In vivo imaging of anaesthetic action in humans: approaches with positron emission tomography (PET) and functional magnetic resonance imaging (fMRI)

W. Heinke*,1 and C. Schwarzbauer2

1 Department of Anaesthesiology and Intensive Care Therapy, University of Leipzig, Liebigstrasse 20a, D-04103 Leipzig, Germany and 2 Cognition and Brain Sciences Unit, Cambridge, UK*Corresponding author

Abstract

Br J Anaesth 2002; 89: 112–22

Keywords: measurement techniques, positron emission tomography; measurement techniques, functional magnetic resonance imaging; theories of anaesthetic action, general mechanisms

Recent progress in brain imaging offers great potential for anaesthesia research.46 Among the most important imaging modalities for mapping human brain activation are positron emission tomography (PET) and functional magnetic resonance imaging (fMRI). Both techniques are based on mapping haemodynamic or metabolic changes that are a direct consequence of alterations in neuronal activity.54 59 In comparison with electroencephalographic techniques, which are able to demonstrate that a given anaesthetic agent has central nervous system activity, PET or fMRI provide information about the pathways and the anatomical localization of the drug effect.

On one hand, fMRI offers a superior spatial resolution to PET. The theoretical limit for PET is expected to be 2 mm isotropic resolution for the human head,17 whereas a recent fMRI study has demonstrated an in-plane resolution of 0.55 mm.28 Furthermore, fMRI does not require the injection of exogenous radioactive tracers. Therefore, a subject can perform a variety of different tasks during various experimental conditions within the same imaging session. Besides these methodological factors, fMRI is of advantage in terms of lower costs and clinical availability. On the other hand, PET offers the unique possibility of in vivo receptor imaging, only shared in part with magnetic resonance spectroscopy, and it provides absolute values of physiological variables. Using the complementary features of both techniques allows us to obtain information about how the brain works from the molecular to the complex neural network level61 and to explore regional drug effects on integrated brain processes at these various levels. Both imaging modalities have the ability to reveal new insights in altered brain processes during anaesthesia and may contribute to our understanding of where, why, and how brain functions collapse in the presence of anaesthetic drugs.

This review will focus mainly on previously reported results obtained with human brain imaging techniques related to the general mechanism of anaesthetic action. It will first discuss some methodological aspects of PET and fMRI and then comment on results obtained mostly from volunteer studies.

Positron emission tomography

PET enables non-invasive measurements of physiological and biochemical processes in any part of the human body. PET is based on the i.v. injection of a radioactive tracer that typically consists of endogenous molecules labelled with positron emitting isotopes. In the tissues, these positrons are wiped out by electrons causing the emission of two collinear gamma rays. These events are detected by PET cameras and subsequent data processing provides three-dimensional images indicating the location and density of the positron sources. Depending on the biochemical and biophysical properties of the tracer, PET can be used to map a variety of different biochemical processes and physiological variables.

Cerebral blood flow and cerebral metabolic rate of glucose
In cognitive neuroscience, functional brain imaging with PET is often used to map changes in regional cerebral blood flow (CBF) or cerebral metabolic rate of glucose (CMRGlu) caused by well-defined stimuli or tasks. These studies are usually called activation studies. The analysis of the measured data is based on statistical comparison of the measurements obtained during different conditions (e.g. performing a task vs resting) reveals regional stimulus or task-related changes in CBF or CMRGlu. In contrast to these cognitive activation studies, PET studies of anaesthetic action in the human brain are often based on comparing physiological variables measured under different states of anaesthesia. Regional alterations in CBF and CMRGlu are assumed to reflect changes in regional neural activity and may provide neuroanatomical evidence for behavioural responses associated with sedation and anaesthesia.

The theoretical basis for regional (r)CBF measurements with PET is the work by Kety,38 39 which allows assessment of rCBF in laboratory animals. The animal tissue autoradiographical methods of rCBF measurements are adapted for use with PET.35 In human studies, the tracer of choice for CBF mapping is 15O-labelled water (H215O).36 This molecule can diffuse freely across the blood–brain barrier. Because of its short half-time (2 min) the tracer allows the performance of repeated measurements in a single study session and the use of a variety of experimental designs. After injection of a single dose of H215O PET images are immediately acquired to obtain the flow-dependent accumulation and disappearance of the radiotracer and to calculate maps of rCBF.

PET measurements of CMRGlu in man are based usually on [18F]fluorodeoxyglucose (18FDG).2 This tracer is taken up by brain neurones depending on their functional state, as if it were glucose. It is phosphorylated in the brain tissue by hexokinase to FDG-6-phosphate. Unlike glucose, FDG-6-phosphate becomes intracellularly trapped for at least 45 min without being further metabolized. Uptake of FDG and metabolic trapping of the FDG-6-phosphate is nearly complete about 30 min after injection of the tracer. After this period PET scans are obtained and blood samples are collected to measure glucose and FDG concentrations in plasma as a function of time. The amount of radioactivity in each region of the brain is related to the glucose uptake and the glucose metabolism in this region. As glucose is the primary substrate for brain cells, the measured radioactivity in any particular region represents the glucose metabolism in this area. However, this technique has some limitations. Because of the long half-time (110 min) and the long uptake period, repeated scans are difficult to perform and a constant neurobehavioural state has to be ensured over a long period. For this reason, many drug studies have been performed using CBF rather than CMRGlu as an index of local cerebral activity.36

Receptor imaging
Because of the biochemical selectivity of PET, neurotransmission processes can be investigated that occur at very low concentrations, typically in the nanomolar to the picomolar range. Specific radioactive tracers can be designed that bind selectively to molecular targets, such as receptors or proteins involved in the synthesis or metabolism of neurotransmitters. A variety of radiolabelled ligands have been designed for tracing different signalling systems in the human brain, including the GABAergic/benzodiazepine, cholinergic, opioid, and monoaminergic system.29 PET therefore offers the unique possibility to identify putative molecular targets of anaesthetics and to probe their action on neurochemical processes in vivo.

Functional magnetic resonance imaging

Most fMRI studies are based on measuring changes in the blood oxygenation level-dependent (BOLD) contrast that arise from the paramagnetic properties of the deoxyhaemoglobin molecule.25 48 A decrease in the regional deoxyhaemoglobin concentration increases the effective transverse relaxation time, T2*, that can be mapped by using T2*-weighted MRI techniques. Neural activation is accompanied by a regional increase in blood volume, blood flow, and oxygen consumption.14 An increase in blood flow produces a decrease in deoxyhaemoglobin concentration, whereas an increase in oxygen consumption or venous blood volume has the opposite effect. The BOLD contrast therefore depends on changes in both cerebral haemodynamics and oxidative metabolism. When periods of neural stimulation and of rest are compared, the BOLD signal typically is increased during periods of neural stimulation indicating that the haemodynamic response to the stimulus is the dominant variable in BOLD-based fMRI.

PET studies of anaesthetic action

Studies of regional cerebral glucose metabolism and rCBF
Several investigators have used PET techniques to measure rCBF15 24 57 and CMRGlu to identify neural structures related to anaesthetic actions in the human brain.3 4 6 Using the 18FDG PET technique Alkire and co-workers studied the effects of isoflurane, halothane, and propofol on brain glucose metabolism.3 4 6 In these studies the anaesthetics were incrementally titrated to the point of unresponsiveness and brain glucose metabolism was investigated during steady-state conditions. At the same clinical endpoint, similar results in whole-brain glucose metabolism were obtained for the three agents. Halothane decreased whole-brain glucose metabolism by 40%, isoflurane by 46%, and propofol by 55%. This general decrease in metabolic activity is believed to reflect the reduced synaptic activity across the brain in the anaesthetic state. This hypothesis is confirmed by a linear correlation of the metabolic reduction in various EEG variables during propofol and isoflurane anaesthesia.5 It is therefore reasonable to assume that the general metabolic reduction associated with the anaesthetic state is a determinant of anaesthetic depth.

However, research by Alkire’s group did not show only a general cerebral depression, it also revealed subtle differences in regional metabolic reduction caused by the volatile agents. In addition, in comparison with volatile anaesthetics these differences were more pronounced for propofol, consistent with the proposal that volatile agents and propofol act through different molecular mechanisms.27 42

Isoflurane anaesthesia caused a nearly uniform cortical and subcortical metabolic reduction.4 Regional shifts in the pattern of brain glucose metabolism were observed for halothane with a more specific metabolic suppression in the thalamus, basal forebrain, cerebellum, occiput, and limbic system.6 In contrast to volatile anaesthetics, the metabolic reduction caused by propofol was not uniform throughout the brain.3 It was more pronounced in the cortex (58%) than in subcortical regions (48%). Furthermore, propofol tended to suppress glucose metabolism in the cerebral cortex, especially in the temporal and occipital regions, to a greater extent than inhalation agents. In addition, compared with halothane, propofol caused significantly less metabolic suppression in the basal ganglia and midbrain regions.

Inhaled anaesthetics may cause unconsciousness by altering neuronal activity in specific regions of the CNS. Therefore, the next logical step taken by Alkire and his colleagues was to compare the regional patterns of metabolic depression caused by isoflurane and halothane.7 Such an analysis was expected to reveal brain regions that were affected to a similar extent. If those brain areas had existed, it would have been reasonable to assume that these regions would mediate the state of inhalation anaesthesia. And, in fact, Alkire’s analysis revealed brain regions, which differed in their functional activity between the awake state and the drug-induced unconsciousness, irrespective of each particular effect of the agent on regional cerebral metabolism. Both agents caused a specific relative reduction of regional cerebral glucose metabolism primarily in the thalamus and also in the midbrain reticular formation, basal forebrain, cerebellum, and occipital cortex. These findings clearly underline the importance of specific neural structures in mediating drug-induced loss of consciousness and reflect the importance of the thalamus as a target in inhibition of the flow of information to the cortex. Similar mechanisms in mediating anaesthesia were proposed by Angel on the basis of in vitro experiments using rats.10 11

The results of recent H215O PET studies further underpin the hypothesis that specific neural networks contribute to the behavioural changes produced by anaesthetics.15 24 Fiset and co-workers investigated rCBF during graded changes of propofol anaesthesia. Consistent with the general reduction in glucose metabolism reported for propofol by Alkire,3 Fiset found a 20.2% overall decrease in absolute CBF, indicating an overall decrease in cortical neural activity. This effect was particularly pronounced in the medial thalamus, posterior cingulate, basal forebrain, and in the occipitoparietal association cortices. In addition, the authors found a strong correlation between the level of consciousness and the reduction in CBF in the thalamus, basal forebrain, and occipitoparietal regions. Furthermore, a significant covariation between the thalamic and midbrain blood flow changes was observed, suggesting a close relationship between these structures. These results indicate that propofol preferentially affects brain activity in areas that are known to be linked to the control of consciousness, associative functions and autonomic control.51

These findings have been confirmed recently by the same group of authors.15 In addition to the previous investigation a stimulus-induced study design was chosen. Changes in rCBF were measured in order to determine whether different levels of propofol anaesthesia would affect subcortical and cortical processing of vibrotactile stimuli in a different way. This study showed a dose-dependent impairment of the processing of vibrotactile stimuli in the brain. Sedative concentrations of propofol already caused a suppression of the rCBF response in the somatosensory cortex, whereas the stimulus-induced thalamic activation was still present. Suppression of rCBF response in this area was merely found when volunteers had lost consciousness. These differential effects may reflect sequential effects on the complex thalamocortical system and may be responsible for dose-dependent behavioural changes. The finding that thalamic activation only disappeared in unconscious subjects provides further evidence for the importance of this particular part of the brain in mediating drug-induced unconsciousness.

Specific changes in rCBF were also reported for midazolam.13 55 56 Beside the well-known global decrease in CBF, midazolam infusion resulted in CBF reductions in discrete brain regions, including the prefrontal cortex, the superior frontal gyrus, the anterior cingulate gyrus, parietal and temporal association areas, the insular cortex, and the thalamus.35 56 These areas subserve arousal and memory processes, particularly the prefrontal cortex.51 It was proposed that the affected regions were most likely involved in mediating amnesia produced by midazolam.56

Most of the brain imaging studies cited above have underscored the importance of the thalamus as a putative macroscopic target of anaesthetic action. It is therefore not surprising that benzodiazepines may also produce their substantial sedative effects via that target. In the study by Veselis the largest decrease in rCBF during deep midazolam sedation was found in the thalamus. Other human PET studies, related to the action of benzodiazepines, also demonstrated large decreases in regional neural activity in the thalamus after lorazepam45 60 or midazolam13 administration.

Like midazolam, low doses of propofol are known to produce substantial anterograde memory impairment. PET imaging revealed significant dose-related CBF decreases in the prefrontal and parietal cortices similar to those obtained for midazolam.57 This may indicate that both drugs, though structurally different, produce their behavioural effects by affecting the same neural networks.

In summary, besides a dose-dependent general decrease in whole-brain activity during anaesthesia, the PET imaging studies indicated similar key brain structures involved in the effects of anaesthetics. Furthermore, specific anatomical regions were identified that appear to be crucial for specific behavioural changes caused by anaesthetics, such as amnesia or loss of consciousness. Very low doses of anaesthetics seemed to affect cortical areas preferentially, mainly in the association cortices, and with a further increase in their dosage the primary sensory cortices. This caused disturbances in attention and memory processes, while subjects were still awake and able to respond to commands. Higher doses of anaesthetics did not only affect cortical but also subcortical brain structures. In particular, the thalamus and the midbrain reticular formation appeared to be key targets for drug-induced loss of consciousness.

Studies of receptor function
Although PET studies provide a macroanatomical picture of cerebral anaesthetic action, they provide no information about why certain brain areas are more or less sensitive to anaesthetics. Indirectly, this information can be obtained by comparing the results of in vivo studies with ex vivo studies of receptor distribution but the use of modern PET ligand technology appears to be a more promising approach.

There is strong evidence from in vitro studies, that anaesthetics may act via GABAergic mechanism.26 27 42 49 50 52 Measuring with PET the binding characteristic of 11C-labelled flumazenil, a specific benzodiazepine antagonist, in the presence of anaesthetics allows a direct assessment of anaesthetic effects on GABAA receptors in the intact brain. Using this technique, isoflurane has been shown to increase receptor-specific radioligand binding, dependent on GABAA receptor density.32 This observation provides strong support for the hypothesis that the GABAA receptor is involved in mediating the action of volatile anaesthetics in humans. In addition, the radioligand binding measured during anaesthesia at 1.5 MAC (minimum alveolar concentration) was significantly greater than at 1.0 MAC, indicating a dose-related effect of isoflurane on GABAA receptor ligand binding (Fig. 1).



View larger version (39K):
[in this window]
[in a new window]
 
Fig 1 Summed PET during 20–80 min after [11C]flumazenil distribution in two representative volunteers during awake conditions (control) and when anaesthetized with 1.0 MAC isoflurane (top row; subject 1 of 1.0 MAC isoflurane group). The plane of both PET and the corresponding magnetic resonance scans is 4 cm (top row) and 2 cm (bottom row) superior to the canthomeatal line and shows the frontal, temporal, and occipital cortical areas, and the thalamus. The colour scale is shown with red corresponding to 99.9 kBq ml–1 brain tissue. From reference 32 reproduced with permission.

 
Alkire and co-workers8 correlated the data from their in vivo studies3 7 reported recently with regional distribution patterns of various human receptor binding site densities obtained from previous ex vivo studies.16 22 64 The regional reduction in glucose metabolism caused by propofol exhibited a significant correlation (r=–0.86, P<0.0005) with the distribution of the benzodiazepine binding site densities, which means that metabolism decreased more in regions with higher receptor levels (Fig. 2).



View larger version (19K):
[in this window]
[in a new window]
 
Fig 2 Regression plot showing a significant linear relationship between the regional metabolic reductions occurring during propofol anaesthesia in humans and the known regional benzodiazepine ([3H]diazepam) receptor densities. The figure shows that brain metabolic metabolism during propofol anaesthesia decreases more in those brain regions that have benzodiazepine receptors. The line through the data is the regression line. From reference 8 reproduced with permission.

 
In contrast to the link between changes in glucose metabolism caused by propofol and the benzodiazepine receptors the metabolic reduction observed during isoflurane anaesthesia revealed no correlation with the distribution of benzodiazepine receptor densities. This implies that the metabolic reduction seen during isoflurane anaesthesia was not mediated through GABAergic mechanism. However, this hypothesis appears to be in disagreement with the aforementioned findings of Gyulai,32 which emphasize the role of GABAergic transmission in mediating isoflurane anaesthesia. However, based on their own experimental results, Alkire and co-workers did not completely rule out GABAergic mechanisms in mediating isoflurane anaesthesia.8

Another finding of Alkire’s study was an isoflurane-induced reduction in glucose metabolism that correlated with muscarinic (acetylcholine) binding density (r=0.85, P=0.03; Fig. 3). Brain metabolism was less decreased during isoflurane anaesthesia in those regions that had more muscarinic receptors. This seems to indicate that regions with higher muscarinic receptor density are less sensitive to isoflurane, suggesting antagonistic actions of isoflurane and acetylcholine on cerebral metabolism.8 The correlation of isoflurane-induced reductions in glucose metabolism with muscarinic binding density is a highly interesting finding, as muscarinic signaling in the central nervous system is known to be involved in modulation of consciousness and tend to enhance wakefulness.20



View larger version (15K):
[in this window]
[in a new window]
 
Fig 3 Regression plot showing a significant relationship between the regional metabolic reductions occurring during isoflurane anaesthesia in humans and the known regional muscarinic acetylcholine ([3H]quinuclidinyl benzilate) receptor densities. From reference 8 reproduced with permission.

 
A recent PET study of the baboon brain provided further evidence for a more complex mechanism mediating inhalation anaesthesia.33 Gyulai and co-workers found that the regional reduction in CMRGlu caused by halothane anaesthesia was not directly linked to GABAA-receptor density. This finding suggested a mechanism of halothane action other than a pure GABAAergic one to be responsible for the depression of glucose metabolism. However, halothane and the GABAA-receptor agonist muscimol caused a nearly similar metabolic reduction in all regions investigated. This metabolic reduction was significantly increased by a combination of both agents. In contrast to the lack of a significant correlation during halothane alone, the changes in regional glucose metabolism induced by either muscimol or muscimol/halothane showed a significant positive correlation with GABAA-receptor density, suggesting a GABAA receptor-associated mechanism in mediating neuronal inhibition by halothane.

In summary, the recently published studies are in support of the hypothesis that in vivo effects of anaesthetics are mediated at least in part through GABAergic mechanisms. Furthermore, some of these studies indicate a more complex mechanism for volatile agents in comparison with propofol in producing metabolic suppression during anaesthesia. With the expected further development in imaging techniques,37 47 PET will be an indispensable tool to get better insights in the molecular mechanism producing anaesthesia in vivo.

PET studies of analgesic drug effects
Opioids and nitrous oxide are administered routinely as a component of clinical anaesthesia. Although these drugs are not anaesthetics, they are commonly used as an analgesic component of balanced anaesthesia. In addition to their analgesic effects, they produce dose-dependent sedation. The clinical importance of these agents is a good reason to illustrate the potential of modern brain imaging techniques in elucidating the supraspinal mechanism of opioid and nitrous oxide analgesia in humans.41

Painful stimulation with noxious stimuli causes activation in a variety of different brain regions including the anterior cingulate cortex, somatic cortex, thalamus, prefrontal areas, parietal areas, and parts of the motor system as discussed in a recent review article by Peyron.53 The neural response in the somatic areas is known to be related to sensory-discriminative aspects of pain processing. Thalamic activation is assumed to be the result of a general arousal reaction induced by pain, whereas the response to noxious stimulation in the anterior cingulate cortex appears to participate in the affective and attentional components of pain sensation. Parts of the parietal and prefrontal cortices were proposed to reflect attentional and memory networks activated by noxious stimulation. Also often found to be activated by pain were motor-related areas such as the striatum, cerebellum, and the supplementary motor area. Activation of theses areas is assumed to reflect pain-associated escape reactions.

Brain regions closely linked to pain processing also demonstrate changes in their functional state during analgesic procedures. In a recent H215O PET study, remifentanil exhibited dose-dependent changes in rCBF in parts of the distributed pain-processing network.62 During low doses (0.05 µg kg–1 min–1), remifentanil induced an increase in rCBF in the prefrontal cortices, inferior parietal cortices, and the supplementary motor area. Increasing the dose (0.15 µg kg–1 min–1) caused further rCBF increases mainly in the anterior cingulate cortex, a region that frequently has been reported to be activated in pain studies,19 53 and in the caudal periventricular grey and the transition zone of the medial part of the occipital lobe.

These findings for remifentanil are partly in agreement with two previous PET studies of fentanyl action.1 23 In both studies fentanyl caused a significant increase in CBF in the anterior cingulate cortex. Besides this consistent finding, various other brain regions associated with pain processing exhibited changes in CBF after fentanyl administration. In addition to anterior cingulate activation, fentanyl caused a significant increase in CBF in parts of the prefrontal cortex, right motor cortex, and caudate nuclei, whereas a significant decrease was found in parts of the left prefrontal cortex, right temporal cortex and bilaterally in the thalamus and in the left posterior cingulate cortex.

These results provide a good description of the effect of fentanyl on specific neural structures. To understand opioid-induced pain relief, however, the drug effect on pain-induced brain activation needs to be investigated. Adler and co-workers1 assumed that fentanyl would cause a suppression of pain evoked cerebral rCBF response. To confirm this hypothesis, they investigated the effect of fentanyl on CBF in the presence or absence of noxious stimulation using H215O PET. In their study painful stimulation caused a CBF increase in the anterior cingulate cortex, right supplementary motor area, left thalamus, and left inferior frontal cortex. Fentanyl augmented the pain-induced activation in the supplementary motor area and the left inferior frontal cortex, whereas alterations in the CBF response of the anterior cingulate and thalamus were not reported. This phenomenon was accompanied by significant pain relief. Therefore, the authors argued that the supplementary motor area and left inferior frontal cortex possibly reflected the neuroanatomical correlates of fentanyl analgesia. It is important to note that these results contradict the assumption of a specific fentanyl-associated suppression of pain-induced brain stimulation.

Despite the current controversies regarding the use of nitrous oxide its antinociceptive and hypnotic potencies are still exploited during general anaesthesia. Breathing of nitrous oxide was reported to be associated with an increase in rCBF in the anterior cingulate cortex.30 In order to further explore the supraspinal structures involved in mediating nitrous oxides antinociceptive properties, the effects of the drug on pain relief were studied.31 Consistent with other studies53 painful stimulation produced cerebral activation in the thalamus, anterior cingulate cortex, and supplementary motor area, reflecting the various aspects of pain processing. Breathing of 20% nitrous oxide abolished the activation in these areas during painful stimulation. In addition, it was associated with a CBF increase in the contralateral infralimbic and orbitofrontal cortices, suggesting a contribution of both regions to the antinociceptive effects of nitrous oxide.

In summary, the studies described above revealed parts of the neuroanatomical pathways that are important for the reduced intensity and response to pain after opioid and nitrous oxide application. Administration of opioids or nitrous oxide during painful stimulation caused deactivation in the thalamus and activation in the anterior cingulate cortex. These areas therefore appear to be targets for opioid and nitrous oxide analgesia. The observed differences in the pattern of rCBF activation caused by the analgesics further suggests that opioids and nitrous oxide modulate pain perception by affecting different neural circuits.

fMRI studies of anaesthetic action
Despite the great potential of fMRI for investigating the effect of anaesthetics on task-induced brain activation in humans only a small number of studies have been reported up to now.12 34 44

Antognini and co-workers12 investigated the effects of isoflurane on processing of somatosensory stimuli. Noxious (electrical shock) and innocuous stimuli (tactile stimulation) were applied in the awake state and during different levels of isoflurane anaesthesia. Both tactile and electric shock stimulation caused a significant activation in the primary and secondary somatosensory cortex while subjects were awake. At a low anaesthetic concentration (0.7%) the response to tactile stimuli and light electric shock was greatly diminished. However, at supramaximal electrical stimulation stimulus processing in the caudate nucleus and the thalamus was still observed. The authors suggested that this activation might reflect an escape reaction to the noxious stimulus, as it is known that the caudate nucleus is possibly involved in coordination and initiation of motor activities. The dose-dependent response to the different stimuli suggests a subcortical site of action wherein isoflurane prevents movements in response to noxious stimulation. The preserved functional response at the thalamic level during low-dose isoflurane anaesthesia is consistent with the above-mentioned findings for propofol.15 It shows that cortical processing of a stimulus is already abolished during low-dose anaesthesia, whereas higher doses are required to shut down the thalamus. This might indicate similar effects of both drugs on thalamo-cortical transmission of sensory information to the cortex.

Martin and co-workers investigated the effects of the i.v. barbiturate thiopental and found a dose-dependent decrease in functional activation evoked by passive visual stimulation in the striate cortex.44 After i.v. administration of 150 mg thiopental, activation in the striate cortex was still present, but considerably smaller in size compared with the awake state. The reduction in BOLD contrast was related to the ratio of amount of thiopental administered and the subject’s body weight. Forty-five minutes after injection of the anaesthetic, cortical activation almost reached baseline. Unfortunately, these finding were not correlated with any behavioural data or indicator of cognitive performance. Therefore, it is not possible to interpret the importance of the remaining activation in the context of perception and processing of the stimulus or of depth of anaesthesia during imaging.

Heinke and Schwarzbauer investigated the effect of subanaesthetic concentrations of isoflurane (0.42 vol%) on brain activation induced by a visual search.34 The main finding was that isoflurane affected task-induced brain activation in specific neural networks rather than causing a global decrease in functional activation. In most brain regions, the study revealed no isoflurane-related changes in task-induced brain activation. In detail, all areas associated with primary information processing (lateral geniculate nucleus, primary visual cortex, primary motor cortex) activated by the task were not affected during breathing of isoflurane (Fig. 4). However, in three focal cortical regions (anterio-superior insula, left intraparietal sulcus, and right intraparietal sulcus), isoflurane significantly decreased task-induced brain activation (Fig. 5). The affected regions in the parietal association areas are known to be involved in visuo-spatial attention,40 63 whereas the anterior insula plays an important role in visual working memory.18 All subjects were conscious and able to respond to the stimulus, however, average reaction time and error rate were significantly increased demonstrating an impairment of cognitive performance. Considering the isoflurane-related decrease in task-induced activation in the parietal association cortex and the insula, the observed deterioration of cognitive performance was interpreted as a selective impairment of attention and working memory processes in the visual domain.



View larger version (43K):
[in this window]
[in a new window]
 
Fig 4 Quantitative analysis of isoflurane-related changes in task-induced brain activation. Representative voxels were selected in six different regions: AI=anterio-superior insula; IPSL=left intraparietal sulcus; IPSR=right intraparietal sulcus; VC=visual cortex; LGN=lateral geniculate nucleus; MC=motor cortex. The plots show the group-specific z-values for each group (isoflurane, control) and condition (BC=baseline condition; IC=isoflurane condition; RC=recovery condition). Comparing the corresponding time courses of the isoflurane and control group reveals a significant isoflurane-related decrease (z >3.1 corresponding to P<0.001) in AI, IPSL, and IPSR. No such effect, however, is visible in VC, LGN, and MC. From reference 34 reproduced with permission.

 


View larger version (61K):
[in this window]
[in a new window]
 
Fig 5 Three-dimensional representation of the functional interaction map showing regions with a significant isoflurane-related decrease (z >3.1 corresponding to P<0.001) in task-induced activation: AI=anterio-superior insula; IPSL=left intraparietal sulcus; IPSR=right intraparietal sulcus.

 
The effect of isoflurane on task-induced activation is dependent on the dose of the anaesthetic agent. At 0.3% Logothetis and co-workers found no significant isoflurane-related changes in the monkey brain during visual stimulation.43 At a comparable concentration Heinke and Schwarzbauer34 observed a significant decrease in task-induced activation in three distinct cortical regions of the human brain; however, most brain regions were not affected by isoflurane (Figs 3 and 4). No activation was found at 1.3% in the human somatosensory cortex during sensory stimulation12 and at 2.0% in the monkey cortex during visual stimulation.

In summary, the results obtained in the different isoflurane-related fMRI studies are consistent with the view that anaesthetics impair cognitive function, stimulus processing, and consciousness in a gradual fashion.9 21 At very low concentrations (0.3%) isoflurane exhibited no effect on cerebral activation in the monkey brain.43 At a comparable concentration, Heinke and Schwarzbauer found a significant decrease in stimulus-induced activation in three distinct cortical regions, and a preserved activation in the primary cortices and at the thalamic level.34 Preserved subcortical activation (thalamus, caudate nucleus) during noxious stimulation was found when isoflurane was administered in a concentration that causes unconsciousness. During higher concentrations, required to achieve immobility to noxious stimuli, no cortical, or subcortical activation was detected using fMRI.12

Conclusions

General anaesthesia is characterized by loss of consciousness, amnesia, and immobility in response to noxious stimuli. The question of how different anaesthetics produce these behavioural changes is of fundamental interest in anaesthesia research. Initial results have been reported in recent PET and fMRI studies. Key brain structures have been identified, which seem to play an important role in understanding the observed behavioural changes during anaesthesia.

Modern PET ligand technology offers the unique opportunity to investigate the question of how anaesthesia modulates different receptor systems in the human brain. For almost all neurotransmitter systems assumed to be involved in anaesthetic action, radioligands have become available paving the way for a rapid progress in this field of research.

Anaesthesia research is also expected to benefit from a combination of PET or fMRI with electro- and magnetoencephalographic methods that provide a temporal resolution in the millisecond range. Such complementary measurements may further clarify whether anaesthetics affect cognitive function in a gradual fashion or if specific changes in the behavioural state are dramatic events occurring over periods of milliseconds as recently suggested.58

Acknowledgements

The authors thank Prof. Dr Derk Olthoff (MD, PhD, Department of Anesthesiology and Intensive Care Therapy, University of Leipzig, Leipzig, Germany) and Prof. Dr Bernd Urban (PhD, Department of Anesthesiology, University of Bonn, Bonn, Germany) for support and valuable comments.

References

1 Adler LJ, Gyulai FE, Diehl DJ, et al. Regional brain activity changes associated with fentanyl analgesia elucidated by positron emission tomography. Anesth Analg 1997; 84: 120–6[Abstract]

2 Aine JC. A conceptual overview and critique of functional neuroimaging techniques in humans: 1. MRI/fMRI and PET. Crit Rev Neurobiol 1995; 9: 229–309[ISI][Medline]

3 Alkire MT, Haier RJ, Barker SJ, et al. Cerebral metabolism during propofol anesthesia in humans studied with positron emission tomography. Anesthesiology 1995; 82: 393–403[ISI][Medline]

4 Alkire MT, Haier RJ, Shah NK, Anderson CT. Positron emission tomography study of regional cerebral metabolism in humans during isoflurane anesthesia. Anesthesiology 1997; 86: 549–57[ISI][Medline]

5 Alkire MT. Quantitative EEG correlations with brain glucose metabolic rate during anesthesia in volunteers. Anesthesiology 1998; 89: 323–33[ISI][Medline]

6 Alkire MT, Pomfrett CJ, Haier RJ, et al. Functional brain imaging during halothane anesthesia in humans: effects of halothane on global and regional cerebral glucose metabolism. Anesthesiology 1999; 90: 701–9[ISI][Medline]

7 Alkire MT, Haier RJ, Fallon JH. Toward a unified theory of narcosis: brain imaging evidence for a thalamocortical switch as the neurophysiologic basis of anesthetic-induced uncon sciousness. Conscious Cogn 2000; 9: 370–86[ISI][Medline]

8 Alkire MT, Haier RJ. Correlating in vivo anaesthetic effects with ex vivo receptor density data supports a GABAergic mechanism of action for propofol, but not for isoflurane. Br J Anaesth 2001; 86: 618–26[Abstract/Free Full Text]

9 Andrade J. Investigations of hypesthesia: using anesthetics to explore the relationships between consciousness, learning, and memory. Conscious Cogn 1996; 5: 562–80[ISI][Medline]

10 Angel A. Central neuronal pathways and the process of anaesthesia. Br J Anaesth 1993; 71: 148–63[ISI][Medline]

11 Angel A. How do anaesthetcis work? Curr Anaesth Crit Care 1993; 4: 37–45

12 Antognini JF, Buonocore MH, Disbrow EA, Carstens E. Isoflurane anesthesia blunts cerebral responses to noxious and innocuous stimuli: a fMRI Study. Life Sci 1997; 61: 349–54[Medline]

13 Bagary M, Fluck E, File SE, et al. Is benzodiazepine-induced amnesia due to deactivation of the left prefrontal cortex? Psychopharmacology 2000; 150: 292–9[ISI][Medline]

14 Bandettini PA, Wong EC. Magnetic resonance imaging of human brain function. Principles, practicalities, and possibilities. Neurosurg Clin N Am 1997; 8: 345–71[ISI][Medline]

15 Bonhomme V, Fiset P, Meuret P, et al. Propofol anesthesia and cerebral blood flow changes elicited by vibrotactile stimulation: a positron emission tomography study. J Neurophysiol 2001; 85: 1299–308[Abstract/Free Full Text]

16 Braestrup C, Albrechtsen R, Squires RF. High densities of benzodiazepine receptors in human cortical areas. Nature 1977; 269: 702–4[ISI][Medline]

17 Budinger TF. PET instrumentation: what are the limits? Semin Nucl Med 1998; 28: 247–67[ISI][Medline]

18 Courtney SM, Ungerleider LG, Keil K, Haxby JV. Transient and sustained activity in a distributed neural system for human working memory. Nature 1997; 386: 608–11[ISI][Medline]

19 Derbyshire SW. Meta-analysis of thirty-four independent samples studied using PET reveals a significantly attenuated central response to noxious stimulation in clinical pain patients. Curr Rev Pain 1999; 3: 265–80[Medline]

20 Durieux ME. Muscarinic signaling in the central nervous system. Recent developments and anesthetic implications. Anesthesiology 1996; 84: 173–89[ISI][Medline]

21 Dwyer R, Bennett HL, Eger EI 2nd, Heilbron D. Effects of isoflurane and nitrous oxide in subanesthetic concentrations on memory and responsiveness in volunteers. Anesthesiology 1992; 77: 888–98[ISI][Medline]

22 Enna SJ, Bennet JP Jr, Bylund DB, et al. Neurotransmitter receptor binding: regional distribution in human brain. J Neurochem 1977; 28: 233–6[ISI][Medline]

23 Firestone LL, Gyulai F, Mintun M, et al. Human brain activity response to fentanyl imaged by positron emission tomography. Anesth Analg 1996; 82: 1247–51[Abstract]

24 Fiset P, Paus T, Daloze T, et al. Brain mechanisms of propofol-induced loss of consciousness in humans: a positron emission tomographic study. J Neurosci 1999; 19: 5506–13[Abstract/Free Full Text]

25 Forster BB, MacKay AL, Whittall KP, et al. Functional magnetic resonance imaging: the basics of blood-oxygen-level-dependent (BOLD) imaging. Can Assoc Radiol J 1998; 49: 320–9[ISI][Medline]

26 Franks NP, Lieb WR. Selective actions of volatile general anaesthetics at molecular and cellular levels. Br J Anaesth 1993; 71: 65–76[ISI][Medline]

27 Franks NP, Lieb WR. Molecular and cellular mechanisms of general anaesthesia. Nature 1994; 367: 607–14[ISI][Medline]

28 Goodyear BG, Menon RS, Brief visual stimulation allows mapping of ocular dominance in visual cortex using fMRI. Hum Brain Mapp 2001; 14: 210–7[ISI][Medline]

29 Grasby P, Malizia A, Bench C. Psychopharmacology—in vivo neurochemistry and pharmacology. Br Med Bull 1996; 52: 513–26[Abstract]

30 Gyulai FE, Firestone LL, Mintun MA, Winter PM. In vivo imaging of human limbic responses to nitrous oxide inhalation. Anesth Analg 1996; 83: 291–8[Abstract]

31 Gyulai FE, Firestone LL, Mintun MA, Winter PM. In vivo imaging of nitrous oxide-induced changes in cerebral activation during noxious heat stimuli. Anesthesiology 1997; 86: 538–48[ISI][Medline]

32 Gyulai FE, Mintun MA, Firestone LL. Dose-dependent enhancement of in vivo GABA(A)-benzodiazepine receptor binding by isoflurane. Anesthesiology 2001; 95: 585–93[ISI][Medline]

33 Gyulai FE, Firestone LL, Mintun MA. Halothane potentiates GABAA receptor-mediated metabolic depression in vivo: a positron emission tomography (PET) study. The Sixth International Conference on Molecular and Basic Mechanisms of Anesthesia, Book of Abstracts, 2001; 5B09: 45

34 Heinke W, Schwarzbauer C. Subanesthetic isoflurane affects task-induced brain activation in a highly specific manner. Anesthesiology 2001; 94: 973–81[ISI][Medline]

35 Herscovitch P, Markham J, Raichle ME. Brain blood flow measured with intravenous H215O. I. Theory and error analysis. J Nuclear Med 1983; 24: 782–9[Abstract]

36 Herscovitch P. Can [15O]water be used to evaluate drugs? J Clin Pharmacol 2001; 41: 11–20

37 Jacobs RE, Cherry SR. Complementary emerging techniques: high-resolution PET and MRI. Curr Opin Neurobiol 2001; 11: 621–9[ISI][Medline]

38 Kety SS. The theory and applications of the exchange of inert gas at the lungs and tissues. Pharmacol Rev 1951; 3: 1–41[ISI][Medline]

39 Kety SS. Measurements of local cerebral flow by the exchange of an inert, diffusible substance. Methods Med Res 1960; 8: 228–36

40 Krause BJ, Schmidt D, Mottaghy FM, et al. Episodic retrieval activates the precuneus irrespective of the imagery content of word pair associates. A PET study. Brain 1999; 122: 255–63[Abstract/Free Full Text]

41 Kurata J, Gyulai FE, Firestone LL. Use of positron emission tomography to measure brain activity responses to fentanyl analgesia. Curr Rev Pain 1999; 3: 359–66[Medline]

42 Lees G. Molecular mechanism of anaesthesia: light at the end of the channel? Br J Anaesth 1998; 81: 491–3[Free Full Text]

43 Logothetis NK, Guggenberger H, Peled S, Pauls J. Functional imaging of the monkey brain. Nat Neurosci 1999; 2: 555–62[ISI][Medline]

44 Martin E, Thiel T, Joeri P, et al. Effect of pentobarbital on visual processing in man. Hum Brain Mapp 2000; 10: 132–9[ISI][Medline]

45 Matthew E, Andreason P, Pettigrew K, et al. Benzodiazepine receptors mediate regional blood flow changes in the living human brain. Proc Natl Acad Sci USA 1995; 92: 2775–9[Abstract]

46 Menon DK. Mapping the anatomy of unconsciousness—imaging anaesthetic action in the brain. Br J Anaesth 2001; 86: 607–10[Free Full Text]

47 Mintun MA. In vivo molecular brain imaging: approaches with positron emission tomography. The Sixth International Conference on Molecular and Basic Mechanisms of Anesthesia, Book of Abstracts, 2001; 5A01: 41

48 Ogawa S, Lee TM, Kay AR, Tank DW. Brain magnetic resonance imaging with contrast dependent on blood oxygenation. Proc Natl Acad Sci USA 1990; 87: 9868–72[Abstract]

49 Orser BA, Wang LY, Pennefather PS, MacDonald JF. Propofol modulates activation and desensitization of GABAA receptors in cultured murine hippocampal neurons. J Neurosci 1994; 14: 7747–60[Abstract]

50 Orser BA, McAdam LC, Roder S, MacDonald JF. General anaesthetics and their effects on GABAA receptor desensitization. Toxicol Lett 1998; 100/101: 217–24

51 Paus T. Functional anatomy of arousal and attention systems in the human brain. Prog Brain Res 2000; 126: 65–77[ISI]

52 Peduto VA, Concas A, Santoro G, et al. Biochemical and electrophysiologic evidence that propofol enhances GABAergic transmission in the rat brain. Anesthesiology 1991; 75: 1000–9[ISI][Medline]

53 Peyron R, Laurent B, Garcia-Larrea L. Functional imaging of brain responses to pain. A review and meta-analysis. Neurophysiol Clin 2000; 30: 263–88[ISI][Medline]

54 Raichle M. Behind the scenes of functional brain imaging: a historical and physiological perspective. Proc Natl Acad Sci USA 1998; 95: 765–72[Abstract/Free Full Text]

55 Reinsel RA, Veselis RA, Dnistrian AM, et al. Midazolam decreases cerebral blood flow in the left prefrontal cortex in a dose-dependent fashion. Int J Neuropsychopharmacol 2000; 3: 117–27[ISI][Medline]

56 Veselis RA, Reinsel RA, Beattie BJ, et al. Midazolam changes cerebral blood flow in discrete brain regions. Anesthesiology 1997; 87: 1106–17[ISI][Medline]

57 Veselis RA, Reinsel R, Feshenko V, et al. Dose related decreases in rCBF in R and L prefrontal cortices (PFC) during memory impairment with midazolam and propofol. NeuroImage 2001; 6: 13(Abstract)

58 Veselis RA, Reinsel RA Feshenko V. Novel insights into drug—memory interactions using functional neuroimaging techniques. The Sixth International Conference on Molecular and Basic Mechanisms of Anesthesia, Book of Abstracts, 2001; 5A02: 41

59 Villringer A, Dirnagel U. Coupling of brain activity and cerebral blood flow: basis of functional neuroimaging. Cerebrovasc Brain Metab Rev 1995; 7: 240–76[ISI][Medline]

60 Volkow ND, Wang GJ, Hitzemann R, et al. Depression of thalamic metabolism by lorazepam is associated with sleepiness. Neuropsychopharmacology 1995; 12: 123–32[ISI][Medline]

61 Volkow ND, Rosen B, Farde L. Imaging the living human brain: magnetic resonance imaging and positron emission tomography. Proc Natl Acad Sci USA 1997; 94: 2787–8[Free Full Text]

62 Wagner KJ, Willoch F, Kochs EF, et al. Dose-dependent regional cerebral blood flow changes during remifentanil infusion in humans. Anesthesiology 2001; 94: 732–9[ISI][Medline]

63 Wojciulik E, Kanwisher N. The generality of parietal involvement in visual attention. Neuron 1999; 23: 747–64[ISI][Medline]

64 Zezula J, Cortes R, Probst A, Palacios JM. Benzodiazepine receptor sites in the human brain: autoradiographic mapping. Neuroscience 1988; 25: 771–95[ISI][Medline]