Skip to main content
NCBI home page
Journal of Cerebral Blood Flow & Metabolism logoLink to Journal of Cerebral Blood Flow & Metabolism
. 2015 Jun 24;35(11):1819–1826. doi: 10.1038/jcbfm.2015.130

Effects of anesthesia on BOLD signal and neuronal activity in the somatosensory cortex

Daniil P Aksenov 1, Limin Li 1, Michael J Miller 1, Gheorghe Iordanescu 1, Alice M Wyrwicz 1,*
PMCID: PMC4635237  PMID: 26104288

Abstract

Most functional magnetic resonance imaging (fMRI) animal studies rely on anesthesia, which can induce a variety of drug-dependent physiological changes, including depression of neuronal activity and cerebral metabolism as well as direct effects on the vasculature. The goal of this study was to characterize the effects of anesthesia on the BOLD signal and neuronal activity. Simultaneous fMRI and electrophysiology were used to measure changes in single units (SU), multi-unit activity (MUA), local field potentials (LFP), and the blood oxygenation level-dependent (BOLD) response in the somatosensory cortex during whisker stimulation of rabbits before, during and after anesthesia with fentanyl or isoflurane. Our results indicate that anesthesia modulates the BOLD signal as well as both baseline and stimulus-evoked neuronal activity, and, most significantly, that the relationship between the BOLD and electrophysiological signals depends on the type of anesthetic. Specifically, the behavior of LFP observed under isoflurane did not parallel the behavior of BOLD, SU, or MUA. These findings suggest that the relationship between these signals may not be straightforward. BOLD may scale more closely with the best measure of the excitatory subcomponents of the underlying neuronal activity, which may vary according to experimental conditions that alter the excitatory/inhibitory balance in the cortex.

Keywords: anesthesia, barrel cortex, BOLD contrast, electrophysiology, functional MRI (fMRI)

Introduction

Blood oxygenation level-dependent (BOLD) functional magnetic resonance imaging (fMRI) is a widely used technique for mapping brain activation in humans. To enable more accurate and sophisticated interpretation of fMRI results, much effort has been devoted in recent years to understanding the neurophysiological basis of the hemodynamic BOLD signal. Typically, this question is posed in terms of the relationship between BOLD and the common electrophysiological measures of neuronal activity: local field potentials (LFP), believed to represent primarily dendritic activity, and spiking as measured by single unit (SU) and multi-unit (MUA) responses. Animal models, which permit simultaneous fMRI and electrophysiolgy, have been vital for studying the spatial and temporal dynamics of this relationship. However, considerable variation has been reported in cortical mapping, vascular response dynamics and in the quantitative coupling between evoked neural and hemodynamic responses. These differences have been ascribed, in part, to the modulatory effects of anesthetics, a variety of which are usually used to limit subject movement in animal experiments. As described in a recent review, these drugs can potentially impact the relationship between the neural and hemodynamic responses through a variety of mechanisms.1

Anesthetics can depress spontaneous and evoked neuronal activity as well as cerebral metabolism,2 and can also alter vascular reactivity and dynamics, modulating the neurovascular coupling.1 General anesthetics typically affect cerebral blood flow and metabolism by depressing excitatory synaptic transmission or enhancing inhibitory transmission,3 with the degree and dynamics of these effects depending on the type and concentration of the drug as well as the regional balance between excitatory and inhibitory synapses.4 Anesthesia can significantly modulate synaptic activity and neuronal response properties, and thereby produce changes in cortical activity. Anesthetics may also affect the neurovascular coupling more directly. Inhalation anesthetics, for example, can induce vasodilation,5 which varies with concentration or depth of anesthesia5, 6 and can complicate the interpretation of hemodynamic measurements. Opioid intravenous anesthetics may reduce blood flow and metabolism without altering the neurovascular coupling, although neurovascular coupling can be affected at very high concentrations.7, 8

To fully understand the impact of anesthesia on evoked neuronal and hemodynamic responses, it is necessary to make a direct comparison with the awake condition. For this purpose we have acquired simultaneous electrophysiological and fMRI measurements during whisker stimulation of rabbits before, during and after anesthesia. The whisker barrel cortex has been characterized extensively and provides a well-controlled environment for quantifying the effects of anesthesia on neuronal and hemodynamic responses. In this study, we have examined the effects of a general anesthetic isoflurane and an analgesic opioid fentanyl. These two drugs represent two different classes of anesthetics and both have been widely used clinically and in previous fMRI animal studies.9, 10, 11 Specifically, opioid anesthesia, alone or combined with isoflurane, has been used in primates both for fMRI and electrophysiological studies. Isoflurane is the prefered anesthetic for longitudinal rodent studies.

We hypothesized that the decrease in neuronal excitation present under anesthesia would produce a corresponding decrease in BOLD and electrophysiological signals. Furthermore, we expected the differential mechanisms of operation of isoflurane and fentanyl would produce differential changes across drug conditions. Although fMRI and optical studies have reported anesthesia-related changes,12, 13 this study is the first to perform simultaneous fMRI and electrophysiology in the same animals within a single experiment under awake, anesthetized, and recovery conditions. This experimental design allows for direct comparison of changes induced by anesthesia in the BOLD signal, SU activity, MUA, and LFPs. Quantifying how these signals change from the predominantly excitatory environment under awake conditions to the greater degree of inhibition under anesthesia enables the quantification of the impact of different anesthetics on the relationship of BOLD and neuronal activity. Our results indicate that anesthesia significantly modulates the BOLD signal as well as baseline and stimulus-evoked neuronal activity, and that this modulation varies considerably with the type of anesthetic. In particular, isoflurane produced an increase in LFPs, yet no corresponding behavior was observed in the BOLD response. These findings show the complexity of the relationship between the BOLD and electrophysiological signals. BOLD may scale more closely with the best measure of the excitatory subcomponents of the underlying neuronal activity, which can vary according to experimental conditions that alter the excitatory/inhibitory balance in the cortex.

Materials and methods

Animal Preparation

Female Dutch-belted rabbits (2–3 kg, 3–4 month of age) were used in these experiments, which were performed in accordance with the National Institutes of Health guidelines, and NorthShore University HealthSystem Institutional Animal Care and Use Committee approved protocol. All the animals were kept in the local animal facility on a 12:12 hours light/dark cycle at the temperature 68–70°F. Animals were kept in single cages and were given ad libitum access to food and water, environmental enrichment was provided, and their health was observed on a daily basis. The animals were acclimated to the environment for at least 1 week before initiating the experiments. The manuscript was written up in accordance with the ARRIVE (Animal Research: Reporting In Vivo Experiments) guidelines.

Animals were anesthetized with a mixture of ketamine (50 mg/kg) and xylazine (10 mg/kg). An incision was made in the scalp and the bone exposed on the top of the skull. Small burr holes were made over the barrel cortex for electrode implantation and a small cut (up to 1 mm) was made through the dura. The recording assembly consisted of a silica tube (Polymicro Technologies, Phoenix, AZ, USA) containing a bundle of four 25-μm diameter gold–silver alloy microwires with formvar insulation (California Fine Wire, Grover Beach, CA, USA) which terminated at different levels within a distance of 100 μm. The microwire bundle was attached to a permanently implanted custom-made nylon microdrive that permitted vertical adjustments of its position. The microwires were connected to a small 6-pin connector that was embedded in dental acrylic. A 150-μm silver wire was placed between the skull and dura mater for grounding. During surgery, lambda was positioned 1.5 mm below bregma and the stereotaxic coordinates for the whisker barrel cortex were as follows: anterior-posterior was 2 mm dorsal to bregma, medial-lateral was 6 mm from midline, and dorsal-ventral was under visual control.

After implantation, the electrode assembly was cemented to the skull using dental acrylic and 6–8 nylon support screws, which were twisted in without full penetration of the skull. A light-weight head restraining device containing four nylon bolts was implanted on top of the skull behind the electrode assemblies. This headbolt was used to secure the radio frequency coil and the animal's head in the same position to obtain a constant imaging angle and slice positioning among subjects, as described previously.14

After a week recovery from surgery, each subject was habituated for 3–5 days to the imaging environment before the experiments. The rabbits were restrained by means of a cloth sleeve, and secured in an acrylic imaging cradle by Velcro (Manchester, NH, USA) straps. The electrodes were advanced to layer IV and their location was confirmed using anatomical MR imaging. All anesthesia-related experiments were conducted 1 month after the surgery to eliminate any possible effects from the ketamine/xylazine anesthesia used for the electrode and headbolt implantation.

Electrophysiological Recording

The SU, MUA, and LFP signals from the microwires were fed through a miniature preamplifier to a multichannel differential amplifier system (Neuralynx, MT, USA). The signals were amplified, band-pass filtered (300 Hz to 3 kHz for SU/MUA and 1–150 Hz for LFP), and digitized (32 kHz per channel) using a Neuralynx data acquisition system. Electrophysiological signals from the neuronal activity were analyzed after removal of blocks of strong interference signals induced by gradient pulses. These blocks were detected by thresholding followed by one-dimensional mathematical morphology15 processing based on erosion and dilation functions. To capture the initial neuronal activity without interfering signals from the gradients, the stimulus onset was delayed for 150 ms from the MR acquisition triggering pulse. Subsequently, unit discrimination was performed offline using threshold detection followed by a cluster analysis of scatter plots of time and amplitude distances between the peak and valley of individual action potential wave shapes. Up to four units could be discriminated on each channel. The discriminated data were processed using Neuralynx and custom software written in Matlab and Visual Basic. Raster and peri-event histograms were constructed for each unit and experiment. Separate histograms were constructed for baseline trials, trials during the anesthesia, and trials during recovery. In each histogram, the baseline firing rate and the magnitude of excitatory changes were computed. Individual normalized cell histograms (spike frequency in Hz) were pooled together for each cell type, anesthesia condition, and period of time to construct average population histograms. LFP and MUA signal energy was computed as the integral of the signal absolute value over nonoverlapping time windows of 50 ms length outside the interference containing windows and averaged over ten trials for each experiment. The modulation-related changes were calculated within the stimulation period for SUs, MUA, and LFP.

fMRI Data Collection and Analysis

Imaging was performed using a Bruker (Billerica, MA, USA) Avance 4.7 T imaging spectrometer operating at a proton frequency of 200 MHz. This system was equipped with an Oxford horizontal magnet and an Acustar actively shielded gradient coil assembly with a clear bore of 15 cm. A flat, circular surface coil (45 mm diameter) was used for radio frequency transmission and reception. A multislice, single-shot gradient-echo echo planar imaging (EPI) pulse sequence, with a repetition time of 2 s and an echo time of 20 ms, was used to map brain activation. Coronal images in a plane perpendicular to the surface coil were collected from four slices (1.0 mm thickness), which included the electrode recording site, using a 64 × 80 matrix size which was zero-filled to 64 × 128, and a 48 × 48 mm field of view, corresponding to an in-plane resolution of 750 × 375 μm. Sixty-five images, including five dummy images, were collected per trial. Anatomical images (512 × 512 matrix, 48 × 48 mm field of view, 94 × 94 μm in-plane resolution) were also obtained using a multislice gradient echo sequence (1.0 mm slice thickness; repetition time, 1.5 seconds; echo time, 20 ms, NA=8). The fMRI data were registered using an affine method implemented in the ITK toolkit.16 Each trial was inspected for remaining head movement after registration, and any trials exhibiting movement were excluded from the analysis. The remaining trials were averaged for each experimental condition in each subject. Activated voxels were detected in the trial-averaged data by cross-correlation to a boxcar function that corresponded to the stimulus duration at a statistical threshold corresponding to P<0.001. BOLD signal response curves were calculated as a mean of signal intensity in the activated pixels which exceeded the statistical threshold for each subject. Activated areas and time courses were then averaged across subjects and expressed as mean ±s.e.m.

Stimulus Presentation

Each experiment consisted of simultaneous recording of electrophysiological and fMRI signals in response to whisker stimulation under awake, anesthetized, and recovery conditions in the same animal. The whisker stimulation paradigm consisted of a nonstimulus baseline period (25 images), a stimulation period (20 images), and a post-stimulus period to allow recovery of the BOLD signal (20 images). Ten trials were acquired for each phase of the experiment. Whisker stimulation was delivered by a magnetic resonance imaging-compatible system that incorporates real-time optical monitoring of the frequency and amplitude to ensure consistency of the vibration stimulus, as described previously.17 A whisker stimulation frequency of 75 Hz with ±1.5 mm deflection was used for all the experiments. This frequency/magnitude was found to be optimal because higher values led to significant animal movement in the awake state and lower values (particularly frequencies lower then 50 Hz) produced unreliable BOLD activation under anesthesia. Two to three whiskers were stimulated in each experiment, although the selection of whiskers depended on the position of the electrodes in the whisker barrel cortex. In all cases, recorded units were located in the slices corresponding to the stimulated whiskers. Stimulus presentation and electrophysiology and fMRI data acquisition were synchronized using a common trigger pulse generated in the PC.

Anesthesia

Three different anesthesia regimens were evaluated for this study; fentanyl alone (n=4 animals), isoflurane alone (n=4), and a combination of fentanyl and isoflurane (n=3) to evaluate their additive effects. Rabbits were randomly assigned to the experimental groups. Fentanyl (Baxter Healthcare, Deerfield, IL, USA) at 0.01 mg/kg per hour was delivered through a catheter placed into a marginal vein of rabbit's ear, isoflurane (Piramal Healthcare, Mumbai, India) was delivered at 0.5% (0.25 minimum alveolar concentration (MAC)) in air via mask (Harvard Apparatus, Holliston, MA, USA) using a calibrated Matrx vaporizer (Midmark, Versailles, OH, USA). Rabbits were habituated to the mask before the experiments. In all cases, anesthesia was delivered for 20 minutes before the stimulation phase of the experiment, and an additional 20 minutes delay was provided before the recovery phase of the experiment to allow sufficient time for achieving stable anesthesia and recovery, respectively.

Statistical Analysis

The statistical analysis of individual parameters of BOLD activated area, BOLD signal magnitude, the firing of neurons, the LFP and MUA, and of population histograms, was conducted using two factorial analysis of variance (ANOVA) (Statistica, StatSoft, Tulsa, OK, USA) with the following factors: drug (three levels: fentanyl; isoflurane; or fentanyl+isoflurane) and time (three levels: before delivery; during delivery; and during recovery). For electrophysiological signals an additional factor of baseline/stimulation was used (three levels: before whisker stimulation (baseline); during whisker stimulation (response); and after whisker stimulation). ANOVA was followed by Newman–Keuls post hoc analysis. The regression and Pearson correlation analysis were applied to indicate the strength and direction of a linear relationship between BOLD signal and electrophysiological data (baseline, and mean, peak and plateau of the evoked response) as well as between baseline and mean response of electrophysiological data. Fisher z-transformation was used to establish the significance (P<0.05) of the correlation coefficients. Mean response was defined by the mean of the signal during stimulation. Plateau was defined as the second half of the response; the peak was defined as the maximum value during the first half of the response. Statistical significance of the changes observed for the different anesthesia conditions was determined by paired Student's t-test in responses normalized to the level before anesthesia. Holm–Bonferroni correction was applied for multiple comparisons. The data are presented as mean ±s.e.m. unless otherwise specified.

Results

Anesthesia and BOLD Signal Response

The area and magnitude of the BOLD response to whisker stimulation decreased during all three anesthesia conditions as compared with the awake state. Examples from a single subject for fentanyl+isoflurane are shown in Figures 1A to 1C. In the awake state, the activated area extended through all layers of the whisker barrel cortex contralateral to the stimulus (Figure 1A). The activation contracted to a small cluster located primarily in layer IV under anesthesia (Figure 1B) and increased approximately to the original size (Figure 1C) after recovery. A similar pattern was observed for fentanyl or isoflurane administered alone. As shown in Figures 1D to 1F, the BOLD response magnitude also decreased substantially during anesthesia (to ~58% for isoflurane+fentanyl) and subsequently recovered to its original level after termination of anesthesia. A similar pattern was observed for fentanyl and isoflurane alone (Supplementary Figure S1, Supplementary Material).

Figure 1.

Figure 1

Open in a new tab

Effect of anesthesia on blood oxygenation level-dependent activation. Functional images of the stimulation of two whiskers (A2 and A3) are shown from a single subject before (A), during (B), and after (C) fentanyl+isoflurane anesthesia. Before anesthesia, a cluster of activation in the whisker barrel cortex contralateral to the stimulated whiskers can be seen extending through all layers of cortex. The activated area in the cortex contracts primarily to layer IV during anesthesia, and subsequently expands during recovery. The color bar represents the magnitude of the correlation coefficient. The blood oxygenation level-dependent time course averaged across the subjects (D) showed a decrease in magnitude during anesthesia (E), followed by recovery of the signal (F). Error bars represent the s.e.m. The horizontal gray bar indicates the timing of the stimulus presentation.

The activated brain area and BOLD signal magnitude were averaged across subjects for each drug condition before, during, and after anesthesia (their mean values are provided in Supplementary Table 1 in Supplementary Data). Analysis of BOLD activated area using two-factor ANOVA showed a significant main effect for the time factor (P<4 × 10−6). Post hoc comparisons revealed that all drugs significantly decreased BOLD area and during recovery BOLD area was significantly restored. Fentanyl and isoflurane administered alone produced similar decrease in the activated area (66–69%) whereas combination of the drugs resulted in a greater decrease (83%). Analysis of BOLD signal magnitude using two-factor ANOVA showed a significant main effect for drug (P=0.0016) and time factors (P<1 × 10−6). One-way ANOVA of BOLD magnitude separately for each type of anesthesia showed a significant main effects for the time factor in the case of isoflurane+fentanyl (P<6 × 10−5) and isoflurane (P<1 × 10−6). Both isoflurane and isoflurane+fentanyl showed similar decrease in BOLD signal magnitude (41–48%), while fentanyl produced a much smaller decrease (25%).

Anesthesia and Neuronal Activity

A total of 46 cells with excitatory response were recorded for these experiments. LFP and MUA activity were normalized to the baseline level before stimulus presentation for the preanesthesia condition to allow comparisons between the different conditions studied. All measures of spontaneous and evoked neuronal activity (SU, MUA, and LFP) decreased during the three anesthesia conditions examined as compared with the awake state. Examples of electrophysiological recordings of neuronal responses to whisker stimulation are shown in Figures 2A to 2I for the case of fentanyl+isoflurane. LFP (Figures 2A to 2C), SU (Figures 2D to 2F) and MUA (Figures 2G to 2I) all showed a transient initial peak in the evoked activity followed by a sustained, lower amplitude plateau, a well-described adaptation phenomenon characteristic of long stimulation periods. This adaptation was similar in onset and duration for SU, MUA, and LFP. The baseline neuronal activity decreased for SUs, MUAs, and LFPs during anesthesia as compared with the awake state. However, whereas the SU and MUA peak and plateau responses decreased, little change was observed in the LFP peak. After termination of anesthesia a full recovery of the baseline and evoked neuronal activity was observed. Fentanyl and isoflurane administered alone also decreased baseline neuronal activity as well as mean SU and MUA response in comparison with the awake state, with isoflurane producing a greater decrease. The mean LFP-evoked response was constant for all three anesthesia conditions, but the LFP baseline activity differed among the anesthetics. The mean values of SU, MUA, LFP-baseline, and -evoked response measured for each drug condition are provided in Supplementary Table S2 (Supplementary Data). Evoked neuronal activity was highly dependent on the spontaneous baseline activity across all the conditions studied as shown in the Supplementary results (Supplementary Figures S2-S3, Supplementary Material). Statistically significant correlations between baseline and response magnitudes were obtained for all three measures of neuronal activity (Supplementary Figure S4, Supplementary Material).

Figure 2.

Figure 2

Open in a new tab

Effect of anesthesia on local field potential, single unit, and multi-unit activity responses. Peri-stimulus histograms of local field potential (A–C), single unit (D–F), and multi-unit activity (G–I) activity in the whisker barrel cortex averaged across subjects are shown before, during, and after fentanyl+isoflurane anesthesia. Responses are shown using both 100 ms (gray) and 1 second (black) bins. Local field potential and multi-unit activity was normalized to the baseline level before stimulus presentation and drug administration. All three measures of neural activity show a biphasic response characterized by initial peak (transient) followed by a plateau (sustained). Although fentanyl+isoflurane anesthesia produced little change in local field potentials peak, both baseline and plateau response decreased and subsequently recovered. Single unit and multi-unit activity baseline, peak, and plateau decreased with anesthesia and later recovered. The gray bar indicates the stimulus presentation.

Analysis of LFP activity by two-factor ANOVA showed a significant main effect for the time and baseline/stimulation factors for all drugs (P<10−6): fentanyl+isoflurane, fentanyl, and isoflurane. Post hoc comparisons revealed that all drugs significantly decreased LFP baseline and the mean response, and during recovery LFP baseline and response were significantly restored.

Analysis of SU activity by two-factor ANOVA showed a significant main effect for the time and baseline/stimulation factors (P<10−6) and interaction between time and baseline/stimulation for all drugs (P<10−6). Post hoc comparisons revealed that all drugs significantly decreased SU baseline and response and during recovery SU baseline and response were significantly restored. Analysis of MUA by two-factor ANOVA showed a significant main effect for the time and baseline/stimulation factors for all drugs (P<10−6) and interaction between time and baseline/stimulation for fentanyl+isoflurane, and fentanyl (P<10−6). Post hoc comparisons revealed that all drugs significantly decreased MUA baseline and response, and during recovery MUA baseline and response were significantly restored. We observed no change in the onset or termination time of the neuronal response analyzed at the scale of 100 ms bins between the three conditions.

Anesthesia and Relationship Between BOLD and Neuronal Activity

The effect of anesthesia on the evoked BOLD signal magnitude and the simultaneously recorded neuronal activity is shown in Figure 3. The BOLD response and neuronal activity were normalized to the preanesthesia level for a direct comparison between drugs. Note that the decrease in BOLD signal is not accompanied by a corresponding change in mean LFP response. No significant differences in the mean LFP response were found for the three anesthesia conditions studied. Significant differences were found for mean MUA responses between fentanyl and isoflurane (P<1 × 10−6), fentanyl and isoflurane+fentanyl (P<1 × 10−6), and isoflurane and isoflurane+fentanyl (P<1 × 10−6). Significant differences were also found for the mean SU responses between fentanyl and isoflurane (P<1 × 10−6), fentanyl and isoflurane+fentanyl (P<1 × 10−6), and isoflurane and isoflurane+fentanyl (P<1 × 10−5). Analysis of mean BOLD responses showed significant differences between fentanyl and isoflurane (P<0.003) and fentanyl and isoflurane+fentanyl (P<0.02). Correlation analysis between BOLD and electrophysiological signals supported these results and indicated that BOLD signal and LFP responses do not change in parallel unlike BOLD and SU/MUA responses (Supplementary Figures S5-S6, Supplementary Material).

Figure 3.

Figure 3

Open in a new tab

Effect of anesthesia on evoked responses. The mean local field potential, single unit, multi-unit activity, and blood oxygenation level-dependent responses during anesthesia are shown for each drug condition. Values are normalized to the level before anesthesia (100%) for each case. The significant differences in single unit, multi-unit activity, and blood oxygenation level-dependent responses between fentanyl (FEN) and isoflurane (ISO) are indicated with brackets. Note that mean multi-unit activity, single unit, and blood oxygenation level-dependent response show significant differences across drug conditions, whereas local field potential remains relatively unchanged.

To understand why the LFP mean response did not change between anesthesia conditions, as shown in Figure 3, we further analyzed the two components of the biphasic response, peak and plateau, separately. A comparison of the changes in LFP peak and plateau with the corresponding components of SU response, which showed the greatest changes across anesthesia condtions, is shown in Figure 4. The peak response magnitude decreased for LFP during fentanyl anesthesia but increased during conditions where isoflurane was present, while the plateau response magnitude did not change. Significant difference between peak and plateau values were found for isoflurane (P<0.008) and for isoflurane+fentanyl (P<0.006) but not fentanyl alone. SU peak and plateau response magnitude decreased in parallel under all anesthesia conditions, with the greatest change occuring under isoflurane+fentanyl. Unlike LFP response, no significant differences were found between SU peak and plateau for the three anesthesia conditions, suggesting that changes in SU peak scaled with changes in SU plateau.

Figure 4.

Figure 4

Open in a new tab

Effect of anesthesia on transient and sustained components of neuronal response. The local field potential (A) and single unit (B) peak and plateau responses to whisker stimulus during anesthesia are shown for each drug condition. Values are normalized to the level before anesthesia (100%) for each case. Peak response magnitude decreased for local field potential during fentanyl anesthesia, but increased in the presence of isoflurane, with the greatest change occurring under isoflurane alone. Local field potential plateau showed a consistent decrease for each drug condition. Single unit peak and plateau magnitude decreased for all anesthesia conditions, with the greatest change occurring under isoflurane+fentanyl. Error bars represent the s.e.m. Significant changes are indicated by brackets.

Discussion

Direct comparison between awake and anesthetized animals within the same experiment, as well as after full recovery, allowed us to examine in detail the modulatory effects of anesthesia on the BOLD and electrophysiological signals. Our results indicate that anesthesia significantly affected these signals, but the effects manifested in different ways that can be explained in terms of different mechanisms of action of these drugs.

Anesthesia and BOLD Signal Response

We observed that anesthesia produced a consistent decrease in activated area, magnitude, and duration of the BOLD signal. The decrease in BOLD activated area corresponds well to the physiology of the whisker barrel cortex. Vibrissal sensory input projects from the VPM thalamic nuclei primarily to layer IV of the cortex.18 The activity propagates first vertically through the layers of the stimulated whisker column and then horizontally to neighboring columns,18 leading to distributed spatial activity and a correspondingly larger BOLD response area, as we reported previously in awake animals.19 Under anesthesia, reduced afferent input and the enhanced contribution of inhibitory systems reduce the spread of stimulus-related activity, resulting in more focal activation and correspondingly smaller BOLD response area. The spread of excitation in the barrel cortex has been shown to be strongly dependent on γ-aminobutyric acid A receptor (GABAA) receptor-mediated inhibition, especially in the supra- and infragranular layers of the cortex.20 Such reduction in the receptive field size under anesthesia has been described for the somatosensory cortical21 and thalamic neurons22 for a number of species including humans23, 24 using different methodology. Using fMRI Austin et al.25 showed changes in the spatial extent of the BOLD response to the electrical stimulation of motor cortex in the rat after switching from halothane to α-chloralose. Peeters et al.13 also showed a smaller area of activated response to forepaw stimulation in the rat under α-chloralose than in the awake state. Thus, a wide variety of studies have showed a reduction in the spatial extent of the stimulus-evoked activation, which can vary with the type of anesthesia used in the experimental protocol.

The decrease in the magnitude of the evoked BOLD response under anesthesia most likely reflects a similar reduction of thalamic input and intra-cortical processing via decreased neuronal excitation. Previous optical imaging spectroscopy (OIS)26 and fMRI13, 27 studies have also shown a decrease in the magnitude of the hemodynamic response to stimulation for anesthetized animals. However, hemodynamic responses measured with OIS were reported to be longer under urethane anesthesia, as compared with the awake condition, for electrically stimulated whisker pad,12 which differs from our findings and likely reflects the differences in the anesthetic agents employed. Comparing results obtained under different anesthetics can be difficult as their mechanisms of action are complex and often not completely understood. Urethane, for example, has an array of effects on multiple receptors, including GABA, glycine, acetylcholine, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA), and N-methyl-D-aspartate (NMDA)28 as well as direct effect on vasodilation.29 Isoflurane, like other inhaled anesthetics, can produce indirect vasoconstriction because of reduction of cerebral metabolism as well as a direct, dose-dependent vasodilatory effect. However, at the low concentration used in our experiments, it has been shown that no direct vasodilation occurs.6, 30 Thus, the coupling between cerebral blood flow and metabolism is preserved under isoflurane concentration used and the hemodynamic changes observed accurately reflect changes in neuronal activity as opposed to vascular effects. These results underscore the general problem that neuronal and hemodynamic changes obtained under a particular anesthetic condition may not be directly comparable to studies performed using other drugs. A more detailed knowledge of the effects of anesthetics, in terms of the interaction with receptors as well as their modulation of neurovascular coupling, should help to account for many of the inconsistencies in hemodynamic results obtained from animal studies performed under a variety of experimental conditions.

Anesthesia and Neuronal Activity

All three electrophysiological signals showed adaptation of neuronal response (i.e., a transient initial peak followed by a sustained plateau) under both awake and anesthetized conditions. Our findings differ from a previous report by Logothetis et al.,10 who measured a much greater adaptation for MUA than LFP in the visual cortex of a macaque under fentanyl+isoflurane anesthesia. Based on these findings, the predominant interpretation has emerged that the BOLD signal is primarily associated with synaptic activity, as measured by LFP which, unlike MUA, was sustained during the stimulation. Although the sensory cortices are largely similar in their basic architecture, it is important to note that subtle differences exist in the layer structure, organization of local circuits and projections, the timing of neuronal responses, 31 and other properties which can potentially translate into functional differences. However, our results are in agreement with a study by Burns et al.9 who reported sustained response of both MUA and LFP to a visual stimulation in the macaque primary visual cortex under opioid anesthesia.

To quantify the changes in neuronal activity induced by each anesthetic, a direct comparison was made between the awake and anesthetized states. The decrease in both SU and MUA-baseline and -evoked activity across the different anesthetic conditions fits well with what is known of the distinct mechanisms of each drug in the brain. Opioid receptors have a complex mechanism of action. For example, they inhibit both glutamate release32 and GABA-ergic interneurons.33 Likely because of this inhibition of both excitatory and inhibitory systems, opioids have been reported to produce relatively little effect either on the latency or magnitude of somatosensory-evoked responses.34 General anesthetics such as isoflurane, in comparison, decrease excitatory cortical synaptic transmission35 and increase inhibitory currents in the neurons.36 Therefore, its effect on both SU baseline and response should be more prominent. The greater decrease in SU and MUA responses induced by the combination of isoflurane and fentanyl suggests that these drugs are able to potentiate each other.

Our results indicate that the stimulus-evoked neuronal response changes in parallel with the spontaneous baseline activity. These findings differ from those reported by Hyder et al.37 who found that lower SU baseline neuronal activity was associated with a greater SU-evoked activity and BOLD signal response under two concentrations of α-chloralose anesthesia. However, Hyder et al. observed in their report that these results may not apply to the awake state or to other anesthetics because α-chloralose can have specific mechanisms of action, particularly with respect to stimulation-dependent metabolic responses.

Compared with the SU and MUA responses, the changes in LFP were more complex. The decrease in LFP-evoked response was approximately the same under all drugs, but the baseline decrease was greater for isoflurane as well as isoflurane+fentanyl. Analysis of the transient and sustained components of the LFP response also revealed differences across the drug conditions, as compared with SU and MUA. Although fentanyl alone decreased both peak and plateau of LFP-evoked response, isoflurane alone and in combination with fentanyl increased the peak relative to the awake state. This behavior can be understood in terms of the effects of isoflurane on the inhibitory systems of the cortex. LFP is generally interpreted as reflecting changes in extracellular electrical field.38 Many factors contribute to this field, but synaptic activity is considered the primary source. As a result LFP often correlates well with MUA or SU activity. However as both excitatory and inhibitory synaptic currents contribute to LFPs, changes in LFPs may not always parallel changes in MUA or SUs, for example, during specific tasks39 or when single units are inhibited by strong inhibitory synaptic currents which themselves contribute significantly to LFP. This complex behavior of LFP responses (i.e., concurrent increase in magnitude of LFP responses and decrease in action potential generation) has been predicted by Buzsáki et al.38 in their review. Inhaled anesthetics have been found to enhance inhibitory Ce- currents by increasing GABAA-mediated inhibition.36 Prolonging inhibitory post-synaptic currents (IPSCs)40 to enhance the net inhibitory current in this manner would contribute to the subthreshold neuronal activity measured by LFP, and may be responsible for the increased LFP modulation that we observed without a corresponding increase in SU/MUA activity.

Anesthesia and Relationship Between BOLD and Neuronal Activity

The BOLD signal and the measured LFP, SU, and MUA responses might be expected to produce parallel changes across all three drug conditions. However, whereas isoflurane anesthesia significantly reduced BOLD, MUA, and SU responses, as compared with fentanyl anesthesia, no such reduction was observed in the LFP responses. This result was surprising, as the BOLD signal is typically thought to depend on LFP rather than on MUA, with few reported exceptions.41

Reducing MUA, either by pharmacological manipulation42 or by changes in stimulus characteristics,43 has been shown to preserve a robust correlation between hemodynamic response and LFP, which supports the interpretation of a secondary role for spiking activity in generating the hemodynamic response, which is primarily driven by LFP. However, in our case, decrease in the BOLD signal in the whisker barrel cortex was not always accompanied by a corresponding decrease in LFP. This diverging behavior of LFP and BOLD response despite a preserved association between MUA, SU, and BOLD suggests a more complex relationship between these various measures of neuronal activity and the BOLD signal. These findings would suggest that over the range of our experimental conditions either the BOLD signal more closely follows SU/MUA activity, or else the relationship between the BOLD signal and LFP is masked by the complex effects of isoflurane. Indeed, the increase in the LFP peak under isoflurane most likely reflects the increase in IPSCs as a result of the potentiation of GABAA receptors by isoflurane. If the increase in LFP response were because of an increase in excitatory post-synaptic potentials during stimulation, we would expect MUA and SU to show the same pattern of change. However, MUA and SU responses decreased along with baseline. Thus, the behavior of LFP responses under isoflurane and isoflurane+fentanyl anesthesia reflects the influence of IPSCs. Under our experimental conditions, the BOLD signal therefore more closely corresponds to excitatory post-synaptic potentials, which initiate action potentials, rather than IPSCs, as it scales with MUA and SU responses and not with LFPs. We would expect that all anesthetics that act by potentiating GABAA receptors in a manner similar to isoflurane will increase the IPSCs, and therefore produce an increase in the LFP response, dependent on the strength of potentiation, which may not necessarily be accompanied by a corresponding increase in the BOLD response. Thus, drawing broad conclusions that the BOLD signal primarily follows either SU/MUA or LFP can be problematic. Our results suggest that the BOLD signal depends specifically on the excitatory component (i.e., action potentials and excitatory post-synaptic potentials) of neuronal activity, which may be reflected at some level in both SU/MUA and LFP and may be modulated to a significant extent by experimental conditions such as anesthesia.

Conclusion

These findings underscore the need for developing a deeper understanding of the relationship between the BOLD signal and the specific excitatory and inhibitory processes that shape it. As we have shown, different classes of anesthetics can affect the baseline and evoked neuronal activity, and consequently the BOLD response generated under these conditions, quite differently. Such effects have not generally been well characterized, in part because of the difficulty of obtaining a direct comparison of the awake and anesthetized states in fMRI animal studies. However, the changes that we observed in this study correspond well to what is understood about the mechanisms of fentanyl and isoflurane, and we would generally expect consistency in the relative scaling of BOLD and neuronal responses within each class of anesthetic even for different stimulation parameters. Further work to characterize these effects, the unique impact of different anesthetics on these signals, and the contributions of the specific excitatory and inhibitory processes that underlie them in the cortex and other structures, will help to provide a more comprehensive understanding of their complex relationship.

Acknowledgments

We thank Dr Chi Wang for assistance in statistical analysis.

Author Contributions

DPA, LL, and AMW designed the study. DPA and LL conducted the experiments. GI and MJM analyzed the data. DPA, MJM, and AMW wrote the manuscript. LL and GI provided input on the manuscript.

The authors declare no conflict of interest.

Footnotes

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

This work was supported by National Institute of Health (R01 NS44617, S10 RR15685).

Supplementary Material

Supplementary Information
Click here for additional data file. (3.1MB, doc)

References

  1. 1Masamoto K, Kanno I. Anesthesia and the quantitative evaluation of neurovascular coupling. J Cereb Blood Flow Metab 2012; 32: 1233–1247. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. 2Ruskin K, Rosenbaum S, Ij R. Fundamentals of neuroanesthesia: a physiologic approach to clinical practice. Oxford Press: Oxford. 2013. [Google Scholar]
  3. 3Richards CD. Actions of general anaesthetics on synaptic transmission in the CNS. Br J Anaesth 1983; 55: 201–207. [DOI] [PubMed] [Google Scholar]
  4. 4Richards CD. Anaesthetic modulation of synaptic transmission in the mammalian CNS. Br J Anaesth 2002; 89: 79–90. [DOI] [PubMed] [Google Scholar]
  5. 5Matta BF, Heath KJ, Tipping K, Summors AC. Direct cerebral vasodilatory effects of sevoflurane and isoflurane. Anesthesiology 1999; 91: 677–680. [DOI] [PubMed] [Google Scholar]
  6. 6Aladj LJ, Croughwell N, Smith LR, Reves JG. Cerebral blood flow autoregulation is preserved during cardiopulmonary bypass in isoflurane-anesthetized patients. Anesth Analg 1991; 72: 48–52. [DOI] [PubMed] [Google Scholar]
  7. 7McPherson RW, Traystman RJ. Fentanyl and cerebral vascular responsivity in dogs. Anesthesiology 1984; 60: 180–186. [DOI] [PubMed] [Google Scholar]
  8. 8Safo Y, Young ML, Smith DS, Greenberg J, Carlsson C, Reivich M et al. Effects of fentanyl on local cerebral blood flow in the rat. Acta Anaesthesiol Scand 1985; 29: 594–598. [DOI] [PubMed] [Google Scholar]
  9. 9Burns SP, Xing D, Shapley RM. Comparisons of the dynamics of local field potential and multiunit activity signals in macaque visual cortex. J Neurosci 2010; 30: 13739–13749. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. 10Logothetis N, Pauls J, Augath M, Trinath T, Oeltermann A. Neurophysiological investigation of the basis of the fMRI signal. Nature 2001; 412: 150–157. [DOI] [PubMed] [Google Scholar]
  11. 11Masamoto K, Kim T, Fukuda M, Wang P, Kim SG. Relationship between neural, vascular, and BOLD signals in isoflurane-anesthetized rat somatosensory cortex. Cerebral Cortex 2007; 17: 942–950. [DOI] [PubMed] [Google Scholar]
  12. 12Martin C, Martindale J, Berwick J, Mayhew J. Investigating neural-hemodynamic coupling and the hemodynamic response function in the awake rat. Neuroimage 2006; 32: 33–48. [DOI] [PubMed] [Google Scholar]
  13. 13Peeters RR, Tindemans I, De Schutter E, Van der Linden A. Comparing BOLD fMRI signal changes in the awake and anesthetized rat during electrical forepaw stimulation. Magn Reson Imaging 2001; 19: 821–826. [DOI] [PubMed] [Google Scholar]
  14. 14Wyrwicz AM, Chen N-K, Li L, Weiss C, Disterhoft JF. fMRI of visual system activation in the conscious rabbit. Magn Res Med 2000; 44: 474–478. [DOI] [PubMed] [Google Scholar]
  15. 15Pratt W. Morphological image processing2nd edn (Wiley-Interscience: New York). 1991. [Google Scholar]
  16. 16Ibanez L, Schroeder W. The ITK software guide. Kitware: Clifton Park, NY. 2005. [Google Scholar]
  17. 17Li L, Weiss C, Talk AC, Disterhoft JF, Wyrwicz AM. A MRI-compatible system for whisker stimulation. J Neurosci Methods 2012; 205: 305–311. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. 18Petersen R, Diamond M. Spatial-temporal distribution of whisker-evoked activity in rat somatosensory cortex and the coding of stimulus location. J Neurosci 2000; 20: 6135–6143. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. 19Song X, Li L, Aksenov D, Miller MJ, Wyrwicz AM. Mapping rabbit whisker barrels using discriminant analysis of high field fMRI data. Neuroimage 2010; 51: 775–782. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. 20Petersen CC, Sakmann B. Functionally independent columns of rat somatosensory barrel cortex revealed with voltage-sensitive dye imaging. J Neurosci 2001; 21: 8435–8446. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. 21Armstrong-James M, George MJ. Influence of anesthesia on spontaneous activity and receptive field size of single units in rat Sm1 neocortex. Exp Neurol 1988; 99: 369–387. [DOI] [PubMed] [Google Scholar]
  22. 22Friedberg MH, Lee SM, Ebner FF. Modulation of receptive field properties of thalamic somatosensory neurons by the depth of anesthesia. J Neurophysiol 1999; 81: 2243–2252. [DOI] [PubMed] [Google Scholar]
  23. 23Kerssens C, Hamann S, Peltier S, Hu XP, Byas-Smith MG, Sebel PS. Attenuated brain response to auditory word stimulation with sevoflurane: a functional magnetic resonance imaging study in humans. Anesthesiology 2005; 103: 11–19. [DOI] [PubMed] [Google Scholar]
  24. 24Plourde G, Belin P, Chartrand D, Fiset P, Backman SB, Xie G et al. Cortical processing of complex auditory stimuli during alterations of consciousness with the general anesthetic propofol. Anesthesiology 2006; 104: 448–457. [DOI] [PubMed] [Google Scholar]
  25. 25Austin VC, Blamire AM, Allers KA, Sharp T, Styles P, Matthews PM et al. Confounding effects of anesthesia on functional activation in rodent brain: a study of halothane and alpha-chloralose anesthesia. NeuroImage 2005; 24: 92–100. [DOI] [PubMed] [Google Scholar]
  26. 26Berwick J, Martin C, Martindale J, Jones M, Johnston D, Zheng Y et al. Hemodynamic response in the unanesthetized rat: intrinsic optical imaging and spectroscopy of the barrel cortex. J Cereb Blood Flow Metab 2002; 22: 670–679. [DOI] [PubMed] [Google Scholar]
  27. 27Sicard K, Shen Q, Brevard ME, Sullivan R, Ferris CF, King JA et al. Regional cerebral blood flow and BOLD responses in conscious and anesthetized rats under basal and hypercapnic conditions: implications for functional MRI studies. J Cereb Blood Flow Metab 2003; 23: 472–481. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. 28Hara K, Harris RA. The anesthetic mechanism of urethane: the effects on neurotransmitter-gated ion channels. Anesth Analg 2002; 94: 313–318, table of contents. [DOI] [PubMed] [Google Scholar]
  29. 29Colantuoni A, Bertuglia S, Intaglietta M. Effects of anesthesia on the spontaneous activity of the microvasculature. Int J Microcirc Clin Exp 1984; 3: 13–28. [PubMed] [Google Scholar]
  30. 30Scheller MS, Tateishi A, Drummond JC, Zornow MH. The effects of sevoflurane on cerebral blood flow, cerebral metabolic rate for oxygen, intracranial pressure, and the electroencephalogram are similar to those of isoflurane in the rabbit. Anesthesiology 1988; 68: 548–551. [DOI] [PubMed] [Google Scholar]
  31. 31Yang Y, Zador AM. Differences in sensitivity to neural timing among cortical areas. J Neurosci 2012; 32: 15142–15147. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. 32Tallent MK. Presynaptic inhibition of glutamate release by neuropeptides: use-dependent synaptic modification. Results Probl Cell Differ 2008; 44: 177–200. [DOI] [PubMed] [Google Scholar]
  33. 33Drake CT, Milner TA. Mu opioid receptors are in somatodendritic and axonal compartments of GABAergic neurons in rat hippocampal formation. Brain Res 1999; 849: 203–215. [DOI] [PubMed] [Google Scholar]
  34. 34Goodarzi M, Shier NH, Grogan DP. Effect of intrathecal opioids on somatosensory-evoked potentials during spinal fusion in children. Spine (Phila Pa 1976) 1996; 21: 1565–1568. [DOI] [PubMed] [Google Scholar]
  35. 35el-Beheiry H, Puil E. Anaesthetic depression of excitatory synaptic transmission in neocortex. Exp Brain Res 1989; 77: 87–93. [DOI] [PubMed] [Google Scholar]
  36. 36Antkowiak B, Helfrich-Forster C. Effects of small concentrations of volatile anesthetics on action potential firing of neocortical neurons in vitro. Anesthesiology 1998; 88: 1592–1605. [DOI] [PubMed] [Google Scholar]
  37. 37Hyder F, Rothman DL, Shulman RG. Total neuroenergetics support localized brain activity: implications for the interpretation of fMRI. Proc Natl Acad Sci USA 2002; 99: 10771–10776. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. 38Buzsaki G, Anastassiou CA, Koch C. The origin of extracellular fields and currents—EEG, ECoG, LFP and spikes. Nat Rev Neurosci 2012; 13: 407–420. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. 39Ekstrom A, Viskontas I, Kahana M, Jacobs J, Upchurch K, Bookheimer S et al. Contrasting roles of neural firing rate and local field potentials in human memory. Hippocampus 2007; 17: 606–617. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. 40Jones MV, Brooks PA, Harrison NL. Enhancement of gamma-aminobutyric acid-activated Cl- currents in cultured rat hippocampal neurones by three volatile anaesthetics. J Physiol 1992; 449: 279–293. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. 41Ekstrom A. How and when the fMRI BOLD signal relates to underlying neural activity: the danger in dissociation. Brain Res Rev 2010; 62: 233–244. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. 42Rauch A, Rainer G, Logothetis NK. The effect of a serotonin-induced dissociation between spiking and perisynaptic activity on BOLD functional MRI. Proc Natl Acad Sci USA 2008; 105: 6759–6764. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. 43Viswanathan A, Freeman RD. Neurometabolic coupling in cerebral cortex reflects synaptic more than spiking activity. Nat Neurosci 2007; 10: 1308–1312. [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

Supplementary Information
Click here for additional data file. (3.1MB, doc)

Articles from Journal of Cerebral Blood Flow & Metabolism are provided here courtesy of SAGE Publications

RESOURCES