Hyperpolarized Xenon Brain MRI

Since hyperpolarized 129Xe MRI was first demonstrated in the lung, air space imaging using hyperpolarized noble gases (129Xe and 3He) has progressed at a rapid rate (Goodson, 2002; Zhou, 2011c). Owing to high lipid solubility, absence of background signal in biological tissue, non-invasiveness, lack of radioactivity, different relaxation to oxygenated and deoxygenated blood, and larger chemical shift to the neighbor environment, hyperpolarized 129Xe magnetic resonance imaging (MRI) has a great potential as a tool for studying the brain, especially for the assessment of cerebral blood flow (CBF) related to the brain function and activities.

phases of Xenon can be easily obtained within an experimentally accessible range of temperatures and pressures (Cook, 1961) (Fig. 1). Furthermore, the large and highly polarizable electron cloud makes xenon highly lipid soluble in solution, without chemically or structurally damage during interactions with other molecules. Ostwald solubility is defined as the ratio of the volume of the gas absorbed to the volume of the absorbing liquid, measured at same temperature and a pressure of 1 atm (101,325 Pa) (Cherubini, 2003;Oros, 2004) (Table 1) Owing to the inter-atomic collisions distortion of xenon electron cloud during the interactions with different chemical environment, the chemical shift of 129 Xe is extremely sensitive. Total solvent effect on the Xe resonance frequency is over 7500 ppm, very large compared with most other NMR sensitive nuclei ( 1 H : 20 ppm、 13 C : 300 ppm).
A few examples are shown in Figure 2, the natural reference point for xenon chemical shift in the gas phase; with respect to the gas resonance (0 ppm), peaks at 70 ppm corresponding to cryptophane-bound xenon; around 197 ppm, five dissolved peaks can be observed, a dominant peak at 194.7 ppm and another discriminable peak at 189 ppm are identified as dissolved hyperpolarized 129 Xe in the brain tissue and non-brain tissue, respectively, two small peaks at 191.6 ppm and 197.8 ppm are still unknown, and a smaller broad peak at 209.5 ppm comes from the dissolved hyperpolarized 129 Xe in the blood (Zhou et al., 2008). As a closed shelled noble gas, xenon has a peculiar large polarizable electron cloud, it can be easily interacted with biological materials, including water 'lipids' and proteins and so on; meanwhile, 129 Xe NMR parameters like relaxation time and chemical shift are very sensitive to local chemical environment, these biochemical physiological characters make xenon a very interesting NMR probe for biological applications. Arising from large increase in sensitivity associated with hyperpolarized 129 Xe, the biological NMR applications have been dramatically extended. Being Chemically Inert, xenon can be safely delivered into living organisms, associated with another advantage of no background signal, thus in vivo MR imaging and spectroscopy is possible using hyperpolarized 129 Xe technique.

Signal enhanced MRI with hyperpolarized 129 Xe
Nuclear magnetic resonance (NMR) has been widely used in most fields of natural sciences, such as physics, chemistry, biology, and medicine. Because of the intrinsic low nuclear spin polarization at the thermal equilibrium, NMR is relatively insensitive. The normal 129 Xe MRI, i.e., at the thermal equilibrium, is not able to get enough signal to visualize tissues or organs. The polarization can be moderately increased by using lower temperature or higher www.intechopen.com Advances in Brain Imaging 130 magnetic fields (Navon et al., 1996), whereas the nuclear spin polarizations of noble gases can be increased by four or five orders of magnitude via the spin-exchange optical pumping (SEOP) techniques (Walker, 1997). Therefore, it enables a very high sensitive detection of hyperpolarized 129 Xe MRI. This subsection will describe how to produce hyperpolarized 129 Xe, which is the basis for the brain imaging.
3 He and 129 Xe are generally selected for hyperpolarized lung gas imaging, because 3 He and 129 Xe are the only noble gas nuclei with nuclear spin 1/2, which results in the longitudinal relaxation time of many hours or even days at the standard temperature and pressure. However, hyperpolarized 3 He could not be used for the brain study due to the extremely low solubility in blood and tissue, while hyperpolarized 129 Xe is a novel contrast agent for the cerebral research. Hyperpolarized 129 Xe is generally generated by employing the technique of SEOP.
For SEOP, the first step is to transfer the anglar momentum of the cicularly polarized laser light to the electronic spin, i.e., optical pumping. In principal, any alkali metal vapor can be optically pumped. Rb is normally used as the corresponding pumping laser diode arrays (LDA), which are routinely manufactured in high power configurations. When the Rb vapor experiences an external magnetic field of 20-30 Gauss, the ground state 5 2 S 1/2 is split into two Zeeman sublevels, M J =-1/2 and M J =1/2, i.e., electronic spin "down" or "up". These two Zeeman states have a nearly equal population at room temperature. After the Rb vapor absorbs the circularly polarized laser light ( +) centered at 795nm, the Rb D1 transition occurs, i.e., 5 2 S 1/2 5 2 P 1/2 . Accordingly, the ground state with a sublevel M J =-1/2 is pumped into the excited state, M J =1/2. The nitrogen gas quenches the excited state back to the ground state. Because the M J =-1/2 sublevel continues to absorb the circularly polarized light ( +) an excess of Rb atoms are optically pumped into the Zeeman sublevel M J =1/2 while the other sublevel M J =-1/2 is depleted. Therefore, an Rb electronic spin polarization of roughly 100% is able to be achieved.
The second step of SEOP is spin exchange, which occurs between the polarized electronic spins of Rb and the xenon nucleus. The collision between polarized Rb atoms and xenon atoms induces the transfer of angular momentum from the electronic spin to the nuclear spin. During this collision, the electron wave function of the Rb overlaps the nuclear wave function of xenon, which results in the spin exchange between the electronic spin and nuclear spin. Binary collisions dominate the spin exchange at high pressure, while threebody collision (by forming a Rb/Xe van der Waals molecule) dominate at low pressure (a few tens of torr).
Generally, Rb atoms and nitrogen gas are employed for optical pumping, however, Cs may be proposed as a better candidate for spin exchange with 129 Xe due to several advantages: the natural abundance of 133 Cs is 100% while Rb has two isotopes ( 85 Rb and 87 Rb), so that Cs is more convenient than Rb for wide applications of hyperpolarized 129 Xe, particularly in the clinic application; optical pumping cells for Cs are operated at lower temperatures with correspondingly fewer chemical corrosion problems; according to the previous experimental results, the spin-exchange rate of Cs-Xe is about 10% higher than the Rb-Xe rate (Zhou et al., 2004a). When the polarization of hyperpolarized 129 Xe is high enough, the observed radiation damping has been reported (Zhou, 2004b). The xenon polarizer with a flow feature can be readily extended to produce larger quantities of hyperpolarized 129 Xe for not only medical imaging but also materials science and biology (Zhou, 2004c;Zhou, 2009a).

Longitudinal relaxation time (T 1 ) of hyperpolarized 129 Xe in the brain
Hyperpolarized 129 Xe magnetization enhanced by SEOP is non-recoverable, and the T 1 of an hyperpolarized gas is the time elapsed for the signal to decay, because its thermal equilibrium polarization is almost zero relative to the hyperpolarized polarization. Usually, small flip angles has to be used to ration the hyperpolarized magnetization, and it is very important to have a T 1 as long as possible to ensure that the signal lasts long enough for the acquisition. Therefore, the T 1 of hyperpolarized 129 Xe in the brain is a critical parameter. This subsection will discuss the aspects that affect T 1 of hyperpolarized xenon, and describe the methods to accurately measure the T 1 in the brain (Zhou et. al, 2008).

T 1 of hyperpolarized 129 Xe
For conventional MRI, the magnetization at thermal equilibrium is induced by the magnetic field, and the longitudinal relaxation time (T 1 ) is the time for the magnetization, i.e. the magnetic resonance signal, to recover back to the thermal equilibrium. However, hyperpolarized noble gas magnetization produced by SEOP is unable to recover back to the hyperpolarized magnetization by itself, and the T 1 of an hyperpolarized noble gas is the decay time of magnetization, because the thermal polarized magnetization is almost zero related to the hyperpolarized magnetization. The longitudinal relaxation time of hyperpolarized 129 Xe in the brain is a critical parameter for developing hyperpolarized 129 Xe brain imaging and spectroscopy and optimizing the pulse sequences, especially in the case of cerebral blood flow measurements. Various studies have produced widely varying estimates of hyperpolarized 129 Xe T 1 in the rat brain (Choquet, 2003;Wakai, 2005).
The hyperpolarized magnetization is generally tipped by a pulse with a small flip angle, and it is very challenging to make the T 1 as long as possible in order to ensure the signal lasts long enough for the acquisition. Therefore, when considering hyperpolarized 129 Xe as a marker for brain perfusion by MRI, evaluation of tissue characterization and pulse sequence optimization, the T 1 of hyperpolarized 129 Xe in the brain is a critical parameter. Previous attempts to measure T 1 in the rat brain have yielded strikingly disparate results. Wilson et al. found that T 1 measured in rat brain homogenates in vitro ranged from 18±1 to 22±2 s (mean ± SD) (Wilson et. al，2009) depending on the oxygenation level of the tissue, and T 1 values from measurements in rat brain in vivo have ranged from 3.6±2.1 (Choquet et al, 2003) to 26±4 s (Wakai et. al, 2005). Part of the discrepancy is believed to be due to the protocols used in T 1 determination. The attempt of Choquet et al. used a multi-pulse protocol during the uptake and washout process by injecting hyperpolarized 129 Xe in a lipid emulsion, whereas the estimation of Wakai et al. used a two-pulse protocol during the washout process after the rat had breathed hyperpolarized 129 Xe gas. Under the condition of typically achieved polarizations (5-21%) (Zook et. al, 2001), low signal-to-noise ratio (SNR) due to the low concentration of the dissolved hyperpolarized 129 Xe in tissue is an important factor in making T 1 measurements in the rat brain (Cherubinia et. al, 2003;Ruppert et. al, 2000). The maximum SNR in the above two measurements in vivo was only 30 (Choquet et. al, 2003) and 46 (Wakai et. al, 2005), and the noise effect was not considered in these studies. When the SNR is low, noise will dominate the measured signal and result in large differences between the true T 1 and the measured T 1 . Thus, low SNR might be a large contributor of error in the published T 1 values.

www.intechopen.com
Advances in Brain Imaging 132

Multi-pulse and two-pulse washout protocols for measurements of T 1
Hyperpolarized 129 Xe transport in the brain has been modeled using appropriate adaptations of the Kety-Schmidt theory (Martin et al., 1997;Peled et al., 1996). Martin and co-workers derived the equation of the cerebral xenon concentration during hyperpolarized 129 Xe delivery to the lungs as follow: where C cereb is the xenon concentration in the cerebral artery, C brain is the xenon concentration in brain parenchyma, F is the tissue perfusion in units of (volume blood)/(volume tissue)/time, p is the brain/blood partition coefficient for xenon, and T 1brain is the longitudinal relaxation time of xenon in the brain. In this equation, the first term on the right describes xenon transport to the brain, and the second term describes the loss of xenon signal due to both perfusion and T 1 decay. Xenon signal observed from the brain is proportional to C brain . During the washout phase of the xenon signal, there is no transport of hyperpolarized 129 Xe by the cerebral artery, and hence C cereb is zero. Accordingly, the xenon concentration in the brain during washout (C brainwashout ) is given by the following equation: This equation can be solved to yield an analytical solution for the concentration of xenon in the brain during the washout of signal. The decay time constant ( ) of hyperpolarized 129 Xe during the washout from the rat brain is given by: Thus, can be calculated from a series of pulse excitations (multi-pulse protocol) after compensating for the hyperpolarized xenon signal losses resulted from radio frequency (RF) excitation, as described below. To compare the results obtained using the multi-pulse protocol, a two-pulse protocol has also been adopted to measure (Wakai et al., 2005). Both protocols were performed on each rat during the washout phase of the 129 Xe signal. Table 2 shows individual T 1 values of hyperpolarized 129 Xe and their mean from the six rat brains using the two protocols, with and without the SNR threshold. The mean T 1 value calculated using the improved two-pulse method is larger than that using its conventional counterpart, whereas the mean T 1 value calculated using the improved multi-pulse method is less than that using its conventional counterpart. These T 1 values were named 'group 1' to 'group 4' for easy reference during discussion.
In this subsection, we investigated the error in T 1 measurement as a result of low SNR of the 129 Xe signal in vivo. Correcting for these errors allowed us to more accurately measure the T 1 of hyperpolarized 129 Xe in the rat brain in vivo. Our calculations produced highly consistent T 1 results independent of the measurement protocol and offer a resolution to the discrepancy between previously reported values.

Rat
Multi-pulse protocol 2-pulse protocol Conventional Improved Conventional Improved T 1 ( s) (group 1)  Table 2. T 1 values of hyperpolarized 129 Xe from six rat brains. The mean T 1 value and standard deviation obtained from the multi-pulse and two-pulse protocols before (conventional method) and after (improved method) setting a threshold of SNR=5.5 are also given. (Zhou et al., 2008)

Stroke MRI with hyperpolarized 129 Xe
Because there is no background signal from xenon in biological tissue, and because the inhaled xenon is delivered to the brain by the blood flow, we would expect a perfusion deficit, such as could be seen in stroke, to reduce xenon concentration in the region of the deficit. Thermal polarization yields negligible xenon signal relative to hyperpolarized xenon; therefore, hyperpolarized xenon can be used as a tracer of cerebral blood flow (CBF). This subsection will describe that hyperpolarized 129 Xe MRI is able to detect, in vivo hypoperfused area of focal cerebral ischemia-i.e., the ischemic core area of stroke, by using a rat permanent right middle cerebral artery occlusion (MCAO) model (Zhou et al., 2011a).
Stroke is the single most common reason for permanent disability and is the third leading cause of death in developed countries. During acute ischemic stroke, a core of brain cells at the center of the affected region dies quickly, and the damage subsequently spreads to surrounding tissue over the next few hours. Because they allow for the delineation of areas of ischemic neuronal injury and hypoperfusion within minutes after the induction of cerebral ischemia, conventional proton MRI, especially diffusion-weighted imaging (DWI) and perfusion weighted imaging (PWI), have been particularly useful in the diagnosis of acute ischemic stroke. The target of acute stroke therapy is the portion of the ischemic region which is still potentially salvageable, that is the ischemic penumbra. MRI operationally defines the ischemic penumbra by the mismatch area of PWI-DWI. DWI detects changes in the apparent diffusion coefficient (ADC) of water molecules associated with early cytotoxic edema in ischemic stroke. Arterial spin labeling (ASL)-based PWI methods provide excellent anatomical information for the measurement of tissue perfusion. The ASL technique shows numerous advantages, such as noninvasive measurements of cerebral blood flow (CBF) quantifiable in standard units of mL/g/min, and is able to image multi-slices and multiregions of the brain. However, in some situations, PWI methods require the injection of gadolinium containing contrast agents to map relative CBF in order to identify the hypoperfused tissue. In addition to the conventional DWI and PWI techniques, van Zijl and coworkers have developed pH-weighted MRI to study stroke and ischemic penumbra. However, proton imaging has a large background signal in biological tissue, and contrast injection is an invasive approach. Moreover, contrast-associated nephrogenic systemic fibrosis has been reported after the use of gadolinium-based agents, and many patients with impaired renal function are not eligible to receive contrast media. In contrast, hyperpolarized 129 Xe MRI shows great potential and advantages for the identification of hypoperfused brain tissue. Xenon is highly lipid soluble and lacks an intrinsic background signal in biological tissue (Albert et al., 1994). Duhamel and co-workers have studied CBF using intra-arterial injection of hyperpolarized 129 Xe dissolved in a lipid emulsion (Duhamel et al., 2000). Alternatively, hyperpolarized 129 Xe can be administered noninvasively by inhalation; following inhalation, 129 Xe is absorbed into the bloodstream and delivered to the brain through the circulation. Because the spin-exchange optical pumping technique can enhance the 129 Xe MR signal 10,000-100,000 times over thermal polarization, the dissolvedphase hyperpolarized 129 Xe signal in the brain can be detected even at low concentrations. Because the xenon signal is proportional to CBF (Zhou et al., 2008), a decrease in the signal is expected to occur in areas of decreased CBF after the inhalation of hyperpolarized xenon gas. Hyperpolarized xenon imaging currently can not achieve a slice as thin as that obtained by ASL. In addition, ASL can be performed with substantially higher spatial resolution than hyperpolarized xenon imaging in brain tissue. However, the ASL technique requires two experiments (arterial spin labeled and controlled) to obtain CBF information. In this subsection, we report, for the first time, that hyperpolarized 129 Xe MRI is able to detect areas of decreased CBF following middle cerebral artery occlusion (MCAO) in a single scan. These findings show the great potential and utility of hyperpolarized 129 Xe MRI for stroke imaging, and further demonstrate that hyperpolarized 129 Xe is a safe and noninvasive signal source for imaging diseases and function of the brain. Figure 3a shows a representative proton ADC map obtained 90 min following MCAO. There is a large ischemic core within the ipsilesional (right) MCA territory, as indicated by ADC values below the critical threshold of 5.3 × 10 -4 mm 2 /s for infarction (Meng et al., 2004). [The normal ADC value of rat brain tissue in the contralesional (left) hemisphere is (7.5±1.8)×10 -4 mm 2 /s.] Figure 3b depicts the corresponding hyperpolarized 129 Xe CSI, indicating signal reduction in large parts of the right hemisphere, consistent with the area typically experiencing decreased CBF following right MCAO in the model used (Duhamel et al., 2002). Figure 3c shows the TTC-stained brain section of the same animal as illustrated in Fig.  3a, b; the black line in this figure delineates the infracted brain tissue. Xenon CSI, shown in Fig. 3b, demonstrates reduced perfusion in brain tissue, ultimately leading to infarction, as shown by TTC staining in Fig. 3c. In Fig. 3d, the blue area represents the difference between the ADC lesion and TTC lesion areas, and the green area shows the nonischemic region. ROI analyses were performed to further characterize the Xe CSI tissue signals within the different observed tissue compartments defined by their respective ADC and TTC signatures.
Xenon signals from each ROI in the contralesional (left) hemisphere were set as a reference (100%), and xenon signals from each ROI in the ipsilesional (right) hemisphere were normalized to these signals. The xenon signal in the ischemic core (ROI 1 ) dropped to 8.4±0.4% of the contralesional side signal, and the xenon signal in normal tissue (ROI 2 ) remained the same. Moreover, the xenon signal in ROI 1 was reduced significantly relative to the corresponding con-tralesional ROIs in the MCAO group, as well as the corresponding ipsilateral ROIs in control animals. Within the control group, no significant differences in the xenon signal were observed between the corresponding ROIs of both hemispheres. Using a rat permanent right middle cerebral artery occlusion model, it has demonstrated that hyperpolarized 129 Xe MRI is able to detect, in vivo, the hypoperfused area of focal cerebral ischemia, that is the ischemic core area of stroke. To the best of our knowledge, this is the first time that hyperpolarized 129 Xe MRI has been used to explore normal and abnormal cerebral perfusion. The study shows a novel application of hyperpolarized 129 Xe MRI for imaging stroke, and further demonstrates its capacity to serve as a complementary tool to proton MRI for the study of the pathophysiology during brain hypoperfusion. More importantly, these results indicate the possibility of the use of in vivo MRI to diagnose brain disease employing inhaled hyperpolarized gas, eliminating the potential adverse effects to the patient resulting from the injection of gadolinium-based contrast agents.

Functional MRI with hyperpolarized 129 Xe
As hyperpolarized xenon is a MRI signal source with properties very different from those generated from water-protons, hyperpolarized 129 Xe MRI may yield structural and functional information not detectable by conventional proton-based MRI methods. This subsection shows that the differential distribution of hyperpolarized 129 Xe in the cerebral cortex of the rat following a pain stimulus evoked in the animal's forepaw. Areas of higher hyperpolarized 129 Xe signal corresponded to those areas previously demonstrated by conventional functional MRI (fMRI) methods as being activated by a forepaw pain stimulus. It demonstrated that the percent increase in hyperpolarized 129 Xe signal over baseline was 13 -28%, which is more sensitive than the conventional fMRI based on the blood oxygen level dependence (BOLD) (2-4%) (Zhou et al., 2011b).
Ex-vivo hyperpolarization of 129 Xe is detectable by magnetic resonance spectroscopy (MRS) and MRI in animals and humans, although the resulting in vivo signal to noise ratio (SNR) of the hyperpolarized 129 Xe signal is not as great as the signal produced by protons in conventional MRI, hyperpolarized 129 Xe has several unique characteristics which may endow it with advantages in some imaging applications, including brain imaging (Zhou et.al, 2011a). The nuclear magnetic resonance frequency range (chemical shift) of hyperpolarized 129 Xe in vivo is large compared to protons (200 ppm vs. 5 ppm respectively) and is also substantially affected by the local chemical environment, providing a means to detect localized physiological changes and biochemical binding events. In particular, the chemical shift experienced by 129 Xe in the presence of oxygen (O 2 ) is substantial and may offer a means to image changes in tissue O 2 concentration that result from changes in neuronal activity. Xenon is also an ideal perfusion tracer (Betz, 1972) and inhaled nonradioactive xenon gas has been used to detect disease induced alterations in cerebral blood flow with high anatomical specificity (Gur, 1982). Because xenon is not intrinsic to biological tissue, hyperpolarized 129 Xe produces virtually no background signal, which, in turn, results in high contrast hyperpolarized 129 Xe MR images (Swanson, 1997).

Rat brain functional hyperpolarized 129 Xe imaging experiment
In the rat brain function study, hyperpolarized 129 Xe MRI was performed in rats to investigate the distribution of the hyperpolarized 129 Xe signal following a well-established paradigm for producing anatomically localized neuronal activity. Six rats were intubated and connected to a ventilator that controlled the delivery of oxygen and hyperpolarized 129 Xe gas. Male Spragur-Dawley rats weighing between 200-250g were placed on animal respirator, with tidal volume of 3ml O 2 (3% isoflurane included) supplied for each breath. Immediately prior to the acquisition of CSI images, the animal was ventilated with alternate breaths of 100% hyperpolarized 129 Xe and 98% O 2 : 2% isoflurane. The breath-hold period during the delivery of each hyperpolarized 129 Xe breath was 2 seconds.
High resolution proton images were taken of the rat head to provide an anatomical reference for hyperpolarized 129 Xe images. In order to evaluate the distribution of hyperpolarized 129 Xe in brain following an external sensory stimulus, we acquired MRS images before and after a pain producing stimulus that has a well-defined functional response that can be measured using traditional fMRI techniques. A baseline hyperpolarized 129 Xe spectroscopic image was acquired from a coronal slice centered at the level of the anatomical reference slice.
The maximal, steady-state 129 Xe brain signal occurred within 15 seconds after starting the ventilation with hyperpolarized 129 Xe. Once verification of the xenon signal in the brain, a baseline 129 Xe chemical shift image (CSI) was acquired that was centred in the plane corresponding to the proton reference image. Low flip angle used for CSI acquisition insured minimal loss of hyperpolarized 129 Xe signal due to RF destruction, and the relatively long TR allowed continuous delivery of hyperpolarized 129 Xe to the tissue, a steady-state concentration of hyperpolarized 129 Xe was maintained in the brain thereby insuring constant signal intensity across the k-space. In a subset of animals (n=3) the animal's left forepaw was injected with a vehicle solution during baseline. Following acquisition of the baseline CSI image, the animal was ventilated for 10 minutes with O 2 (isoflurane to allow for complete clearance of 129 Xe magnetization from the brain). After that, the chemical irritant capsaicin (20 ul of 3 mg/ml) was injected into the animal's right forepaw (n= 6), and a second CSI was acquired.

Experiment results and discussion
A robust hyperpolarized 129 Xe spectroscopic signal with one primary peak at 194.7 ppm developed within 15 seconds after starting the ventilation with hyperpolarized 129 Xe. In order to determine the extent of hyperpolarized 129 Xe distribution throughout the rat brain, a magnetic resonance spectroscopic image was acquired of the primary peak during the administration of hyperpolarized 129 Xe (Figure 4). The four smaller resonances did not have sufficient SNR to produce spectroscopic images. Figure 4a shows an hyperpolarized 129 Xe image taken in the axial plane. Addition of a color look-up table (Figure 4b) aided in visually delineating areas of low and high SNR. Figure 4c show a 1 mm proton slice in which the olfactory bulbs and cerebellum are visible. Overlay of the hyperpolarized 129 Xe spectroscopic image onto the proton reference image (Figure 4d) revealed that the steadystate hyperpolarized 129 Xe signal originated from within the brain tissue and further demonstrated a pattern of hyperpolarized 129 Xe distribution throughout the brain with varying signal intensity in different brain regions.
Three of the six animals studied received a vehicle injection (saline) to the left forepaw immediately prior to the acquisition of the baseline image. 10 minutes after acquisition of the baseline hyperpolarized 129 Xe MRS, the animal's right forepaw was injected with the chemical irritant capsaicin (20 ml of 3 mg/ml), and a second hyperpolarized 129 Xe spectroscopic image was acquired. Responses from three individual animals are shown in Figure 5. Whereas baseline images showed some hyperpolarized 129 Xe signal intensity in cortical and sub-cortical brain regions ( Figure 5, left panel), images acquired following administration of capsaicin showed both higher hyperpolarized 129 Xe signal intensity and an increased area of distribution within the brain ( Figure 5, right panel). Superimposition of a rat brain atlas (Figure 5a) revealed that areas of hyperpolarized 129 Xe signal increase occurred both bilaterally and contralaterally in areas of the brain known to be involved in the processing of forepaw pain information, including the anterior cingulate and somatosensory cortices.  In spite of the as yet unrefined nature of this imaging modality, our results indicate that hyperpolarized 129 Xe MRI may have use as a probe for brain physiology and function. Because xenon is not inherent in the body, the substantial challenges resulting from high background signal in 1 H fMRI may be somewhat reduced. Extracting meaningful data from 1 H fMRI experiments is labour intensive, and requires a large number of subjects and image acquisitions. Extensive image post-processing is required and the influence that different post-processing steps play on the final data set achieved is actively debated. Conversely, hyperpolarized 129 Xe MRI showed patterns of brain activation consistent with those obtained using 1 H fMRI, using only a single set of images (one baseline and one post stimulus image) obtained from six animals. The magnitude of the signal difference between baseline and stimulus conditions for hyperpolarized 129 Xe (13-28%) was comparable to differences typically obtained with conventional BOLD fMRI (2 to 29%) (Bock et al., 1998;Silva et al., 1999;Mandeville et al., 1999;Tuor et al., 2000) using a rat forepaw activation paradigm.

Outlook of hyperpolarized 129 Xe brain MRI
This subsection will summarize the current progress as previously described, and comment the future research directions and applications in the brain imaging. Conventional MRI focuses mainly on the nuclear spin of the proton because it is ubiquitous in most parts of the human body. However, certain organs have a low proton spin density attributable to the large volume of air dispersed throughout the tissue. The low sensitivity of traditional magnetic resonance has motivated the development of techniques using hyperpolarized noble gases for NMR and MRI. Xenon has the unique characteristic of being soluble in many fluids and biological tissues, such as water, blood, lung tissue, and white and gray matter. Being a trace element in the atmosphere, xenon has no natural background signal in the human body. Therefore, dissolved-phase xenon MRI and molecular imaging could provide rich information related to biological and physiological changes beyond void space lung imaging. Efforts have demonstrated the value of dissolved xenon MRI in the study of lung gas exchange (Swanson et.al, 1999), and brain perfusion (Swanson et al., 1997;Kilian et al., 2004;Zhou et al., 2008Zhou et al., , 2011aZhou et al., , 2011b and function (Mazzanti et al., 2011). Recently, xenonbased molecular imaging has been demonstrated by using cryptophane-containing biosensors (Hilty et al., 2006). Sensitivity enhancement using a chemical amplification technique, hyperpolarized xenon chemical exchange saturation transfer (Hyper-CEST) (Schröder et al., 2006), allows imaging at low concentrations; however, for in vivo applications the small filling factor of a region of interest in the body relative to the NMR coil is a significant factor limiting sensitivity. In such cases remote detection methods (Hilty et.al, 2005) can provide dramatic improvements in sensitivity. In remote detection, the normal NMR coil that contains the full region of interest is used to encode spectroscopic and spatial information, then stores it as longitudinal magnetization. These encoded spins then flow into a second coil with an optimized filling factor for detection.
Remote detection can overcome the filling factor issue of dissolved xenon MRI, although a low concentration of xenon in solution can be another significant impediment to highly sensitive detection. It has been shown that the solvated xenon signal can be amplified by xenon polarization transfer contrast, in which the dissolved-phase xenon from either lungs or brains is selectively saturated, and through exchange, the gas-phase signal is attenuated.
This method is able to indirectly image dissolved-phase xenon, but is limited to tissue in direct exchange with the air in the lungs. The gas exchange process could be similarly exploited for direct signal amplification of dissolved xenon with the remote detection technique, which extends the study area from lung to brain. Xenon gas can be extracted from the dissolved solutions and concentrated in the gas phase for detection. Furthermore, with the long longitudinal relaxation time of gas-phase xenon, extracted xenon gas from solution can be compressed or liquefied while preserving the encoded information. The xenon density in the liquid state is approximately four orders of magnitude higher than in aqueous solutions, which in principle could result in up to 10,000 times enhancement of spin density, thus allowing substantial signal amplification We have demonstrated the hyperpolarized xenon signal amplification by gas extraction (Hyper-SAGE) method (Zhou et al., 2009b) with enhanced NMR spectra and time-of-flight (TOF) images by using recently commercialized membrane technology for high-efficiency xenon dissolution (Baumer et.al, 2006). The Hyper-SAGE technique relies on physical amplification by exploiting a phase change and is completely distinct from chemical amplification. In combination with additional amplification techniques such as Hyper-CEST, this method promises to dramatically decrease the detection threshold of MRI and has the potential to benefit molecular imaging applications and brain imaging.
Recent innovations in the production of highly polarized 129 Xe and novel method of signal enhancement should make feasible the emergence of hyperpolarized 129 Xe MRI as a viable adjunct method to conventional MRI for the study of brain function and disease. The high sensitivity of hyperpolarized noble gas signal and non-background noise in biological tissue offer xenon as an important and promising contrast agent to study the brain. Because the polarization of hyperpolarized xenon does not depend on the magnetic field strength, the technique for brain imaging could also be applied for use with low field portable MRI devices (Appelt, 2007;Blümich, 2008;Paulsen, 2008).

Acknowledgments
This work was supported by the Chinese Academy of Sciences (the 100 talents program and KJCX2-EW-N06-04), Natural Science Foundation of China (11004228), and Innovation Method Fund of China (2010IM030600).