P=0.0012 for MTT, P=0.0082 for CTT and, P=0.0026 for rCBVlw). The error bars represent one standard deviation from the mean.

Variation in Relative Cerebral Blood Volume of Labeled Water, Relative Cerebral Blood Flow, and Perfusion Coefficient During Neuronal Activation

Table 2 shows the values obtained for the change in rCBVlw during neuronal activation. The rCBVlw was found to increase on average from 0.086±0.003 to 0.096±0.005 arbitrary units, which corresponds to an increase in rCBVlw during neuronal activation by a factor of 1.123±0.034. This result is illustrated in Figure 5C, in which a two-tailed paired t-test was used to test for statistical significance in the magnitude of the increase in the rCBVlw. A significant increase in the rCBVlw in the S1FL region during neuronal activation was discovered with P=0.0026. The increase in rCBF and P resulting from the measured changes in MTT, CTT, and rCBVlw in the activated S1FL region were calculated using Equations (6) and (7), respectively, with the results for each animal presented in Table 3. The rCBF was found to increase on average by a factor of 1.353±0.078 and P was found to increase on average by a factor of 1.479±0.148 during neuronal activation.

Table 2. Calculation of the increase in rCBVlw in the S1FL during neuronal activation.

Animal C0 α rCBVlw(act)/rCBVlw(ctrl)
  Control Activation Control Activation  
1 0.075±0.010 0.088±0.008 0.871±0.051 0.912±0.039 1.121±0.158
2 0.076±0.012 0.086±0.014 0.910±0.027 0.921±0.052 1.118±0.222
3 0.070±0.008 0.079±0.010 0.847±0.032 0.854±0.041 1.119±0.167
4 0.073±0.016 0.085±0.012 0.840±0.087 0.865±0.058 1.131±0.255
5 0.085±0.010 0.094±0.016 0.957±0.044 0.939±0.069 1.127±0.203
Open in a new tab

rCBV is calculated from the amplitude fitting parameter, C0, normalized by the degree of inversion, α. Errors are calculated from the standard error of the curve fitting results by applying Gaussian error propagation.

Table 3. Calculation of factor by which rCBF and the perfusion coefficient, P, increases in the S1FL during neuronal activation.

Animal MTTctrl/ MTTact CTTact/ CTTctrl rCBVact/ rCBVctrl Pact/Pctrl
1 1.204±0.071 0.686±0.068 1.350±0.173 1.250±0.186
2 1.284±0.099 0.793±0.095 1.435±0.243 1.634±0.261
3 1.121±0.084 0.910±0.082 1.254±0.187 1.430±0.204
4 1.153±0.126 0.889±0.121 1.305±0.285 1.513±0.310
5 1.264±0.108 0.772±0.108 1.425±0.233 1.567±0.257
Open in a new tab

Ratios of control and activation rCBF and P are calculated using Equations (6) and (7), respectively. Errors are calculated from the standard error of the curve fitting results by applying Gaussian error propagation.

Discussion

The functional btASL technique introduced in this study provides a unique insight into the hemodynamic response to neuronal activation. The findings show the ability of the technique to simultaneously quantify variations in MTT, CTT, rCBF, rCBVlw, and the perfusion coefficient, P, during focal increases in neural activity. The measured decrease in both transit times and increase in the rCBVlw corresponds to an increase in rCBF and P in the somatosensory cortex of the rat brain during electrical stimulation of the forepaw.

The increases in rCBF and rCBVlw (35.3±7.8% and 12.3±1.1%, respectively) were expected, as increases in these parameters during neuronal activation have been measured earlier by both contrast agent techniques and conventional ASL approaches. In a contrast agent study by Mandeville et al, 1998, an increase in focal CBV of 24% was reported in the somatosensory cortex during electrical stimulation of the rat forepaw. A human study comparing the hemodynamic response measured by contrast agent bolus-tracking, ASL, and BOLD measured increases in rCBV and rCBF increases of 19% and 35%, respectively, in the motor cortex with the contrast agent technique during finger-tapping stimulation of the motor cortex (Li et al, 2000). In the same study, an increase in rCBF of 36% was also measured by the flow-sensitive alternating recovery pulsed ASL technique, with the original ASL perfusion model (Detre et al, 1992) used for quantification. More recently, a comparison of the change in CBF as measured by flow sensitive alternating recovery ASL and positron emission tomography reported a 22% increase in CBF in the visual cortex, which was slightly lower than the 28% change measured by positron emission tomography (Chen et al, 2008). A continuous ASL fMRI study of perfusion changes during a finger-tapping task reported a 31% increase in rCBF in the primary motor cortex (Garraux et al, 2005).

Most quantitative ASL techniques are based on the theory of the early perfusion experiments carried out by Kety and Schmidt, 1948. These approaches are intrinsically based on Fick's principle, which assumes that magnetically labeled water protons act as a freely diffusible tracer. When a freely diffusible tracer is assumed, the CBV cannot be estimated and either an intravascular tracer-based measure of CBV or the application Grubb's power law (Grubb et al, 1974) with a constant exponent (α=0.38) is required to estimate the change in CBV during neuronal activation. An alternative to these methods is desirable as the former relies on the invasive injection of a contrast agent and the latter is known to produce systematic errors as it has been established that α can vary depending on the experimental conditions (Kida et al, 2007).

The btASL quantitative fMRI technique is capable of quantifying the change in both CBV and CBF during neuronal activation. The rCBVlw is calculated from the area under the ASL concentration–time curve (or zeroth moment), which allows the rCBF to be calculated from the decrease in the MTT. The increase in the perfusion coefficient, P, can also provide some novel information about the hemodynamic response to neuronal activation. One mechanism that may explain an increase in the parameter P is the so-called capillary recruitment hypothesis (Shockley and LaManna, 1988). The recruitment of additional, tortuous capillaries during neuronal activation would lead to increased pseudo-diffusion of the labeled bolus and result in an increase in the parameter, P. However, a series of studies have suggested that under resting conditions, all cerebral capillaries are continuously perfused with plasma and consequently the recruitment of latent capillaries is not possible (Gobel et al, 1989). Alternatively, if we consider that P describes the diffusion of the labeled bolus at the ROI, we can hypothesize that an increase in this parameter during neuronal activation is due to an increase in the exchange of water from the capillary bed to the surrounding tissue in the activated region.

The btASL technique has earlier been applied to an aging study in the rat brain (Kelly et al, 2009b), in which both the MTT and CTT were found to increase in a group of aged rats compared with young and middle aged group. In the absence of a change in the rCBVlw, the increase in the transit times corresponded to a decrease in rCBF and P with age. When coupled with the neuronal activation results presented here, it can be concluded that the technique holds much potential for quantifying changes in the hemodynamic response to neural activity in aged and or diseased cohorts. By combining this quantitative approach with the anesthetic protocol described above and considering the signal stability offered by ASL fMRI compared with BOLD (Siewert et al, 1996), the btASL technique is ideal for longitudinal animal studies of both healthy and diseased aging.

The btASL fMRI technique has a number of shortcomings in its present form. First, the temporal resolution required to acquire the ASL signal–time curves (Figure 2) necessitates long stimulation periods of approximately 1 min duration. Although 11 time points were used to generate these curves, the aim was to provide proof of concept and assess the applicability of the btASL technique to an fMRI study. As only three fitting parameters are necessary to perform the least-squares fit, conceivably four time points would be sufficient for quantification of the perfusion parameters. This would greatly reduce the stimulation duration. In addition, the technique is restricted to single-slice acquisition because of magnetization transfer effects. The use of a dedicated labeling coil (Zhang et al, 1995) or a more sophisticated labeling scheme such as the double adiabatic inversion approach (Alsop and Detre, 1998) would facilitate multi-slice imaging and allow more comprehensive mapping of the activated region. Finally, the calculation of the rCBVlw requires the DOI to be estimated. The method used in this study involves measuring the signal intensity in the middle cerebral arteries and includes a partial volume error. The DOI was not found to vary significantly between the control and activation experiments (Table 2) and as a result, when considering the relative change in rCBVlw between the control and activation states, the DOI error is mainly cancelled out. However, a more accurate quantification of the inversion efficiency, such as the phase contrast magnetic resonance angiography method (O'Gorman et al, 2006), would improve the accuracy of the results.

In conclusion, the results of this study show the ability of the btASL technique to provide a novel quantitative assessment of the hemodynamic response to neuronal activation in terms of a unique set of perfusion parameters (MTT, CTT, rCBVlw, rCBF, and P). The sensitivity of the btASL technique to age-related changes in cerebral perfusion and its suitability for longitudinal fMRI studies in small animals should give rise to numerous future applications of the technique.

Acknowledgments

This work was supported in part by the Irish Research Council for Science, Engineering and Technology (IRCSET), Trinity College Institute of Neuroscience, Health Research Board (HRB), and Higher Education Authority (HEA).

The authors declare no conflict of interest.

Footnotes

Supplementary Information accompanies the paper on the Journal of Cerebral Blood Flow & Metabolism website (http://www.nature.com/jcbfm)

Supplementary Material

Supplementary Figure 1
Click here for additional data file. (51.2KB, pdf)
Supplementary Figure 2
Click here for additional data file. (97.8KB, pdf)

References

  1. Alsop DC, Detre JA. Multisection cerebral blood flow MR imaging with continuous arterial spin labeling. Radiology. 1998;208:410–416. doi: 10.1148/radiology.208.2.9680569. [DOI] [PubMed] [Google Scholar]
  2. Bandettini PA, Wong EC, Hinks SR, Tikofsky RS, Hyde JS. Time course EPI of human brain function during task activation. Magn Reson Med. 1992;25:390–397. doi: 10.1002/mrm.1910250220. [DOI] [PubMed] [Google Scholar]
  3. Belliveau JW, Kennedy DN, Jr, McKinstry RC, Buchbinder BR, Weisskoff RM, Cohen MS, Vevea JM, Brady TJ, Rosen BR. Functional mapping of the human visual cortex by magnetic resonance imaging. Science. 1991;254:716–719. doi: 10.1126/science.1948051. [DOI] [PubMed] [Google Scholar]
  4. Bulte DP, Drescher K, Jezzard P. Comparison of hypercapnia-based calibration techniques for measurement of cerebral oxygen metabolism with MRI. Magn Reson Med. 2009;61:391–398. doi: 10.1002/mrm.21862. [DOI] [PubMed] [Google Scholar]
  5. Buxton RB, Frank LR. A model for the coupling between cerebral blood flow and oxygen metabolism during neural stimulation. J Cereb Blood Flow Metab. 1997;17:64–72. doi: 10.1097/00004647-199701000-00009. [DOI] [PubMed] [Google Scholar]
  6. Chen JJ, Wieckowska M, Meyer E, Pike GB. Cerebral blood flow measurement using fMRI and PET: a cross-validation study. Int J Biomed Imaging. 2008;2008:516359. doi: 10.1155/2008/516359. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. D'Esposito M, Deouell LY, Gazzaley A. Alterations in the BOLD fMRI signal with ageing and disease: a challenge for neuroimaging. Nat Rev Neurosci. 2003;4:863–872. doi: 10.1038/nrn1246. [DOI] [PubMed] [Google Scholar]
  8. Davis TL, Kwong KK, Weisskoff RM, Rosen BR. Calibrated functional MRI: mapping the dynamics of oxidative metabolism. Proc Natl Acad Sci USA. 1998;95:1834–1839. doi: 10.1073/pnas.95.4.1834. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Detre JA, Leigh JS, Williams DS, Koretsky AP. Perfusion imaging. Magn Reson Med. 1992;23:37–45. doi: 10.1002/mrm.1910230106. [DOI] [PubMed] [Google Scholar]
  10. Dixon WT, Du LN, Faul DD, Gado M, Rossnick S. Projection angiograms of blood labeled by adiabatic fast passage. Magn Reson Med. 1986;3:454–462. doi: 10.1002/mrm.1910030311. [DOI] [PubMed] [Google Scholar]
  11. Garraux G, Hallet M, Talagala SL. CASL fMRI of subcortico-cortical perfusion changes during memory-guided finger sequences. Neuroimage. 2005;25:122–132. doi: 10.1016/j.neuroimage.2004.11.004. [DOI] [PubMed] [Google Scholar]
  12. Gobel U, Klein B, Schrock H, Kuschinsky W. Lack of capillary recruitment in the brains of awake rats during hypercapnia. J Cereb Blood Flow Metab. 1989;9:491–499. doi: 10.1038/jcbfm.1989.72. [DOI] [PubMed] [Google Scholar]
  13. Grubb RL, Jr, Raichle ME, Eichling JO, Ter-Pogossian MM. The effects of changes in PaCO2 on cerebral blood volume, blood flow, and vascular mean transit time. Stroke. 1974;5:630–639. doi: 10.1161/01.str.5.5.630. [DOI] [PubMed] [Google Scholar]
  14. Hock C, Villringer K, Muller-Spahn F, Wenzel R, Heekeren H, Schuh-Hofer S, Hofmann M, Minoshima S, Schwaiger M, Dirnagl U, Villringer A. Decrease in parietal cerebral hemoglobin oxygenation during performance of a verbal fluency task in patients with Alzheimer's disease monitored by means of near-infrared spectroscopy (NIRS)--correlation with simultaneous rCBF-PET measurements. Brain Res. 1997;755:293–303. doi: 10.1016/s0006-8993(97)00122-4. [DOI] [PubMed] [Google Scholar]
  15. Holsinger AE, Riederer SJ. The importance of phase-encoding order in ultra-short TR snapshot MR imaging. Magn Reson Med. 1990;16:481–488. doi: 10.1002/mrm.1910160315. [DOI] [PubMed] [Google Scholar]
  16. Hyder F, Shulman RG, Rothman DL. A model for the regulation of cerebral oxygen delivery. J Appl Physiol. 1998;85:554–564. doi: 10.1152/jappl.1998.85.2.554. [DOI] [PubMed] [Google Scholar]
  17. Kelly M, Blau C, Bechara R, Lynch M, Kerskens C. Bolus-tracking arterial spin labeling; a new marker for age and age related diseases. J Cereb Blood Flow Metab. 2009b;29:S87. [Google Scholar]
  18. Kelly ME, Blau CW, Kerskens CM. Bolus-tracking arterial spin labelling: theoretical and experimental results. Phys Med Biol. 2009a;54:1235–1251. doi: 10.1088/0031-9155/54/5/009. [DOI] [PubMed] [Google Scholar]
  19. Kerskens CM, Hoehn-Berlage M, Schmitz B, Busch E, Bock C, Gyngell ML, Hossmann KA. Ultrafast perfusion-weighted MRI of functional brain activation in rats during forepaw stimulation: comparison with T2-weighted MRI. NMR Biomed. 1996;9:20–23. doi: 10.1002/(SICI)1099-1492(199602)9:1<20::AID-NBM381>3.0.CO;2-R. [DOI] [PubMed] [Google Scholar]
  20. Kety SS, Schmidt CF. The nitrous oxide method for the quantitative determination of cerebral blood flow in man: theory, procedure and normalvalues. J Clin Invest. 1948;27:476–483. doi: 10.1172/JCI101994. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Kida I, Rothman DL, Hyder F. Dynamics of changes in blood flow, volume, and oxygenation: implications for dynamic functional magnetic resonance imaging calibration. J Cereb Blood Flow Metab. 2007;27:690–696. doi: 10.1038/sj.jcbfm.9600409. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Kim T, Kim SG. Quantification of cerebral arterial blood volume using arterial spin labeling with intravoxel incoherent motion-sensitive gradients. Magn Reson Med. 2006;55:1047–1057. doi: 10.1002/mrm.20867. [DOI] [PubMed] [Google Scholar]
  23. Kwong KK, Belliveau JW, Chesler DA, Goldberg IE, Weisskoff RM, Poncelet BP, Kennedy DN, Hoppel BE, Cohen MS, Turner R, Cheng H-M, Brady TJ, Rosen BR. Dynamic magnetic resonance imaging of human brain activity during primary sensory stimulation. Proc Natl Acad Sci USA. 1992;89:5675–5679. doi: 10.1073/pnas.89.12.5675. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Leniger-Follert E, Hossmann KA. Simultaneous measurements of microflow and evoked potentials in the somatomotor cortex of the cat brain during specific sensory activation. Pflugers Arch. 1979;380:85–89. doi: 10.1007/BF00582617. [DOI] [PubMed] [Google Scholar]
  25. Li TQ, Haefelin TN, Chan B, Kastrup A, Jonsson T, Glover GH, Moseley ME. Assessment of hemodynamic response during focal neural activity in human using bolus tracking, arterial spin labeling and BOLD techniques. Neuroimage. 2000;12:442–451. doi: 10.1006/nimg.2000.0634. [DOI] [PubMed] [Google Scholar]
  26. Mandeville JB, Marotta JJ, Kosofsky BE, Keltner JR, Weissleder R, Rosen BR, Weisskoff RM. Dynamic functional imaging of relative cerebral blood volume during rat forepaw stimulation. Magn Reson Med. 1998;39:615–624. doi: 10.1002/mrm.1910390415. [DOI] [PubMed] [Google Scholar]
  27. Meier P, Zierler KL. On the theoryof the indicator-dilution method for measurement of blood flow and volume. J Appl Physiol. 1954;6:731–744. doi: 10.1152/jappl.1954.6.12.731. [DOI] [PubMed] [Google Scholar]
  28. Ogawa S, Lee TM, Nayak , Glynn P. Oxygenation-sensitive contrast in magnetic resonance image of rodent brain at high magnetic fields. Magn Reson Med. 1990;14:68–78. doi: 10.1002/mrm.1910140108. [DOI] [PubMed] [Google Scholar]
  29. O'Gorman RL, Summers PE, Zelaya FO, Williams SC, Alsop DC, Lythgoe DJ. In vivo estimation of the flow-driven adiabatic inversion efficiency for continuous arterial spin labeling: a method using phase contrast magnetic resonance angiography. Magn Reson Med. 2006;55:1291–1297. doi: 10.1002/mrm.20864. [DOI] [PubMed] [Google Scholar]
  30. Parkes LM, Tofts PS. Improved accuracy of human cerebral blood perfusion measurements using arterial spin labeling: accounting for capillary water permeability. Magn Reson Med. 2002;48:27–41. doi: 10.1002/mrm.10180. [DOI] [PubMed] [Google Scholar]
  31. Petersen ET, Lim T, Golay X. Model-free arterial spin labeling quantification approach for perfusion MRI. Magn Reson Med. 2006;55:219–232. doi: 10.1002/mrm.20784. [DOI] [PubMed] [Google Scholar]
  32. Sanganahalli BG, Herman P, Blumenfeld H, Hyder F. Oxidative neuroenergetics in event-related paradigms. J Neurosci. 2009;29:1707–1718. doi: 10.1523/JNEUROSCI.5549-08.2009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Shockley RP, LaManna JC. Determination of rat cerebral cortical blood volume changes by capillary mean transit time analysis during hypoxia, hypercapnia and hyperventilation. Brain Res. 1988;454:170–178. doi: 10.1016/0006-8993(88)90816-5. [DOI] [PubMed] [Google Scholar]
  34. Siewert B, Bly BM, Schlaug G, Darby DG, Thangaraj V, Warach S, Edelman RR. Comparison of the BOLD- and EPISTAR-technique for functional brain imaging by using signal detection theory. Magn Reson Med. 1996;36:249–255. doi: 10.1002/mrm.1910360212. [DOI] [PubMed] [Google Scholar]
  35. Silva AC, Williams DS, Koretsky AP. Evidence for the exchange of arterial spin-labeled water with tissue water in rat brain from diffusion-sensitized measurements of perfusion. Magn Reson Med. 1997;38:232–237. doi: 10.1002/mrm.1910380211. [DOI] [PubMed] [Google Scholar]
  36. Wang J, Aguirre GK, Kimberg GK, Roc AC, Li L, Detre JA. Arterial spin labeling perfusion fMRI with very low task frequency. Magn Reson Med. 2003;49:796–802. doi: 10.1002/mrm.10437. [DOI] [PubMed] [Google Scholar]
  37. Weber R, Ramos-Cabrer P, Wiedermann D, van Camp N, Hoehn M. A fully noninvasive and robust experimental protocol for longitudinal fMRI studies in the rat. Neuroimage. 2006;29:1303–1310. doi: 10.1016/j.neuroimage.2005.08.028. [DOI] [PubMed] [Google Scholar]
  38. Zhang W, Silva AC, Williams DS, Koretsky AP. NMR measurement of perfusion using arterial spin labeling without saturation of macromolecular spins. Magn Reson Med. 1995;33:370–376. doi: 10.1002/mrm.1910330310. [DOI] [PubMed] [Google Scholar]
  39. Zhang W, Williams DS, Koretsky AP. Measurement of rat brain perfusion by NMR using spin labeling of arterial water: in vivo determination of the degree of spin labeling. Magn Reson Med. 1993;29:416–421. doi: 10.1002/mrm.1910290323. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Figure 1
Click here for additional data file. (51.2KB, pdf)
Supplementary Figure 2
Click here for additional data file. (97.8KB, pdf)

Articles from Journal of Cerebral Blood Flow and Metabolism: Official Journal of the International Society of Cerebral Blood Flow and Metabolism are provided here courtesy of SAGE Publications

RESOURCES