The effect of different anesthetics on neurovascular coupling
Introduction
Since its introduction in 1991 the role of functional magnetic resonance imaging (fMRI) in basic and clinical neuroscience has grown rapidly, with 2500 papers published in 2005 alone (Bandettini, 2007). Although fMRI is now widely used for non-invasive investigations of human brain function, we still do not have a clear understanding of how accurately these images based on vascular changes reflect neural activity (Bandettini, 2007, Iadecola, 2004, Shibasaki, 2008). Because of the growing number of functional studies with fMRI, quantification of the relationship between the hemoglobin signal and the underlying neural activity is becoming increasingly important.
Recently, significant effort has been devoted to invasive animal studies (Berwick et al., 2008, Devor et al., 2005, Mathiesen et al., 1998, Sheth et al., 2004). Such studies are typically performed using different anesthetics, which may alter the coupling between neural and vascular responses in different ways. While the effects of different anesthetics on electrical, metabolic, or vascular responses alone have been reported, only a few studies have investigated the effect of anesthetics on the relationship between the electrical and vascular functional imaging signals measured during neuronal activity (Austin et al., 2005, Maandag et al., 2007, Martin et al., 2006, Masamoto et al., 2009, Ueki et al., 1992). During event-related parametric electrical forepaw stimulation in rats, we investigated the neurovascular coupling using six different anesthetics (alpha-chloralose, isoflurane, pentobarbital, propofol, ketamine–xylazine, and fentanyl–droperidol).
Different anesthetics act differently on neurotransmitters and neuronal membrane polarization thresholds (Hyder et al., 2002, Maandag et al., 2007, Sibson et al., 1998, Sicard et al., 2003), and as a result modulate the measured EEG evoked signals differently (Antunes et al., 2003a, Antunes et al., 2003b). The main properties/characteristics of the anesthetics we chose for this study are briefly listed here.
Alpha-chloralose, pentobarbital, isoflurane, and propofol are mainly GABAergic anesthetics and prolong the evoked inhibitory postsynaptic currents mediated by γ-aminobutyric acid A (GABAA) by increasing channel conductance or channel open time (Belelli et al., 1999, Franks and Lieb, 1994). Of these four GABAergic anesthetics, alpha-chloralose is the most commonly used for functional hemodynamic studies because of its weaker effects on cardiovascular, respiratory, and reflex functions (Nakao et al., 2001), and more importantly, the evoked hemodynamic responses are larger with alpha-chloralose than with other anesthetics (Austin et al., 2005). Isoflurane is commonly used in electrophysiology studies due to its ease of use, even though it partially reduces neuronal excitation and cerebral metabolism like most volatile anesthetics. At doses higher than 1.6% isoflurane increases cerebral blood flow (CBF) (Eger, 1984) and for this reason it is not commonly used for hemodynamic functional studies. Pentobarbital and propofol are usually not used in functional studies as they depress the central nervous system as well as cortical and subcortical structures, and produce large decreases in EEG responses (Antognini et al., 2006, Crosby et al., 1983). Pentobarbital, a barbiturate, in contrast with other anesthetics, potentiates not only inhibitory but also excitatory postsynaptic receptors (Franks and Lieb, 1994). Propofol is the intravenous anesthetic of choice in surgery because of its favorable operating conditions and associated rapid recovery. It reduces heart rate (Wang et al., 2004) and baseline cerebral blood flow (Cenic et al., 2000, Veselis et al., 2005). With propofol, a dose-dependent gradual reduction of SEP amplitude and prolonged latency has been measured in rats (Logginidou et al., 2003).
The other two combinations of anesthetics we used are not GABAergic. Ketamine, in particular, does not interact with GABA receptors (Franks and Lieb, 1994) but mainly inhibits excitatory glutamatergic neurotransmission, blocking N-methyl-d-aspartate (NMDA) receptors. Ketamine also acts on opioid, monoaminergic and muscarinic receptors (Hirota and Lambert, 1996). Ketamine is often used in functional electrophysiology studies because it does not suppress neural activity (Kochs and Bischoff, 1994), and also because animals do not need to be intubated for its use. In fact, with ketamine, there is minimal cardio-respiratory depression. Ketamine causes both increases and decreases in cerebral metabolism (glucose utilization) depending on the brain region. Decreases in metabolism occur in the somatosensory and auditory cortices (Crosby et al., 1982). It produces a dose-related clinical state of dissociative anesthesia in combination with analgesic properties, and is commonly used in conjunction with xylazine. Xylazine is a sedative and a muscle relaxant. It minimizes the side effects produced by ketamine alone, such as tremor, muscle rigidity, and excitement during recovery (Wright, 1982). Xylazine inhibits noradrenergic neurotransmission by activating presynaptic adrenergic alpha-2 receptors (Oria et al., 2008). Because it is an agonist for alpha-2-adrenoreceptors, xylazine decreases the heart rate, causes hypotension, decreases venous cerebral blood volume and intracranial pressure, and depresses the central nervous system (Greene and Thurmon, 1988). In addition, xylazine induces a reduction in CBF (Lei et al., 2001).
Fentanyl is a synthetic opiate analgesic used routinely in anesthesia procedures in humans. Fentanyl binds with high affinity to the morphine μ-opioid receptors (Villiger et al., 1983). Fentanyl produces a dose-related decrease in both CBF and cerebral metabolic rate of oxygen (CMRO2) (Carlsson et al., 1982). It is commonly used in conjunction with droperidol to induce neuroleptanesthesia (Bissonnette et al., 1999). Droperidol is a neuroleptic drug and belongs to the butyrophenones; it has antipsychotic effects, which are produced by blocking dopamine receptors (Bissonnette et al., 1999). One of the advantages of droperidol is its lack of EEG effects, and although it reduces cerebral blood flow by vasoconstricting the cerebral vessels, the cerebral metabolic rate of oxygen remains unchanged. Clinical doses of droperidol decrease systemic blood pressure and cause reductions in tidal volume, airway resistance and functional residual capacity.
We used scalp electroencephalography (EEG) and diffuse optical imaging (DOI) for our functional measurements. These two modalities allow for non-invasive scalp measurements, interrogate large volumes of tissue, and can be easily integrated together for simultaneous measurements (Franceschini et al., 2008). While EEG offers only an indirect measure of the cascade of neuronal events during neural activity, it is the only modality other than MEG that can be used to monitor neural activity non-invasively in humans, and facilitates the translation of results obtained in animals using the same techniques to human studies. Much effort over the past 30 years has focused on determining the relationship between somatosensory evoked potentials (SEP), local field potentials (LFP), and neural activity. Anatomical data from the sites of specific thalamic inputs (White, 1979, White and Hersch, 1982) and the projection patterns of pyramidal axons between cortical layers (Thomson and Bannister, 2003) have confirmed that activation of the primary somatosensory cortex (SI) starts with the thalamic input in layers IV and VI and is followed by activation in layers III and II and then layer V. Response latencies increase in a systematic fashion from middle to superficial to deep cortical layers (Simons, 1978). In addition, when synaptic activity reaches the superficial layers, it propagates horizontally with a large amount of synapses between layers I and III. Evoked potential responses across primary cortices in rats, primates and humans have a similar structure, consisting first of a large and narrow positive component, P1, followed within 10 ms by a large negative component, N1, and then by two slow components, P2 and N2, tens of ms after N1 (Allison et al., 1989, Arezzo et al., 1981, Di and Barth, 1991, Kulics and Cauller, 1986). Current source density (CSD) analysis of laminar profiles of LFP link P1 to the largest and earliest current sinks in layers IV and VI. P1 is the primary evoked potential directly originated from SI-specific thalamocortical inputs (Mitzdorf, 1985), and reflects the initial depolarization of layer II and V pyramidal cells. Following this initial depolarization, population spikes are generated in layer Vb infragranular cells. Axon-collaterals of layer Vb pyramidal cells produce an enhanced activation of the supragranular pyramidal cells in layer I–II, which generates the secondary evoked potential N1 (Agmon and Connors, 1991, Jellema et al., 2004, Kulics and Cauller, 1986). P2 and N2 arise from activation of cortico-cortical connections originating in the superficial layers of the central column (Kublik et al., 2001, Wrobel et al., 1998), and derive from a combination of both inhibitory and repolarization processes (Steriade, 1984). While the neural origin of P1 and N1 is well established, the inconsistent source-sink patterns and absence of multi-unit activity (MUA) at longer times make the functional significance of P2 and N2 unclear (Kulics and Cauller, 1986).
While most invasive animal studies have attempted to correlate the hemodynamic responses to the primary evoked potential, we believe that a substantial contribution of the hemodynamic response is derived from secondary and late cortical transmission. In fact, in a previous study using the same measurement modalities (EEG and DOI) and parametric electrical forepaw stimulation (Franceschini et al., 2008), we have found that the primary SEP component (P1) exhibits a weaker correlation with the hemodynamic response than the secondary (N1) and late (P2) SEP components. Similarly in humans, using DOI and MEG during median nerve stimulation experiments, we have obtained better hemoglobin response predictions using late (> 30 ms) neural components (Ou et al., 2009). Parametric stimulation alone does not produce enough differences in the SEP components to distinguish the individual contributions of N1 and P2. Here, we tested whether the pharmacological manipulation produced by the different anesthetics on the SEP is sufficient to disentangle individual contributions of the secondary and late components to the hemodynamic response.
Determining whether the hemodynamic response from the SEP is driven by secondary or late cortico-cortico transmission rather than by afferent inputs to layer IV would have a profound effect on the clinical role of BOLD fMRI and DOI. In fact, the presence of a hemodynamic response to a sensory stimulus itself would confirm the integrity of the sensory system to the level of the cortico-cortical responses and indicate that sensory information had arrived and was processed to the point that it could pass to other areas of the cortex. However, if the presence of a hemodynamic response only confirms the integrity of the system to the level of the thalamic inputs, the hemodynamic response would only confirm that sensory information had reached the cortex; it would provide no information as to the potential for further processing. Therefore, if the hemodynamic response is a marker not just of information arrival from non-cortical areas but of local information processing, the hemodynamic response can serve as a biomarker for the functional integrity of cortical regions.
As discussed above, anesthetics also act on baseline cerebral blood flow, cerebral blood volume (CBV) and vascular reactivity (Hyder et al., 2002, Maandag et al., 2007, Sibson et al., 1998, Sicard et al., 2003). To account for the resulting effects on neurovascular coupling, in our experiments, in addition to functional electrical and vascular responses, we measured baseline blood flow and response to a hypercapnia challenge with each anesthetic. These measures allow us to determine the influence of baseline blood flow on the neurovascular coupling.
Section snippets
Animal preparation
Six groups of male Sprague–Dawley rats (5–6 rats in each anesthetic group), a total of 33 animals (298 ± 18 g), were included in this study. During all surgical procedures the animals were anesthetized with isoflurane (2–2.5%) administered via a facemask in a gas mixture of 80% air and 20% oxygen. After tracheotomy and cannulation of the femoral artery and vein, animals were mounted on a stereotactic frame. Heating blankets maintained core temperature at 36.5–37.5 °C. Six different anesthetics were
Baseline CBF and CBF changes during hypercapnia for different anesthetics
As part of the protocol, we measured cerebral blood flow using the DCS system to evaluate the effects of the different anesthetics on the baseline vascular state, as well as the possible effects of different baseline vascular states on neurovascular coupling. As expected, isoflurane provided the largest baseline BFi, and ketamine–xylazine the lowest because of the action of xylazine on systemic blood flow (see Table 2). In all rats, in addition to baseline BFi, we measured BFi changes during
Discussion and conclusions
These results suggest that the hemodynamic response is not solely driven by thalamic afferent inputs (P1) but it is largely controlled by secondary and late cortical transmission and influenced by baseline blood flow. In fact, when using a linear regression model, the coupling between the thalamic afferent component P1 and the hemodynamic responses changes across anesthetics. In order to maintain the same neurovascular coupling relationship across different anesthetics we need to add secondary
Acknowledgments
We would like to thank Anna Devor, Ellen Grant and John Marota for valuable discussions and Gary Boas for careful editing of the manuscript. This research is supported by the US National Institutes of Health (NIH) grants R01-EB001954 and R01-EB006385.
References (99)
- et al.
Thalamocortical responses of mouse somatosensory (barrel) cortex in vitro
Neuroscience
(1991) - et al.
Comparison of electroencephalogram activity and auditory evoked responses during isoflurane and halothane anaesthesia in the rat
Vet. Anaesth. Analg.
(2003) - et al.
Topography and intracranial sources of somatosensory evoked potentials in the monkey. II. Cortical components
Electroencephalogr. Clin. Neurophysiol.
(1981) - et al.
Confounding effects of anesthesia on functional activation in rodent brain: a study of halothane and alpha-chloralose anesthesia
NeuroImage
(2005) Functional MRI today
Int. J. Psychophysiol.
(2007)- et al.
General anaesthetic action at transmitter-gated inhibitory amino acid receptors
Trends Pharmacol. Sci.
(1999) - et al.
Optical coherence tomography (OCT) reveals depth-resolved dynamics during functional brain activation
J. Neurosci. Methods
(2009) - et al.
Topographic analysis of field potentials in rat vibrissa/barrel cortex
Brain Res.
(1991) - et al.
Coupling between somatosensory evoked potentials and hemodynamic response in the rat
NeuroImage
(2008) - et al.
Ketamine: its mechanism(s) of action and unusual clinical uses
Br. J. Anaesth.
(1996)