Clinical Electroencephalography for Anesthesiologists: Part I: Background and Basic Signatures

Abstract

The widely used electroencephalogram-based indices for depth-of-anesthesia monitoring assume that the same index value defines the same level of unconsciousness for all anesthetics. In contrast, we show that different anesthetics act at different molecular targets and neural circuits to produce distinct brain states that are readily visible in the electroencephalogram. We present a two-part review to educate anesthesiologists on use of the unprocessed electroencephalogram and its spectrogram to track the brain states of patients receiving anesthesia care. Here in part I, we review the biophysics of the electroencephalogram and the neurophysiology of the electroencephalogram signatures of three intravenous anesthetics: propofol, dexmedetomidine, and ketamine, and four inhaled anesthetics: sevoflurane, isoflurane, desflurane, and nitrous oxide. Later in part II, we discuss patient management using these electroencephalogram signatures. Use of these electroencephalogram signatures suggests a neurophysiologically based paradigm for brain state monitoring of patients receiving anesthesia care.

Conflict of interest statement

CONFLICTS OF INTERST

Masimo has signed an agreement with Massachusetts General Hospital to license the signal processing algorithms developed by Drs. Brown and Purdon for analysis of the electroencephalogram to track the brain states of patients receiving general anesthesia and sedation for incorporation into their brain function monitors.

Figures

Fig. 1

The neurophysiological origins of the electroencephalogram. A. Normal communication in the brain through neuronal spiking activity induces oscillatory extracellular electrical currents and potentials that are one of the ways information exchange is modulated and controlled in the central nervous system. B. The geometry of the neurons in the cortex favors the production of large extracellular currents and potentials. C. The electroencephalogram recorded on the scalp is a continuous measure of the electrical potentials produced in the cortex. D. Because the cortex (orange region) is highly interconnected with subcortical regions, such as the thalamus (yellow region), and the major arousal centers in the basal forebrain, hypothalamus, midbrain and pons, profound changes in neural activity in these areas can result in major changes in the scalp electroencephalogram. Panel A is reproduced with permission from Hughes and Crunelli, Neuroscientist, 2005. Panel B is reproduced with permission from Rampil, Anesthesiology, 1998.

Fig. 2

Unprocessed electroencephalogram signatures of propofol-induced sedation and unconsciousness. A. Awake eyes open electroencephalogram pattern. B. Paradoxical excitation. C. Alpha and beta oscillations commonly observed during propofol-induced sedation (Fig. 5). D. Slow-delta and alpha oscillations commonly seen during unconsciousness. E. Slow oscillations commonly observed during unconsciousness at induction with propofol (fig. 6) and sedation with dexmedetomidine (fig. 11) and with nitrous oxide (fig. 13). F. Burst suppression, a state of profound anesthetic-induced brain inactivation commonly occurring in elderly patients, anesthetic-induced coma and profound hypothermia (fig. 6B and 6D) G. Isoelectric electroencephalogram pattern commonly observed in anesthetic-induced coma and profound hypothermia. With the exception of the isoelectric state, the amplitudes of the electroencephalogram signatures of the anesthetized states are larger than the amplitudes of the electroencephalogram in the awake state by a factor of 5 to 20. All electroencephalogram recordings are from the same subject.

Fig. 3

Construction of the spectrogram. A. A ten-second electroencephalogram trace recorded under propofol-induced unconsciousness. B. The electroencephalogram trace in A filtered into its two principal oscillations: the blue curve, an alpha (8 to 12 Hz) oscillation and the green curve, a slow-delta (0.1 to 4 Hz) oscillation. C. The spectrum provides a decomposition of the electroencephalogram in A in to power by frequency for all of the frequencies in a specified range. The range here is 0.1 to 30 Hz. Power at a given frequency is defined in decibels as the 10 times the log base 10 of the squared amplitude (10 log10 (amplitude)2). The green horizontal line underscores the slow-delta frequency band and the blue horizontal line underscores the alpha frequency band used to compute the filtered signals in B. The median frequency, 3.4 Hz, (dashed vertical line) is the frequency that divides the power in the spectrum in half. The spectral edge frequency, 15.9 Hz (solid vertical line) is the frequency such that 95% of the power in the spectrum lies below this value. D. The three-dimensional spectrogram (compressed spectral array) displays the successive spectra computed on a 32-minute electroencephalogram recording from a patient anesthetized with propofol. Each spectrum is computed on a 3-second interval and adjacent spectra have 0.5 seconds of overlap. The black curve at minute 24 is the spectrum in C. D. The spectrogram in C plotted in two-dimensions (density spectral array). The black vertical curve is the spectrum in C. The lower white curve is the time course of the median frequency and the upper white curve is the time course of the spectral edge frequency. Panels A–E were adapted from Purdon and Brown, Clinical Electroencephalography for the Anesthesiologist (2014), with permission from the Partners Healthcare Office of Continuing Professional Development.

Fig. 4

Neurophysiological mechanisms of propofol’s actions in the brain. Propofol enhances GABAA-mediated inhibition in the cortex, thalamus and brainstem. Shown are three major sites of action: post-synaptic connections between inhibitory interneurons and excitatory pyramidal neurons in the cortex; the GABAergic neurons in the thalamic reticular nucleus (TRN) of the thalamus; and post-synaptic connections between GABAergic projections from the pre-optic area (POA) of the hypothalamulus, and the monoaminergic nuclei which are the tuberomammillary nucleus (TMN) that releases histamine (His), the locus ceruleus (LC), that releases norepinephrine (NE), the dorsal raphé (DR) that releases serotonin (5HT); and the cholinergic nuclei which are the basal forebrain (BF), pedunculopontine tegmental (PPT) nucleus and the lateral dorsal tegmental (LDT) nucleus that release acetylcholine (Ach). This figure is reproduced with permission from Brown, Purdon and Van Dort, Annual Review of Neuroscience, 2011.

Fig. 5

Spectrogram and the time-domain signature of propofol-induced sedation. A. Spectrogram shows slow-delta oscillations (0.1 to 4 Hz) and alpha-beta (8 to 22 Hz) oscillations in a volunteer subject receiving a propofol infusion to achieve and maintain a target effect-site concentration of 2 mcg/ml, starting at time zero. The subject was responding correctly to the verbal, but not to click train auditory stimuli delivered every 4-seconds for the entire 16 minutes suggesting that she was becoming sedated. The lower and upper white curves are the median and the spectral edge frequencies respectively. B. Ten-second electroencephalogram trace recorded at minute 6 of the spectrogram in A.

Fig. 6

Spectrogram and time-domain electroencephalogram signatures of two patients receiving propofol for induction and maintenance of unconsciousness. A. High slow-delta power following the 200mg propofol bolus at minute 3 is evident between minutes 3 and 5. The electroencephalogram transitions to robust slow-delta and alpha oscillations maintained by a propofol infusion at 100mcg/kg/min. The lower and upper white curves are the median and the spectral edge frequencies respectively. B. Following bolus doses of propofol the patient’s electroencephalogram transitions between 3 different states: slow oscillations (minutes 5 to 8) following the 100 mg propofol bolus at minute 3; burst suppression (minutes 8 to 15) following two additional 50 mg propofol boluses; and slow-delta and alpha oscillations from minutes 15 to 25. Beginning at minute 24 the alpha band power decreases and broadens to the beta band. The slow-delta oscillation power decreases after minute 24. The dissipation of the slow-delta and alpha oscillation power as the patient emerges gives the appearance of a zipper opening. C. Ten-second electroencephalogram traces recorded at minute 5.5 (slow-delta oscillations) and minute 24 (slow-delta and alpha oscillations) of the spectrogram in A. D. Ten-second electroencephalogram traces showing slow oscillations at minute 7.1, burst suppression at minute 11.5 and slow-delta and alpha oscillations at minute 17 for the spectrogram in B. Panels A–D were adapted from Purdon and Brown, Clinical Electroencephalography for the Anesthesiologist (2014), with permission from the Partners Healthcare Office of Continuing Professional Development.

Fig. 7

Spatio-temporal characterization of electroencephalogram alpha and slow oscillations observed during induction of and recovery from propofol-induced unconsciousness. A. In the volunteer subject lying awake with eyes closed, spatially coherent alpha oscillations are observed over the occipital area. The alpha oscillations shift to the front of the head with loss of consciousness (LOC) where they intensify and become spatially coherent during unconsciousness. The alpha oscillations dissipate anteriorly and return to the occipital area during return of consciousness (ROC) where they re-intensify and are spatially coherent in the eyes-closed awake state. B. During consciousness there is broadband communication between the thalamus and the frontal cortex with beta and gamma activity in the electroencephalogram. Modeling studies suggest during propofol-induced unconsciousness that the spatially coherent alpha oscillations are highly-structured rhythms in thalamocortical circuits. C. Slow oscillations recorded from grid electrodes implanted in a patient with epilepsy, 30 seconds after bolus induction of general anesthesia with propofol. The slow oscillations at nearby electrodes (red and green dots) are in phase (red and green traces) whereas the slow oscillation recorded at an electrode 2 centimeters away (blue dot) is out of phase (blue trace) with those at the other two locations. Neurons spike only (histograms) in a limited time window governed by the phase of the local slow oscillations. These slow oscillations are a marker of intracortical fragmentation with propofol as communication through spiking activity is restricted to local areas. The spatially coherent alpha oscillations and the disruption of neural spiking activity associated with the slow oscillations are likely to be two of the mechanisms through which propofol induces unconsciousness. Panel A is reproduced from Purdon et al. PNAS, 2013 and Panel B is adapted from Lewis et al. PNAS, 2012 with permission.

Fig. 8

Characterization of burst suppression. A. The spectrogram in fig. 6B from minute 4 to minute 20. Burst suppression in the spectrogram shows as periods of blue (isoelectric activity) interspersed with periods of red-yellow (slow-delta and alpha oscillations). The horizontal red line shows the principal period of burst suppression. B. Unprocessed electroencephalogram recordings corresponding to the spectrogram in A. The horizontal red lines at ±5 microvolts are the thresholds which separate burst events (amplitude ≥ 5 microvolts) from suppression events (amplitude

Fig. 9

Neurophysiology and electroencephalogram signatures of ketamine. A. At low doses, ketamine blocks preferentially the actions of glutamate NMDA receptors on GABAergic inhibitory interneurons in the cortex and subcortical sites such as the hippocampus and the limbic system. The antinociceptive effect of ketamine is due in part to its blockade of glutamate release from peripheral afferent (PAF) neurons in the dorsal root ganglia (DRG) at their synapses on to projection neurons (PN) in the spinal cord. B. Spectrogram showing the beta-gamma oscillations in the electroencephalogram of a sixty-one year-old woman who received ketamine sedation for a vacuum dressing change. Blocking the inhibitory action of the interneurons in cortical and subcortical circuits helps explain why ketamine produces a high-frequency beta oscillation as its electroencephalogram signature. C. Ten-second electroencephalogram trace recorded at minute 5 from the spectrogram in B. Arrows indicates times of ketamine doses. Panel A is reproduced with permission from Brown, Lydic and Schiff, New England Journal of Medicine, 2010. Panels B and C were adapted from Purdon and Brown, Clinical Electroencephalography for the Anesthesiologist (2014), with permission from the Partners Healthcare Office of Continuing Professional Development.

Fig. 10

Neurophysiology of dexmedetomidine. A. Dexmedetomidine acts pre-synaptically to block the release of norepinephrine (NE) from neurons projecting from the locus coeruleus (LC) to the basal forebrain (BF), the pre-optic area (POA) of the hypothalamus, the intralaminar nucleus (ILN) of the thalamus and the cortex. Blocking the release of NE in the POA leads to activation of its inhibitory GABAergic (GABA) and galanergic (Gal) projections to dorsal raphé (DR) which releases serotonin (5HT), the tuberomamillary nucleus (TMN) which releases histamine (His), the LC, the ventral periacqueductal gray (PAG) which releases dopamine, the lateral dorsal tegmental (LDT) nucleus and the pedunculopontine tegmental (PPT) nucleus which release acetylcholine (Ach). These actions lead to decreased arousal by inhibition of the arousal centers. B. Ten-second electroencephalogram segment showing spindles, intermittent 9 to 15 Hz oscillations (underlined in red), characteristic of dexmedetomidine sedation. C. The spindles are most likely produced by intermittent oscillations between the cortex and thalamus (light green region). Panel A is reproduced with permission from Brown, Purdon and Van Dort, Annual Review of Neuroscience, 2011.

Fig. 11

Spectrograms and time-domain electroencephalogram signatures of dexmedetomidine-induced sedation. A. Spectrogram of a 59 kg patient receiving a 0.65 mcg/kg/hour dexmedetomidine infusion to maintain sedation. The spectrogram shows spindles (9 to 15 Hz oscillations) and slow-delta oscillations. B. Ten-second electroencephalogram trace recorded at minute 60 from the spectrogram in A emphasizing spindles (red underlines). C. Spectrogram of a 65 kg patient receiving a 0.85 mcg/kg/hour dexmedetomidine infusion to maintain sedation. D. Ten-second electroencephalogram trace recorded at minute 40 from the spectrogram in C showing the slow-delta oscillations. Panels A–D were adapted from Purdon and Brown, Clinical Electroencephalography for the Anesthesiologist (2014), with permission from the Partners Healthcare Office of Continuing Professional Development.

Fig. 12

Spectrograms and time-domain electroencephalogram signatures of sevoflurane, isoflurane and desflurane at surgical levels of unconsciousness. The inspired concentration of the anesthetics is the blue trace in the upper part of each panel. Green arrows below each panel are propofol bolus doses. A. At sub-MAC concentrations (minute 40 to minute 60) the spectrogram of sevoflurane resembles that of propofol (fig. 6, A and B). As the concentration of sevoflurane is increased (minute 100 to minute 120), theta (5 to 7Hz) oscillations appear. The theta oscillations dissipate when the sevoflurane concentration (blue curve) is decreased. B. Ten-second electroencephalogram trace of sevoflurane recorded at minute 39.8 of the spectrogram in A. C. The spectrogram of sevoflurane shows constant the alpha, slow, delta and theta oscillations at a constant concentration of 3%. D. Ten-second electroencephalogram trace of sevoflurane recorded at minute 30 of the spectrogram in C. E. At sub-MAC concentrations (minute 16 to minute 26) the spectrogram of isoflurane resembles that of propofol (fig. 6A, B) and sub-MAC sevoflurane (panel A). Theta oscillations strengthen as the isoflurane concentration increases towards MAC. F. Ten-second electroencephalogram trace of desflurane recorded at minute 40 of the spectrogram in E. G. At the sub-MAC concentrations shown here the spectrogram of desflurane resembles propofol with very low theta oscillation power. H. Ten-second electroencephalogram trace of desflurane recorded at minute 40 of the spectrogram in G. Panels A, C, E, and G were adapted from Purdon and Brown, Clinical Electroencephalography for the Anesthesiologist (2014), with permission from the Partners Healthcare Office of Continuing Professional Development.

Fig. 13

Slow-delta and beta-gamma oscillations associated with nitrous oxide. A. In anticipation of emergence, a patient was maintained on 0.5% isoflurane and 58% oxygen. At minute 82, the composition of the anesthetic gases was changed to 0.2% isoflurane (blue curve) in 75% nitrous oxide (green curve) and 24% oxygen. The total gas flow was increased from 3 to 7 liters per minute. The alpha, theta and slow oscillation power decreased between minutes 83 to 85. At minute 86 the power in the theta to beta bands decreased considerably (blue area) as the slow-delta oscillation power increased. At minute 88 the slow-delta oscillation power decreased and the beta-gamma oscillations appeared at minute 90. The flow rates and anesthetic concentrations were maintained constant between minutes 82 and 91. Isoflurane was turned off at minute 91. B. Ten-second electroencephalogram traces of the slow-delta oscillation at minute 86.7 and the beta-gamma oscillations at minute 90.8. Panels A and B were adapted from Purdon and Brown, Clinical Electroencephalography for the Anesthesiologist (2014), with permission from the Partners Healthcare Office of Continuing Professional Development.

Fig. 14

Different anesthetics (propofol, sevoflurane, ketamine and dexmedetomidine), different electroencephalogram signatures and different molecular and neural circuit mechanisms. A. Anesthetic-specific differences in the electroencephalogram are difficult to discern in unprocessed electroencephalogram waveforms. B. In the spectrogram, it is clear that different anesthetics produce different electroencephalogram signatures. The dynamics the electroencephalogram signatures can be related to the molecular targets and the neural circuits at which the anesthetics act to create altered states of arousal. Panels A and B were adapted from Purdon and Brown, Clinical Electroencephalography for the Anesthesiologist (2014), with permission from the Partners Healthcare Office of Continuing Professional Development.