Historical and modern evidence for the role of reward circuitry in emergence

Abstract

Increasing evidence supports a role for brain reward circuitry in modulating arousal along with emergence from anesthesia. Emergence remains an important frontier for investigation, since no drug exists in clinical practice to initiate rapid and smooth emergence. In this review we discuss (1) clinical evidence and (2) preclinical evidence indicating a role for two brain regions classically considered integral components of the mesolimbic brain reward circuitry: the ventral tegmental area and the nucleus accumbens, in emergence from propofol and volatile anesthesia. We then describe (3) modern systems neuroscience approaches to neural circuit investigations that will help span the large gap between preclinical and clinical investigation with the shared aim of developing therapies to promote rapid emergence without agitation or delirium. We propose that neuroscientists include models of whole brain network activity in future studies to inform the translational value of preclinical investigations and foster productive dialogues with clinician anesthesiologists.

I. Introduction

Emergence from general anesthesia is a dynamic time of transition from the anesthetized to awake state that continues to be an unpredictable and fragile period for patients in perioperative care^1–3. Some evidence suggests that emergence is regulated by additional neural processes independent of drug clearance^4–6, and is not simply the reverse process of anesthetic induction, as described by models of anesthetic hysteresis^7–9. Many surgical cases necessitate rapid arousal for assessment of patients’ postoperative cognitive and motor abilities. So, rapid and smooth emergence is desirable for both patient safety and perioperative efficiency. Why some patients emerge quickly from higher anesthetic concentrations than others and why subsets of patients undergo an agitated, combative state during the process of emergence is currently not well understood at multiple levels. Together, this unpredictability and the lack of an established therapy to facilitate emergence highlights the need for a better understanding of the basic neuropharmacological mechanisms mediating emergence from anesthesia.

Emergence agitation can be dangerous, with patients manifesting combative behaviors that can result in self-injury, harm to providers, catheter removal, self-extubation and airway obstruction. Postoperative delirium can also result in longer hospital length of stay and worsened clinical outcomes^10,11,12. Efforts to mitigate adverse emergence phenomena, like agitation and delirium, are currently focused on avoiding inhalational anesthetics or supplementing inhalational agents with intravenous sedatives like the alpha-2 receptor agonist dexmedetomidine^13–18.

However, the results from these studies vary widely with patient population^19–20, and the basic neuronal mechanisms modulating emergence under different anesthetic conditions is still unclear. Mechanisms of emergence from volatile or propofol anesthesia, which can directly bind to gamma aminobutyric acid (GABA) receptors, can be inconsistent with mechanisms found to mediate the effects of ketamine, which blocks glutamatergic neurotransmission^22,23. However, common anesthetic substrates in the brain exist, such as the activation of hypothalamic neurons²⁴. Here we will focus our discussion on studies that use propofol, sevoflurane or isoflurane for maintenance general anesthesia, since they form much of the literature investigating mesolimbic circuitry in emergence. Further mechanistic basic science research is needed to examine whether findings hold true across disparate anesthetic conditions. There is a clinical need for therapeutic interventions targeting emergence and the postanesthetic period to improve the predictability, speed and safety of anesthesia care. By prioritizing a translational and multidisciplinary approach, basic neuroscientists can help to uncover these gaps in knowledge.

Decades of accumulating literature supports a role for dopaminergic signaling through the brain reward circuitry in promoting arousal (for reviews see ref^25–27). Here we summarize key clinical and preclinical evidence supporting a central role of the brain’s mesolimbic dopaminergic reward circuitry in modulating emergence from general propofol and volatile anesthesia, with a focus on the ventral tegmental area and the nucleus accumbens regions. The same neural circuitry may be important for the pathophysiology of emergence agitation, given the essential role of reward circuitry in regulating emotional and arousal-related behavioral states. We then discuss systems neuroscience approaches for bridging preclinical and clinical studies of brain reward circuitry in emergence to promote the therapeutic application of preclinical investigations (Figure 1).

Figure 1.

Schematic highlighting the need to bridge the gap between preclinical and clinical studies of anesthesia emergence with translational research. The numbered boxes highlight common methodologies in either clinical or preclinical research. The insets show midbrain slices from the human and mouse brain as well as the chemical structures of propofol and sevoflurane, two of the most common anesthetics under investigation. Figure created with BioRender.com

Until recently, high resolution approaches for careful examination within the intact brain did not exist to enable discrete cell-type and region-specific investigations of circuit dynamics. However, now we can harness viral-mediated and genetic approaches to deliver engineered photoactivatable compounds and perform whole brain imaging at the single-cell resolution. This more intricate systems neuroscience approach can be used to understand brain circuitry in preclinical models of awake, behaving rodents and even non-human primates^28–31. In order to place the granularity of these investigations in a clinically useful framework, mechanistic cellular-level studies must then be examined in the context of changes to whole brain activity.

While there is a broad and vast literature describing the role of dopaminergic circuitry in mediating arousal and reward-reinforcing behaviors, this review is limited to systems neuroscience studies of particular relevance to the clinical practicing anesthesiologist (Table 1). We also refer the reader to comprehensive reviews of mesolimbic circuitry^32–36 and to manuscripts discussing transcriptomic tools for studying the brain under anesthesia^37,38 such as single-cell RNA sequencing and CRISPR/Cas9 approaches^39–41 that are beyond the scope of this review.

Table 1.

Selected list of primary preclinical manuscripts investigating mesocorticolimbic reward circuitry in emergence from general anesthesia.

Year	Citation	Brain Region / Neurotransmitter	Anesthetic	Species	Major Finding
Historical References
1961	Eckenhoff et al, Anesthesiology1	--	Thiopental, Halothane, Ether, Cyclopropane, Nitrous oxide, Spinal	Human	First descriptions of “emergence excitement” and differential effects of anesthetic drug treatment in 14,436 patients
1994	Mantz et al, Anesthesiology2	Striatum / dopamine	Halothane, Isoflurane, Thiopental, Ketamine	Rat	Anesthetics significantly alter spontaneous and evoked dopamine release in striatal synaptosomes
1997	Irifune, et al, Anesthesiology3	Nucleus accumbens / dopamine	Isoflurane	Mouse	HPLC assays show increased dopamine turnover, hyperlocomotion during emergence
1999	Tsukada et al, Brain Res4	Striatum / dopamine	Isoflurane	Rhesus Monkey	PET and microdialysis show enhanced DAT-inhibition and D2 receptor binding under isoflurane
1999	Fiset et al., J Neurosci5	Medial thalamus, midbrain	Propofol	Human	PET shows cerebral blood flow changes in midbrain and thalamus
Brain Region-Specific Manipulations
2008	Kelz et al, PNAS6	Hypothalamus / orexin	Isoflurane, Sevoflurane	Mouse	Ablation of orexinergic neurons or an orexin-1 antagonist delays emergence, not induction
2010	Mhuircheartaigh, et al J Neurosci7	Putamen, Thalamus, Cortex	Propofol	Human	fMRI BOLD shows changes in subcortical connectivity
2011	Shirasaka et al, J Anesth8	Prefrontal cortex / orexin	Propofol	Rat	ICV injection of orexin speeds emergence, increases NE and DA release in PFC
2014	Solt et al, Anesthesiology9	Ventral tegmental area, substantia nigra / dopamine?	Isoflurane, Propofol	Rat	Electrical stimulation of the VTA speeds emergence
2014	McCarren, et al, J Neurosci10	Ventrolateral preoptic nucleus / norepinephrine	Isoflurane, Dexmedetomidine	Mouse	Single-cell RT-PCR of VLPO neurons and role of adrenergic manipulation
2014	Vazey et al, PNAS11	Locus coeruleus / norepinephrine	Isoflurane	Rat	Chemogenetic activation of LC neurons speeds emergence
2015	Zhou, et al, PLoS One12	Ventral tegmental area / dopamine	Propofol, Isoflurane, Ketamine	Rat	Lesioning VTA dopamine neurons with 6-OHDA prolongs emergence from propofol, not isoflurane
2016	Taylor, et al, PNAS13	Ventral tegmental area / dopamine	Isoflurane	Mouse	Optogenetic activation of VTA dopamine neurons speeds emergence
2016	Muindi et al, Behav Brain Res14	Parabrachial nucleus / glutamate?	Isoflurane	Mouse	Electrical stimulation of parabrachial nucleus speeds emergence
2017	Fu et al, J Neurochem15	Central medial thalamus / norepinephrine	Propofol	Rat	Norepinephrine microinjection in central medial thalamus speeds emergence
2018	Du, et al, Cell Reports16	Locus coeruleus / norepinephrine	Propofol, Etomidate	Zebrafish	Deletion of dopamine-β-hydroxylase in LC neurons delays emergence from intravenous anesthesia
2019	Yin et al, Front Neural Circuits17	Ventral tegmental area, hypothalamus / GABA	Isoflurane	Mouse	Activation of VTA GABA to hypothalamus slows emergence
2019	Wang, et al, Anesthesiology18	Parabrachial nucleus / glutamate	Sevoflurane	Mouse	Activation of PBN glutamate neurons speeds emergence
2019	Zhang et al, FASEB J19.	Reticular thalamus / norepinephrine	Propofol	Mouse	LC to TRN norepinephrine projections delay emergency by activating α1 adrenergic receptor
2019	Torturo, et al, eNeuro20	Ventral tegmental area / dopamine	Isoflurane	Rat	Isoflurane inhibits exocytosis in cultured rat dopamine neurons by a distinct calcium-mediated mechanism
2019	Li, et al, Br J Anaesth21	Ventral tegmental area / orexin	Isoflurane	Rat	Microinjection of orexin in the VTA promotes emergence by activating dopamine neurons
2020	Zhang, et al, J Neurophysiol22	Prefrontal cortex / acetylcholine, adenosine, norepi	Isoflurane	Mouse	Microdialysis studies showing neurotransmitter roles in anesthetized to awake state transition
2020	Luo et al, Front Neurosci23	Basal forebrain / acetylcholine	Isoflurane, Propofol	Mouse	Chemogenetic activation of cholinergic neurons speeds emergence
2020	Gretenkord, et al Eur J Neurosci24	Ventral tegmental area, prefrontal cortex / dopamine	Urethane	Rat	Stimulation of VTA and D1 receptors in PFC promotes arousal
2021	Ao, et al, Brain Behav25	Paraventricular thalamus / dopamine	Isoflurane	Mouse	PVT c-fos activity increases after emergence, enhanced by a D2-agonist
2021	Zhang et al, Brain Behav26	Nucleus accumbens shell / dopamine	Isoflurane	Mouse	D1 receptor agonist accelerates emergence in young but not aged mice
2021	Bao et al, Current Biology27	Nucleus accumbens/dopamine	Sevoflurane	Mouse	Chemogenetic activation of D1 receptors delays induction and accelerates emergence
Recommended Literature Reviews on Emergence
2008	Franks NP28	--	--	Human/rodent	Molecular targets of arousal
2010	Brown, et al, NEJM29	--	--	Human	Relationship of anesthesia to sleep and coma
2016	Tarnal, et al, J Neurosurg Anesthesiol30	--	--	Human	Hysteresis, neural inertia, and active emergence
2019	Kelz et al, Anesth Analg31	--	--	Rodent	Neurotransmitter modulators of emergence

List of Table References:

^1.Eckenhoff JE, Kneale DH, Dripps RD: The incidence and etiology of postanesthetic excitment. A clinical survey. Anesthesiology 1961; 22:667–73

^2.Mantz J, Varlet C, Lecharny JB, Henzel D, Lenot P, Desmonts JM: Effects of volatile anesthetics, thiopental, and ketamine on spontaneous and depolarization-evoked dopamine release from striatal synaptosomes in the rat. Anesthesiology 1994; 80:352–63

^3.Irifune M, Sato T, Nishikawa T, Masuyama T, Nomoto M, Fukuda T, Kawahara M: Hyperlocomotion during Recovery from Isoflurane Anesthesia Is Associated with Increased Dopamine Turnover in the Nucleus Accumbens and Striatum in Mice. Anesthesiology 1997; 86:464–75

^4.Tsukada H, Nishiyama S, Kakiuchi T, Ohba H, Sato K, Harada N, Nakanishi S: Isoflurane anesthesia enhances the inhibitory effects of cocaine and GBR12909 on dopamine transporter: PET studies in combination with microdialysis in the monkey brain. Brain Res 1999; 849:85–96

^5.Fiset P, Paus T, Daloze T, Plourde G, Meuret P, Bonhomme V, Hajj-Ali N, Backman SB, Evans AC: Brain Mechanisms of Propofol-Induced Loss of Consciousness in Humans: a Positron Emission Tomographic Study. J Neurosci 1999; 19:5506–13

^6.Kelz MB, Sun Y, Chen J, Cheng Meng Q, Moore JT, Veasey SC, Dixon S, Thornton M, Funato H, Yanagisawa M: An essential role for orexins in emergence from general anesthesia. Proc Natl Acad Sci U S A 2008; 105:1309–14

^7.Mhuircheartaigh RN, Rosenorn-Lanng D, Wise R, Jbabdi S, Rogers R, Tracey I: Cortical and Subcortical Connectivity Changes during Decreasing Levels of Consciousness in Humans: A Functional Magnetic Resonance Imaging Study using Propofol. J Neurosci 2010; 30:9095–102

^8.Shirasaka T, Yonaha T, Onizuka S, Tsuneyoshi I: Effects of orexin-A on propofol anesthesia in rats. J Anesth 2011; 25:65–71

^9.Solt K, Van Dort CJ, Chemali JJ, Taylor NE, Kenny JD, Brown EN: Electrical stimulation of the ventral tegmental area induces reanimation from general anesthesia. Anesthesiology 2014; 121:311–9

^10.McCarren HS, Chalifoux MR, Han B, Moore JT, Meng QC, Baron-Hionis N, Sedigh-Sarvestani M, Contreras D, Beck SG, Kelz MB: α2-Adrenergic Stimulation of the Ventrolateral Preoptic Nucleus Destabilizes the Anesthetic State. J Neurosci 2014; 34:16385–96

^11.Vazey EM, Aston-Jones G: Designer receptor manipulations reveal a role of the locus coeruleus noradrenergic system in isoflurane general anesthesia. Proc Natl Acad Sci U S A 2014; 111:3859–64

^12.Zhou X, Wang Y, Zhang C, Wang M, Zhang M, Yu L, Yan M: The Role of Dopaminergic VTA Neurons in General Anesthesia. PLoS ONE 2015; 10:e0138187

^13.Taylor NE, Van Dort CJ, Kenny JD, Pei J, Guidera JA, Vlasov KY, Lee JT, Boyden ES, Brown EN, Solt K: Optogenetic activation of dopamine neurons in the ventral tegmental area induces reanimation from general anesthesia. Proceedings of the National Academy of Sciences of the United States of America 2016; 113:12826–31

^14.Muindi F, Kenny JD, Taylor NE, Solt K, Wilson MA, Brown EN, Van Dort CJ: Electrical stimulation of the parabrachial nucleus induces reanimation from isoflurane general anesthesia. Behav Brain Res 2016; 306:20–5

^15.Fu B, Yu T, Yuan J, Gong X, Zhang M: Noradrenergic transmission in the central medial thalamic nucleus modulates the electroencephalographic activity and emergence from propofol anesthesia in rats. J Neurochem 2017; 140:862–73

^16.Du W, Zhang R, Li J, Zhang B, Peng X, Cao S, Yuan J, Yuan C, Yu T, Du J: The Locus Coeruleus Modulates Intravenous General Anesthesia of Zebrafish via a Cooperative Mechanism. Cell Reports 2018; 24:3146–3155.e3

^17.Yin L, Li L, Deng J, Wang D, Guo Y, Zhang X, Li H, Zhao S, Zhong H, Dong H: Optogenetic/Chemogenetic Activation of GABAergic Neurons in the Ventral Tegmental Area Facilitates General Anesthesia via Projections to the Lateral Hypothalamus in Mice. Front Neural Circuits 2019; 13:73

^18.Wang T-X, Xiong B, Xu W, Wei H-H, Qu W-M, Hong Z-Y, Huang Z-L: Activation of Parabrachial Nucleus Glutamatergic Neurons Accelerates Reanimation from Sevoflurane Anesthesia in Mice. Anesthesiology 2019; 130:106–18

^19.Zhang Y, Fu B, Liu C, Yu S, Luo T, Zhang L, Zhou W, Yu T: Activation of noradrenergic terminals in the reticular thalamus delays arousal from propofol anesthesia in mice. FASEB J 2019; 33:7252–60

^20.Torturo CL, Zhou Z-Y, Ryan TA, Hemmings HC: Isoflurane Inhibits Dopaminergic Synaptic Vesicle Exocytosis Coupled to CaV2.1 and CaV2.2 in Rat Midbrain Neurons. eNeuro 2019; 6

^21.Li J, Li H, Wang D, Guo Y, Zhang X, Ran M, Yang C, Yang Q, Dong H: Orexin activated emergence from isoflurane anaesthesia involves excitation of ventral tegmental area dopaminergic neurones in rats. Br J Anaesth 2019; 123:497–505

^22.Zhang X, Baer AG, Price JM, Jones PC, Garcia BJ, Romero J, Cliff AM, Mi W, Brown JB, Jacobson DA, Lydic R, Baghdoyan HA: Neurotransmitter networks in mouse prefrontal cortex are reconfigured by isoflurane anesthesia. J Neurophysiol 2020; 123:2285–96

^23.Luo T-Y, Cai S, Qin Z-X, Yang S-C, Shu Y, Liu C-X, Zhang Y, Zhang L, Zhou L, Yu T, Yu S-Y: Basal Forebrain Cholinergic Activity Modulates Isoflurane and Propofol Anesthesia. Front Neurosci 2020; 14:559077

^24.Gretenkord S, Olthof BMJ, Stylianou M, Rees A, Gartside SE, LeBeau FEN: Electrical stimulation of the ventral tegmental area evokes sleep-like state transitions under urethane anaesthesia in the rat medial prefrontal cortex via dopamine D1 -like receptors. Eur J Neurosci 2020; 52:2915–30

^25.Ao Y, Yang B, Zhang C, Li S, Xu H: Application of quinpirole in the paraventricular thalamus facilitates emergence from isoflurane anesthesia in mice. Brain Behav 2021; 11:e01903

^26.Zhang Y, Gui H, Hu L, Li C, Zhang J, Liang X: Dopamine D1 receptor in the NAc shell is involved in delayed emergence from isoflurane anesthesia in aged mice. Brain Behav 2021; 11:e01913

^27.Bao W-W, Xu W, Pan G-J, Wang T-X, Han Y, Qu W-M, Li W-X, Huang Z-L: Nucleus accumbens neurons expressing dopamine D1 receptors modulate states of consciousness in sevoflurane anesthesia. Current Biology 2021; 31:1893–1902.e5

^28.Franks NP: General anaesthesia: from molecular targets to neuronal pathways of sleep and arousal. Nat Rev Neurosci 2008; 9:370–86

^29.Brown EN, Lydic R, Schiff ND: General anesthesia, sleep, and coma. N Engl J Med 2010; 363:2638–50

^30.Tarnal V, Vlisides PE, Mashour GA: The Neurobiology of Anesthetic Emergence. J Neurosurg Anesthesiol 2016; 28:250–5

^31.Kelz MB, García PS, Mashour GA, Solt K: Escape From Oblivion: Neural Mechanisms of Emergence From General Anesthesia. Anesth Analg 2019; 128:726–36

II. Brain Reward Circuitry

The brain reward circuitry, also known as the mesolimbic dopamine system or mesolimbic circuitry, is composed of interconnected subcortical and cortical brain regions. This circuitry is evolutionarily conserved across mammals to mediate reinforcing behaviors important for survival, like sex^42,43 and food consumption^44–54. The same brain regions also modulate sleep/wake transitions and states of arousal essential for executing these reward-related behaviors^55–59. Dysregulation of reward seeking is concomitant with dysregulated arousal⁶⁰. For example, disordered sleep is an important feature of illnesses characterized by anhedonia and dysregulated reward circuit functioning like depression, addiction, schizophrenia and Parkinson’s disease^61,62.

Dopamine neurons in the ventral tegmental area (VTA) project to the nucleus accumbens (NAc) forming a key projection in the mesolimbic dopamine system, the circuitry that guides reward-related behaviors and promotes arousal^32,63–72 (Figure 2). The NAc is a central processing hub in ventral striatum that integrates inputs from the VTA with inputs from myriad brain regions in the reward circuitry which has been shown to guide a variety of behavioral responses and emotional states^{32,36,73–79}. Distinct neurochemical markers in the striatum divide its neurons into either the direct (“go”) or indirect (“no-go”) pathways³². These markers include the dopamine receptors, which are G-protein coupled receptors classified as either D1 (Gs-coupled) that signal through the direct pathway, or D2 (Gi-coupled) receptors that signal through the indirect pathway. The direct pathway, named for its direct projection to the VTA in the midbrain, expresses the neuropeptides dynorphin and substance P, while indirect pathway peptide expression includes enkephalin and adenosine A2A receptors³². This region is very heterogenous in anatomical and functional properties, is enriched in numerous neuropeptides and modulators include acetylcholine, among others^{32,44,71,73,74,80}. The dorsal striatum expresses similar dopamine receptor subtypes as ventral striatum, but projects to the substantia nigra, instead of the VTA, to guide motor responses to stimuli. In reality, the canonical role of dorsal striatum as confined to pure sensorimotor processing and the role of ventral striatum as confined to emotional processing is less explicitly segregated than was once believed (for review see⁸¹).

Figure 2.

Left: Systems neuroscience toolbox for dissecting mesolimbic circuitry with numbered boxes highlighting key methodologies for circuit-level and whole brain investigations. Right: Highly simplified schematic of mesolimbic reward circuitry with relevant glutamate (green), dopamine (purple), GABA (blue), and other neuropeptide (dashed) projections shown. Abbreviations: PFC-prefrontal cortex, HIP-hippocampus, DStr-dorsal striatum, NAc-nucleus accumbens, AMY-amygdala, LH-lateral hypothalamus, TH-thalamus, VTA-ventral tegmental area, DR-dorsal raphe, LDTg-lateral dorsal tegmentum, PBN-parabrachial nucleus, LC-locus coeruleus. Serotonergic projects from DR, noradrenergic projections from LC, and orexinergic projections from LH (not shown) are also important for emergence. Below inset shows a magnified view of VTA to NAc projections with up arrows indicating the increased activity of dopaminergic projects and D1-receptor activation in emergence. The role of other VTA to NAc projections in emergence is unclear. Figure created with BioRender.com

GABAergic medium spiny neurons are the principal neurons in the NAc, comprising over 95% of the neuronal population, and form local inhibitory synapses between medium spiny neurons as well as long range GABA projections to other brain regions in mesolimbic reward circuitry (Figure 2). Approximately 3% of NAc neurons are cholinergic interneurons releasing acetylcholine, while less than 2% are inhibitory interneurons releasing GABA with either somatostatin or parvalbumin^32,82. As a result, the inhibitory GABA receptors are ubiquitously expressed in NAc in addition to neuronal expression of dopamine receptors, mu opioid receptors, glutamate receptors, enkephalin, dynorphin, and other neuropeptide signaling substrates. The mechanism of action of propofol and volatile anesthetics, like sevoflurane and isoflurane, is known to involve direct activation of GABA-A receptors (for review see ref⁸³). Binding of propofol or volatile anesthetic to GABA receptors presumably may occur at both the medium spiny neuron and interneuron, thus modulating dopaminergic signaling within the NAc microcircuit in addition to affecting long range GABAergic projections in mesolimbic circuitry. Dopamine signaling also changes calcium currents and NMDA-induced currents studied in striatal slices^84–87. Changes in dopamine release, dopamine D1 receptor activation and transcriptional activation of deltaFosB in NAc are shown to be necessary for the behavioral and abuse liability properties of propofol administration^88–91. Further research is needed to define GABA, glutamate and dopaminergic interactions during behavioral emergence.

VTA dopamine neuron firing patterns signal errors in reward prediction, help to guide and modify behavior, and reinforce the motivation to seek rewards^{44–48,92–95}. VTA dopamine neurons respond in two different modes: single spike tonic firing to maintain dopamine tone^96–98 and phasic burst firing that is thought to signal an unexpected reward or salient event^46,99–105. Dopamine neuron bursting activity increases during transitions from sleep to wakefulness^55,56. Increased burst firing from VTA neurons results in dopamine release to downstream interconnected brain regions, including the NAc, prefrontal cortex (PFC), hypothalamus, and amygdala^{68,72,106,107} (Figure 2).

Classically attributed to orchestrating dopamine signaling, VTA circuitry is modulated by numerous other neuropeptides, as well as local and long-range GABAergic, glutamatergic, serotonergic, and cholinergic projections^72,108,109. For example, VTA GABA neurons send long range projections to synapse directly on NAc cholinergic neurons, forming one specialized circuit important for reward reinforcement⁷⁴ and associative learning¹¹⁰. VTA GABA neurons are also engaged in sleep arousal and modulated by anesthesia^111–113. The role of VTA GABA to NAc cholinergic neuron projections in emergence is unclear. The VTA connects directly with the thalamus, basal forebrain, orexinergic neurons in the hypothalamus (HYP), noradrenergic neurons in the locus coeruleus (LC) and serotonergic neurons in the dorsal raphe (DR), that are each individually important for mediating arousal and differentially affected by anesthesia^4,114,115.

Importantly, the state of general anesthesia, a drug-induced reversible coma, is distinct from natural sleep (for comprehensive reviews on this topic see ref^116,117). While insight into brain reward circuitry is gained from studies of sleep arousal, the same mechanisms should not be expected to correlate directly with emergence from anesthesia. This important caveat must be considered when comparing studies of sleep arousal and anesthesia as reviewed below.

III. Human Studies of Reward Circuitry in Emergence

Pharmacological, neuroimaging and genetic manipulations support the role of monoaminergic circuits in both arousal from sleep and emergence from general anesthesia. Dopamine and norepinephrine are important neurotransmitter mediators of arousal, as evidenced by impaired arousal seen in mice missing the dopamine beta hydroxylase^27,118,119 and dopamine transporter¹²⁰ genes. In humans, single nucleotide gene polymorphisms affecting the dopamine transporter and dopamine D2-receptor genes are associated with variations in self-reported sleep duration¹²¹. Treatment with a tyrosine hydroxylase inhibitor, with the end result of decreasing dopaminergic tone, increases sleepiness in studies of healthy adults^122,123. Dopamine D2 receptor levels also decrease specifically in the human ventral striatum after sleep deprivation, as assayed by PET imaging¹²⁴.

In contrast, stimulants that increase dopaminergic and catecholaminergic tone have strong effects on arousal. Dopamine-enhancing medications heighten arousal and accelerate emergence, as shown by studies of D1 receptor agonist treatment^125,126. In contrast, dopamine antagonism with droperidol slows emergence by deepening sevoflurane anesthesia¹²⁷. Together, these studies indicate a role for dopaminergic tone in promoting arousal, with a specific role for activation of dopamine receptors in ventral striatum.

Human neuroimaging under anesthesia consistently demonstrates thalamic deactivation and disruption of thalamocortical connectivity in states of general anesthesia^128–131, along with deactivation of the basal forebrain and basal ganglia¹³², important components of the brain reward circuitry. Studies using functional magnetic resonance imaging (fMRI) infer changes in brain activity by correlating changes in cerebral blood flow. However, they are not able to directly measure neuronal activity, an important limitation when interpreting results of fMRI studies. Invasive electrocorticography can be used to obtain direct recordings from the cortex of patients undergoing neurosurgery for intractable epilepsy. These studies demonstrate thalamocortical suppression with induction of general anesthesia, while recovery from anesthesia reflects a progressive increase in cortical activity, decrease in reticulothalamic activity and return of tonic activity in the thalamus¹³³. General anesthesia also inhibits auditory processing in higher-order auditory association areas while maintaining local field potential neuronal activity in the primary auditory cortex, suggesting anesthesia may selectively affect higher order signaling to disrupt cortical circuits¹³⁴.

Anesthesia research in both rodent models and the human clinical population utilize EEG changes and perioperative EEG monitoring as a tool for monitoring depth of anesthesia. General anesthetics are well known to produce distinct EEG patterns, with a shared pattern of increased delta oscillations (for a recent review see ref¹¹⁷). Interestingly, a recent study links increased delta oscillations to dopamine depletion and loss of D2 receptor activation in mouse striatum independent of anesthesia exposure¹³⁷. Studies of D1 dopamine receptor activation during emergence from isoflurane anesthesia in mice show reduced delta and increased gamma oscillations, accelerating emergence^125,138. EEG changes do not always reliably correlate with behavioral arousal. For example, optogenetic stimulation of VTA dopamine neurons promoted behavioral arousal with minimal change in EEG¹³⁹.

EEG studies of the brain are useful as a noninvasive and easily translatable method but have limitations for interpretation. A recent multicenter study enrolled 60 healthy volunteers to evaluate frontal-parietal EEG dynamics in recovery from anesthesia, independent of surgery¹⁴⁰. Results from this study support a model of early return of prefrontal cortical dynamics and executive function. However, EEG dynamics do not predict cognitive recovery after anesthesia. Burst suppression in the EEG is considered to reflect a very deep state of anesthesia that may be desirable to avoid¹³⁶. In a study of 27 healthy human volunteers, EEG burst suppression does not change the time to emergence nor affect the degree of cognitive impairment after isoflurane exposure, using a computational model to predict time of emergence¹⁴¹. The ENGAGES randomized clinical trial also finds that EEG-guided administration of general anesthesia does not reduce the incidence of postoperative delirium, compared to usual care, in adults aged 60 years and older¹⁴². Thus, while EEG is a useful readout of brain oscillatory arousal states, it is only one tool for evaluating clinical effects of emergence in the perioperative setting. A multidisciplinary systems neuroscience approach, additionally informed by preclinical research, is needed for a holistic view of anesthesia emergence and post-anesthetic cognitive sequelae.

IV. Preclinical Studies of Reward Circuitry in Emergence

Current basic neuroscience understanding of arousal is derived primarily from rodent studies of sleep/wake states and general anesthesia^{59,111,113,114,139,143–147}. Sleep studies in rodents consistently support a central role for dopaminergic signaling, and specifically VTA neuron activity in arousal^{59,27,111,113}. Emergence from anesthesia, defined as arousal and return of awareness, is assayed behaviorally in the rodent by restoration of the righting reflex response, a reflex that develops shortly after birth to maintain the prone position. In these studies, the mouse or rat is turned on its back while anesthetized and upon emergence, the animal will right itself to having its paws on the ground^148–152.

Preclinical research supports an important role for dopaminergic signaling in mediating emergence from anesthesia. Isoflurane anesthesia inhibits synaptic vesicle exocytosis from dopamine neurons, in cultured rat VTA dopamine neurons^153,154. Subanesthetic propofol exposure causes an increase in spontaneous VTA dopamine neuron firing recorded from rat brain slices, and propofol potentiates evoked postsynaptic excitatory synaptic currents recorded downstream in the NAc¹⁵⁵. The stimulant drug amphetamine also causes presynaptic release of dopamine and inhibits dopamine reuptake in the striatum in recordings from striatal brain slices¹⁵⁶. Systemic methylphenidate and amphetamine administration, which increase catecholaminergic tone by inhibiting norepinephrine and dopamine reuptake, speed emergence from both isoflurane and propofol general anesthesia in rodents, as measured behaviorally by restoration of the righting reflex^157–160. Similarly, other reports show that intravenous caffeine administration also accelerates emergence from isoflurane general anesthesia in both mice and humans^161–164. These indirect pharmacological studies support a general role for increased dopaminergic tone influencing emergence.

Studies directly manipulating the VTA under general anesthesia demonstrate the sufficiency of VTA neuron activity in promoting emergence. Direct stimulation of VTA neurons using an electrode inserted above the VTA in the rat results in faster emergence from both isoflurane and propofol anesthesia¹⁶⁵. Further, cell-type-specific stimulation of only dopamine neurons in the VTA using optogenetics in transgenic mice promotes emergence from isoflurane anesthesia¹³⁹. Together, these findings support a working model of reduced VTA dopamine neuron activity under general anesthesia, with emergence characterized by a resurgence of dopamine activity as brought about by direct neuronal stimulation, or stimulant drug administration.

Multiple reports indicate a critical role for the engagement and activation of dopamine receptors during emergence. Early studies of phenobarbital anesthesia demonstrate a role for activation of both dopamine D1 and dopamine D2 receptors in promoting emergence using systemic receptor agonist treatment in rats^166,167 and rabbits¹⁶⁸. Dopamine D1 receptor agonists promote emergence from isoflurane and propofol anesthesia¹⁶⁹. A2A receptor agonist administration, which also activates medium spiny neurons expressing D2-receptors, modulates depth of propofol anesthesia and activates the nucleus accumbens (NAc) in mice as measured by increased c-fos expression¹⁷⁰. However, these studies all use systemic administration of dopamine receptor agonists, so the neural circuit mechanism and sites of their action in the brain are unknown. Direct microinjection of D1-receptor agonist or antagonist into the NAc supports a bidirectional regulation of time to emergence with dopamine receptor activation specifically within the NAc¹³⁸. In addition, selective chemogenetic activation of NAc D1-receptor-expressing neurons accelerates emergence and delays induction with sevoflurane¹⁷¹.

Together, these findings support a working model of the anesthetized state as marked by a reduction of dopaminergic tone, with emergence from anesthesia promoted by increased VTA dopamine neuron activity which subsequently causes activation of downstream D1-type dopamine receptors within the NAc (Figure 2). It is unclear whether the increase in VTA activity is driven by increases in tonic dopamine neuron firing or phasic discharge during emergence. While VTA dopamine appears to be necessary for emergence, it is unclear if VTA stimulation alone is sufficient to drive emergence. Optogenetic studies of VTA dopamine neurons in emergence used repeated stimulation over 30 minutes to increase probability of righting¹³⁹. Additional mechanisms may be engaged within VTA circuitry with repeated stimulation over time. VTA dopamine neurons project to numerous target brain regions to form the mesolimbic reward circuit, as discussed previously. Outside of the VTA, manipulations of the parabrachial nucleus, which directly projects to VTA, the locus coeruleus and the thalamus also promote emergence from general anesthesia^172–176. It is possible that additional dopaminergic pathways are further engaged in these studies. In addition to dopamine, the VTA contains numerous neuropeptide-containing neurons⁴⁴, as well as GABAergic and glutamatergic cells that can send long-range projections. The effects of increased VTA neuron activity during emergence on heterogeneous downstream circuitry remains to be fully described (Figure 2 inset). Many investigations of brain circuitry also largely ignore the contribution of non-neuronal cell types, while there is a new study indicating an important role for astrocytes in emergence¹⁷⁷. Future research must be aimed at comprehensively evaluating all cell types in target brain circuit regions during emergence to form a complete mechanistic understanding and provide new therapeutic targets.

V. Preclinical Neuroscience Methods for Neural Circuit Investigation

Modern neuroscience tools enable a detailed dissection of neural circuits in the awake-behaving animal with high temporal and spatial resolution using optical manipulation and behavioral modeling. Neural circuit investigation is strengthened by an interrogation at multiple levels of analysis: from molecular/cellular to systems to behavioral. Beginning with the revolutionary introduction of optogenetics^178–180, the optical tools available for interrogating brain circuit connectivity now extend from light-activated ion channels to optically active G-protein coupled receptors, like parapinopsin¹⁸¹ and the optogenetically activated mu-opioid receptor^69,146 and beta-2 adrenergic receptor¹⁴⁷. Genetically encoded fluorescent sensors of neuropeptide and neurotransmitter release such as the dLight sensor that detects dopamine release and the GRAB_NE sensor that detects norepinephrine release¹⁸⁵, among many others¹⁸⁶, are used together with calcium imaging to better elucidate the dynamics of neural circuit action during behavior. Optofluidic devices also enable the wireless light-evoked delivery of drugs into the brain for pharmacologic studies with high temporal and regional specificity^187–190. Coupled to transcriptomic manipulations at the single-cell level, the investigation of novel receptor-mediated signaling mechanisms in specific brain circuits is possible with exquisite detail¹⁹¹.

Advances in microscopy now allow for imaging across the whole brain at single cell resolution after brain clearing using light sheet microscopy¹⁹². Some studies of general anesthesia are beginning to take advantage of the whole-brain approach to investigating neural circuits, like the reticular activating system¹⁹³. In addition, calcium dynamics can be imaged at the individual neuronal level within a specified circuit using in vivo two-photon microscopy in a head-fixed animal, or in vivo one-photon imaging after implanting a miniature microscope (GRIN lens) in freely moving mice^194–197. Calcium imaging can then be paired with optogenetic studies to dissect the effects of circuit activation or inhibition on neuronal activity. These newer imaging modalities provide high single cell and spatial resolution, enabling detailed cellular-level preclinical investigations, compared to approaches with poorer spatial resolution like fMRI and PET^198,199.

Multi-region, high-density recordings of neuronal activity using advanced physiology methods like implanted silicone probes, called Neuropixels, can be used to decipher circuit activity during emergence¹⁸⁵. Neuropixels do not utilize genetically encoded sensors and thus lack cell-type specificity as well as tracking of the same neurons across long term temporal domains²⁰⁰. However, Neuropixels can be paired with opto-tagging, in which a neuron is optogenetically activated to determine its identity, and the high-density nature of Neuropixels recordings affords a more system-wide view of a given series of brain regions.

Neuronal recordings and optogenetic manipulation can then be paired with computational neuroethology for closed loop stimulation or analytical studies²⁰¹. Closed loop deep brain stimulation therapy improved depression symptoms in one individual with major depression²⁰², and similar approaches could be adopted for modifying emergence. Open source toolkits for high-throughput behavioral analysis using machine learning approaches, like DeepLabCut²⁰³, SimBA²⁰⁴, or DeepSqueak²⁰⁵, can be applied to studying emergence from anesthesia. Machine learning approaches are useful for identifying previously unknown behavioral repertoires within simple behaviors, such as grooming²⁰⁶ and subtle pharmacologic effects on behavior in rodents²⁰⁷. Behavioral models of emergence such as spontaneous restoration of the righting reflex (RORR) are currently analyzed as a binary output interpreted by visual manual scoring (either positive when the rodent is aroused and upright, or negative when the rodent is lying unconscious on its back). The binary RORR model as currently analyzed is suggested to variably correlate with cortical signatures of arousal assayed by EEG and local field potential analysis²⁰⁸. Even a simple experimental model like RORR presents an array of behavioral features (e.g. whisker movement, tail curling, side rolling, increased chest movement, then righting). In-depth behavioral classification using pose estimation and machine learning classification methods helps to remove experimenter subjectivity and provides an automated analysis pipeline to facilitate data comparisons across experiments, investigators, and research centers. There are several studies applying machine learning approaches to assessing depth of anesthesia in human subjects^209,210. These approaches may yield further mechanistic insights when also translated to the preclinical model for concurrent neural circuit interrogations.

VI. Conclusion: Bridging the Preclinical: Clinical Divide

In order to develop a better understanding of anesthetic emergence and work towards new clinical strategies to promote smooth emergence, existing studies of network state changes in humans might be used as working templates for further mechanistic dissection of brain arousal circuitry in preclinical animal models. By layering relevant clinically translational endpoints onto preclinical models, such as EEG analysis and fMRI imaging to identify shared areas of activation, a comprehensive view of brain circuitry during emergence may develop. The preclinical model can then be used to develop a more granular mechanistic analysis of neuronal changes, taking advantage of high-resolution single cell approaches in the context of whole brain dynamics.

Emergence is likely a convergence of the activity of multiple distributed transmitters, receptors and circuits, for example, a unification of orexinergic, dopaminergic and noradrenergic systems⁴. Further investigations are needed to understand the effects of different anesthetic conditions, like ketamine as compared to sevoflurane or propofol, on neural circuits. In order to study emergence, the aggregate brain network must then be examined using network-wide manipulations. The patterns of neuronal circuit activity that regulate emergence can be directly controlled and modified by utilizing closed loop approaches, as discussed above. Additional tools to elucidate behaviorally activated brain-wide circuits include utilizing transgenic mice, like Fos-CreER^{T2 211}, together with viral approaches to access activity-regulated neuronal ensembles^197,212,213 (for review see ref²¹⁴). At this level of whole brain analysis and neuronal activity, it is then possible to generate neuronal decoders for predicting and modifying emergence. Overall, an improved understanding of brain circuitry changes during emergence will help to facilitate predictable transitions from anesthetized to awake state that will in turn improve patient safety and satisfaction with anesthesia care.

While it is not easy to reconcile preclinical and clinical approaches, innovative new tools exist for studying brain circuitry that can be applied strategically to heighten the translational value of preclinical anesthesia investigations. We must build dialogue and collaborative studies between basic neuroscientists and clinician researchers, appreciate the limitations of each scientific approach, and compare parallel findings from the preclinical and clinical literature as guides for future shared investigation.

Summary Statement:

This review explores the integration of advanced systems neuroscience approaches into translational anesthesia research to elucidate the important role of mesolimbic brain reward circuitry in emergence from general anesthesia.