Rapid magnetic resonance measurement of global cerebral metabolic rate of oxygen consumption in humans during rest and hypercapnia

Introduction

Carbon dioxide (CO₂) is a byproduct of oxidative metabolism and is an important cellular signaling molecule in all organisms. The arterial partial pressure of carbon dioxide (pCO₂) is one of the major regulators of cerebral circulation. Increased pCO₂ (hypercapnia) causes vasodilatation that, in turn, results in a drop in vascular resistance and a concomitant increase in cerebral blood flow (CBF) when autoregulation pathways are intact (Brian, 1998). Carbon dioxide likely mediates its effects by lowering perivascular pH which, in turn, is known to drive vascular tone. However, the exact mechanism by which this modulation occurs is not completely understood (Sharabi et al, 2009).

Since hypercapnia is a common pathophysiologic condition accompanying various cardiorespiratory diseases (e.g., chronic obstructive pulmonary disease, hypoventilation syndromes), understanding its effects on CBF and metabolism are of pivotal clinical importance. The effect of hypercapnia-induced hyperperfusion is used as a diagnostic tool to measure cerebrovascular reactivity (CVR), providing information on the integrity of cerebral autoregulation and is used to assess intracerebral hemodynamics (Muller and Schimrigk, 1996). Cerebrovascular reactivity also serves as a predictor of cerebrovascular disease progression (Liem et al, 2009), and its impairment is regarded as a risk factor for future vascular events such as stroke.

While the vascular effects of hypercapnia on CBF have long been recognized (Kety and Schmidt, 1948), its effects upon central nervous system activity and cerebral metabolism are not well established. Diagnostic and experimental applications of hypercapnia to the study of CVR are based on the assumption that mild alterations of pCO₂ have negligible effect on cerebral metabolism.

The first quantitative report of the effect of hypercapnia on cerebral metabolic rate of oxygen consumption (CMRO₂), which appeared almost 60 years ago, suggested conservation of cerebral metabolism upon inhalation of 5% to 7% CO₂ (Kety and Schmidt, 1948). Since then, numerous groups have investigated neuronal and metabolic implications of hypercapnia and reported a wide spectrum of responses ranging from increase (Jones et al, 2005), decrease (Kliefoth et al, 1979; Xu et al, 2011), and stasis of CMRO₂ (Chen and Pike, 2010; Kety and Schmidt, 1948; Massik et al, 1989). Electrophysiological studies investigating the effect of hypercapnia have not been conclusive either, with results ranging from extreme cortical depression to extreme arousal (with some even reporting seizures) (Woodbury and Karler, 1960). Differences in methodologies such as strength and duration of CO₂ stimulus, use of anesthetic agents, and species studied, etc., have confounded direct comparison and interpretation of these results.

Even though there is no complete consensus regarding the metabolic effects of hypercapnia, constant CMRO₂ is often assumed during increased pCO₂ for normalizing blood oxygen-level dependent response in functional magnetic resonance imaging (Davis et al, 1998; Kim et al, 1999). The functional blood oxygen-level dependent effect (increase in signal amplitude during functional activation due to decreased transverse relaxation rate, R₂*) is a complex function of CBF, cerebral blood volume (CBV), and CMRO₂. With the assumption of constant CMRO₂, the decrease in R₂* during hypercapnia results solely from an increase in CBF and commensurate reduction in oxygen extraction fraction. On the other hand, functional activation paradigms, such as visual stimulation, have been shown to cause a smaller relative decrease in R₂*, the difference being ascribed to greater oxygen extraction due to increased CMRO₂ (Kim et al, 1999). Hence, hypercapnia has been used to ‘calibrate’ blood oxygen-level dependent signal against purely hemodynamic effects without concomitant metabolic changes.

The primary motivation for the present research was to further elucidate the effect of hypercapnia on CMRO₂ using a more advanced methodology. The current work builds on some of the authors' recently conceived method for simultaneously measuring venous oxygen saturation (S_vO₂) and total cerebral inflow (tCBF) to ensure efficient, accurate, and reliable CMRO₂ quantification, with a temporal resolution of about two measurements per minute (Jain et al, 2010). The method measures venous blood oxygen saturation in the superior sagittal sinus (SSS) via magnetic resonance (MR) susceptometry, yielding the intravascular bulk magnetic susceptibility relative to surrounding tissue, and simultaneously measures arterial flow in the major inflow vessels (internal carotid and vertebral arteries) by phase-contrast MR imaging (MRI).

Materials and methods

Ten nonsmoking, healthy volunteers (males, age: 28 ± 5 years) were recruited for the study after giving informed consent. The subjects were judged to be healthy on the basis of their medical history and physical examination. The particular population demographic was chosen to ensure maximal subject compliance to the hypercapnia challenge protocol. The protocol was approved by the Institutional Review Board of University of Pennsylvania.

Measurement Principle

The S_vO₂ quantification relies on the relative magnetic susceptibility difference, Δ_χ, of intravascular blood and surrounding tissue. The vessel of interest is modeled as a long paramagnetic cylinder (Fernández-Seara et al, 2006; Haacke et al, 1997) and oxygen saturation (%HbO₂) is determined as

% H b O 2 = [ 1 - 2 | Δ ϕ | γ Δ χ do B o Δ T E ( cos 2 θ - 1 / 3 ) H c t ] × 100

(1)

where Δφ is the average phase difference between intravascular blood and surrounding tissue and is measured as

Δ ϕ = γ Δ B ⋅ Δ T E

(2)

ΔB is the incremental field inside the vessel relative to that of the adjacent tissue and ΔTE is the interecho spacing between two gradient-recalled echoes. Δ_χdo = 4π(0.27) (Spees et al, 2001) is the susceptibility difference in SI units between fully deoxygenated and oxygenated erythrocytes, Hct is the volume fraction of the packed erythrocytes in whole blood, and θ is the tilt angle of the vessel with respect to the main field B_o.

Total CBF was quantified by using phase-contrast MRI. The method utilises motion sensitizing gradient waveforms to encode information about velocity into the phase of the MR signal. Specifically, the pulse sequence involves two interleaves; both having null zeroth gradient moment along the direction of blood flow (M₀ = ∫ G(t)dt = 0) but nonzero first gradient moments (M₁ = ∫ G(t)tdt ≠ 0). The latter determines the sensitivity of the accrued phase difference between the two interleaves, Δφ, to the velocity of the flowing spins and is expressed as,

Δ ϕ = γ Δ M 1 v

(3)

where ΔM₁ is the difference in first moment between the two interleaves and is dictated by a user-defined parameter VENC, defined as

Δ M 1 = π / γ ⋅ V E N C

(4)

VENC represents the velocity that causes a net phase accumulation of π radians and is typically chosen to be 30% higher than the maximum velocity expected.

Finally, CMRO₂ was calculated via Fick's equation,

C M R O 2 = C a C B F ( S a O 2 - S v O 2 )

(5)

where CMRO₂ is in μmol/min per 100 g, CBF is the total cerebral blood inflow to the brain in mL/100g per minute, S_aO₂ and S_vO₂ represent the arterial and venous oxygen saturation levels, respectively, C_a is the oxygen concentration in moles of O₂ per 100 mL of blood and was determined in each subject based on their hemoglobin level assuming an O₂ carrying capacity of 1.39 mL of O₂ per gram of hemoglobin (West, 2007).

Magnetic Resonance

An interleaved gradient-echo pulse sequence was used for simultaneous measurement of S_vO₂ and tCBF at two different anatomic locations. Specifically, S_vO₂ was quantified at the level of SSS using MR susceptometry based oximetry and tCBF was determined in the internal carotid and vertebral arteries via phase-contrast MR described above. Additional details on the pulse sequence is given in some of the authors' previous work (Jain et al, 2010).

Images were acquired on a 3T Siemens Tim Trio system (Siemens Medical Solutions, Erlangen, Germany) using a 12-channel head and neck coil. A vendor provided sagittal FLASH (fast low angle shot) sequence was used to obtain an angiogram of head and neck vessels. The sequence is based on a fast rephased-dephased gradient-echo sequence and generates two echoes per excitation to provide images where static tissue is eliminated while signal from flowing blood is preserved. Sequence parameters: 280 × 280, resolution: 1.5 × 1.5 × 40 mm³, slices = 3, TR = 38 milliseconds, TE = 7.2 milliseconds, flip angle = 15°, total scan time = 22 seconds. This, in conjunction with gradient-echo axial localizer scans was used to choose the straightest section of vessels of interest (SSS, internal carotid and vertebral arteries) and estimate the tilt angle θ of the SSS with respect to B_o from the coordinates of the centroid of the vessel (Δz ∼ 4 cm) (Langham et al, 2009).

Scan parameters: Voxel size = 1 × 1 × 5 mm³, flip angle = 25°, TR = 35 milliseconds, echo spacing = 3.5 milliseconds for S_vO₂ quantification and voxel size = 0.85 × 0.85 × 5 mm³, VENC = 60 cm/s (normocapnia) and 100 cm/s (hypercapnia) for tCBF quantification, total scan time = 28 seconds. Additionally, a T1-weighted 3D magnetization-prepared rapid gradient-echo image data set (voxel size = 1 × 1 × 1 mm³) was used to estimate intracranial volume (Mugler and Brookeman, 1990). Total brain volume was obtained using a semiautomated region-growing algorithm (ITKSnap; Yushkevich et al, 2006).

It is noted that the typical tilt angles of SSS in our study were < 15°. Since the induced field outside a cylinder falls off as the inverse square of the distance and the square of the sine of the tilt angle (Haacke et al, 1999), the magnetic field outside the SSS was negligibly perturbed (note in Figures 2C–2E the sharp drop in the net accumulated phase outside the SSS and the near constancy of the field in the reference tissue). A region of interest was chosen within the SSS and in the surrounding homogenous brain tissue at a distance on the order of the diameter of the vessel and Δφ was calculated by subtracting the net accumulated phase in the two regions.

A linear model for cerebral vascular response to hypercapnia was assumed and CVR was estimated as the slope of the %ΔtCBF versus ΔEtCO₂ curve. Analysis of variance was used to test for differences in group means of S_vO₂, tCBF, and CMRO₂ between the three test conditions (baseline, hypercapnia, and recovery).

Results

Inhalation of 5% CO₂ gas mixture produced similar changes in EtCO₂ across all subjects. Figure 1 shows the time course of various physiological parameters in one of the subjects (heart rate, S_aO₂, and EtCO₂) recorded during the experiment. Group means for all subjects are listed in Table 1. Figure 2 displays the magnitude and phase-difference images at the level of the SSS (Figures 2B–2E) and velocity maps of the neck vessels (Figures 2F–2I). Notice the expected decrease in phase contrast for the SSS, indicating increased S_vO₂ levels and the enhanced blood velocity in neck vessels during hypercapnia.

OPEN IN VIEWER

Figure 1

Time course plots of various physiological parameters: heart rate (HR), arterial oxygen saturation (S_aO₂), and end-tidal CO₂ (EtCO₂) during baseline, hypercapnia (5% CO₂ shaded), and recovery in a representative subject.

OPEN IN VIEWER

Figure 2

Images from which venous oxygen saturation (S_vO₂) and total cerebral blood flow (tCBF) during baseline, hypercapnia, and recovery were derived. (A) Sagittal localizer angiogram indicating the locations for S_vO₂ (top) and tCBF (bottom) measurements; (B) axial magnitude; (C-E) phase-difference images for S_vO₂ quantification highlighting superior sagittal sinus (SSS); (F) axial magnitude and (G-I) velocity images of the neck showing the major cerebral inflow vessels (internal carotid and vertebral arteries). Note the change in contrast in vessels during hypercapnia.

OPEN IN VIEWER

Table 1

Summary data of various parameters

	S a O 2 (%)	S v O 2 (%)	tCBF (mL/100g per minute)	CMRO 2 (μmol/100g per minute)	HR (b.p.m.)	EtCO 2 (mmHg)
Baseline	97 ± 1	65 ± 6	49 ± 6	134 ± 10	65 ± 13	39 ± 1
Hypercapnia	97 ± 1	78 ± 5*	80 ± 13*	130 ± 13	73 ± 13	47 ± 2*
Recovery	97 ± 1	64 ± 5	47 ± 6	132 ± 8	66 ± 13	38 ± 2

CMRO₂, cerebral metabolic rate of oxygen consumption; EtCO₂, end-tidal CO₂; HR, heart rate; S_aO₂, arterial oxygen saturation; S_vO₂, venous oxygen saturation; tCBF, total cerebral blood flow.

Group means and s.d. of various measured and physiological parameters.

*

Indicates statistically significant changes (P < 0.005).

The initial tCBF and S_vO₂ response to hypercapnia is estimated to be < 30 seconds, consistent with previously reported results. The response stabilized after ∼1 to 2 minutes for all subjects (Figure 3). Additionally, an average undershoot of 6.7% ± 4% and 17.1% ± 7% was observed in S_vO₂ and tCBF values upon recovery from hypercapnia in six subjects (Figure 3).

OPEN IN VIEWER

Figure 3

Plot of time-resolved measurements of total cerebral blood flow (tCBF) and venous oxygen saturation (S_vO₂) during normocapnia and hypercapnia (shaded) in a representative subject. Note the undershoot during recovery from hypercapnia.

The group averages for the various measured parameters during the three experimental conditions are given in Table 1. A significant linear correlation was observed between percent ΔtCBF and ΔEtCO₂ (R² = 0.72, P < 0.005), and CVR, computed as the slope of the above curve, was 6.1% ± 1.6%/mmHg (Figure 4A). Further, CVR was negatively correlated with baseline tCBF, even though this relationship was only marginally significant (R² = 0.44, P = 0.052; Figure 4B). Baseline S_vO₂ and tCBF were correlated as expected (R² = 0.72, P < 0.005; Figure 4C). Finally, significant changes in both S_vO₂ and tCBF were observed during hypercapnia (both P < 0.005; Figures 5A and 5B), whereas global CMRO₂ remained unchanged (P = 0.78; Figure 5C).

OPEN IN VIEWER

Figure 4

(A) Percent change in total cerebral blood flow (CBF) versus change in end-tidal CO₂ (ΔEtCO₂) in response to hypercapnia. The slope of the curve represents cerebrovascular reactivity (CVR); relationship between (B) intersubject CVR and baseline tCBF and (C) baseline venous oxygen saturation (S_vO₂) versus tCBF.

OPEN IN VIEWER

Figure 5

Scatter plot of functional parameters for all subjects during baseline, hypercapnia (shaded), and recovery: (A) total cerebral blood flow (tCBF); (B) venous oxygen saturation (S_vO₂); (C) cerebral metabolic rate of oxygen consumption (CMRO₂). The vertical span of each diamond represents the 95% confidence interval (CI) of the group means, symbols refer to individual subjects.

Discussion

The effect of increased blood CO₂ on neuronal activity and cerebral oxidative metabolism has been a subject of intensive investigation during the past six decades. In this study, we report simultaneous measurement of metabolic (S_vO₂) and hemodynamic (tCBF) parameters in response to mild hypercapnia (5% inspired CO₂), in healthy, awake, adult subjects by means of quantitative MRI. Our results support the notion of negligible change in CMRO₂ with a mild hypercapnic stimulus. Previous human studies with different methodologies, but also invoking Fick's principle for CMRO₂ quantification (Chen and Pike, 2010; Kety and Schmidt, 1948; Massik et al, 1989), as well as in vivo small animal studies measuring neuronal metabolites of oxidative phosphorylation (adenosine triphosphate, adenosine diphosphate, and adenosine monophosphate), have reported similar results upon exposure to mild hypercapnia (Buckweitz et al, 1980; Folbergrova et al, 1974).

In distinction to prior work, our experimental approach allows for rapid, simultaneous measurement of S_vO₂, tCBF, and EtCO₂. The ability to simultaneously quantify two key parameters in Fick's equation (equation 5) gives us particular confidence in the veracity of our CMRO₂ estimates as opposed to all prior work requiring sequential measurements (Chen and Pike, 2010; Xu et al, 2011). Additionally, our method enables point-by-point temporal analysis of vascular and metabolic responses to hypercapnia with a temporal resolution of 30 seconds. This improves the experimental sensitivity to visualize short time scale responses to pCO₂ changes (e.g., detection of temporal lags in achieving a steady-state response due to abrupt induction and cessation of hypercapnia). Even the most recently conducted hypercapnia studies have precluded such an evaluation (Chen and Pike, 2010; Xu et al, 2011).

The observed experimental CVR value of 6.1% ± 1.6%/mm Hg, obtained using a linear-fit model, is in the range of previously reported values (4% to 7%/mm Hg) (Chen and Pike, 2010; Noth et al, 2006; Pollock et al, 2009). The between-subject variability of ∼21% in the individual CVR measurements is in good agreement with coefficients of variation reported previously (Kassner et al, 2010; Totaro et al, 1998). While the underlying mechanism for this variation in CVR is not fully understood, we conjecture that it might be explainable, at least in part, by intersubject variations in baseline tCBF. Support for such an interpretation is the observation of a moderate negative correlation, approaching statistical significance, between vascular response to hypercapnia (CVR) and baseline tCBF (R² = 0.44, P = 0.052). This association suggests that the capacity for vasodilatation by a hypercapnic stimulus might be a function of baseline cerebral flow. A similar negative correlation (though not statistically significant) between regional cerebral vascular response to hypercapnia and baseline CBF values in the occipital cortex had been noted in a positron emission tomography-based study by Ito et al (2008).

Of interest, our experimental data demonstrate concordance with many known principles of human physiology. For example, a strong positive correlation was observed between baseline S_vO₂ and tCBF, analogous to prior reports by Xu et al (2009), and more recently by some of the present authors (Jain et al, 2010). Under stable resting cerebral physiologic conditions, changes in flow are unlikely to change the rate of metabolism. This concept is based on the assumption that in a resting awake healthy human brain, the metabolic demand should be constant (Lajtha et al, 2007). As expected during hypercapnia, the observed increase in S_vO₂ was accompanied by a commensurate increase in tCBF, suggesting cerebral activity to remain invariant.

Our data indicate that the steady-state tCBF hypercapnic response is attained with some delay. Prior work has shown that the bulk of this delay is caused by the time required for the build-up of CO₂ in arterial blood; which in turn is a function of the level of ventilation, CO₂ diffusion time and pulmonary blood flow (Shapiro et al, 1965, 1966). However, studies have also noted that tCBF equilibration lags behind arterial pCO₂ stabilization. Therefore, we posit that the delay we observed might also be due to CO₂ mediating its physiologic effects via alterations in perivascular pH. Many animal studies have noted that decreases in pH result in vasodilatation and, conversely, increases cause vasoconstriction. These effects have been thought to be mediated by secondary messengers regulating smooth muscle vascular tone (Kontos et al, 1977).

In our experimental paradigm it is plausible to assume that as arterial pCO₂ increases during presentation of the hypercapnic stimulus, CO₂ diffuses into the brain parenchyma, eventually attaining a steady-state value. The initial changes in pH are likely offset, at least in part, by local tissue buffering at the intracellular and extracellular level (Celotto et al, 2008). However, as the hypercapnic stimulus persists and the steady-state parenchymal pCO₂ value is reached, these local homoeostatic measures are overwhelmed. The ensuing decrease in pH likely then triggers the production of secondary intracellular mediators at end arterioles, thereby lowering cerebral vascular resistance and producing the observed increase in blood flow (Brian, 1998). As these events depend on a series of processes occurring, it is reasonable to assume that the observed steady-state response would develop over a period of minutes rather than instantaneously. In our study, the initial changes in CBF were observed in < 30 seconds; however, the time to achieve steady-state CBF was on the order of 1 to 2 minutes. These time scales are consistent with previous reports (Poulin et al, 1996; Shapiro et al, 1965, 1966).

We also noted an average of 17% decrease in CBF below the baseline after cessation of the hypercapnic stimulus in 6 of the 10 subjects (Figure 3). Other investigators have also reported similar undershoots in tCBF during recovery from hypercapnia. Previous work by Claassen et al (2007) and Poulin et al (1996), using transcranial doppler ultrasound, as well as more recent work by Chen and Pike (2010), using arterial spin labeling (ASL), indicate similar trends in tCBF data.

Although the exact physiological mechanism of this undershoot in CBF is unclear, we hypothesize that this response might have been mediated by a combination of rapid respiratory egress of pCO₂ via hyperventilation, in conjunction with a slower reequilibration of cellular buffering mediators (such as tissue bicarbonate and inorganic bases). So, even after cessation of the hypercapnic stimulus, arterial pCO₂ likely remains above the normal physiologic range. This interpretation is supported by the presence of transiently elevated EtCO₂ (above baseline) even upon abrupt removal of the hypercapnic stimulus. Consequently, hyperventilation would persist as central pH-sensitive chemoreceptors in the respiratory centers of the pons and medulla, as well as peripheral receptors for pH and CO₂ in the carotid and aortic bodies continue to fire (Speck, 1992). However, as the respiratory clearance of CO₂ quickly ensues (Shapiro et al, 1965, 1966), the cellular chemical buffering mediators (initiated at the onset of hypercapnia) are likely to take longer to equilibrate, resulting in a brief period of tissue alkalosis. The proposed brief period of tissue alkalosis during recovery from hypercapnia has also been reported in a ³¹P-Nuclear Magnetic Spectroscopy-based rat study (Nishimura et al, 1989). This transient increase in pH may produce vasoconstriction at local end arterioles resulting in the observed undershoot in tCBF. It is noted that hyperventilation-induced hypocapnia is unlikely to be the reason for the observed undershoot as a similar trend in EtCO₂ data was not observed.

As pointed out earlier, numerous studies have investigated the vascular, neuronal, and metabolic implications of hypercapnia. While a review of the hypercapnia literature in its entirety is beyond the scope of the current article, two recent studies utilising MRI-based methods for CMRO₂ quantification and a study protocol somewhat similar to ours are of particular relevance, even though they are based on a different methodology for quantifying S_vO₂ and flow. Our results are in good agreement with those reported by Chen and Pike (2010). Specifically, they reported a conservation of CMRO₂ during mild hypercapnia and a CVR value very similar to ours (5.9 versus 6.1). Additionally, an undershoot in tCBF values was also observed in the presented tCBF time series plots. However, our results are at variance with those reported by Xu et al (2011), who observed a decrease in CMRO₂ during hypercapnia with the effect being dose dependent.

A few methodological differences could potentially have contributed to the discrepant results. Typically, including in our study, CMRO₂ is calculated as the product of whole brain blood flow and difference in oxygen saturation between arterial and SSS blood. However, Xu et al measured CBF in the SSS as representative of tCBF. Under physiological alterations such as hypercapnia, flow through the SSS would be a valid alternative for tCBF as long as the proportion of cerebral venous drainage remained constant. However, it is conceivable that in order to accommodate the dramatic increases in flow prompted by hypercapnia, other venous conduits might be recruited beyond the SSS. It is noted that the cerebral venous system is valveless; and several routes for drainage exist. The extent of venous drainage might vary under physiological challenges as the cerebral veins and sinuses are low compliance vessels with a fixed capacity for drainage. Hence, there might be great variability in the distribution of venous return when faced with increases in CBF such as the present experimental paradigm. The above could lead to underestimation of flow changes in response to hypercapnia when approximating tCBF from a measurement in the SSS.

We note that quantification of S_vO₂ in the SSS as representative of global S_vO₂ to be of lesser concern. Positron emission tomography data demonstrated little interregional variation in S_vO₂ during resting conditions in human subjects. Hence the SSS, which receives a large portion of the cerebral venous drainage, is likely to be a good indicator of global S_vO₂ (Lebrun-Grandie et al, 1983). Direct spectro-photometric measurements in anesthetized cats have given similar results (Buckweitz et al, 1980).

While our experiment lends support for negligible change in CMRO₂ in response to hypercapnia, our measurements do not preclude the possibility of regional variations in CMRO₂. In fact, evidence for such regional heterogeneities in neuronal response to hypercapnia have been reported previously (Biswal et al, 1997; Chen and Pike, 2010; St John, 1981). Such observations withstanding, we sought to evaluate only global metabolic response to hypercapnia in our experiment. Additionally, we investigated only the effects of mild and acute exposure to increased CO₂ levels. It is probable that a stronger hypercapnic stimulus, or a longer exposure might alter the CMRO₂. There is some evidence for such CMRO₂ changes based on previous animal studies (Zappe et al, 2008). Finally, it is possible that the small population size studied might have limited our ability to evaluate changes in CMRO₂ in response to hypercapnia. However, while designing the study, our analysis indicated that a subject size (N) of 10 would allow detection of 7% to 8% relative difference in any of the measured parameters, with 80% power and 5% significance. Reproducibility of CMRO₂ at 3T was previously found to be on the order of 8% (Jain et al, 2010).

In summary, this study sought to investigate the effect of mild hypercapnia (∼9 mm Hg) on global CMRO₂ by simultaneously measuring tCBF and global S_vO₂. The simultaneous measurement scheme is likely to improve accuracy and reliability of measuring CMRO₂ as opposed to performing sequential measurements. Additionally, the achieved temporal resolution (∼0.5 minutes) allows probing the temporal dynamics of vascular and metabolic effects of hypercapnia.