- Email: [email protected]
Effect of propofol anaesthesia on the event-related potential mismatch negativity and the auditory-evoked potential N1
British Journal of Anaesthesia 89 (3): 382-8 (2002)
Effect of propofol anaesthesia on the event-related potential mismatch negativity and the auditory-evoked potential Nl T. P. Simpson1, A. R. Manara1*, N. M. Kane2, R. L. Barton3, C. A. Rowlands3 and S. R. Butler3 department of Anaesthesia and 2Department of Neurophysiology Frenchay Hospital, Frenchay Park Road, Bristol BS16 1LE, UK 3The Burden Neurological Institute, Institute, Stoke Lane, Bristol BS16 1QT, UK ^Corresponding author
Methods. Twenty-one patients received I fig ml"1 stepped increases in the target concentration of propofol using Diprifusor™ until a maximum of 6 [ig ml"1 was achieved or the patient had lost consciousness. Neurophysiological responses (MMN and N I) and the patients' level of consciousness were recorded before the administration of propofol and at a target effector site concentration of propofol of 1,2, 3, 4, and 6 [ig ml"1. Grand average evoked potentials were computed at baseline, before the administration of propofol (A); at the highest propofol concentration at which each patient was responsive (B); and at the concentration of propofol at which the patient became unconscious (C). Results. Patients lost consciousness at different target concentrations of propofol, all being unresponsive by 4 u,g ml"1. The response to the deviant stimuli used to elicit duration-shift MMN was significantly more negative than to the standard stimuli at A (mean difference 2.58 |iV, P=0.0011) but this difference was virtually abolished at point B, before the patients lost consciousness (mean difference 0.63 uV, P=ns). The amplitude of NI evoked by standard stimuli was negative compared with electrical baseline at both point A (mean amplitude -3.81 uV, P<0.00l) and at point B (mean amplitude -2.2 uV, P=0.002), but was no longer significantly different to baseline at point C (mean amplitude 0.51 uV, P=ns). The change in the mean amplitude of N l from last awake (point B) to first unconscious (point C) was also significant (mean difference in amplitude 1.69 uV, P=0.02). Conclusions. MMN is unlikely to be a clinically useful tool to detect awareness in surgical patients. In contrast, the loss of NI may identify the transition from consciousness to unconsciousness and deserves further study. BrJ Anaesth 2002; 89: 382-8 Keywords: anaesthesia, depth, awareness; anaesthetics i.v., propofol; monitoring, depth of anaesthesia Accepted for publication: March 14, 2002
As the sense of hearing has been reported to be the last to be lost and the first to be recovered during the induction of anaesthesia and the return of consciousness,1 it is not surprising that the auditory-evoked potential (AEP) has been investigated as a potential monitor of the depth
of anaesthesia. Most studies have concentrated on the middle latency (ML)AEP. The midlatency components are quantified in terms of their latency and amplitude or a ratiometric index.2"5 The MLAEP are 'exogenous' evoked potentials components as their features are determined
© The Board of Management and Trustees of the British Journal of Anaesthesia 2002
Downloaded from http://bja.oxfordjournals.org/ at Yale University on May 17, 2015
Background. Studies on the effects of anaesthesia on event-related potentials and long latency auditory-evoked potentials (AEP) are sparse. Both provide information on cortical processing and may have potential as monitors of awareness. We studied the effect of propofol on the event-related potential mismatch negativity (MMN) and the long-latency AEP N I .
Effect of propofol on MMN and Nl
Patients and methods We studied 21 ASA I or II unpremedicated patients aged between 31 and 65 yr who were to undergo elective general surgical procedures under general anaesthesia. The hospital ethics committee approved the study, and written informed consent was obtained from each patient. All subjects were free from neurological or otological disease. The study was carried out in the anaesthetic room before the start of the surgical procedure. The patients' arterial pressure, electrocardiogram, and peripheral oxygen saturation were monitored non-invasively. All patients received supplemental oxygen via a Hudson facemask at 4 litre min"1. The only drug administered for the duration of this study was propofol which was infused using a target-controlled infusion (TCI) system (Diprifusor™ TCI, Zeneca Ltd, Cheshire, UK).12 The TCI system consists of a microprocessor-driven infusion pump that adjusts the infusion rate to achieve the desired target concentration of propofol as predicted by a three-compartment pharmacokinetic model.13 The model assumes immediate mixing in blood and takes no account of the delay required to achieve drug equilibration in the brain. Thus, a delay occurs between the time the TCI system indicates that a target concentration has been achieved and the maximum pharmacodynamic effect. Using a computer simulation model of the pharmacokinetic variables, it was assumed that 5 min would be sufficient to achieve steady-state conditions between the blood and effector site compartments, and that after this period there would be minimal change in the clinical effect of the propofol. After a baseline recording of the EEG and MMN had been made, the TCI pump was started and set to achieve a target blood concentration of 1 |ig ml"1 and, after a period of 5 min for equilibration, EEG and MMN recordings were repeated. The level of consciousness was assessed at the end of the 5-min equilibration period and again after the recordings had been taken. The patients were deemed to be unconscious and unaware if they no longer obeyed commands and the eyelash reflex was absent. The procedure of infusion, equilibration, neurophysiological data collection, and assessment of consciousness was repeated at target blood concentrations of propofol of 2, 3, 4, and 6 fig ml"1. The study was stopped if the mean arterial pressure was reduced by more than 20% of the baseline value. Only when all the required recordings had been made were further drugs administered and the airway manipulated to allow the planned surgical procedure to begin.
Neurophysiology The EEG was recorded from 10 scalp electrodes (F3, Fz, F4, C3, C4, Al, A2, P3, P4, and Cz) referred to a nose electrode, with a time constant of 1.2 s and a high-frequency filter at 70 Hz. The recordings were subsequently referenced to averaged mastoids ((Al+A2)/2). The signals were amplified
383
Downloaded from http://bja.oxfordjournals.org/ at Yale University on May 17, 2015
mainly by the physical nature of the evoking stimulus. Therefore, whilst their presence establishes the integrity of the afferent auditory pathway they may not necessarily provide information on perceptual processing in the cerebral cortex or the capacity to lay down memory of the auditory experience, both of which would be important in a method for determining awareness during anaesthesia. The long latency components (>80 ms) of the AEP (LLAEP) including PI, Nl, P2, N2, and P3 are more sensitive to the psychological state of the subject. These 'endogenous' components are thought to reflect activity in areas of the cerebral cortex that are involved in the perceptual and cognitive processing of the auditory signal and have an established role in the study of human cognition.6 They are best observed when a longer stimulus is used such as a tone burst of 80-100 ms duration. As they are generated by areas of cortex at and beyond the primary projection area, they may be better indicators of the effect of anaesthetic agents on mental state than MLAEP. The Nl component of the LLAEP is not a single phenomenon but a summation of electrical activity from a number of anatomically and functionally discrete generators.7 These can be distinguished in terms of their scalp distribution, their sensitivity to the physical properties of the stimulus (frequency and duration) and the subject's attention to it. However, subcomponents of the Nl potential are independent of attention8 and this is clearly a prerequisite for a neurophysiological monitor of awareness. Mismatch negativity (MMN) is an automatic response to changes in auditory input, generated primarily in auditory areas of the temporal cortex.9 If a series of identical 'standard' tones is used to evoke Nl, an infrequent 'deviant' tone (one differing in frequency, duration, loudness, or some other physical property from the 'standard') evokes an additional negative potential that overlaps and extends beyond the negativity of Nl itself. The response is almost unaffected by the subject's attention to the evoking stimulus. Naatanen and Lyytinen10 postulated that MMN correlates with the laying down of echoic memory and that it serves to identify stimuli to which conscious attention should be directed. As it is not generated in the primary auditory cortex, but beyond in the auditory and probably frontal association cortex, it may identify the capacity of cerebral cortex to mediate mental activity. Indeed, the presence of MMN has been shown to be a sensitive predictor of the return of consciousness in patients with traumatic head injury.11 There are no studies on the effects of anaesthesia on MMN. The presence of MMN may identify auditory awareness and the capacity to lay down auditory memories during anaesthesia. The present study was, therefore, carried out to observe how Nl and MMN changed with a decreasing level of consciousness achieved by stepped increases in effector site concentration of propofol, and to determine if either measure had potential as a monitor of awareness during anaesthesia.
Simpson et al.
Statistical methods MMN corresponds to a period of greater negativity in the response to the deviant than the standard stimuli in the period following Nl. To assess whether MMN was reliably present, the difference in the mean voltage of the responses to standards and deviants was assessed using paired Student's Mests. The period for analysis was chosen by inspection of the grand average responses recorded before the administration of propofol, which is at stage A. This showed that frequency-shift MMN was present during the period 100-250 ms after the stimulus and 140-250 ms in the case of duration-shift MMN. The onset of duration-shift MMN is later than that of frequency-shift MMN because detection of the deviant cannot begin until the (longer) tone has ended. To determine whether the amplitude of Nl itself differed significantly from electrical baseline, the mean voltage of the response to standards in the period 90-135 ms after the stimulus was assessed using paired Student's Mests. The period was selected because it spans the peak of Nl in the grand average recorded before the administration of propofol.
Results Propofol dose and level of consciousness Patients lost consciousness at different target concentrations of propofol. Eight patients became unresponsive for the first time at 2 \ig ml"1, 11 at 3 (Ig ml"1, and two at 4 (0,g ml"1. In no patient was there a change in level of consciousness between the end of the equilibration period and the next increase in the dose of propofol. Thus, the target blood
concentration of propofol did not accurately predict the point at which consciousness was lost. MMN The changes in duration-shift MMN recorded from electrode Fz with respect to mastoids at times A, B, and C are shown in Figure 1. In the baseline recording, before the administration of propofol, the response to the deviants was significantly more negative than that to standards (mean difference=2.58 (iV; standard error (SE) of the difference 0.68 |iV; P=0.0011). At the target concentration of propofol at which the subjects were last aware, duration-shift MMN was already virtually abolished, that is its amplitude was no longer significantly different from zero (0.63 uV; SE=0.64 (J.V; P=ns). Once the subjects became unconscious, the response to deviants became more positive than that to standards though the difference did not reach significance. The amplitude of the MMN when the subjects were last aware did not differ significantly from its amplitude when they were first unconscious (mean difference 2.1 |iV; SE=1.11; P=ns). Frequency-shift evoked a MMN, which, as expected, began slightly earlier than that to duration-shift. Figure 2 shows the MMN to frequency-shift recorded from electrode Fz with respect to mastoids at three levels of consciousness. In the baseline recording, the response to the deviants was significantly more negative than that to the standards (mean difference=2.31 (iV; SE=0.73 UV; />=0.005). Frequencyshift evoked a MMN that was still present in the last recording made while the subjects were last conscious (mean difference^.38 uV; SE=0.56 uV; P=0.02). The response to the deviants remained more negative than that to standards after loss of consciousness but the difference was no longer significant. The variance in the amplitude of the MMN increased considerably with the onset of unconsciousness. Although its amplitude was lower than when the subjects were last conscious, the difference between the two conditions was not significant (mean difference=0.84 juV; SE=3.11;/)=ns).
Nl Nl is the peak in the response to the standard tones. For example, the 700 Hz 75 ms tones, used as standards to evoke the duration-shift MMN, evoke an Nl potential that has a latency of approximately 110 ms (Fig. 3). In the baseline recording, Nl had a mean amplitude of-3.81 (J.V (significantly above baseline: SE=0.54 U.V; P<0.0001) in the latency window 90-135 ms. Figure 3 also shows a systematic decrease in the mean amplitude of Nl with increasing target blood concentrations of propofol. The amplitude of Nl was still negative with respect to baseline at 2 ug ml"1 (1.5 \TV; SE=0.69 uV; P=0.044) but the trace became significantly more positive than baseline during the same period at 3 |ig ml"1.
384
Downloaded from http://bja.oxfordjournals.org/ at Yale University on May 17, 2015
by a Nihon Koden 7300F EEG machine and digitized at 1000 Hz and 12-bit resolution. Responses to deviants and standards were averaged separately over epochs of 512 ms until responses to 32 deviants had been collected. Frequency-shift MMN was evoked using 990 Hz standard tones and 500 Hz deviants, each with a duration of 84 ms. Duration-shift MMN was evoked with 700 Hz tones, the standards lasting 75 ms and deviants 25 ms. Fifteen per cent of stimuli were deviants, 85% standards. All stimuli were presented at an intensity of 60 dB nl and with an interstimulus interval of 750 ms. To observe the neurophysiological changes associated with the transition from consciousness to unconsciousness we calculated grand averages of the Nl and MMN responses (duration-shift and frequency-shift) in each patient at the following times: 1. A, baseline, before administration of propofol; 2. B, at the target concentration of propofol each patient was last conscious; 3. C, at the target concentration at which they were first unconscious.
Effect of propofol on MMN and Nl
140-250 ms -3-I -2-
\y
2J
100 ms Fig 1 The mean amplitude of duration-shift MMN 140-250 ms after the stimuli at different levels of consciousness. (Left) Grand average waveforms. The thin lines are responses to standard stimuli, thick lines responses to deviant stimuli. (Right) Mean amplitude of MMN with SE. A=at baseline, B=at the highest target concentration at which patients were last conscious, and C=at target concentration at which consciousness was first lost. *Responses to deviant stimuli significantly more negative than responses to standards.
100-250 ms
-3-i
10
100 ms Fig 2 The mean amplitude of frequency-shift MMN 100-250 ms after the stimuli at different levels of consciousness. Caption as Figure 1.
The mean amplitude of Nl at the highest concentration of propofol at which subjects were still conscious fell to -2.2 \iW (Fig. 4) but this remained significantly above the
electrical baseline (SE=0.65 (iV; P=0.002). Once subjects were unconscious, the mean amplitude dropped to -0.51 JIV a value not significantly different from zero (SE=0.42 (J,V,
385
Downloaded from http://bja.oxfordjournals.org/ at Yale University on May 17, 2015
10
Simpson et al.
90-135 ms Baseline/
1\
Baseline * mh 1
1 ng
ml-1y
2 [ig mh1
3 ng ml-1 ^ 4 \ig mh
—
,
1
100 \is Fig 3 The mean amplitude of Nl 90-135 ms after standard stimuli (75 ms duration 700 Hz), at increasing predicted blood concentrations of propofol. *Responses whose amplitudes are significantly different from the electrical baseline.
90-135 ms
-3-
A
B
C
10
100 ms Fig 4 The amplitude of Nl 90-135 ms after standard stimuli at different levels of consciousness. Caption as Figure 3.
/^ns). The change in the mean amplitude of Nl from last awake to first unconscious was significant (mean difference 1.69 |iV; SE=0.50 fiV; P=0.02).
The Nl, in response to the 900 Hz tones, used as standards to evoke the frequency-shift MMN showed similar changes in mean amplitude with the level of consciousness. The mean amplitude of Nl at baseline was
-4.3 JIV in the period between 90 and 135 ms after the stimulus. Although it was reduced to -1.8 |iV at the concentration of propofol when patients were last conscious, this was still significantly different from zero (SE=0.52 (iV; P=0.003). Once consciousness was lost, the mean amplitude of the signal in the 90-135 ms window became positive at +1.1 (iV but not significantly different
386
Downloaded from http://bja.oxfordjournals.org/ at Yale University on May 17, 2015
10 uV
Effect of propofol on MMN and Nl
from zero (t=l.91, P=ns). The variance in the amplitude of Nl did not increase during the same period as it did for MMN. The change in Nl amplitude from last awake to unconsciousness was significant (mean difference=2.9 |J,V; SE=0.71 uV; P=0.001).
Discussion
387
Downloaded from http://bja.oxfordjournals.org/ at Yale University on May 17, 2015
In this study we showed that both duration-shift and frequency-shift MMN were abolished by concentrations of propofol insufficient to produce unconsciousness. There were several patients in whom a clear MMN could not be elicited despite them being conscious and obeying commands. The variability of the frequency-shift MMN waveform after loss of consciousness would make it difficult to determine if it is present or absent in individual patients. Though the variability of the waveforms was somewhat less in the response to duration-shift, the MMN evoked by this form of stimulation was abolished before consciousness was lost. MMN, as recorded in this investigation, does not appear to discriminate the level of consciousness as we hypothesized on the basis of its role in cognitive processing. No tests of implicit or explicit learning or memory were used in this study, but it has been repeatedly documented that patients who are lightly anaesthetized can obey commands yet have no explicit memory of the event.14 15 If MMN is representative of short-term auditory echoic memory, then one could reasonably expect it to be present if explicit memory is being laid down. It is possible that the concentration of propofol at which MMN disappears is that at which the ability to lay down explicit or implicit auditory memory is lost, as opposed to that at which consciousness is lost. While the relationship between MMN and the formation of long-term memory has. not been studied, the disappearance of MMN before loss of consciousness would be in accord with the dissociation between awareness and the ability to establish memories. Clearly this hypothesis would require further study. The technical problems in recording MMN in individual patients also need to be addressed. Several factors contribute to the variability in the MMN responses. To evoke MMN it is necessary to use deviant stimuli that are, by definition, relatively infrequent. If recordings are to be clinically useful they must be made in the shortest possible time to enable changing levels of awareness to be detected. Almost 3 min were required to accumulate 32 epochs into the average response to deviants. An average containing so few responses may still be affected by noise induced from equipment in the theatre environment and the increases in amplitude of the background EEG as anaesthesia is induced. Increasing the number of responses to deviant stimuli would allow better resolution of the MMN response but at the cost of increasing the time required between recordings to a clinically unacceptable extent. Furthermore, since this study began, there have been preliminary reports of test-retest variation in serial recordings of MMN. 1617 These indicate
that MMN may occasionally be unrecordable even in awake, unanaesthetized subjects. Although the factors responsible for this have not been determined, the finding casts further doubt on the value of MMN as a monitor of depth of anaesthesia. In contrast, Nl evoked by the standard stimuli was less vulnerable to these sources of variance because many more responses were included in the average (approximately 200). This component stayed significantly above baseline while the patients remained conscious with both the 990 Hz 84 ms (used to evoke frequency-shift MMN) and the 700 Hz 75 ms standard tone bursts (used for duration-shift MMN). Once consciousness was lost, the waveform evoked by the 700 Hz tones fell to baseline in the period occupied previously by the Nl and became positive in response to 900 Hz tones. This is in keeping with the findings of previous reports of the effect of anaesthesia on Nl. However, in these studies the induction of anaesthesia was by a rapid bolus of i.v. anaesthetic agent and the transition period from consciousness to unconsciousness was not studied in detail.18"20 In the present study, slowly increasing the depth of anaesthesia using the TCI system enabled us to see how Nl varied with predicted blood propofol concentration and the level of consciousness. It may be argued that a longer period of equilibration than that chosen for this study is required to achieve a steadystate concentration of propofol at the effector site compartment.21 Therefore, it is possible that the concentration of propofol may have continued to increase from the end of the equilibration period to the end of data collection and may have had some effect on the neurophysiological variables being measured. While this is possible, the effects on our results should be minimal for two reasons. First, as we were using only 3- to 4-min epochs of data and the sequence of recordings was the same at each concentration of propofol, then the effect should have been consistent on each of the variables measured. Secondly, and more importantly, there was no change in the level of consciousness between the end of the equilibration period and the end of data collection in any patient. The study was primarily concerned with neurophysiological changes occurring at the transition from consciousness to unconsciousness, and not the correlation of neurophysiological variables to the predicted plasma concentration of anaesthetic agent. However, in the light of our findings it would be prudent if in future studies the time period for equilibration were increased and the incremental increase in the target concentration of propofol reduced to 0.2 or 0.5 |ig ml"1 increments, particularly in the range where consciousness is lost, that is 2-A |ag ml"1. The mean amplitude of Nl was a more sensitive predictor of the transition from consciousness to unconsciousness than was the predicted blood concentration of propofol. This is not surprising given the inter-individual pharmacokinetic and haemodynamic variability that will markedly affect each patient's response to propofol and the fact that the
Simpson et al.
Acknowledgements This research was supported by The Burden Trust and by BIOMED-2 contract BMH4-CT96-0819-COBRAIN of the European Union. Part of this work was presented at the 4th International Symposium on Memory and Awareness in Anaesthesia, Northwick Park, June 1998 and has been published in the proceedings of that meeting.
frequency, median frequency and auditory evoked potential index. Br J Anaesth 1997; 78: I8CM 6 Hillyard SA, Kutas M. Electrophysiology of cognitive processing. Annu Rev Psychol 1983; 34: 33-61 7 Naatanen R, Picton TW. The N I wave of the human electricand magnetic response to sound: a review and analysis of the component structure. Psychophysiolog/ 1987; 24: 375—425 8 Winter O, Kok A, Kenemans JL, Elton M. Auditory event related potentials to deviant stimuli during drowsiness and stage 2 sleep. Electroencephalogr Clin Neurophysiol 1995; 96: 398-412 9 Naatanen R. The mismatch negativity: a powerful tool for cognitive neuroscience. Ear Hearing 1995; 16: 6-18 10 Naatanen R, Lyytinen H. Mismatch negativity in sleep. In: Ogilvie RD, Harsh JR, eds Sleep Onset, Normal and Abnormal Processes. Washington: American Psychological Association, 1995; 339—49 I I Kane NM, Curry SH, Rowlands CA, Cummins B, Manara AR, Butler SR. Event related potentials—neurophysiological tools for predicting emergence and early outcome from traumatic coma. Intensive Care Med 1996; 22: 3 9 ^ 6 12 Kenny GNC, White M. A portable target controlled propofol infusion system. Int J Clin Monit Comput 1992; 9: 179-82 13 Marsh B, White M, Morton N, Kenny GNC. Pharmacokinetic model driven infusion of propofol in children. Br j Anaesth 1991; 67:41^8 14 Andrade J, Munglani R,* Jones JG, Baddeley AD. Cognitive performance during anaesthesia. Consciousn Cogn 1994; 3: 148-65 15 Ghoneim MM, Block Rl, Dhanaraj VJ. Interaction of a subanaesthetic concentration of isoflurane with midazolam: effects on responsiveness, learning and memory. Br J Anaesth 1998; 80: 581-7 16 Morlet D, Bertrand O, Giard MH, Bouchet P, Fischer C. Longlasting continuous monitoring of MMN. In: Abstracts of the First International Workshop on Mismatch Negativity and it Clinical Applications. Helsinki: University of Helsinki, Mulktiprint: 1998; 30 17 Morlet D, Fischer C. MMN System for coma monitoring: development and parameter optimisation. In: Summary of the Fourth COBRAIN Meeting—Progress Report to the European Commission Biomed 2 Programme. 1997; 20 18 Plourde G, Joffe D, Villemure C, Trahan M. The P3a wave of the auditory event-related potential reveals registration of pitch change during sufentanil anesthesia for cardiac surgery. Anesthesiology 1993; 78: 498-509 19 Plourde G, Picton TW. Long-latency auditory evoked potentials during general anesthesia: N I and P3 components. Anesth Analg
References
1991; 72: 342-50
1 Jones JG, Konieczko K. Hearing and memory in anaesthetised patients. Br Med J 1986; 292: 1291-2 2 Thornton C, Barrowcliffe MP, Konieczko KM, Ventham P, Dore CJ, Newton DE. The auditory evoked response as an indicator of awareness. Br J Anaesth 1989; 63: I 13-5 3 Thornton C, Sharpe RM. Evoked responses in anaesthesia. Br J Anaesth 1998; 8 1 : 771-81 4 Schwender D, Daunderer M, Mulzer S, Klasing S, Finsterer U, Peter K. Midlatency auditory evoked potentials predict movements during anesthesia with isoflurane or propofol. AnesthAnalg 1997; 85: 164-73
22 Swinhoe CF, Peacock JE, Glen JB, Reilly CS. Evaluation of the predictive performance of a 'Diprifusor' TCI system. Anaesthesia 1998; 53: 61-7
5 Doi M, Gajraj RJ, Mantzaridis H, Kenny GN. Relationship between calculated blood concentration of propofol and electrophysiological variables during emergence from anaesthesia: comparison of bispectral index, spectral edge
23 de Beer NA, van Hooff JC, Brunia CH, Cluitmans PJ, Korsten HH, Beneken JE. Midlatency auditory evoked potentials as indicators of perceptual processing during general anaesthesia. BrJ Anaesth 1996; 77: 617-24
388
20 vanHooff JC, de Beer NA, Brunia CH, Cluitmans PJ, Korsten HH. Event-related potential measures of information processing during general anaesthesia. Electroencephalogr Clin Neurophysiol
1997; 103: 268-81 21 White M, Schenkels MJ, Engbers FHM, et al. Effect-site modelling of propofol using auditory evoked potentials. Br j Anaesth 1999; 82: 333-9
Downloaded from http://bja.oxfordjournals.org/ at Yale University on May 17, 2015
Diprifusor™ system has been shown to have a median performance error of 16%, even if allowed to equilibrate.22 The later part of these waveforms becomes increasingly positive with reducing levels of consciousness. This phenomenon can be observed in the recordings of long latency event-related potentials published by previous investigators.2023 Although the waveforms are different from those which occur during normal consciousness, they indicate that some form of stimulus-related activity is still taking place in the cerebral cortex at these levels of anaesthesia. We hope to investigate the reproducibility and significance of this phenomenon in future studies. Our study may be criticized for studying mean values as opposed to individual patient data. The study was designed to investigate the effect of loss of consciousness on the selected neurophysiological variables recorded and to that end multiple variables were studied for short periods only and grand averages compared. As we believe that in Nl we have identified a possibly useful neurophysiological marker of the transition from consciousness to unconsciousness, future studies of this variable should collect data over longer periods and compare changes in individual patients. In conclusion, we have shown that the event-related potential MMN is not a robust measure of loss of consciousness and have postulated that it may represent the loss of ability to lay down auditory memory. We also demonstrated that the disappearance of the LLAEP Nl identifies the transition from consciousness to unconsciousness and warrants further investigation of its suitability of a monitor of depth of anaesthesia.
Effect of propofol anaesthesia on the event-related potential mismatch negativity and the auditory-evoked potential Nl T. P. Simpson1, A. R. Manara1*, N. M. Kane2, R. L. Barton3, C. A. Rowlands3 and S. R. Butler3 department of Anaesthesia and 2Department of Neurophysiology Frenchay Hospital, Frenchay Park Road, Bristol BS16 1LE, UK 3The Burden Neurological Institute, Institute, Stoke Lane, Bristol BS16 1QT, UK ^Corresponding author
Methods. Twenty-one patients received I fig ml"1 stepped increases in the target concentration of propofol using Diprifusor™ until a maximum of 6 [ig ml"1 was achieved or the patient had lost consciousness. Neurophysiological responses (MMN and N I) and the patients' level of consciousness were recorded before the administration of propofol and at a target effector site concentration of propofol of 1,2, 3, 4, and 6 [ig ml"1. Grand average evoked potentials were computed at baseline, before the administration of propofol (A); at the highest propofol concentration at which each patient was responsive (B); and at the concentration of propofol at which the patient became unconscious (C). Results. Patients lost consciousness at different target concentrations of propofol, all being unresponsive by 4 u,g ml"1. The response to the deviant stimuli used to elicit duration-shift MMN was significantly more negative than to the standard stimuli at A (mean difference 2.58 |iV, P=0.0011) but this difference was virtually abolished at point B, before the patients lost consciousness (mean difference 0.63 uV, P=ns). The amplitude of NI evoked by standard stimuli was negative compared with electrical baseline at both point A (mean amplitude -3.81 uV, P<0.00l) and at point B (mean amplitude -2.2 uV, P=0.002), but was no longer significantly different to baseline at point C (mean amplitude 0.51 uV, P=ns). The change in the mean amplitude of N l from last awake (point B) to first unconscious (point C) was also significant (mean difference in amplitude 1.69 uV, P=0.02). Conclusions. MMN is unlikely to be a clinically useful tool to detect awareness in surgical patients. In contrast, the loss of NI may identify the transition from consciousness to unconsciousness and deserves further study. BrJ Anaesth 2002; 89: 382-8 Keywords: anaesthesia, depth, awareness; anaesthetics i.v., propofol; monitoring, depth of anaesthesia Accepted for publication: March 14, 2002
As the sense of hearing has been reported to be the last to be lost and the first to be recovered during the induction of anaesthesia and the return of consciousness,1 it is not surprising that the auditory-evoked potential (AEP) has been investigated as a potential monitor of the depth
of anaesthesia. Most studies have concentrated on the middle latency (ML)AEP. The midlatency components are quantified in terms of their latency and amplitude or a ratiometric index.2"5 The MLAEP are 'exogenous' evoked potentials components as their features are determined
© The Board of Management and Trustees of the British Journal of Anaesthesia 2002
Downloaded from http://bja.oxfordjournals.org/ at Yale University on May 17, 2015
Background. Studies on the effects of anaesthesia on event-related potentials and long latency auditory-evoked potentials (AEP) are sparse. Both provide information on cortical processing and may have potential as monitors of awareness. We studied the effect of propofol on the event-related potential mismatch negativity (MMN) and the long-latency AEP N I .
Effect of propofol on MMN and Nl
Patients and methods We studied 21 ASA I or II unpremedicated patients aged between 31 and 65 yr who were to undergo elective general surgical procedures under general anaesthesia. The hospital ethics committee approved the study, and written informed consent was obtained from each patient. All subjects were free from neurological or otological disease. The study was carried out in the anaesthetic room before the start of the surgical procedure. The patients' arterial pressure, electrocardiogram, and peripheral oxygen saturation were monitored non-invasively. All patients received supplemental oxygen via a Hudson facemask at 4 litre min"1. The only drug administered for the duration of this study was propofol which was infused using a target-controlled infusion (TCI) system (Diprifusor™ TCI, Zeneca Ltd, Cheshire, UK).12 The TCI system consists of a microprocessor-driven infusion pump that adjusts the infusion rate to achieve the desired target concentration of propofol as predicted by a three-compartment pharmacokinetic model.13 The model assumes immediate mixing in blood and takes no account of the delay required to achieve drug equilibration in the brain. Thus, a delay occurs between the time the TCI system indicates that a target concentration has been achieved and the maximum pharmacodynamic effect. Using a computer simulation model of the pharmacokinetic variables, it was assumed that 5 min would be sufficient to achieve steady-state conditions between the blood and effector site compartments, and that after this period there would be minimal change in the clinical effect of the propofol. After a baseline recording of the EEG and MMN had been made, the TCI pump was started and set to achieve a target blood concentration of 1 |ig ml"1 and, after a period of 5 min for equilibration, EEG and MMN recordings were repeated. The level of consciousness was assessed at the end of the 5-min equilibration period and again after the recordings had been taken. The patients were deemed to be unconscious and unaware if they no longer obeyed commands and the eyelash reflex was absent. The procedure of infusion, equilibration, neurophysiological data collection, and assessment of consciousness was repeated at target blood concentrations of propofol of 2, 3, 4, and 6 fig ml"1. The study was stopped if the mean arterial pressure was reduced by more than 20% of the baseline value. Only when all the required recordings had been made were further drugs administered and the airway manipulated to allow the planned surgical procedure to begin.
Neurophysiology The EEG was recorded from 10 scalp electrodes (F3, Fz, F4, C3, C4, Al, A2, P3, P4, and Cz) referred to a nose electrode, with a time constant of 1.2 s and a high-frequency filter at 70 Hz. The recordings were subsequently referenced to averaged mastoids ((Al+A2)/2). The signals were amplified
383
Downloaded from http://bja.oxfordjournals.org/ at Yale University on May 17, 2015
mainly by the physical nature of the evoking stimulus. Therefore, whilst their presence establishes the integrity of the afferent auditory pathway they may not necessarily provide information on perceptual processing in the cerebral cortex or the capacity to lay down memory of the auditory experience, both of which would be important in a method for determining awareness during anaesthesia. The long latency components (>80 ms) of the AEP (LLAEP) including PI, Nl, P2, N2, and P3 are more sensitive to the psychological state of the subject. These 'endogenous' components are thought to reflect activity in areas of the cerebral cortex that are involved in the perceptual and cognitive processing of the auditory signal and have an established role in the study of human cognition.6 They are best observed when a longer stimulus is used such as a tone burst of 80-100 ms duration. As they are generated by areas of cortex at and beyond the primary projection area, they may be better indicators of the effect of anaesthetic agents on mental state than MLAEP. The Nl component of the LLAEP is not a single phenomenon but a summation of electrical activity from a number of anatomically and functionally discrete generators.7 These can be distinguished in terms of their scalp distribution, their sensitivity to the physical properties of the stimulus (frequency and duration) and the subject's attention to it. However, subcomponents of the Nl potential are independent of attention8 and this is clearly a prerequisite for a neurophysiological monitor of awareness. Mismatch negativity (MMN) is an automatic response to changes in auditory input, generated primarily in auditory areas of the temporal cortex.9 If a series of identical 'standard' tones is used to evoke Nl, an infrequent 'deviant' tone (one differing in frequency, duration, loudness, or some other physical property from the 'standard') evokes an additional negative potential that overlaps and extends beyond the negativity of Nl itself. The response is almost unaffected by the subject's attention to the evoking stimulus. Naatanen and Lyytinen10 postulated that MMN correlates with the laying down of echoic memory and that it serves to identify stimuli to which conscious attention should be directed. As it is not generated in the primary auditory cortex, but beyond in the auditory and probably frontal association cortex, it may identify the capacity of cerebral cortex to mediate mental activity. Indeed, the presence of MMN has been shown to be a sensitive predictor of the return of consciousness in patients with traumatic head injury.11 There are no studies on the effects of anaesthesia on MMN. The presence of MMN may identify auditory awareness and the capacity to lay down auditory memories during anaesthesia. The present study was, therefore, carried out to observe how Nl and MMN changed with a decreasing level of consciousness achieved by stepped increases in effector site concentration of propofol, and to determine if either measure had potential as a monitor of awareness during anaesthesia.
Simpson et al.
Statistical methods MMN corresponds to a period of greater negativity in the response to the deviant than the standard stimuli in the period following Nl. To assess whether MMN was reliably present, the difference in the mean voltage of the responses to standards and deviants was assessed using paired Student's Mests. The period for analysis was chosen by inspection of the grand average responses recorded before the administration of propofol, which is at stage A. This showed that frequency-shift MMN was present during the period 100-250 ms after the stimulus and 140-250 ms in the case of duration-shift MMN. The onset of duration-shift MMN is later than that of frequency-shift MMN because detection of the deviant cannot begin until the (longer) tone has ended. To determine whether the amplitude of Nl itself differed significantly from electrical baseline, the mean voltage of the response to standards in the period 90-135 ms after the stimulus was assessed using paired Student's Mests. The period was selected because it spans the peak of Nl in the grand average recorded before the administration of propofol.
Results Propofol dose and level of consciousness Patients lost consciousness at different target concentrations of propofol. Eight patients became unresponsive for the first time at 2 \ig ml"1, 11 at 3 (Ig ml"1, and two at 4 (0,g ml"1. In no patient was there a change in level of consciousness between the end of the equilibration period and the next increase in the dose of propofol. Thus, the target blood
concentration of propofol did not accurately predict the point at which consciousness was lost. MMN The changes in duration-shift MMN recorded from electrode Fz with respect to mastoids at times A, B, and C are shown in Figure 1. In the baseline recording, before the administration of propofol, the response to the deviants was significantly more negative than that to standards (mean difference=2.58 (iV; standard error (SE) of the difference 0.68 |iV; P=0.0011). At the target concentration of propofol at which the subjects were last aware, duration-shift MMN was already virtually abolished, that is its amplitude was no longer significantly different from zero (0.63 uV; SE=0.64 (J.V; P=ns). Once the subjects became unconscious, the response to deviants became more positive than that to standards though the difference did not reach significance. The amplitude of the MMN when the subjects were last aware did not differ significantly from its amplitude when they were first unconscious (mean difference 2.1 |iV; SE=1.11; P=ns). Frequency-shift evoked a MMN, which, as expected, began slightly earlier than that to duration-shift. Figure 2 shows the MMN to frequency-shift recorded from electrode Fz with respect to mastoids at three levels of consciousness. In the baseline recording, the response to the deviants was significantly more negative than that to the standards (mean difference=2.31 (iV; SE=0.73 UV; />=0.005). Frequencyshift evoked a MMN that was still present in the last recording made while the subjects were last conscious (mean difference^.38 uV; SE=0.56 uV; P=0.02). The response to the deviants remained more negative than that to standards after loss of consciousness but the difference was no longer significant. The variance in the amplitude of the MMN increased considerably with the onset of unconsciousness. Although its amplitude was lower than when the subjects were last conscious, the difference between the two conditions was not significant (mean difference=0.84 juV; SE=3.11;/)=ns).
Nl Nl is the peak in the response to the standard tones. For example, the 700 Hz 75 ms tones, used as standards to evoke the duration-shift MMN, evoke an Nl potential that has a latency of approximately 110 ms (Fig. 3). In the baseline recording, Nl had a mean amplitude of-3.81 (J.V (significantly above baseline: SE=0.54 U.V; P<0.0001) in the latency window 90-135 ms. Figure 3 also shows a systematic decrease in the mean amplitude of Nl with increasing target blood concentrations of propofol. The amplitude of Nl was still negative with respect to baseline at 2 ug ml"1 (1.5 \TV; SE=0.69 uV; P=0.044) but the trace became significantly more positive than baseline during the same period at 3 |ig ml"1.
384
Downloaded from http://bja.oxfordjournals.org/ at Yale University on May 17, 2015
by a Nihon Koden 7300F EEG machine and digitized at 1000 Hz and 12-bit resolution. Responses to deviants and standards were averaged separately over epochs of 512 ms until responses to 32 deviants had been collected. Frequency-shift MMN was evoked using 990 Hz standard tones and 500 Hz deviants, each with a duration of 84 ms. Duration-shift MMN was evoked with 700 Hz tones, the standards lasting 75 ms and deviants 25 ms. Fifteen per cent of stimuli were deviants, 85% standards. All stimuli were presented at an intensity of 60 dB nl and with an interstimulus interval of 750 ms. To observe the neurophysiological changes associated with the transition from consciousness to unconsciousness we calculated grand averages of the Nl and MMN responses (duration-shift and frequency-shift) in each patient at the following times: 1. A, baseline, before administration of propofol; 2. B, at the target concentration of propofol each patient was last conscious; 3. C, at the target concentration at which they were first unconscious.
Effect of propofol on MMN and Nl
140-250 ms -3-I -2-
\y
2J
100 ms Fig 1 The mean amplitude of duration-shift MMN 140-250 ms after the stimuli at different levels of consciousness. (Left) Grand average waveforms. The thin lines are responses to standard stimuli, thick lines responses to deviant stimuli. (Right) Mean amplitude of MMN with SE. A=at baseline, B=at the highest target concentration at which patients were last conscious, and C=at target concentration at which consciousness was first lost. *Responses to deviant stimuli significantly more negative than responses to standards.
100-250 ms
-3-i
10
100 ms Fig 2 The mean amplitude of frequency-shift MMN 100-250 ms after the stimuli at different levels of consciousness. Caption as Figure 1.
The mean amplitude of Nl at the highest concentration of propofol at which subjects were still conscious fell to -2.2 \iW (Fig. 4) but this remained significantly above the
electrical baseline (SE=0.65 (iV; P=0.002). Once subjects were unconscious, the mean amplitude dropped to -0.51 JIV a value not significantly different from zero (SE=0.42 (J,V,
385
Downloaded from http://bja.oxfordjournals.org/ at Yale University on May 17, 2015
10
Simpson et al.
90-135 ms Baseline/
1\
Baseline * mh 1
1 ng
ml-1y
2 [ig mh1
3 ng ml-1 ^ 4 \ig mh
—
,
1
100 \is Fig 3 The mean amplitude of Nl 90-135 ms after standard stimuli (75 ms duration 700 Hz), at increasing predicted blood concentrations of propofol. *Responses whose amplitudes are significantly different from the electrical baseline.
90-135 ms
-3-
A
B
C
10
100 ms Fig 4 The amplitude of Nl 90-135 ms after standard stimuli at different levels of consciousness. Caption as Figure 3.
/^ns). The change in the mean amplitude of Nl from last awake to first unconscious was significant (mean difference 1.69 |iV; SE=0.50 fiV; P=0.02).
The Nl, in response to the 900 Hz tones, used as standards to evoke the frequency-shift MMN showed similar changes in mean amplitude with the level of consciousness. The mean amplitude of Nl at baseline was
-4.3 JIV in the period between 90 and 135 ms after the stimulus. Although it was reduced to -1.8 |iV at the concentration of propofol when patients were last conscious, this was still significantly different from zero (SE=0.52 (iV; P=0.003). Once consciousness was lost, the mean amplitude of the signal in the 90-135 ms window became positive at +1.1 (iV but not significantly different
386
Downloaded from http://bja.oxfordjournals.org/ at Yale University on May 17, 2015
10 uV
Effect of propofol on MMN and Nl
from zero (t=l.91, P=ns). The variance in the amplitude of Nl did not increase during the same period as it did for MMN. The change in Nl amplitude from last awake to unconsciousness was significant (mean difference=2.9 |J,V; SE=0.71 uV; P=0.001).
Discussion
387
Downloaded from http://bja.oxfordjournals.org/ at Yale University on May 17, 2015
In this study we showed that both duration-shift and frequency-shift MMN were abolished by concentrations of propofol insufficient to produce unconsciousness. There were several patients in whom a clear MMN could not be elicited despite them being conscious and obeying commands. The variability of the frequency-shift MMN waveform after loss of consciousness would make it difficult to determine if it is present or absent in individual patients. Though the variability of the waveforms was somewhat less in the response to duration-shift, the MMN evoked by this form of stimulation was abolished before consciousness was lost. MMN, as recorded in this investigation, does not appear to discriminate the level of consciousness as we hypothesized on the basis of its role in cognitive processing. No tests of implicit or explicit learning or memory were used in this study, but it has been repeatedly documented that patients who are lightly anaesthetized can obey commands yet have no explicit memory of the event.14 15 If MMN is representative of short-term auditory echoic memory, then one could reasonably expect it to be present if explicit memory is being laid down. It is possible that the concentration of propofol at which MMN disappears is that at which the ability to lay down explicit or implicit auditory memory is lost, as opposed to that at which consciousness is lost. While the relationship between MMN and the formation of long-term memory has. not been studied, the disappearance of MMN before loss of consciousness would be in accord with the dissociation between awareness and the ability to establish memories. Clearly this hypothesis would require further study. The technical problems in recording MMN in individual patients also need to be addressed. Several factors contribute to the variability in the MMN responses. To evoke MMN it is necessary to use deviant stimuli that are, by definition, relatively infrequent. If recordings are to be clinically useful they must be made in the shortest possible time to enable changing levels of awareness to be detected. Almost 3 min were required to accumulate 32 epochs into the average response to deviants. An average containing so few responses may still be affected by noise induced from equipment in the theatre environment and the increases in amplitude of the background EEG as anaesthesia is induced. Increasing the number of responses to deviant stimuli would allow better resolution of the MMN response but at the cost of increasing the time required between recordings to a clinically unacceptable extent. Furthermore, since this study began, there have been preliminary reports of test-retest variation in serial recordings of MMN. 1617 These indicate
that MMN may occasionally be unrecordable even in awake, unanaesthetized subjects. Although the factors responsible for this have not been determined, the finding casts further doubt on the value of MMN as a monitor of depth of anaesthesia. In contrast, Nl evoked by the standard stimuli was less vulnerable to these sources of variance because many more responses were included in the average (approximately 200). This component stayed significantly above baseline while the patients remained conscious with both the 990 Hz 84 ms (used to evoke frequency-shift MMN) and the 700 Hz 75 ms standard tone bursts (used for duration-shift MMN). Once consciousness was lost, the waveform evoked by the 700 Hz tones fell to baseline in the period occupied previously by the Nl and became positive in response to 900 Hz tones. This is in keeping with the findings of previous reports of the effect of anaesthesia on Nl. However, in these studies the induction of anaesthesia was by a rapid bolus of i.v. anaesthetic agent and the transition period from consciousness to unconsciousness was not studied in detail.18"20 In the present study, slowly increasing the depth of anaesthesia using the TCI system enabled us to see how Nl varied with predicted blood propofol concentration and the level of consciousness. It may be argued that a longer period of equilibration than that chosen for this study is required to achieve a steadystate concentration of propofol at the effector site compartment.21 Therefore, it is possible that the concentration of propofol may have continued to increase from the end of the equilibration period to the end of data collection and may have had some effect on the neurophysiological variables being measured. While this is possible, the effects on our results should be minimal for two reasons. First, as we were using only 3- to 4-min epochs of data and the sequence of recordings was the same at each concentration of propofol, then the effect should have been consistent on each of the variables measured. Secondly, and more importantly, there was no change in the level of consciousness between the end of the equilibration period and the end of data collection in any patient. The study was primarily concerned with neurophysiological changes occurring at the transition from consciousness to unconsciousness, and not the correlation of neurophysiological variables to the predicted plasma concentration of anaesthetic agent. However, in the light of our findings it would be prudent if in future studies the time period for equilibration were increased and the incremental increase in the target concentration of propofol reduced to 0.2 or 0.5 |ig ml"1 increments, particularly in the range where consciousness is lost, that is 2-A |ag ml"1. The mean amplitude of Nl was a more sensitive predictor of the transition from consciousness to unconsciousness than was the predicted blood concentration of propofol. This is not surprising given the inter-individual pharmacokinetic and haemodynamic variability that will markedly affect each patient's response to propofol and the fact that the
Simpson et al.
Acknowledgements This research was supported by The Burden Trust and by BIOMED-2 contract BMH4-CT96-0819-COBRAIN of the European Union. Part of this work was presented at the 4th International Symposium on Memory and Awareness in Anaesthesia, Northwick Park, June 1998 and has been published in the proceedings of that meeting.
frequency, median frequency and auditory evoked potential index. Br J Anaesth 1997; 78: I8CM 6 Hillyard SA, Kutas M. Electrophysiology of cognitive processing. Annu Rev Psychol 1983; 34: 33-61 7 Naatanen R, Picton TW. The N I wave of the human electricand magnetic response to sound: a review and analysis of the component structure. Psychophysiolog/ 1987; 24: 375—425 8 Winter O, Kok A, Kenemans JL, Elton M. Auditory event related potentials to deviant stimuli during drowsiness and stage 2 sleep. Electroencephalogr Clin Neurophysiol 1995; 96: 398-412 9 Naatanen R. The mismatch negativity: a powerful tool for cognitive neuroscience. Ear Hearing 1995; 16: 6-18 10 Naatanen R, Lyytinen H. Mismatch negativity in sleep. In: Ogilvie RD, Harsh JR, eds Sleep Onset, Normal and Abnormal Processes. Washington: American Psychological Association, 1995; 339—49 I I Kane NM, Curry SH, Rowlands CA, Cummins B, Manara AR, Butler SR. Event related potentials—neurophysiological tools for predicting emergence and early outcome from traumatic coma. Intensive Care Med 1996; 22: 3 9 ^ 6 12 Kenny GNC, White M. A portable target controlled propofol infusion system. Int J Clin Monit Comput 1992; 9: 179-82 13 Marsh B, White M, Morton N, Kenny GNC. Pharmacokinetic model driven infusion of propofol in children. Br j Anaesth 1991; 67:41^8 14 Andrade J, Munglani R,* Jones JG, Baddeley AD. Cognitive performance during anaesthesia. Consciousn Cogn 1994; 3: 148-65 15 Ghoneim MM, Block Rl, Dhanaraj VJ. Interaction of a subanaesthetic concentration of isoflurane with midazolam: effects on responsiveness, learning and memory. Br J Anaesth 1998; 80: 581-7 16 Morlet D, Bertrand O, Giard MH, Bouchet P, Fischer C. Longlasting continuous monitoring of MMN. In: Abstracts of the First International Workshop on Mismatch Negativity and it Clinical Applications. Helsinki: University of Helsinki, Mulktiprint: 1998; 30 17 Morlet D, Fischer C. MMN System for coma monitoring: development and parameter optimisation. In: Summary of the Fourth COBRAIN Meeting—Progress Report to the European Commission Biomed 2 Programme. 1997; 20 18 Plourde G, Joffe D, Villemure C, Trahan M. The P3a wave of the auditory event-related potential reveals registration of pitch change during sufentanil anesthesia for cardiac surgery. Anesthesiology 1993; 78: 498-509 19 Plourde G, Picton TW. Long-latency auditory evoked potentials during general anesthesia: N I and P3 components. Anesth Analg
References
1991; 72: 342-50
1 Jones JG, Konieczko K. Hearing and memory in anaesthetised patients. Br Med J 1986; 292: 1291-2 2 Thornton C, Barrowcliffe MP, Konieczko KM, Ventham P, Dore CJ, Newton DE. The auditory evoked response as an indicator of awareness. Br J Anaesth 1989; 63: I 13-5 3 Thornton C, Sharpe RM. Evoked responses in anaesthesia. Br J Anaesth 1998; 8 1 : 771-81 4 Schwender D, Daunderer M, Mulzer S, Klasing S, Finsterer U, Peter K. Midlatency auditory evoked potentials predict movements during anesthesia with isoflurane or propofol. AnesthAnalg 1997; 85: 164-73
22 Swinhoe CF, Peacock JE, Glen JB, Reilly CS. Evaluation of the predictive performance of a 'Diprifusor' TCI system. Anaesthesia 1998; 53: 61-7
5 Doi M, Gajraj RJ, Mantzaridis H, Kenny GN. Relationship between calculated blood concentration of propofol and electrophysiological variables during emergence from anaesthesia: comparison of bispectral index, spectral edge
23 de Beer NA, van Hooff JC, Brunia CH, Cluitmans PJ, Korsten HH, Beneken JE. Midlatency auditory evoked potentials as indicators of perceptual processing during general anaesthesia. BrJ Anaesth 1996; 77: 617-24
388
20 vanHooff JC, de Beer NA, Brunia CH, Cluitmans PJ, Korsten HH. Event-related potential measures of information processing during general anaesthesia. Electroencephalogr Clin Neurophysiol
1997; 103: 268-81 21 White M, Schenkels MJ, Engbers FHM, et al. Effect-site modelling of propofol using auditory evoked potentials. Br j Anaesth 1999; 82: 333-9
Downloaded from http://bja.oxfordjournals.org/ at Yale University on May 17, 2015
Diprifusor™ system has been shown to have a median performance error of 16%, even if allowed to equilibrate.22 The later part of these waveforms becomes increasingly positive with reducing levels of consciousness. This phenomenon can be observed in the recordings of long latency event-related potentials published by previous investigators.2023 Although the waveforms are different from those which occur during normal consciousness, they indicate that some form of stimulus-related activity is still taking place in the cerebral cortex at these levels of anaesthesia. We hope to investigate the reproducibility and significance of this phenomenon in future studies. Our study may be criticized for studying mean values as opposed to individual patient data. The study was designed to investigate the effect of loss of consciousness on the selected neurophysiological variables recorded and to that end multiple variables were studied for short periods only and grand averages compared. As we believe that in Nl we have identified a possibly useful neurophysiological marker of the transition from consciousness to unconsciousness, future studies of this variable should collect data over longer periods and compare changes in individual patients. In conclusion, we have shown that the event-related potential MMN is not a robust measure of loss of consciousness and have postulated that it may represent the loss of ability to lay down auditory memory. We also demonstrated that the disappearance of the LLAEP Nl identifies the transition from consciousness to unconsciousness and warrants further investigation of its suitability of a monitor of depth of anaesthesia.