Sage Journals: Discover world-class research

Abstract

Blood oxygen level-dependent (BOLD) functional magnetic resonance imaging studies have shown the existence of ongoing blood flow fluctuations in the absence of stimuli. Although this so-called ‘resting-state activity’ appears to be correlated across brain regions with apparent functional relationship, its origin might be predominantly vascular and not directly representing neuronal signaling. To investigate this, we simultaneously measured BOLD and perfusion signals on healthy human subjects (n = 11) and used their ratio (BOLD/perfusion ratio or BPR) as an indicator of metabolic demand. BPR during rest and sleep was compared with that during a visual task (VT) and a breath-holding task (BH), which are challenges with substantial and little metabolic involvement, respectively. Within the visual cortex, BPR was 3.76 ± 1.23 during BH, which was significantly higher than during the VT (1.76 ± 0.27) and rest (1.56 ± 0.41). Meanwhile, BPR values during VT and rest were not significantly different, suggesting a similar metabolic involvement. Eight subjects showed stage 1 and 2 sleep, during which temporally correlated BOLD and perfusion activity continued. In these subjects, there was no significant difference in BPR between the sleep and waking conditions (1.79 ± 0.54 and 1.66 ± 0.67, respectively), but both were lower than the BPR during BH. These data suggest that resting-state activity, at least in part, represents a metabolic process.

Keywords

BOLD connectivity fMRI resting state sleep

Introduction

Mapping of brain function with blood oxygen level-dependent (BOLD) functional magnetic resonance imaging (fMRI) relies on the detection of focal changes in cerebral blood flow (CBF) in response to external stimuli. These changes are believed to support the changes in metabolic demand related to task-induced changes in electrical and chemical signaling (Logothetis et al, 2001) at neuronal synapses.

In the absence of stimuli, the brain continues to show temporal fluctuations in blood flow, as measured with BOLD and perfusion-based MRI (Biswal et al, 1995, 1997). This phenomenon of so-called ‘resting-state activity’ resembles the very low-frequency (< 0.1 Hz) fluctuations of spontaneous neuronal and metabolic activity, which were observed in animals that were deprived of external stimuli (Allers et al, 2002; Dora and Kovach, 1981; Leopold et al, 2003; Ruskin et al, 2003). The observed synchronized fluctuation across distributed regions is suggestive of the underlying functional relationship (Biswal et al, 1995; Cordes et al, 2000; Greicius et al, 2003; Hampson et al, 2002; Lowe et al, 2000). In addition, the regions showing correlated fluctuations are highly consistent within and across subjects (Beckmann et al, 2005; van de Ven et al, 2004), suggesting that they might allow mapping and classification of the networks that underlie human brain function without the need for carefully conditioned stimuli.

Despite their potential functional significance for neuroscience research, the origin of blood flow fluctuations in the absence of stimuli has not yet been established. Potential contributors that involve modulation in electrical and chemical signaling at neuronal synapses include cognitive processes, fluctuations in vigilance and conscious awareness, and homeostatic (restorative) and plasticity processes. An alternative possibility is that, rather than having a neuronal origin, the fluctuations are caused by physiologic fluctuations (e.g., vasomotor effect (Johansson and Bohr, 1966) or fluctuations in endtidal CO₂ (Wise et al, 2004)), without a substantial metabolic or neuronal component. As such, these fluctuations might bear little relevance to the mapping of functional networks.

Because of these confounding sources, the metabolic contribution to resting-state activity cannot be determined from BOLD or perfusion-based fMRI measurements alone. Nevertheless, several previous studies have shown that the analysis of simultaneously acquired BOLD and perfusion signals during a variety of tasks might reveal the level of the underlying metabolic activity (Chiarelli et al, 2007a; Davis et al, 1998; Hoge et al, 1999). In the context of this work, we designed an experiment to simultaneously acquire perfusion and BOLD signals during resting-state activity and tasks with varying metabolic demand, to estimate relative levels of metabolic activity.

Methods

Theory

A number of studies have compared the relationship between BOLD and perfusion signals during metabolically demanding tasks, such as stimulation-evoked activity, and tasks with reduced or negligible metabolic involvement, such as hypercapnia through controlled CO₂ inhalation or breath-holding (BH) (Kety and Schmidt, 1948; Novack et al, 1953). Hoge et al (1999) found that the BOLD perfusion relationship could be accurately modeled by

Δ B O L D = M \cdot (1 - (1 + Δ C M R_{O_{2}})^{β} \cdot (1 + Δ C B F)^{α - β})

(1)

with ΔBOLD, ΔCBF, and ΔCMR_O₂ normalized to baseline levels; M, a scaling constant dependent on MRI acquisition parameters and physiologic baseline state; and the physiologic parameters α and β estimated at 0.38 and 1.5, respectively (Davis et al, 1998). The relationship between ΔCMR_O₂ and ΔCBF was found to be close to linear (Chiarelli et al, 2007b; Hoge et al, 1999) with a proportionality constant k dependent on metabolic demand, leading to

Δ B O L D = M \cdot (1 - (1 + k \cdot Δ C B F)^{β} \cdot (1 + Δ C B F)^{α - β})

(2)

In the absence of metabolic fluctuations, k = 0, whereas metabolic processes evoked by a visual stimulation task (VT) or motor task are associated with nonzero, positive k values in the range of 0.24 to 0.51 (Ances et al, 2008; Chiarelli et al, 2007a; Hoge et al, 1999; Kastrup et al, 2002; Stefanovic et al, 2005; Uludag et al, 2004). Equation (2) indicates that k directly affects the BOLD—perfusion relationship, as illustrated in Figure 1, for a range of k values. Figure 1 suggests that for a certain level of perfusion fluctuations, the expected BOLD fluctuations are highest in the absence of a metabolic component (k = 0). Furthermore, the relative size of BOLD and perfusion fluctuations (henceforth referred to as BOLD/perfusion ratio or BPR) varies substantially in the observed range of k = 0 to 0.51. For example, in the approximation ΔCBF ≪ 1, equation (2) simplifies to

Δ B O L D \approx M \cdot (β \cdot (k + 1) + α) \cdot Δ C B F

(3)

with BPR = M (β (k + 1) + α), which, for α = 0.38 and β = 1.5, ranges from 1.12 to 0.36 M in the range of k = 0 to 0.51. According to these numbers, the proportionality constant between BOLD and perfusion signals can be up to 3.1 times higher during BH than VT. This indicates that BPR values can be used to estimate the metabolic involvement in resting-state activity.

Figure 1

Simulated BOLD—perfusion relationship based on equation (2) using α = 0.38, β = 1.5, M = 0.2, and a range of physiologic k values. Note that increasing k (increasing metabolic contribution) shifts the relationship to a lower (less steep) curve. Perfusion changes are expressed as a fraction of baseline perfusion levels. BOLD, blood oxygen level-dependent; CBF, cerebral blood flow.

Experimental Design

To determine whether spontaneous fMRI activity during the resting state has a substantial metabolic component, we performed simultaneous BOLD and perfusion MRI studies on 11 volunteers (7 males, 4 females, age: 22 to 38 years, average: 28 years) during a 40-min long experiment. Informed consent was obtained from all subjects before conducting the study, which was approved by the Institutional Review Board at the National Institutes of Health under protocol number 00-N-0082.

The experiment had three separate conditions (Figure 2): a VT, a BH task, and a resting state (REST). Visual and BH tasks were designed to strongly modulate fMRI signals with substantial and small or negligible fluctuation in metabolic demand, respectively. A BH task was chosen over controlled CO₂ inhalation for the ease of implementation. If fluctuations during REST occur in the absence of metabolic fluctuations, their associated fluctuations in cerebral metabolic rate of oxygen (CMR_O₂) should be significantly lower than CMR_O₂ fluctuations during VT, and similar to CMR_O₂ fluctuations during BH.

Figure 2

Timeline of simultaneous blood oxygen level-dependent and perfusion functional magnetic resonance imaging experiment. The experiment starts with a 5 mins and 20 secs task protocol, in which the central and peripheral portions of the visual field are alternatingly stimulated for 40-sec periods (visual task or VT). One minute after the VT, a breath-holding (BH) task follows, with five repeated cycles. Each cycle consists of taking in a single deep breath, followed by 40 secs of BH and 80 secs of normal breathing. The final part of the experiment consists of approximately 23 mins of rest, during which subjects close their eyes and are allowed to fall asleep (REST).

Stimulus Presentation

Before the start of each study, the subject was informed of the details of the experiment. In addition, subjects were asked to pay attention during VT, avoid head motion, and were encouraged to fall asleep during REST.

The visual stimulus, presented during VT at the initial part of the experiment (see Figure 2), lasted 320 secs and consisted of a contrast-reversing checkerboard (7.5 Hz) stimulating either the central (0 to 5°) or peripheral (5 to 15°) part of the visual field in eight blocks of 40 secs duration each. To ensure alertness during this part of the experiment, and minimize saccades, subjects were instructed to focus on a small dot that was presented in the center of the checkerboard display, and push a button on an fORP MRI compatible response box (Current Design Inc., Philadelphia, PA, USA) when the color of the dot changed. Color changes were randomized in time and occurred on average at the rate of 0.05/sec. The visual stimuli were back-projected onto a translucent screen with a U2-1200DLP data projector (PLUS Vision Corp., Tokyo, Japan). Button presses were recorded using Presentation 0.80 software (Neurobehavioral Systems Inc., Albany, CA, USA), which also recorded the timing of the MRI acquisition.

The BH task consisted of five cycles (120 secs each) of a BH paradigm, starting 1 min after the end of the VT. At the beginning of each cycle, subjects were prompted to take a deep breath, followed by 40 secs of breath-hold. During the remaining 80 secs of each cycle, the subjects were instructed to breathe freely.

The resting period that followed BH lasted 23 mins and 40 secs. Room lights were dimmed and subjects were instructed to close their eyes and relax. They were allowed to sleep.

Magnetic Resonance Imaging Scanning

All experiments were performed on a SIGNA VH/i 3T MRI scanner (GE, Milwaukee, WI, USA), retrofit with a 16-channel receiver system. For optimal sensitivity, a 16-channel detector array was used, designed in-house, and constructed by Nova Medical Inc. (Wilmington, MA, USA). The MRI acquisition method was based on a technique described by van Gelderen et al (2005), which allowed simultaneous measurement of BOLD and perfusion signals. This method allows for minimizing the contribution of stationary tissue and BOLD signals to the perfusion measurement. Suppression of stationary tissue signal was achieved with a double-inversion labeling method. Five axial-oblique slices covering the primary visual cortex (VC) were acquired using the following parameters: resolution, 3.75 mm × 3.75 mm × 5.0 mm; inversion delays, 1,320 ms (TI1) and 450 ms (TI2); repetition time, 4 secs; echo times, 20 and 35 ms for perfusion and BOLD acquisition, respectively. For optimal temporal resolution, perfusion measurements were performed without reference acquisition. On three subjects, additional data were acquired (using an additional VT condition) with perfusion reference to allow the estimation of changes relative to baseline perfusion. Signal-to-noise ratio in the individual BOLD images ranged from 200:1 in deep brain structures to 500:1 in cortical areas.

Physiologic Monitoring

To monitor the physiologic and cognitive state of the subjects, a number of physiologic signals were recorded in synchrony with MRI acquisition. These signals included electroencephalography (EEG), cardiac and respiratory cycles, and button presses (subjects' response during VT).

Electroencephalographic data, collected using a 64-channel Neuroscan Synamps2 system (Compumedics USA Ltd, El Paso, TX, USA), served primarily to measure the subjects' alertness and to detect sleep. Signals were collected using 40 channels of a sintered Ag/AgCl electrode cap connected via carbon fibers and RF filters to amplifiers and digitizers. Electrodes were placed to cover most of the scalp. Two of the channels were bipolar and were used to allow monitoring of the electrooculogram and electrocardiogram. Acquisition of EEG was performed continuously, in DC mode and with a digitization rate of 10 kHz and analog low-pass filtering at 2 kHz. At each MRI repetition interval (TR), a trigger pulse from the scanner was recorded and used for localization of the gradient artifact (see below).

Magnetic Resonance Imaging Data Analysis

All MRI data analyses were performed using IDL 6.2 software (ITT Visual Information Solutions, Boulder, CO, USA). After EPI image reconstruction, to normalize the BOLD and perfusion signal time course to the resting baseline, both BOLD and perfusion images were divided by the first image in the BOLD-weighted time series. Subsequently, linear drift correction and band-pass filtering between 0.01 and 0.1 Hz were performed on both BOLD and perfusion data. The specific choice of pass-band was made to reduce nonlinear drift and thermal noise, while minimally affecting the predominant frequencies of BOLD and perfusion signals during VT, BH, and REST. After these preprocessing steps, VT, BH, and REST data segments were analyzed separately.

Based on the VT data, a functional (activated) region of interest (ROI) was selected by comparing the signal level during active stages to that during nonactive stages of visual stimuli. This was carried out for foveal and peripheral stimuli separately. This ROI roughly covered most of the lower visual areas. Pixels were considered activated when significant (t > 4.5) increases were observed in both BOLD and perfusion signals.

Subsequently, potential time shifts between BOLD and perfusion signals were determined by correlating these signals at different time shifts (lags). This was done to compensate for the difference in acquisition time of these two signals, and time shifts related to inherent contrast mechanisms (e.g., vascular delays). A perfusion lag of −1 time point (−1.77 secs) was found to be optimal for all subjects. This value was used to correct (time shift) all perfusion data (VT, BH, and REST) before further analysis. For the REST data, global intensity (average of whole-brain signal) fluctuations were removed by using the brain-averaged signal as a regressor in a pixel-by-pixel manner. This was done for BOLD and perfusion data separately.

Finally, to determine the proportionality constant between BOLD and perfusion signals (see equation (2)), linear regression was performed on the average signals in the functional ROI in the VC for VT, BH, and REST periods. For VT, foveal and peripheral signals were averaged after time-shifting to account for the differences in stimulation paradigm between these regions. In addition, to estimate the fluctuation level, root-mean-square (RMS) of the averaged signal time course in the VC was measured for each period separately. Since this fluctuation is a spontaneous phenomenon, because of lack of the reference for regression analysis as in task-evoked activity, RMS value was employed in this study. To determine the differences between sleep and wake states, this analysis was also performed separately on the eight subjects who fell asleep, as determined from EEG (see below).

To evaluate the presence of ongoing activity, the through-plane spatial correlation was determined on a pixel-by-pixel basis. This was carried out for each experimental condition. Through-plane spatial correlation was chosen rather than in-plane correlation to avoid potential bias (spatial blurring) inherent to the image acquisition process.

In addition, further analysis was performed using ROI selection based on spatial correlation during REST. Similar to the methodology used in earlier studies of resting-state activity (Biswal et al, 1995; Fox and Raichle, 2007), a ‘seed’ voxel was placed in a specific anatomic location, and an ROI was generated by selecting all pixels that had a correlation of 0.3 or higher with the time course of the seed voxel in both BOLD and perfusion data. Regions of interest were placed in a number of locations, including VC, lateral temporo-parietal cortex, sensory-motor cortex, and prefrontal cortex based on the anatomic MRI data. Linear regression was performed on the average signals in each ROI.

Electroencephalographic Data Analysis

Electroencephalographic data were analyzed using Scan 4.3.2 software (Neuroscan; Compumedics USA Ltd, El Paso, TX, USA). After low-pass filtering of the data at 250 Hz (IIR, 48 dB/oct), gradient artifacts were removed based on template subtraction (Allen et al, 2000). The parameters of template subtraction were selected as follows: a window from −50 to 3,955 ms around the scanner's trigger, phase shift of 99 points, and window size of 3 events. Data were then downsampled to 200 Hz and subsequently band-pass-filtered (IIR, 24 dB/oct) to select a 0.5 to 28 Hz interval. A template of the cardioballistic artifact was constructed for each recording and used to extract principal components. These components were removed from the filtered data set.

Electroencephalographic data sets were visually inspected by a sleep expert (TJB) blinded to the fMRI data. A score (0: wake, 1: stage 1 sleep, 2: stage 2 sleep, 3/4: slow-wave sleep, 5: REM (rapid eye movement) sleep) was assigned for each 30-sec interval (Rechtschaffen and Kales, 1968). These values were then averaged over 2-min intervals and used to discriminate sleep from waking conditions. Sleep was defined as an average score greater than 0 over a 2-min interval. The longest single sleep interval was chosen for each (sleep) subject.

Results

All subjects actively performed both BH task and VT, as indicated from button responses and the signal recorded from the respiratory bellows. Eight of the 11 studies showed extended periods of sleep (stage 1 and 2 sleep only) lasting at least 5 mins. No slow-wave (stage 3 or 4) or REM sleep was detected.

In all subjects, a significant amount of spatially correlated activity was found outside the VC region that was active during the task. This was the case for both VT and REST periods, with correlation levels (averaged over covered brain area) during REST (0.20 ± 0.03) significantly (P < 0.05) higher than during VT (0.16 ± 0.04). In addition, during REST, BOLD and perfusion signals were correlated in most part of the brain (Figure 3). Significant correlation was observed primarily in gray matter, with little involvement of white matter and cerebrospinal fluid. Within the VC of most subjects, large fluctuations in both BOLD and perfusion signals were observed. From the additional perfusion reference data acquired on three subjects, it was estimated that during VT, perfusion and BOLD signals increased about 39% ± 6.7% and 2.19% ± 0.34% above baseline values leading to M = 0.23 for α = 0.38, β = 1.5, and k = 0.51 (Hoge et al, 1999). The estimated M value was higher than that found in a recent study (Chiarelli et al, 2007a). This difference might relate to difference in pulse sequences and parameters, and the assumed k value. An example of the time course of BOLD and perfusion signals in the VC is shown in Figure 4. The fluctuation level during REST approached the level of task-related signal change observed during VT. Average RMS fluctuation levels were 1.09 ± 0.28, 1.31 ± 0.22, and 0.54 ± 0.14 for VT (RMS-B_VT), BH (RMS-B_BH), and REST (RMS-B_REST), respectively, for BOLD, and 0.54 ± 0.18 (RMS-P_VT), 0.31 ± 0.10 (RMS-P_BH), and 0.22 ± 0.09 (RMS-P_REST), respectively, for perfusion (all values are a percentage of the average BOLD-weighted image intensity; see also Figure 5). RMS-B_VT and RMS-B_BH were significantly higher than RMS-B_REST (P = 5.2 times 10⁻⁶ and P = 9.1 times 10⁻⁹, t-test with Bonferroni correction for multiple comparisons). RMS-P_VT was significantly higher than RMS-P_BH and RMS-P_REST (P = 5.0 times 10⁻³ and P = 5.3 times 10⁻⁶). In most subjects, a substantial correlation between BOLD and perfusion signals was observed during all experimental stages, including the waking state as well as stage 1 and 2 sleep.

Figure 3

Example (single subject data) of correlated activity (see Methods) in one of the five slices studied. Images in each of the three columns indicate (in color) spatial correlation of BOLD (blood oxygen level-dependent) signal (A, D), spatial correlation of perfusion signal (B, E), and correlation between BOLD and perfusion signals (C, F). The top and bottom rows show correlation during VT (visual task) (A to C) and REST (D to F), respectively. For all images, a correlation threshold of ± 0.4 was used.

Figure 4

Example of time course EEG signals (parietal electrode), sleep stage, and BOLD and perfusion signals in the visual cortex (region activated during VT). Perfusion signals were normalized to baseline BOLD signal (see Methods). Image segments (top to bottom) show experimental timeline, EEG spectral amplitude, sleep stage (see Methods), and average fMRI signal evolution in functional region of interest. BOLD (black line) and perfusion fMRI (red line) signals show substantial correlated activity during each of the experimental stages, including VT, BH, and REST. This activity persists during sleep periods (sleep stage > 0), which is characterized by a reduction in alpha power (8 to 12 Hz), and an increase in extended theta power (2 to 7Hz). BH, breath-holding; BOLD, blood oxygen level-dependent; EEG, electroencephalography; fMRI, functional magnetic resonance imaging; VT, visual task.

Figure 5

Perfusion and BOLD signals in visual cortex during three experimental stages. (A) Relationship between BOLD and perfusion signals (single subject data). The dashed line represents a linear fit to the data. The relative amplitude of BOLD with respect to perfusion fluctuations (BPR) is similar during REST and VT, and increases during BH. (B) Mean fluctuation level of BOLD and perfusion signals (%RMS ± s.d.) and BPR in visual cortex for all subjects. *P < 0.005 indicate significant differences. BH, breath-holding; BOLD, blood oxygen level-dependent; BPR, BOLD/perfusion ratio; RMS, root-mean-square; VT, visual task.

The average ratios of BOLD and perfusion signals in VC were 1.76 ± 0.27, 3.76 ± 1.23, and 1.56 ± 0.41 for VT, BH, and REST periods, respectively (Figure 5). Normalized to the ratio during VT, the relative BOLD/perfusion ratios during BH and REST (BPR_BH and BPR_REST) averaged 2.16 ± 0.73 and 0.90 ± 0.25, respectively. BPR_BH was significantly higher than BPR_VT and BPR_REST (P = 2.7 times 10⁻⁶ and P = 5.0 times 10⁻⁷), whereas BPR_VT was not significantly different from BPR_REST (Figure 5). Normalized BPR_BH was somewhat lower than the theoretical value (equation (3)) of 3.1. This is possibly because of inaccurate model assumptions or parameter estimates.

Within VC of sleep subjects, BPR_VT, BPR_BH, and BPR_REST averaged 1.76 ± 0.30, 3.75 ± 1.44, and 1.60 ± 0.47, respectively. Again, BPR_BH was significantly higher than BPR_VT and BPR_REST (P = 6.4 times 10⁻⁴ and P = 2.7 times 10⁻⁴), whereas BPR_VT was not significantly different from BPR^REST. During the individual sleep and wake segments of REST, BPR_SLEEP and BPR_WAKE averaged 1.79 ± 0.54 and 1.66 ± 0.67, respectively, which were significantly lower than BPR_BH (P = 8.6 times 10⁻⁴ and P = 5.0 times 10⁻³), whereas BPR_SLEEP was not significantly different from BPR_WAKE.

During REST, a substantial correlation between perfusion and BOLD signals was also observed in the ROI derived from seed-point analysis (see Methods). The ROIs in VC, lateral temporo-parietal cortex, sensory-motor cortex, and prefrontal cortex all showed BPR_REST; in VC, BPR_VT values were also shown, which were significantly higher than BPR_BH (Figure 6 and Table 1). In general, for brain regions outside VC, BOLD and perfusion signals did not consistently show sufficient fluctuation above noise level to reliably determine BPR during VT. In addition, BPR in these regions might reflect a mixture of ongoing and task-evoked activity. In fact, BPR_VT was often unreliable because of the low level of correlation outside VC during this stage. This is attributed to a reduced level of signal fluctuation, possibly because of a reduction in ongoing activity owing to increased attention demands during VT (Nir et al, 2006). For this reason, BPR_VT values in these regions were not included in the comparison between the various conditions.

Table 1

Mean fluctuation level (%RMS) of BOLD and perfusion signals and BPR in several brain regions

	Visual (n = 11)		Temporo-parietal (n = 7)		Sensory-motor (n = 6)		Prefrontal (n = 6)
	Mean	s.d.	Mean	s.d.	Mean	s.d.	Mean	s.d.
BOLD
VT	0.841	0.147	0.276	0.115	0.225	0.006	0.254	0.088
BH	1.431	0.277	1.108	0.239	1.064	0.191	0.922	0.182
REST	0.505	0.142	0.418	0.088	0.416	0.123	0.335	0.078
Perfusion
VT	0.431	0.100	0.134	0.051	0.127	0.063	0.123	0.034
BH	0.299	0.121	0.300	0.070	0.316	0.098	0.281	0.076
REST	0.175	0.076	0.215	0.050	0.211	0.056	0.170	0.025
BPR
VT	1.684	0.300	—	—	—	—	—	—
BH	3.897	1.475	3.200	0.757	2.960	0.596	2.820	1.069
REST	1.275	0.498	1.337	0.549	1.448	0.539	1.392	0.219

Abbreviations: BH, breath-holding; BOLD, blood oxygen level-dependent; BPR, BOLD/perfusion ratio; RMS, root-mean-square; VT, visual task.

Figure 6

Relationship between BOLD and perfusion signals during each of the experimental stages in a number of regions selected based on seed-point correlation analysis (single subject data). Regions include visual cortex (A), lateral temporo-parietal cortex (B), sensory-motor cortex (C), and prefrontal cortex (D). The dashed lines represent a linear fit to the data. No fit was performed on data obtained during VT from regions outside the VC, as negligible task-evoked activity was expected. In all regions, the BPR increases during the BH task. BH, breath-holding; BOLD, blood oxygen level-dependent; BPR, BOLD/perfusion ratio; VT, visual task.

Discussion

Major Findings

This is the first human study to investigate metabolic involvement in the phenomenon of spatially correlated BOLD signal fluctuations in the absence of stimuli. Using the fluctuation level of BOLD signal relative to simultaneously acquired perfusion signal, an indicator was derived for relative CMR_O₂ during rest, and compared with values found during tasks with substantial (VT) and insignificant (BH) metabolic involvement.

The main finding of this study is that in VC during REST, BOLD fluctuations are accompanied by perfusion fluctuations to an extent that is very similar to what occurs during VT, and significantly dissimilar to a BH task. The latter was true in the regions with strong ongoing activity outside VC as well.

This finding suggests that ‘resting-state activity’ is not a purely vascular effect, but has a metabolic component, and possibly supports a functional role. Among the possible candidates are neuronal signaling, homeostatic processes, and plasticity processes.

Additionally, it was found that resting-state metabolic activity does not only occur during waking rest but also during light (stage 1 and 2) sleep. This continued metabolic activity during sleep indicates that the processes underlying BOLD resting-state activity neither exclusively support cognitive activity nor activity underlying conscious awareness and monitoring of the environment, all of which are presumably reduced during sleep.

These findings are consistent with the presence of slow fluctuations (< 0.5 Hz) in cortical oxidative metabolism, which was observed by reflectance spectrophotometry in animals (Vern et al, 1988). Furthermore, as observed in animal studies (Vern et al, 1997, 1988), the observed fluctuations were synchronized over both hemispheres, and maintained during both wake and sleep.

In addition to establishing a metabolic origin of resting-state BOLD fluctuations, the current findings support the notion of a persistent level of brain connectivity during sleep, as suggested previously (Fukunaga et al, 2006; Horovitz et al, 2007). To some extent, this appears contradictory to a recent study that reports a reduced effect of transcranial magnetic stimulation excitation on the spatial extent of EEG activity during sleep as compared with waking rest conditions (Massimini et al, 2005). However, it must be noted that the study reports on slow-wave sleep, and measured effective cortico-cortical connectivity on relatively fast time scales (hundreds of milliseconds), which enabled the identification of a causal relationship (Massimini et al, 2005). In contrast, the current MRI study reports on functional connectivity on a multisecond time scale that might not be causal and/or originate or be guided by secondary nonneocortical regions, such as the thalamus (Steriade, 1999).

Limitations

Although the present study suggests the presence of metabolic processes with apparent functional relevance during rest and sleep, interpretation of the results is subject to a number of limitations.

One of the main limitations of the current study is that it does not allow the determination of absolute levels of cerebral perfusion or CMR_O₂. This is primarily because of the lack of a perfusion reference scan (scan without perfusion label), which was omitted to improve temporal resolution and accurately detect and capture perfusion fluctuation levels. It is therefore difficult to relate the current findings to positron emission tomographic CBF measurements of sleep, which suggest minimal or no reduction in VC activity during light sleep, and a substantial reduction during stage 3 and 4 non-REM sleep (Braun et al, 1997; for review see Maquet and Phillips, 1998).

Another limitation is the low sensitivity of the methodology, which makes it difficult to derive reliable measures of relative CMR_O₂ in focal regions outside the VC under all conditions (including a VT). It is hoped that transition to a higher magnetic field (7T), which is currently in progress in our laboratory, will alleviate this problem. Conversely, it will remain difficult to assess CMR_O₂ on a pixel-by-pixel basis, because of the fact that MRI measures of perfusion and BOLD do not originate from precisely the same location. This is because BOLD signals originate primarily from the venous side of the vasculature, whereas perfusion signals originate primarily from the arterial side.

Lastly, there are limitations to the model assumptions that might affect the comparison of relative level of metabolic fluctuations during the different stages of the experiment. The calculation of BPR values used for this comparison was based on an assumed linearity between BOLD and perfusion signals. With the actual size of perfusion changes occurring in the experiments (estimated at about 40% during VT; see Results), a substantial departure from linearity is expected (see Figure 1). Evidence of this nonlinearity can be seen in Figures 5A and 6, particularly with the BH challenge. In the presence of these nonlinearities, the BOLD—perfusion relationship cannot be accurately described with a single BPR value. This might partly explain the fact that the measured average BPR_BH value of 2.16 (relative to BPR_VT) was lower than the theoretical estimate of 3.1 derived from equation (3) for small perfusion changes. Other factors that might affect relative BPR levels are the differences in actual and assumed values for α, β, and k. Lastly, the relatively low value of BPR_BH might indicate that this task evokes a nonzero level of metabolic activity (Martin et al, 2006). Nevertheless, these limitations do not negate the main finding of this work, the observed similarity in BOLD—perfusion relationship (as characterized from BPR values) between VT and REST conditions, and the dissimilarity between these conditions and the BH condition.

Implications

The suggestion that resting-state BOLD activity represents metabolic and, therefore, most likely, functionally relevant processes has potentially important implications for the study of brain function. The findings suggest that the spatial clusters found with resting-state BOLD fMRI represent functional connectivity that is at least in part neuronal, rather than purely vascular in origin. However, they do not exclude the potential for a substantial nonmetabolic contribution, the size of which will depend on experimental conditions. The current findings validate the assumption that systematic identification of correlation clusters might provide an insight into the functional architecture under normal and pathologic conditions. The study of functional connectivity during sleep might provide new insights into the role and mechanistic aspects of sleep. In this context, it appears particularly interesting to use fMRI to study deep (slow-wave) sleep, which is believed to be functionally relevant for recovery and memory consolidation processes.

Footnotes

Acknowledgements

We thank Alan P Koretsky for helpful suggestions and Susan C Fulton for her assistance with the experiments.

References

Allen

Josephs

Turner

(2000) A method for removing imaging artifact from continuous EEG recorded during functional MRI. Neuroimage 12:230–9.

Allers

Ruskin

Bergstrom

Freeman

Ghazi

Tierney

Walters

(2002) Multisecond periodicities in basal ganglia firing rates correlate with theta bursts in transcortical and hippocampal EEG. J Neurophysiol 87:1118–22.

Ances

Leontiev

Perthen

Liang

Lansing

Buxton

(2008) Regional differences in the coupling of cerebral blood flow and oxygen metabolism changes in response to activation: implications for BOLD-fMRI. Neuroimage 39:1510–21.

Beckmann

DeLuca

Devlin

Smith

(2005) Investigations into resting-state connectivity using independent component analysis. Philos Trans R Soc Lond B Biol Sci 360:1001–13.

Biswal

Yetkin

Haughton

Hyde

(1995) Functional connectivity in the motor cortex of resting human brain using echo-planar MRI. Magn Reson Med 34:537–41.

Biswal

Van Kylen

Hyde

(1997) Simultaneous assessment of flow and BOLD signals in resting-state functional connectivity maps. NMR Biomed 10:165–170.

Braun

Balkin

Wesenten

Carson

Varga

Baldwin

Selbie

Belenky

Herscovitch

(1997) Regional cerebral blood flow throughout the sleep—wake cycle. An H2(15)O PET study. Brain 120(Pt7):1173–97.

Chiarelli

Bulte

Gallichan

Piechnik

Wise

Jezzard

(2007a) Flow-metabolism coupling in human visual, motor, and supplementary motor areas assessed by magnetic resonance imaging. Magn Reson Med 57:538–47.

Chiarelli

Bulte

Wise

Gallichan

Jezzard

(2007b) A calibration method for quantitative BOLD fMRI based on hyperoxia. Neuroimage 37:808–20.

10.

Cordes

Haughton

Arfanakis

Wendt

Turski

Moritz

Quigley

Meyerand

(2000) Mapping functionally related regions of brain with functional connectivity MR imaging. AJNR Am J Neuroradiol 21:1636–44.

11.

Davis

Kwong

Weisskoff

Rosen

(1998) Calibrated functional MRI: mapping the dynamics of oxidative metabolism. Proc Natl Acad Sci USA 95:1834–9.

12.

Dora

Kovach

(1981) Metabolic and vascular volume oscillations in the cat brain cortex. Acta Physiol Acad Sci Hung 57:261–75.

13.

Fox

Raichle

(2007) Spontaneous fluctuations in brain activity observed with functional magnetic resonance imaging. Nat Rev Neurosci 8:700–11.

14.

Fukunaga

Horovitz

Van Gelderen

De Zwart

Jansma

Ikonomidou

Chu

Deckers

Leopold

Duyn

(2006) Large-amplitude, spatially correlated fluctuations in BOLD fMRI signals during extended rest and early sleep stages. Magn Reson Imaging 24:979–92.

15.

Greicius

Krasnow

Reiss

Menon

(2003) Functional connectivity in the resting brain: a network analysis of the default mode hypothesis. Proc Natl Acad Sci USA 100:253–8.

16.

Hampson

Peterson

Skudlarski

Gatenby

Gore

(2002) Detection of functional connectivity using temporal correlations in MR images. Hum Brain Mapp 15:247–62.

17.

Hoge

Atkinson

Gill

Crelier

Marrett

Pike

(1999) Linear coupling between cerebral blood flow and oxygen consumption in activated human cortex. Proc Natl Acad Sci USA 96:9403–8.

18.

Horovitz

Fukunaga

De Zwart

Van Gelderen

Fulton

Balkin

Duyn

(2007) Low frequency BOLD fluctuations during resting wakefulness and light sleep: a simultaneous EEG—fMRI study. Hum Brain Mapp (in press).

19.

Johansson

Bohr

(1966) Rhythmic activity in smooth muscle from small subcutaneous arteries. Am J Physiol 210:801–6.

20.

Kastrup

Kruger

Neumann-Haefelin

Glover

Moseley

(2002) Changes of cerebral blood flow, oxygenation, and oxidative metabolism during graded motor activation. Neuroimage 15:74–82.

21.

Kety

Schmidt

(1948) The effects of altered arterial tensions of carbon dioxide and oxygen on cerebral blood flow and cerebral oxygen consumption of normal young men. J Clin Invest 27:484–92.

22.

Leopold

Murayama

Logothetis

(2003) Very slow activity fluctuations in monkey visual cortex: implications for functional brain imaging. Cereb Cortex 13:422–33.

23.

Logothetis

Pauls

Augath

Trinath

Oeltermann

(2001) Neurophysiological investigation of the basis of the fMRI signal. Nature 412:150–7.

24.

Lowe

Dzemidzic

Lurito

Mathews

Phillips

(2000) Correlations in low-frequency BOLD fluctuations reflect cortico-cortical connections. Neuroimage 12:582–7.

25.

Maquet

Phillips

(1998) Functional brain imaging of human sleep. J Sleep Res 7(Suppl 1):42–7.

26.

Martin

Jones

Martindale

Mayhew

(2006) Haemodynamic and neural responses to hypercapnia in the awake rat. Eur J Neurosci 24:2601–10.

27.

Massimini

Ferrarelli

Huber

Esser

Singh

Tononi

(2005) Breakdown of cortical effective connectivity during sleep. Science 309:2228–32.

28.

Nir

Hasson

Levy

Yeshurun

Malach

(2006) Widespread functional connectivity and fMRI fluctuations in human visual cortex in the absence of visual stimulation. Neuroimage 30:1313–24.

29.

Novack

Shenkin

Bortin

Goluboff

Soffe

(1953) The effects of carbon dioxide inhalation upon the cerebral blood flow and cerebral oxygen consumption in vascular disease. J Clin Invest 32:696–702.

30.

Rechtschaffen

Kales

(1968) A manual of Standardized Terminology, Techniques and Scoring System for SLEEP Stages of Human Subjects. Bethesda, MD: US Department of Health, Education, and Welfare Public Health Service. National Institutes of Health. National Institute of Neurological Disorders and Stroke. Neurological Information Network.

31.

Ruskin

Bergstrom

Tierney

Walters

(2003) Correlated multisecond oscillations in firing rate in the basal ganglia: modulation by dopamine and the subthalamic nucleus. Neuroscience 117:427–38.

32.

Stefanovic

Warnking

Kobayashi

Bagshaw

Hawco

Dubeau

Gotman

Pike

(2005) Hemodynamic and metabolic responses to activation, deactivation and epileptic discharges. Neuroimage 28:205–15.

33.

Steriade

(1999) Coherent oscillations and short-term plasticity in corticothalamic networks. Trends Neurosci 22:337–45.

34.

Uludag

Dubowitz

Yoder

Restom

Liu

Buxton

(2004) Coupling of cerebral blood flow and oxygen consumption during physiological activation and deactivation measured with fMRI. Neuroimage 23:148–55.

35.

Van de Ven

Formisano

Prvulovic

Roeder

Linden

(2004) Functional connectivity as revealed by spatial independent component analysis of fMRI measurements during rest. Hum Brain Mapp 22:165–78.

36.

Van Gelderen

CWH

De Zwart

Cohen

Hallett

Duyn

(2005) Resolution and reproducibility of BOLD and perfusion functional MRI at 3.0 Tesla. Magn Reson Med 54:569–76.

37.

Vern

Leheta

Juel

LaGuardia

Graupe

Schuette

(1997) Interhemispheric synchrony of slow oscillations of cortical blood volume and cytochrome aa3 redox state in unanesthetized rabbits. Brain Res 775:233–9.

38.

Vern

Schuette

Leheta

Juel

Radulovacki

(1988) Low-frequency oscillations of cortical oxidative metabolism in waking and sleep. J Cereb Blood Flow Metab 8:215–26.

39.

Wise

Ide

Poulin

Tracey

(2004) Resting fluctuations in arterial carbon dioxide induce significant low frequency variations in BOLD signal. Neuroimage 21:1652–64.

Metabolic Origin of Bold Signal Fluctuations in the Absence of Stimuli

Abstract

Keywords

Introduction

Methods

Theory

Experimental Design

Stimulus Presentation

Magnetic Resonance Imaging Scanning

Physiologic Monitoring

Magnetic Resonance Imaging Data Analysis

Electroencephalographic Data Analysis

Results

Discussion

Major Findings

Limitations

Implications

Footnotes

Acknowledgements

References