Electroencephalogram Dynamics of Etomidate-Induced Loss of Consciousness

Erwei Gu (  ay_guew_mz@163.com ) Department of Anesthesiology,The First a iliated hospital of Anhui medical university Lei Zhang department of anesthesiology,the rst a liated hospital of Anhui medical university Shunqin Fan the rst medical college of Anhui medical university Jiawei Zhang the rst medical college of Anhui medical university Kun Fang the rst medical college of Anhui medical university Lei Wang The First A liated Hospital of Anhui Medical University Yuanyuan Cao The First A liated Hospital of Anhui Medical University Lijian Chen The First A liated Hospital of Anhui Medical University Xuesheng Liu The First A liated Hospital of Anhui Medical University


Introduction
Etomidate is a non-barbiturate intravenous anesthetic that has a rapid onset of action and induces a stable, quiet, comfortable, and non-excitable transition during the induction period. Enhancement of γaminobutyric acid A receptor (GABAAR) activity is considered to be the primary mechanism mediating etomidate-induced anesthesia [1,2]. Etomidate has little effects on cardiovascular and respiratory systems. Based on these characteristics, etomidate is widely used in the induction of anesthesia in patients with impaired hemodynamics, such as in the elderly and those with cardiovascular disease or critical illness [3,4]. General anesthesia is a transient unconscious state [5]. General anesthetics are well described at the molecular level and cellular pharmacological level, but the neural circuit mechanisms that cause loss of consciousness remain unclear [6,7]. Electroencephalography (EEG) is considered to be the most direct indicator of central nervous system activity [8,9]. The quantitative evaluation of the depth of anesthesia by EEG has made an important contribution to the practice of clinical anesthesia [8]. A recent study reported the effect of general anesthesia on neuronal data recorded in the frontal cortex, and found that frontal-lobe electrical activity is the precursor to a loss of consciousness [10].
Etomidate-induced general anesthesia produces dynamic changes in EEG data, but the characteristics of these changes have not been well elucidated. Furthermore, the relationship between etomidate-induced loss of consciousness and characteristic changes in EEG data is not well understood. Hence, the aim of our present study was to clarify the relationship between etomidate-induced loss of consciousness and electroencephalogram dynamics of etomidate-induced loss of consciousness.

Methods
This study followed the Declaration of Helsinki and was approved by the Ethics Committee of the First A liated Hospital of Anhui Medical University.Standard monitoring techniques (noninvasive blood pressure, electrocardiography [ECG], and pulse oximetry) were applied. Hemodynamic variables were recorded every 60 sec. Written informed consent was obtained from the included patients, who were between 18 and 65 years old (ASA: ~ , No gender limit) and required general anesthesia. Exclusion criteria were as follows: pregnancy, hearing impairments, mental disorders, lack of coordination during induction of anesthesia, or taking drugs that may interfere with the accuracy of EEG recordings.
Involuntary myoclonus often occurs during the induction of etomidate [11,12] anesthesia. If obvious myoclonus occurred in a patient, the patient's EEG data were not included in the data analysis. Case selection is shown in Figure 1. Table 1 presents the basic demographic and clinical characteristics of the included subjects. Time of LOC (min), mean (±SD) 168 (7)

* LOC=Loss of consciousness
We used a four-channel Sedline brain function monitor (Masimo, Irvine, CA USA) to record the frontal-lobe EEG data. EEG data were recorded in patients undergoing elective surgery (n = 20) during a baseline of 3 min (with awake periods of closed eyes), etomidate-induced loss of consciousness, and 3 min after the unconsciousness. Etomidate (0.06 mg/kg/min) was the only anesthetic used [13]. The EEG data were recorded with a preampli er bandwidth of 0.5-92 Hz, a sampling rate of 178 Hz, 16 bits, and a resolution of 29 nV. The standard Sedline-Sedtrace electrode array records were from electrodes located roughly at positions Fp1, Fp2, F7, and F8, with the ground electrode at Fpz and the reference electrode at roughly 1 cm above Fpz. The electrode impedance was less than 5 kΩ in each channel. The time record in the case report form was required to match the time on the EEG recorder in order to mark key events (e.g., induction start, loss of consciousness) during the analysis.
Criteria for the loss of consciousness At present, in the induction of clinical anesthesia, the determination of the loss of consciousness is assessed by aimless movements after harmful stimulation [14]. In the present study, the auditory stimulation assessment [15,16] was supplemented by the disappearance of the eyelash re ex to con rm the loss of consciousness. Before inducing anesthesia, we instructed patients not to open or move their eyes.

Data preprocessing
A researcher with experience in reading electroencephalograms manually browsed the EEG data of each patient to manually remove artifacts. The investigator used the recorded information in the case report form to select the appropriately timed EEG data segment. For each case, the EEG segment representing 60 s of consciousness and closed eyes was carefully selected during the perioperative period, as was the EEG segment corresponding to 60 s after the loss of consciousness, for data analysis.

Spectral analysis
The power spectrum quanti es the frequency distribution of power or energy within a signal. The spectrogram was computed using the multitaper method achieved in the Chronux toolbox in MATLAB [17] . And the group-level spectrogram computed by taking the median of all patients. The spectrum of the selected EEG epochs was also calculated by us. Then, for all epochs, the resulting power spectra were averaged, and by way of multitaper-based jackknife techniques [17], 95% con dence intervals (CIs) were computed. Parameters for spectral analysis are as follows: window length T = 2 s with a 1.95 s overlap; time-bandwidth product of TW = 3; number of tapers, K = 5; and spectral resolution = 3 Hz. The timefrequency analysis is re ected in Figure 2.

Coherence analysis
Coherence graphs are coherent time-varying versions, which are estimated using continuous windows of EEG data. Between two signals, x and y, the coherence Cxy (f) function, is determined as follows: Sxy (f) is the cross-spectrum between the signals x (t) and y (t), Sxx (f) is the power spectrum of the signal x (t), and Syy (f) is the power spectrum of the signal y (t) [17]. In order to acquire the appraised coherence, based on the Chronux toolbox in MATLAB, the coherence was computed between F7 and F8, the two frontal electrodes [18] . The electrode position is shown in Figure 3. By taking the median across subjects, the group-level coherograms were computed. The coherence for selected EEG epochs was also calculated. The resulting coherence estimates were averaged for all epochs, and by way of multitaperbased jackknife techniques ,95% CIs were computed [17]. Parameters for the coherence analysis were similar to spectral analysis and spectral resolution of 2 W = 3 Hz. The peak coherence and its frequency of the frontal alpha oscillation for each individual patient were estimated by us. Then, in order to acquire the group-level peak coherence and frequency for these oscillations, we averaged across subjects.

Statistical analysis
So as to emulate spectral and coherence computes between groups, we utilized jackknife-based methods [17], the two-group test for spectra, and the two-group test for coherence, as performed in the Chronux toolbox routine [19]. This method takes into account the frequency spectrum and the basic spectral resolution of the coherence estimation, and only when the difference occurs at a continuous frequency on a frequency band wider than the spectral resolution of 2 W, the difference is considered signi cant. To be speci c, for frequencies f > 2 W, the negative assumption was rejected only if the test statistic surpassed the signi cance threshold over a contiguous frequency range ≥2 W. For frequencies 0 ≤ f ≤ 2 W, in order to illustrate the capabilities of multitaper spectral estimation when the frequency is near zero, the negative assumption was rejected only if the test statistic surpassed the signi cance threshold over a contiguous frequency range from 0 to max (f,W) ≤ 2 W. A signi cance threshold of P < 0.001 was con rmed for comparisons between groups.

Results
We collected EEG data from 40 cases of etomidate-induced general anesthesia but excluded 12 patients due to the coordination of their EEG records in the awake period being poor. Ultimately, of the remaining 28 patients, 20 patients were considered t for analysis. Subjects which with poor-quality data were excluded, most likely due to poor electrode contacts.

Etomidate power-spectra analysis
We observed the EEG spectra during both the awake period before etomidate administration and during etomidate-induced loss of consciousness. oscillations. Theta-wave (4.0-8.0 Hz) and alpha-wave (8.0-13.0Hz) oscillations were relatively insigni cant. Next, we performed time-frequency analysis of EEG data in the two periods and found that after etomidate-induced loss of consciousness, the oscillation in each band increased. We also observed that the two periods were clearly different in time and frequency between 0-22.97 Hz and 27.28-40.00 Hz. The result of power-spectra analysis is shown in Figure 4.

Etomidate coherence analysis
We next analyzed the similarities and differences in the correlation patterns between the awake period and the period during etomidate-induced loss of consciousness. We found that the awake-phase coherence map had a speci c coherence in the theta wave and alpha wave. In contrast, the phase coherence map during the etomidate-induced loss of consciousness was signi cantly enhanced for the theta wave and alpha wave compared to that during the awake phase. Hence, some coherence was produced. We also observed that there was a signi cant difference in the coherence between the waking period and the period of etomidate-induced loss of consciousness in the 1.86-3.17 Hz band. The result of coherence analysis is shown in Figure 5.

Discussion
In the present study, we analyzed the characteristic changes in the EEG bands before and during etomidate-induced anesthesia via spectral analysis and coherence analysis. From the awake period to the loss of consciousness period, etomidate general anesthesia induced an increase in oscillation in each frequency band within the EEG data. Additionally, when etomidate induced loss of consciousness, the coherence of the EEG signal in the theta wave and alpha wave was obviously enhanced, and the delta wave also showed obvious coherence.
Etomidate exerts its anesthetic actions through potentiation of GABAARs containing β2 and β3 subunits. It has recently been shown that the β2 subunit contributes to the sedative properties of etomidate, whereas the β3 subunit is responsible for its anesthetic properties [20,21]. In our present study, etomidate induced anesthesia and enhanced theta-wave oscillations. In our coherence analysis, etomidate produced strong coherence in the theta wave; interestingly, β3-subunit-containing receptors play an important role in theta-wave oscillations. A previous study analyzed the effects of etomidate on rhythmic population activity by recording local eld potentials (LFPs) [22]. In slices, which derived from wild-type mice, etomidate (200 nM) ampli ed the oscillatory population activity in the theta-frequency band, but this effect was not seen in slices, which derived from β3-knockin mice, and this phenomenon was also observed in vivo [23]. These ndings indicate that the neuronal mechanism of etomidate-induced loss of consciousness is closely related to the β3 subunit of GABAAR functional activity. Furthermore, previous evidence has shown that the overall effect of etomidate re ects a balance between enhancement and inhibition produced by GABAARs containing β2 and β3 subunits [22]. Etomidate enhances theta oscillations by acting on β3-containing GABAA receptors but depresses these oscillations via β2-subunitcontaining receptors. In our present study of etomidate-induced loss of consciousness, we systematically observed the enhancement of EEG theta oscillations. However, inhibition of EEG theta oscillations and a balance of enhancement and inhibition of EEG theta oscillations was not re ected in our present study. Therefore, future studies should be conducted to investigate these additional phenomena.
In the present study, etomidate induced a loss of consciousness and changed the EEG theta rhythm.
Some studies have indicated that hippocampal interactions with the prefrontal cortex, another signi cant memory-associated structure, are coordinated by theta-rhythm plasticity [24,25]. Amnesia is a signi cant symptom of general anesthesia. Several different brain areas participate in memory formation, including the prefrontal cortex, amygdala, and hippocampus. Anesthesia may be associated with memory impairment, as has been indicated by many studies in the hippocampus [26,27], and evidence suggests that different frequency oscillations are related to changes in information coding and synaptic weights [28]. Etomidate-induced amnesia arises by means of GABAAR modulation, which highly depends on α-5 subunit modulation; this process likely occurs principally within the hippocampus [29,30]. Anesthetics may disturb the greatly organized rhythmic activity patterns that are thought to be indispensable for hippocampal learning [31]. For instance, changes in theta frequency and theta power may contribute to amnesia by means of anesthetic actions on a variety of molecular targets [32]. Theta rhythms act as internal clocks that synchronize large networks and serve as a reference mechanism for internal and external synchronization, which is greatly coherent throughout the medial temporal lobe. These ndings suggest that etomidate-induced loss of consciousness and changes in hippocampal theta rhythms are closely related. Collectively, these potential associations have led us to hypothesize that etomidateinduced theta oscillations may be an indicator of a functional disconnection between the hippocampus and cerebral cortex.
The EEG data analyzed in this study were all derived from the frontal 4-channel pathway, so our analysis was unable to assess other reported cortical kinetic connectivity associated with anesthetic-induced loss of consciousness. Our observations need to be further validated in future high-density EEG studies [33] . At present, exploring the relationship between the neural circuit mechanism of etomidate-induced loss of consciousness and the characteristic changes of EEG can only be carried out on animal experimental models. Therefore, the clinical observed 'the connection' still needs to be veri ed by animal models.
According to our analysis and discussion, from the awake period to the loss of consciousness, the neural circuit mechanism of etomidate-induced loss of consciousness is closely related to the enhancement of delta-wave, alpha-wave, and theta-wave coherence. The characteristic transformation of the EEG that we described can be calculated and displayed in real time, providing a good reference for the monitoring of the depth of anesthesia and the evaluation of the level of sedation and consciousness by an anesthesiologist.