General Anesthesia Causes Telltale Brain Activity Patterns

Starting in the 1990s, researchers developed algorithms to consolidate the signals recorded from several EEG electrodes into a single number that provided a simplified measurement of arousal level. More recently, direct observation of the raw EEG signals and their breakdown in time by frequencies, the spectrogram, is gaining traction for monitoring patients during general anesthesia. Learning to interpret the raw brain activity and its spectrogram, rather than relying on a single-number summary, has allowed anesthesiologists to assess how different anesthetics affect brain activity and produce the anesthetic state.¹

By tracking brain activity during general anesthesia, researchers are also uncovering a wealth of new information that helps them understand the biological basics of how brain function is altered in an anesthetized state. In addition, general anesthesia has provided new options to treat a range of ailments, from sleep problems to depression.

Ether follies

Humans have practiced surgical procedures for thousands of years, and the search for ways to minimize pain and discomfort during invasive interventions is probably just as old. Wine and opium are among the first substances known to have been tried. Opium is a potent analgesic and mild sedative, and the ethyl alcohol in wine is a sedative, but neither of these drugs succeeds in making patients unaware of the trauma their bodies undergo during surgery.

In the first half of the 19th century, dentists stumbled upon two promising leads: nitrous oxide, which soon after its discovery became widely used in the US and Europe to perform tooth extractions, and chloroform, which was used for both veterinary and human surgeries for a few decades before it fell out of favor due to safety concerns. In the 1840s, Boston dentist William Morton was looking for ways to perform pain-free dental procedures and considered using nitrous oxide. But, 

Charles Jackson, a chemist at Harvard Medical School, advised him to try another option: ether.

At that time, it was common in academic and other social circles to hold parties, called “ether follies,” where people would inhale ether for its exhilarating properties. Jackson had seen a man sustain a considerable leg injury during one such escapade. The man, who had been high on ether, showed no signs of pain. Morton took Jackson’s advice and proceeded to experiment with ether on himself and his dog, and subsequently performed several dental procedures on his patients after administering the drug to them.

Anesthetic-induced oscillations dramatically alter when neurons can spike, and impede commu­nication between brain regions that play a role in consciousness.

Morton contacted Harvard Medical School surgeon Henry Bigelow, and together they organized what would become known as the first public demonstration of surgery performed under general anesthesia. On October 16, 1846, in the operating theater now known as the Ether Dome at Massachusetts General Hospital, John Collins Warren, the founding dean of Harvard Medical School and the hospital’s chief surgeon, removed a tumor from the neck of patient Edward Gilbert Abbott, while Morton held a glass flask containing an ether-soaked sponge that spouted ether vapor through a glass tube that was attached to Abbott’s nose. Several prominent surgeons and physicians watched from the viewing area of the theater.

Warren performed the surgery with the patient showing minimal signs of pain;² at the end of the procedure, Warren famously declared: “Gentlemen, this is no humbug.” Afterward Abbott did state that he had experienced sensations, though not pain, during surgery. The following day, flaws in ether administration were corrected, and a second tumor removal patient declared that she had felt and known nothing. Surgeries using ether-induced anesthesia were soon performed at nearby hospitals, and within a couple of months it began to change medical practice the world over.

The modern definition of general anesthesia requires that five endpoints are achieved. (See Box above.) Ether provides all of these endpoints to some degree. Most modern inhaled anesthetics, such as isoflurane, desflurane, and sevoflurane, are chemical derivatives of ether but are more potent, less flammable, and are delivered using modern vaporizers and techniques. Improvements to the hypodermic needle achieved in the second half of the 19th century made possible the development of intravenous anesthesia, and physicians began combining anesthetic drugs with opioids to more effectively achieve analgesia. Later, muscle relaxants were added to ensure immobility.

The modern practice of general anesthesia, known as balanced anesthesia, uses combinations of drugs with the goal of distancing the patient from the trauma the body is undergoing while minimizing side effects. Recently, researchers have detailed the mechanism underlying the action of modern anesthetics, identifying links between the neural receptors on which the drugs act and patterns of overall brain activity that are linked to changes in neuronal firing. These connections allow anesthesiologists to track brain activity patterns during general anesthesia to improve patient experiences and outcomes, as well as to learn more about how the anesthetized brain functions.

How is general anesthesia achieved?

One of the most conspicuous features of general anesthesia is the profound state of unconsciousness that it produces. Up until the 1980s, the prevailing hypothesis on how the unconscious state was achieved was influenced by the observation that anesthetic potency directly correlated with solubility in olive oil, suggesting a hydrophobic site of action such as the lipid bilayer membranes of neurons. Researchers speculated that the drugs disrupted normal membranes function and prevented the conduction of action potentials. This idea was known as the lipid hypothesis. Then, Nicholas Franks and William Lieb of Imperial College London showed that the true targets of anesthetic drugs were neuronal receptors embedded in the membrane.³

Neuronal receptors regulate the probability that neurons will fire action potentials, often by acting to control channels for specific ions to go into and out of nerve cells. Activating excitatory receptors increases a neuron’s firing potential, while activating inhibitory receptors decreases it. Hence, anesthetic drugs could, in principle, be grouped into two main classes: those that activate inhibitory receptors and those that inactivate excitatory receptors. (See illustration.)

Inhaled ether derivatives and intravenous propofol, the most widely used anesthetic drug, bind to the inhibitory GABA_A receptor. Under normal, physiological conditions, the receptor is activated by gamma-aminobutyric acid (GABA) released from inhibitory neurons, and it allows the flow of chloride ions into the cell, dropping the relative voltage of the neuron’s interior and thereby decreasing the probability of firing an action potential. Anesthetic drugs that target this receptor act as agonists to promote the influx of chloride ions, further suppressing the cell’s ability to fire.

Other anesthetics such as ketamine, which was synthesized in 1962, and nitrous oxide block the channel of the N-methyl D-aspartate (NMDA) glutamate receptor. Normally activated by the neurotransmitter glutamate released from excitatory neurons, the NMDA receptor allows the flow of potassium ions out of the cell and calcium and sodium ions in, increasing the relative voltage of the neuron’s interior and thereby increasing the probability of firing an action potential. Anesthetic drugs that target this receptor act as antagonists to block these ions fluxes, decreasing the ability of the cell to fire.

Anesthetics & Neuronal Receptors General anesthetics work by altering the activity of specific neurons in the brain. One main class of these drugs, which includes propofol and the ether-derivative sevoflurane, work primarily by increasing the activity of inhibitory GABAA receptors, while a second class that includes ketamine primarily blocks excitatory NMDA receptors.  Propofol and sevoflurane The GABAA receptor is a channel that allows chloride ions to flow into the neuron, decreasing the voltage within the cell relative to the extracellular space. Such hyper­polarization decreases the probability that the neuron will fire. Propofol and sevoflurane increase the chloride current going into the cell, making the inhibition more potent. © LUCY READING-IKKANDA Ketamine The NMDA receptor allows sodium and calcium ions to flow into the cell, while letting potassium ions out, increasing the voltage within the cell relative to the extra­cellular space and increasing the probability of neural firing. Ketamine blocks this receptor, decreasing its excitatory actions. See full infographic: WEB | PDF © LUCY READING-IKKANDA

Knowing the actions of anesthetic drugs on the receptors still does not fully explain how unconsciousness occurs, however. Both GABA and NMDA receptors are found on the excitatory and inhibitory neurons that make up neuronal circuits. The function of these circuits and their relationship to behavior can be understood within the framework of systems neuroscience: the changes in ion fluxes produced by anesthetic binding to receptors dramatically alters neuronal activity across the brain, eliciting highly structured oscillations. In humans, these oscillations are readily visible in the EEG readouts. They are of high amplitude and lie within well-defined frequency bands lower than those of unstructured, low-amplitude oscillations seen in the brain of a conscious person.

The observed waves depend critically on which receptors are bound and how the targeted regions are connected to other areas of the brain. The oscillations change systematically with anesthetic drug class, drug dose, and patient age. For example, alpha oscillations (8–12 Hz) produced by GABAergic anesthetics depend critically on excitatory and inhibitory connections between the thalamus and the cortex.⁴ The beta/gamma oscillations (15–50 Hz) produced by ketamine⁵ might depend on blocking the NMDA receptors of inhibitory and excitatory neurons in the cortex, while the slow oscillations produced by both GABA agonists^6,7 and NMDA antagonists might depend on inhibition of the brainstem and its projections to the thalamus and cortex. In elderly patients the oscillations have lower amplitudes across all frequency bands. These oscillations also dramatically alter when neurons can spike, and impede communication between brain regions that play a role in consciousness.

Oscillations in the Anesthetized Brain Anesthetics’ interactions with neural receptors alter how neurons work, and as a consequence, how different brain regions communicate. These alterations manifest as highly structured oscillations in brain activity that are associated with the dramatic behavioral changes characteristic of general anesthesia. See full infographic: WEB | PDF © LUCY READING-IKKANDA

The characteristics of the oscillations produced by anesthetics suggest that they are a significant part of the mechanism of anesthetic action, and explain how the brain state of a patient under general anesthesia can be reliably tracked using the EEG. However, monitoring brain activity has not been standard in anesthesiology practice. The early attempts to use EEG as an additional piece of information to monitor patients and inform the dose and rate of anesthetic delivery focused on developing an index that would provide a single readout of anesthetic state. However, EEG activity observed during general anesthesia is different across people of different ages, and these indices can be misleading when used in children or elderly people. Furthermore, EEGs captured under general anesthesia differ across drugs, and aggregated indices cannot take these differences into consideration.

For those reasons, in the last few years anesthesiologists have begun to monitor EEG readouts of brain signals during procedures involving general anesthesia. The oscillation patterns for common anesthetics are identifiable to the trained eye, and the assessment of their frequencies can be performed in real time with computer aid, providing a more nuanced picture of a patient’s brain state. This has allowed clinicians to manage anesthetic drug dosing in a more nuanced way and to reduce the amount of anesthesia required to achieve the same anesthetic state.⁸ As we understand more about how anesthetics work, and gain more experience directly observing EEGs generated during anesthesia, the practice will continue to be improved.

Drug	Primary Receptor	Anesthetic-specific Oscillations	EEG readouts
Propofol	GABAA	Alpha (8–12 Hz) oscillations result from synchronization of neural activity in the cortex and thalamus.
Ketamine	NMDA	Beta/gamma (25–50 Hz) oscillations, perhaps due to an increased spiking rate of excitatory neurons in the cortex following ketamine-induced reduction of activity in nearby inhibitory neurons

General anesthesia as treatment

In the last several decades, research on brain activity patterns and general anesthesia has yielded insights not only into the effects of anesthetics themselves, but also into neural processes related to conditions in which brain oscillations are altered, such as aging⁹ and pathological conditions including autism.¹⁰ In addition, the anesthetics methylhexital and alfentanil have proven useful in stimulating seizure activity in the brains of epilepsy patients, helping neurosurgeons to precisely locate the problematic tissue to resect. Advances in this field also point to the possibility of using anesthesia as a treatment for a handful of brain-related conditions.

This concept is not entirely new. For example, during deep general anesthesia and coma, an EEG pattern known as burst-suppression is observed. This pattern consists of bursts of electrical activity alternating with flat periods of inactivity. Neurologists frequently use anesthetics to induce a medical coma in patients with intractable seizures or raised intracranial pressure to arrest the seizure activity or decrease brain swelling. The comatose state is maintained by observing EEGs and titrating the anesthetic infusion rate to maintain a specific number of bursts per minute. This procedure ordinarily requires a human to assess the burst-suppression rate and manually adjust the anesthetic dose, but our research suggests that full automation of the process is possible.¹¹

The brain state of a patient under general anesthesia can be reliably tracked using the EEG. 

Another area where insights into the neural mechanisms of anesthesia might improve treatment options is sleep. Sleep has two main stages: rapid eye movement (REM) and non-REM sleep. Non-REM sleep is a state of profound unconsciousness that scientists consider to be most important for achieving properly restful sleep. 

Non-REM sleep is characterized by two main oscillatory patterns in brain activity: sleep spindles (10–15 Hz) and slow oscillations. Most sleep aid medications do not produce oscillatory brain activity that closely resembles the activity observed during natural sleep. However, the anesthetic dexmedetomidine, which affects circuits in the brainstem that are involved in control of wakefulness,¹² EEG patterns that are similar to those that occur during non-REM sleep.¹³ Clinical trials are currently under way to test its efficacy as a sleep aid.

Anesthetics such as ketamine, xenon, and nitrous oxide have already been shown to have acute antidepressant effects. There is evidence that other anesthetics, such isoflurane and propofol, when dosed to the level of producing burst suppression indicative of a medical coma, have long-lasting antidepressant effects without the short-term cognitive impairment and amnesia associated with electroconvulsive therapy. Ketamine might exert its antidepressant effect by increasing the number of synaptic receptors and other synaptic signaling proteins, and even increasing the number of synapses in the brain.¹⁴

More than 170 years after its first public demonstration, general anesthesia allows millions of painless surgeries to be performed daily across the world and is still the bedrock on most surgical procedures are performed. At the same time, the study of the effects of anesthetics in brain function is opening many exciting opportunities for the development of novel anesthetic paradigms and for research on other questions in clinical neuroscience.

Emery N. Brown is the Warren M. Zapol Professor of Anaesthesia at Harvard Medical School, a practicing anesthesiologist at Massachusetts General Hospital, and the Edward Hood Professor of Medical Engineering and Computational Neuroscience at MIT. Francisco J. Flores is an instructor in Anaesthesia at Massachusetts General Hospital and Harvard Medical School.

References

1. P.L. Purdon et al., “Clinical electroencephalography for anesthesiologists: Part I: Background and basic signatures,” Anesthesiology, 123:937–60, 2015.
2. H.J. Bigelow, “Insensibility during surgical operations produced by inhalation,” Boston Med Surg J, 35:309–17, 1846.
3. N.P. Franks, W.R. Lieb, “Do general anaesthetics act by competitive binding to specific receptors?” Nature, 310:599–601, 1984.
4. F.J. Flores et al., “Thalamocortical synchronization during induction and emergence from propofol-induced unconsciousness,” PNAS, 114:E6660–68, 2017.
5. O. Akeju et al., “Electroencephalogram signatures of ketamine anesthesia-induced unconsciousness,” Clin Neurophysiol, 127:2414–22, 2016.
6. P.L. Purdon et al., “Electroencephalogram signatures of loss and recovery of consciousness from propofol,” PNAS, 110:E1142–51, 2013.
7. K.J. Pavone et al., “Nitrous oxide-induced slow and delta oscillations,” Clin Neurophysiol, 127:556–64, 2016.
8. E.N. Brown et al., “Multimodal general anesthesia: Theory and practice,” Anesth Analg, 127:1246–58, 2018.
9. P.L. Purdon et al., “The ageing brain: Age-dependent changes in the electroencephalogram during propofol and sevoflurane general anaesthesia,” Br J Anaesth, 115 (Suppl 1):i46–57, 2015.
10. E.C. Walsh et al., “Age-dependent changes in the propofol-induced electroencephalogram in children with autism spectrum disorder,” Front Syst Neurosci, 12:23, 2018.
11. M.M. Shanechi et al., “A brain-machine interface for control of medically-induced coma,” PLOS Comput Biol, 9:e1003284, 2013.
12. V. Breton-Provencher, M. Sur, “Active control of arousal by a locus coeruleus GABAergic circuit,” Nat Neurosci, 22:218–28, 2019.
13. O. Akeju et al., “Dexmedetomidine promotes biomimetic non-rapid eye movement stage 3 sleep in humans: A pilot study,” Clin Neurophysiol, 129:69–78, 2018.
14. N. Li et al., “mTOR-dependent synapse formation underlies the rapid antidepressant effects of NMDA antagonists,” Science, 329:959–64, 2010.