Episodic waveforms in the electroencephalogram during general anaesthesia: a study of patterns of response to noxious stimuli.
The relationship between the states of general anaesthesia and sleep is complex and poorly understood. Many of the EEG patterns seen in patients under [gamma]-amino butyric acid (GABA)-ergic general anaesthesia are broadly similar to those seen in slow wave sleep, but differ in the details of frequency and duration. For example, natural sleep spindle frequency tends to be higher (~14 Hz) than the distorted 'sleep spindle-like' waveforms generated during anaesthesia (~10 Hz) (2,9,10). There may also be significant differences in the neurobiological processes that generate these EEG patterns (11). Nevertheless, part of the action of general anaesthetic drugs is to activate natural sleep processes (12). During anaesthesia, episodic 10 Hz activity and episodes of burst suppression or near-burst suppression occur commonly. Both these patterns are very likely to be due to anaesthetic drug effects on ion channels, similar to sleep spindles and the 'up-down' oscillations seen during some phases of natural sleep. Because these episodic EEG waveforms are abolished with arousal from natural sleep, it is possible that these EEG phenomena may be indicators of the balance between anaesthetic-induced inhibition and nociceptive-induced activation of the brain. The ability to maintain these sleep-like EEG patterns in the face of nociceptive stimuli might be taken as a reasonable indication of a sufficient level of anaesthesia and thus potentially aid in accurate anaesthesia titration.
In this preliminary observational study, we quantified the effect of noxious stimuli on these episodic EEG waveforms (10 Hz spindle-like waves and slow oscillations in high frequency power) as well as the non-episodic delta waves, in a cohort of healthy patients undergoing general anaesthesia. The three subgroups of subjects were defined by the dose of fentanyl used during the induction of anaesthesia.
MATERIALS AND METHODS
Approval was received from the Northern Y Regional Ethics Committee (NTY/06/08/071). Written informed consent to participate in this study was acquired from all patients. The study group included 30 patients (17 female, 13 male) who were classified as American Society of Anesthesiologists physical status I or II, aged 23 to 82 years and scheduled for elective gynaecologic, orthopaedic or abdominal surgery. Patients were excluded for the following reasons: obesity, severe reflux (requiring rapid intubation), ischemic heart disease, chronic or recent opioid use, heavy alcohol or marijuana use, history of neurologic disease and use of regional or local anaesthetic. A standardised anaesthetic was given. Each of the first 20 patients was randomised to receive either 1.0 [micro]g/kg or 2.5 [micro]g/kg of fentanyl intravenously (IV) on induction of anaesthesia. To define another, higher dose fentanyl group, the final 10 patients received 4.0 [micro]g/kg of fentanyl IV at induction. Anaesthesia was induced with IV propofol (70 to 220 mg), titrated to loss of eyelash reflex. Neuromuscular blockade was established with IV rocuronium (30 to 100 mg), atracurium (12.5 to 50 mg) or vecuronium (7 to 10 mg), and laryngoscopy and tracheal intubation achieved. Routine monitoring consisted of electrocardiogram, pulse oximetry, non-invasive blood pressure, inhaled vapour and gas analysis, and EEG (S5 GE DatexOhmeda Fairfield, CT, USA) using the three-electrode commercial sensor strip and standard prefrontal montage. Inspired sevoflurane was titrated to aim for a state entropy index (M-Entropy, GE Datex-Ohmeda Fairfield, CT, USA) of 30 to 60 prior to the initial surgical incision. The raw EEG signal (100 samples/seconds), heart rate, end-tidal sevoflurane concentration and state and response entropy data (sampled at five second intervals) were obtained from the monitor using the GE Collect software and saved to a computer. The recordings were continued for two minutes after the surgical incision, at which point the study ceased and the operation and anaesthesia continued as per clinical routine.
We defined two different events of noxious stimulation: 1) laryngoscopy/intubation ('ETT') and 2) surgical incision ('incision'). Due to the bolus dosing of fentanyl and propofol and the often long delay (range 14 to 40 minutes) between induction and surgical incision, the incision stimulus occurred during a period of steady-state anaesthesia (as defined in the previous paragraph) consisting of primarily sevoflurane, with low effect-site concentrations of fentanyl. In contrast, the ETT stimulus occurred during a period of anaesthesia that consisted of propofol and fentanyl. To evaluate the changes caused by these noxious stimuli, we analysed and compared two segments of the EEG before and after the stimulus. The pre-noxious stimulus period consisted of the one minute EEG segment 70 to 10 seconds before initial laryngoscopy or before incision. The post-noxious stimulus period was the one-minute segment beginning 40 seconds after successful intubation or after incision. When laryngoscopy or intubation was attempted more than once (n=3), pre-noxious stimulus was taken to be before the first laryngoscopy and post-noxious stimulus was defined as after successful intubation. We used the changes in heart rate over the same pre- and post-stimulus periods as a simple indicator of sympathetic nervous system activation induced by the noxious stimuli. Effect-site fentanyl, propofol and sevoflurane concentrations were estimated using the pharmacokinetic models derived from Shafer, Willens, Scott, White and McKay (13-18).
The analysis was performed using custom software written in MatLab (The MathWorks, Natick, MA, USA). We chose to quantify the pattern of EEG in five second segments, using three parameters. Namely, 1) 10 Hz-score: the amount of each EEG segment in which 'spindle-like' patterns were detected; 2) high frequency variability index (HFVI): alterations between hyperpolarised (so-called 'down') states and high firing ('up') states; and 3) delta: the absolute amount of delta power. This is not an episodic waveform. The derivations of these parameters are described below.
Because there are no widely accepted methods for the analysis of spindles, we developed a simple way to quantify the 'spindle-like' episodic 10 Hz activity present in the EEG by examining the signal for generic patterns to generate a 10 Hz-score. We used this method in preference to simply obtaining the power at 10 Hz in the power spectrum of the EEG signal because the 10 Hz power conflates amplitude and duration of the signal and does not distinguish between a true narrowband 10 Hz oscillation and underlying broad-band 1/f background noise.
A sleep spindle may be defined as a section of waxing and waning 8 to 16 Hz oscillation in the EEG. In sleep physiology it was originally strictly defined as only 12 to 14 Hz and lasting >0.5 seconds. In general anaesthesia, spindle-like patterns are often seen but the frequency is commonly slightly lower than 12 to 14 Hz and the waxing and waning envelope less obvious. The Matlab code for this function ('Spindle_score.m') is included in the Appendix. The score is the proportion of time in which the EEG showed a 10 Hz peak-trough pattern, in runs lasting at least two and a half cycles. The function moves, one sample step at a time, through the EEG data segment; searching for six consecutive data points whose differences changed from positive to negative--two 'peaks' and two 'troughs' (i.e. a '[??]'pattern). Consecutive points of the raw EEG at a sampling rate of 100 /second would show spindle activity at about 100/2=50 Hz (note: a complete sinewave will consist of 10 ms up and 10 ms down, making a total of 20 ms [=50 /second] per complete cycle). To search for spindle-like patterns at lower frequencies, we need to look at longer cycles. For example, a spindle at a frequency of 10 Hz requires a 5 data-point lag (in the code in the Appendix, this is the parameter 'tau'=5). Thus at each time point in the EEG series, the spindle finding function assigned a value of 'zero' if no spindle sequence was detected in the next 6xlag data points, or 'one' if a spindle-sequence was detected. This is returned as the parameter 'tag' in the Appendix code. An example of the how the algorithm classifies EEG spindle-like patterns is shown in Figure 1. When we adapted the method to look only at long spindle sequences (e.g. lengths of 9 or 15 peaks and troughs), there was no significant difference in the results. In our analysis, we examined 'spindle-like' activity at a frequency of 10 Hz.
[FIGURE 1 OMITTED]
There is no clear gold standard for validation of this index. In an effort to confront this issue, an expert EEG evaluator was given 20 second segments of each pre-noxious stimulus raw EEG and asked to subjectively score the segment as being either a high or low 10 Hz-score. (He was blinded to the actual 10 Hz-score.) Using a threshold 10 Hz-score of 0.5, there was 79% agreement between the subjective observer and the objective 10 Hz-score. Acknowledging that this is a relatively coarse method of validation, we felt that our 10 Hz-score was probably a reasonable measure of the amount of 'spindle-like' activity present on the EEG (Figures 1 and 3).
High-frequency variability analysis
Recently Mukovski and co-workers have validated a simple method using the surface EEG for detecting the alternating periods of synchronous neocortical neuronal activity and silence that occur during slow wave sleep (19). They are therefore able to use the scalp EEG to gain some information about the membrane potential of populations of cortical pyramidal neurons. Their method is somewhat paradoxical, because it uses high-frequency information (>20 Hz) to tell us about very low-frequency fluctuations. In essence it uses the periods of vigorous high-frequency activity to identify the 'up' states of neuronal depolarisation and periods of almost no high-frequency activity to identify the 'down' states of neuronal hyperpolarisation. Their methods appear to be more reliable than simply identifying positive and negative half-waves in the EEG. The method is insensitive to pre-filtering and free from the constraints of the temporal resolution of electrode drift. Mukovski and co-workers found that the differences between power spectra from active and silent states are most obvious in the high-beta-gamma range (20 to 100 Hz) (19); yet, because a filtered scalp EEG has no reliable EEG signal >50 Hz, we used a modified method based on the principles of Mukovski and co-workers. We ran a fifth order band-pass Butterworth filter over our raw EEG to extract 20 to 45 Hz activity. In our study the patients were profoundly paralysed, so there was minimal electromyogram interference at these frequencies. We then estimated the standard deviation of this band-limited, high-frequency signal for 100 millisecond, non-overlapping segments. The fluctuation in the standard deviation of this high-frequency component of the signal formed the basis of identifying the active ('up') and inactive ('down') states. There are well-described correlations between the presence of high frequencies in the scalp EEG and the neuronal membrane potential in states of true burst suppression. However, it is less clear that high frequency fluctuations correspond with 'up' and 'down' neuronal states during moderate levels of general anaesthesia. Therefore we did not use the final part of the Mukovski algorithm to make a binary separation of up and down states. The running standard deviation was then smoothed (10 point median filter) and the upper and lower envelopes extracted. Our index of the 'amplitude' of the 'up-down' states was calculated each five seconds and defined by the equation:
HFVI = (upper envelope--lower envelope)/lower envelope
It was chosen because it made the index sensitive to the periodic very quiet periods (the 'down' states) and was resistant to the continuous nature of the 'up' state in the awake state. If the 'down' state is prolonged and has no underlying delta oscillation, it is the same as the 'suppression' phase of the well-known burst-suppression pattern found in deep anaesthesia.
The power spectrum of the EEG signal was obtained using the 'psd.m' Matlab function (frequency resolution of 1 Hz and Hanning window) and the delta was defined as the mean power in 1 to 4 Hz frequency band averaged over each five second segment.
Unless otherwise stated, the descriptive data are presented as median (range). Two way repeated-measures analysis of variance was used to detect a change in mean EEG indices (10 Hz-score, HFVI and delta). The model included the one minute pre-noxious stimulus segment (-70 to -10 seconds) to the one minute (40 to 100 seconds) post-noxious stimulus segment as the within-subject effect, the fentanyl dosage as the between-subject effect and the fentanyl-noxious stimulus interaction term. Scatterplots showed linear relationships between some of the pre-incision values and the change in values, and the significance of these relationships was quantified with Pearson's linear correlation coefficient (r). If the noxious stimulation caused a mean (SD) change in heart rate of 10 (10) bpm, recruitment of 10 patients in each treatment arm would be required to achieve a statistical power of 0.81.
Of our 30 patients, all were included in the analysis of changes due to ETT, however three were excluded from the analysis of changes due to incision because the entropy monitor detached or because local anaesthetic was given by the surgeon prior to incision (one from the low dose fentanyl group and two from the high dose fentanyl group). The mean age of the patients was 51 years and the operations were gynaecological (n=10), orthopaedic (n=11) or general surgery (n=9), and five patients were on preoperative beta-blocking drugs.
Patterns of EEG response to noxious stimulus
General anaesthesia induced various stereotypical EEG patterns amongst the different patients. Nevertheless, within most patients there were observable changes in EEG pattern in response to the noxious stimuli. As a first step we present two examples from different patients that serve to illustrate some of the variety of 'high-frequency variability', 'spindle-like' and delta EEG waveforms seen (Figures 2 and 3). Like the sympathetic nervous system response, the cortical response was typically delayed 10 to 40 seconds after the incision and intubation.
[ILLUSTRATION 2 OMITTED]
[ILLUSTRATION 3 OMITTED]
The raw changes induced by the noxious stimuli in heart rate and in the EEG indices are shown in Table 1. The change in heart rate was not strongly correlated with change in 10 Hz-score (ETT r=0.09, incision r=0.06), HFVI (ETT r=-0.09, incision r=0.13), delta activity (ETT r=-0.22, incision r=-0.20) or entropy (ETT r=0.08, incision r=0.12).
The repeated measures analysis of variance of these data demonstrated the following:
1. Significant differences between patients (P <0.001) for all variables and stimuli.
2. Both ETT and incision resulted in a significant decrease in 10 Hz-score: ETT decreased from 0.25 (0.02) to 0.20 (0.02, P=0.01) and incision decreased from 0.33 (0.02) to 0.27 (0.02, P=0.01). This is the classical arousal pattern.
3. Incision, but not intubation, resulted in a significant decrease in HFVI from 2.8 (0.2) to 2.0 (0.2, P=0.02).
4. No statistically significant change in delta for either incision or intubation.
5. No effects of fentanyl, either directly or in the fentanyl-noxious stimulus interaction term.
Within each individual, the pre-noxious stimulus EEG pattern was relatively consistent between the predominantly propofol anaesthesia (prior to ETT), compared to the predominantly sevoflurane anaesthesia (prior to incision). Pre-ETT 10 Hz-score correlated well with pre-incision 10 Hz-score (r=0.58, P=0.008) and pre-ETT delta correlated with pre-incision delta (r=0.71, P <0.0001). HFVI however, showed no correlation (r=0.07, P=0.78).
Association of pre-noxious EEG patterns with change in EEG pattern
For each patient, the correlation between the value of each parameter before the noxious stimulus (x-axis), and its subsequent change in value (y-axis), is shown in the scatterplots in Figure 4. These plots enable the noxious-induced change in the EEG index for each individual patient to be seen. Patients with a high 10 Hz-score or HFVI (i.e. those points towards the right of each graph) showed a loss of these patterns in response to the noxious stimulus and reverted to a simpler delta pattern (change in spindles vs change in delta power, r=-0.46, P=0.02). It was rare for 10 Hz-score or HFVI to increase with noxious stimulus, and these increases were very small in comparison to the large nociceptive-induced decreases in these parameters in some patients. In contrast, the delta almost doubled in four patients in response to ETT--the paradoxical arousal pattern.
Table 2 details how the subgroup of patients who showed exaggerated patterns before the noxious stimuli altered their EEG patterns in response to these stimuli.
[FIGURE 4 OMITTED]
Effect of fentanyl and other drugs
The estimated median (range) effect-site fentanyl concentrations were 3.1 ng/ml (1.0 to 5.6 ng/ml) at ETT and 0.86 ng/ml (0.2 to 2.9 ng/ ml) at the time of incision. The estimated effect-site propofol concentration at ETT was 6.6 mg/ml (3.4 to 13.3 mg/ml) and estimated effect-site sevoflurane concentration at incision was 2.1% (1.1 to 3.6%). The patterns of changes in EEG in response to noxious stimulation described above were not significantly correlated to drug dosage. Neither sevoflurane nor propofol effect-site concentrations were correlated to pre-stimulus 10 Hz-score, delta HFVI, nor changes in any of these parameters. However, post-hoc analysis demonstrated that the 1 [micro]g/kg fentanyl group had a greater change in 10 Hz-score (-0.14) as compared to the high-dose groups (-0.01, P=0.01, t-test, see top right panel Figure 4). As has been previously reported, fentanyl modulated the brainstem control of the sympathetic nervous system--as exhibited by its effect on heart rate--more than it modulated the cortex. High-dose fentanyl limited the post-ETT increase in heart rate from mean (SD) of 22 (10) bpm in the low-dose fentanyl group to 7 (7) bpm in the high-dose group (P=0.002, t-test). Fentanyl dose was unrelated to both pre-noxious stimulus response entropy and nociception-induced change in response entropy.
In this preliminary study we have described some novel EEG analysis methods to quantify episodic EEG phenomena during anaesthesia, and how they change in response to noxious stimuli. The long-term clinical goal would be to have an EEG-based 'nociception index', to which anaesthesia could be titrated to facilitate adequate pain relief during recovery. However, due to our current incomplete understanding of the relationship between EEG measures and nociceptive effects, this goal remains elusive. The numbers of patients in the present exploratory study are small and we plan further studies for the more detailed quantification of pharmacological effects. Nevertheless, our main findings may be summarised as follows:
1. A substantial proportion of the patients had EEGs which clearly demonstrated episodic phenomena (either 'spindle-like' [10 Hz-score] or 'high-frequency variability' patterns) prior to noxious stimulation (Table 2). In about half of these cases, the effect of nociceptive input was to ablate or reduce these episodic phenomena in the EEG. This reduction in episodic EEG patterns tended to be greatest in those who showed the most florid pre-noxious stimulus episodic patterns (Figure 4) and was more pronounced in patients who had a lower dose of fentanyl.
2. The pre-noxious stimulus EEG patterns were not strongly correlated with the calculated effect-site concentrations of fentanyl, propofol or sevoflurane.
3. Increased fentanyl dose clearly modified the tachycardic response to intubation, but had less consistent effects on the episodic phenomena seen on the EEG. The effects of fentanyl in reducing tachycardia did not correlate with the reduction in episodic EEG patterns. Hence, we infer that the EEG changes--which measure cortical activity--are poor indicators of responsiveness in the brainstem vasoactive centres. In powering the study, we had anticipated that the cortical effects of the fentanyl might be similar in magnitude to its effects in modulating the tachycardic response to the noxious stimulation. However, the cortical responses are clearly more complex, which requires a larger, differently designed study to tease out proper dose-response surfaces between the noxious stimulus and hypnoticopioid interactions.
Patients look different in appearance--so do their EEGs. There was marked variation in EEG pattern between patients even before the noxious stimulus. About half the patients showed obvious episodic patterns prior to the noxious stimulus and most (but not all) lost them in response to the noxious stimulus (Table 2). However, it is important to take note that about half the patients did not show pronounced episodic changes--even before noxious stimulation. The question arises--why do some patients' EEGs fail to develop pronounced episodic patterns? Although the numbers in our study are relatively small, this group of patients did not differ grossly in regard to dose of sevoflurane or fentanyl. We are unable to exclude pharmacokinetic or pharmacodynamic explanations, but it is our suspicion that the observed inter-individual variations reflect exogenous and genetic neuromodulatory factors that influence the state of intrinsic neuronal ion channels in that particular patient (20-24) and are worthy of further study.
The original method devised to estimate depolarised/active ('up') and hyperpolarised/inactive ('down') states was applied to the slow oscillation of natural sleep. Exactly what the fluctuation in high-frequency power (HFVI) is measuring during general anaesthesia is less well-defined, but is at least partially correlated with neuronal firing rate and hence to changes in cortical neuronal membrane potential (11,25,26). The most extreme active and inactive states in general anaesthesia are associated with the burst-suppression pattern27 which are nicely detected by the HFVI. Whether the less extreme values of HFVI represent a transition to burst-suppression is not known. Waves at spindle frequency are thought to be co-ordinated by the slow oscillation during natural slow wave sleep (28,29). similarly, under general anaesthesia we noted that 'spindle-like' waveforms were associated with the active phase of burst-suppression (Figure 2), suggesting that these two phenomena are not entirely separate processes.
In this study, we have shown that it is possible to quantify a variety of stereotypical episodic patterns found in the EEG under general anaesthesia and the subsequent changes that occur in response to noxious stimuli. Would it be beneficial to try to devise an anaesthetic plan that effectively opposes these EEG changes and thus maintains an EEG pattern that is similar to that found in the unstimulated state? At this stage we are not able to answer this question. But the concept of a patient who has achieved an anaesthetised state which continues to show features of natural sleep, even in the presence of a surgical nociceptive barrage, remains an intuitively appealing therapeutic goal.
APPENDIX Matlab spindle finding function function [p, tag] = spindle_score(y, tau, amp); % finds number of order six spindles in EEG segment % Inputs % the segment of EEG for analysis (y), % the lag (tau), % the minimum amplitude threshold (amp) % Outputs % the fraction of the segment that is assigned as part of a spindle (p), % an index of whether each EEG datapoint is assigned as part of a spindle or not (tag) ord=6; % length of spindle lenn=length (y); global tag; p=; % Pass the spindle pattern over y cs=0; % setup variable for number of spindle motifs i=1; while i <lenn-tau*ord; tag (i) = 0; yy=y (i:tau:i+ord*tau-1); da=diff (yy); % ord=6 if (da(1)>0)&(da(2)<=0)&(da(3)>0)&(da(4)<=0) &(da(5)>0)&(abs(min(da))> = amp); % i.e. 0-1-0 pattern [??] tag (i:i+tau*ord-1) = 1; % record this in the tag vector i=i+tau*ord-1; % jump beyond the end of the spindle cs=cs+tau*ord; % add these points to the spindles count end % if i=i+1; % look at each data point and see if it is associated with spindle end % i p=cs/(lenn); return
This work was funded by the Waikato Medical Research Foundation and the Braemar Charitable Trust.
Accepted for publication on June 4, 2009.
(1.) Hagihira S, Takashina M, Mori T, Ueyama H, Mashimo T. Electroencephalographic bicoherence is sensitive to noxious stimuli during isoflurane or sevoflurane anesthesia. Anesthesiology 2004; 100:818-825.
(2.) Morimoto Y, Hagihira S, Yamashita S, iida Y, Matsumoto M, Tsuruta S et al. Changes in electroencephalographic bicoherence during sevoflurane anesthesia combined with intravenous fentanyl. Anesth Analg 2006; 103:641-645.
(3.) Morimoto Y, Matsumoto A, Koizumi Y, Gohara T, Sakabe T, Hagihira S. Changes in the bispectral index during intraabdominal irrigation in patients anesthetized with nitrous oxide and sevoflurane. Anesth Analg 2005; 100:1370-1374.
(4.) Kox WJ, von Heymann C, Heinze J, Prichep LS, John ER, Rundshagen i. Electroencephalographic mapping during routine clinical practice: cortical arousal during tracheal intubation? Anesth Analg 2006; 102:825-831.
(5.) Rundshagen I, Schroder T, Prichep LS, John ER, Kox WJ. Changes in cortical electrical activity during induction of anaesthesia with thiopental/fentanyl and tracheal intubation: a quantitative electroencephalographic analysis. Br J Anaesth 2004; 92:33-38.
(6.) Bischoff P, Kochs E, Haferkorn D, Schulte am Esch J. intraoperative EEG changes in relation to the surgical procedure during isoflurane-nitrous oxide anesthesia: hysterectomy versus mastectomy. J Clin Anesth 1996; 8:36-43.
(7.) Muthuswamy J, Thakor NV Spectral analysis methods for neurological signals. J Neurosci Methods 1998; 83:1-14.
(8.) Hobson JA, Pace-Schott EF, Stickgold R. Dreaming and the brain: toward a cognitive neuroscience of conscious states. Behav Brain Sci 2000; 23:793-842.
(9.) Hayashi K, Sawa T, Matsuura M. Anesthesia depth-dependent features of electroencephalographic bicoherence spectrum during sevoflurane anesthesia. Anesthesiology 2008; 108:841850.
(10.) De Gennaro L, Ferrara M. Sleep spindles: an overview. Sleep Med Rev 2003; 7:423-440.
(11.) Crunelli V, Cope DW, Hughes SW. Thalamic T-type Ca2+ channels and NREM sleep. Cell Calcium 2006; 40:175-190.
(12.) Nelson LE, Guo TZ, Lu J, Saper CB, Franks NP, Maze M. The sedative component of anesthesia is mediated by GABA(A) receptors in an endogenous sleep pathway. Nat Neurosci 2002; 5:979-984.
(13.) Shafer SL, Varvel JR, Aziz N, Scott JC. Pharmacokinetics of fentanyl administered by computer-controlled infusion pump. Anesthesiology 1990; 73:1091-1102.
(14.) Willens JS, Myslinski NR. Pharmacodynamics, pharmacokinetics, and clinical uses of fentanyl, sufentanil, and alfentanil. Heart Lung 1993; 22:239-251.
(15.) Scott JC, Cooke JE, Stanski DR. Electroencephalographic quantitation of opioid effect: comparative pharmacodynamics of fentanyl and sufentanil. Anesthesiology 1991; 74:34-42.
(16.) Shafer SL. Advances in propofol pharmacokinetics and pharmacodynamics. J Clin Anesth 1993; 5:14S-21S.
(17.) White M, Schenkels MJ, Engbers FH, Vletter A, Burm AG, Bovill JG et al. Effect-site modelling of propofol using auditory evoked potentials. Br J Anaesth 1999; 82:333-339.
(18.) McKay iDH, Voss LJ, Sleigh JW, Barnard JP, Johannsen EK. Pharmacokinetic-pharmacodynamic modeling the hypnotic effect of sevoflurane using the spectral entropy of the electroencephalogram. Anesth Analg 2006; 102:91-97.
(19.) Mukovski M, Chauvette S, Timofeev I, Volgushev M. Detection of active and silent states in neocortical neurons from the field potential signal during slow-wave sleep. Cereb Cortex 2007; 17:400-414.
(20.) Chen X, Sirois JE, Lei Q, Talley EM, Lynch C 3rd, Bayliss DA. HCN subunit-specific and cAMP-modulated effects of anesthetics on neuronal pacemaker currents. J Neurosci 2005; 25:5803-5814.
(21.) Lee KH, McCormick DA. Modulation of spindle oscillations by acetylcholine, cholecystokinin and 1S,3R-ACPD in the ferret lateral geniculate and perigeniculate nuclei in vitro. Neuroscience 1997; 77:335-350.
(22.) Luthi A, McCormick DA. Modulation of a pacemaker current through Ca(2+)-induced stimulation of cAMP production. Nat Neurosci 1999; 2:634-641.
(23.) McCormick DA, Pape HC. Noradrenergic and serotonergic modulation of a hyperpolarization-activated cation current in thalamic relay neurones. J Physiol 1990; 431:319-342.
(24.) Ying S-W, Abbas SY, Harrison NL, Goldstein PA. Propofol block of I(h) contributes to the suppression of neuronal excitability and rhythmic burst firing in thalamocortical neurons. Eur J Neurosci 2006; 23:465-480.
(25.) Hughes SW, Cope DW, Blethyn KL, Crunelli V. Cellular mechanisms of the slow (<1 Hz) oscillation in thalamocortical neurons in vitro. Neuron 2002; 33:947-958.
(26.) Steriade M, Nunez A, Amzica F. A novel slow (<1 Hz) oscillation of neocortical neurons in vivo: depolarizing and hyperpolarizing components. J Neurosci 1993; 13:3252-3265.
(27.) Kroeger D, Amzica F. Hypersensitivity of the anesthesia-induced comatose brain. J Neurosci 2007; 27:10597-10607.
(28.) Steriade M. Grouping of brain rhythms in corticothalamic systems. Neuroscience 2006; 137:1087-1106.
(29.) Molle M, Marshall L, Gais S, Born J. Grouping of spindle activity during slow oscillations in human non-rapid eye movement sleep. J Neurosci 2002; 22:10941-10947.
Address for correspondence: Dr J. W. Sleigh, Waikato Clinical School, Waikato Hospital Private Bag 3200, Hamilton 3240, New Zealand.
E. C. MACKAY *, J. W. SLEIGH ([dagger]), L. J. VUSS ([double dagger]), J. P. BARNARD ([section]) Department of Anaesthesia, Waikato Hospital, Hamilton, New Zealand
* B.A., Research Student.
([dagger]) F.A.N.Z.C.A., M.D., Professor of Anaesthesia.
([double dagger]) Ph.D., Research Scientist.
([section]) F.A.N.Z.C.A., Senior Lecturer.
TABLE 1 Changes in heart rate and electroencephalogram parameters with noxious stimulus Pre-ETT Post-ETT All patients Heart rate (bpm) 64 (17) 79 (24) Response entropy 43 (20) 43 (18) 10 Hz-score 0.25 (0.18) 0.20 (0.19) Delta power ([mu]v) 315 (121) 362 (170) HFVI 3.1 (1.9) 2.7 (1.8) Low dose fentanyl patients Heart rate (bpm) 67 (11) 91 (15) Response entropy 41 (21) 47 (22) 10 Hz-score 0.22 (0.19) 0.17 (0.20) Delta power ([mu]V) 374 (166) 371 (192) HFVI 3.6 (2.4) 2.4 (0.7) Medium dose fentanyl patients Heart rate (bpm) 64 (25) 67 (34) Response entropy 44 (26) 38 (20) 10 Hz-score 0.32 (0.21) 0.26 (0.23) Delta power ([mu]V) 323 (87) 349 (158) HFVI 2.9 (1.7) 2.8 (2.3) High dose fentanyl patients Heart rate (bpm) 61 (13) 67 (14) Response entropy 43 (12) 43 (11) 10 Hz-score 0.20 (0.13) 0.17 (0.12) Delta power ([mu]V) 247 (56) 367 (176) HFVI 2.9 (1.7) 2.7 (2.0) Pre-incision Post-incision All patients Heart rate (bpm) 66 (20) 70 (23) Response entropy 39 (13) 38 (12) 10 Hz-score 0.33 (0.19) 0.27 (0.18) Delta power ([mu]v) 351 (134) 364 (130) HFVI 2.7 (2.0) 2.0 (0.5) Low dose fentanyl patients Heart rate (bpm) 74 (12) 82 (17) Response entropy 38 (10) 39 (8) 10 Hz-score 0.37 (0.21) 0.23 (0.19) Delta power ([mu]V) 347 (149) 382 (147) HFVI 2.9 (2.6) 2.0 (0.5) Medium dose fentanyl patients Heart rate (bpm) 67 (30) 78 (33) Response entropy 36 (16) 34 (14) 10 Hz-score 0.34 (0.17) 0.35 (0.15) Delta power ([mu]V) 369 (100) 346 (97) HFVI 1.9 (0.4) 1.8 (0.4) High dose fentanyl patients Heart rate (bpm) 57 (11) 57 (9) Response entropy 41 (11) 41 (14) 10 Hz-score 0.31 (0.19) 0.28 (0.19) Delta power ([mu]V) 337 (158) 365 (168) HFVI 3.3 (2.4) 2.1 (0.5) ETT = laryngoscopy/intubation, HFVI = high frequency variability index. TABLE 2 Changes in predominant EEG pattern, induced by noxious stimulus No. EEG patterns Total change Increase Decrease High 10 Hz-score 25 15 1 9 (>0.10) High HFVI (>2) 16 5 0 11 High delta (>0.03) 26 21 2 2
|Printer friendly Cite/link Email Feedback|
|Author:||MacKay, E.C.; Sleigh, J.W.; Voss, L.J.; Barnard, J.P.|
|Publication:||Anaesthesia and Intensive Care|
|Article Type:||Clinical report|
|Date:||Jan 1, 2010|
|Previous Article:||Monitoring of extubated patients: are routine arterial blood gas measurements useful and how long should patients be monitored in the intensive care...|
|Next Article:||Preoperative estimated glomerular filtration rate and RIFLE-classified postoperative acute kidney injury predict length of stay post-coronary bypass...|