1Department of Psychology, University of Sheffield, Western Bank, Sheffield S10 2TP, UK. 2University Department of Anaesthetics, Northern General Hospital, Sheffield S5 7AU, UK*Corresponding author
Accepted for publication: September 28, 2000
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Br J Anaesth 2001; 86: 18995
Keywords: memory; anaesthetics i.v., propofol; monitoring, bispectral index
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
This study investigated the relative importance of dose of anaesthetic and stimulation for learning during propofol infusion. We tested patients immediately before their scheduled operations, assuming that they would be more anxious and therefore more likely to learn than volunteers who, in a previous study, had shown no memory for words presented during propofol sedation.3 We manipulated the depth of anaesthesia experimentally by presenting words to one group of patients receiving a small, sedative dose of propofol and to another group receiving a light anaesthetic dose of propofol. We used the bispectral index4 to check that these depth manipulations were effective. We assessed the effect of physical stimulation by comparing memory for words played during anaesthesia without stimulation with memory for words played during tracheal intubation with the same dose of anaesthetic. Memory for the words was assessed on recovery using recognition tests plus a category generation test and a preference rating task. These latter tests have been shown to reveal memory even in the absence of conscious recognition.5 6 Category generation was used for comparison with previous demonstrations of learning during general anaesthesia.7 8 Preference ratings have also been used to demonstrate learning during sedation9 and anaesthesia,10 but were chosen specifically because they are sensitive to emotional changes caused by levels of learning that are insufficient to produce the cognitive changes that are picked up by more conventional tests of implicit memory.11 12
![]() |
Methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The stimuli for the category generation task were 30 repetitions of the words yellow banana green pear, which have been used in previous demonstrations of learning during anaesthesia.7 8 These stimuli were played during the propofol infusion. In the test (on recovery), patients were asked to say the first three fruits and first three colours that came to mind. Implicit memory is revealed in this task as an increased tendency to respond with the previously presented examples of fruits and colours.
The stimuli for the preference rating task were 12 repetitions of 10 nonsense (Finnish) words. One list of 10 Finnish words was presented before propofol infusion and a second list, containing a different set of 10 Finnish words, was presented during propofol infusion. Implicit memory increases liking or preference, a phenomenon known as the mere exposure effect;12 patients were therefore asked to rate their liking or preference for presented and distractor Finnish words. Pilot-testing with a sample of undergraduates validated the preference rating test as a test of implicit memory and provided an estimate of effect size for the test. Explicit, or conscious, memory was tested by asking patients a standardized set of questions about their recall of peri-operative events, and by asking them to rate their recognition of the Finnish words.
The different memory tests necessitated different control procedures. For the preference rating test, baseline measures were provided by asking patients to rate distractor words and presented words, the use of word lists as distractor or presented being fully counterbalanced across patients. For the category generation task, a baseline measure was provided by asking a control group of patients to generate examples of fruits and colours without prior exposure to the stimulus words.
Pre-operative procedure
Approval was granted from the North Sheffield Local Research Ethics Committee. We obtained written informed consent from 37 females, ASA status I and II, attending the Northern General Hospital for laparoscopy. Exclusion criteria were perceptible hearing loss, language difficulties, neurological dysfunction, treatment with drugs affecting the central nervous system, and the ability to speak or understand Finnish.
Patients were unpremedicated and entered the anaesthetic room approximately 25 min before the scheduled start of their operation. Bispectral index (BIS) monitoring was commenced with an Aspect-1000 monitor (software version 2.51; Aspect Medical Systems, Framingham, MA, USA) with bifrontal montage (F7, F8, reference Fp2). BIS was monitored continually throughout the study period. Patients were asked to lie comfortably and the experimenter explained that a list of Finnish words would shortly be played and that they should try hard to remember as many words as possible. They were reassured that they were not expected to understand the words and that they should concentrate on the sounds of the words. They were told that their memory for the words, and for some more words presented during anaesthesia, would be assessed when they came round from the anaesthetic. The first list of Finnish words was played via headphones (Sony MDR-CD170) using a portable compact disk player (Sanyo MCD-Z37L). This pre-operative list of words served to make the post-operative memory tests more meaningful to patients by ensuring that they remembered at least some of the stimulus words. The experimenters (LE and CH) were blinded to the content of the Finnish tracks.
An i.v. cannula was then inserted for the infusion of propofol and a second i.v. cannula was inserted into the contralateral arm for the collection of blood samples.
Patients were randomized to one of three groups. In one patient, memory testing could not be carried out and she was replaced. The preference rating task was designed to be more sensitive than tests used previously in studies of learning during anaesthesia. As it had not been used before with patients, we conducted a pilot study (described in the Appendix) to estimate its power to detect memory under conditions of altered awareness. Pilot data gave an effect size of 0.74 for the preference rating memory test under conditions of divided attention; we therefore estimated power for detecting implicit memory as 0.99 for the whole patient sample and 0.73 for single groups.
Group 1
Patients in group 1 received a sedating dose of propofol by computer-controlled infusion using a Graseby Diprifuser pump to a target concentration of 2 µg ml1. After 10 min, a blood sample was taken for determination of plasma propofol concentration. The experimenter then instructed patients to listen carefully and try to remember the Finnish words, and then played the second list of Finnish words. After a break of approximately 1 min, the experimenter repeated the instruction to listen carefully and remember the words, then played the list of category examples. After this, the target propofol infusion was increased to 5 µg ml1 to achieve full anaesthesia. Patients then received vecuronium 5 mg, their tracheas were intubated and the patients were transferred to the operating theatre for surgery. All patients then received fentanyl 1 µg ml1 i.v. and anaesthesia was continued with propofol to a target concentration of 5 µg ml1 and artificial ventilation with 60% nitrous oxide and 40% oxygen. At the end of surgery, residual neuromuscular block was reversed with neostigmine and glycopyrollate. Propofol was turned off and patients were woken as normal before being taken to the recovery ward.
Groups 2 and 3
Patients in groups 2 and 3 received a higher initial dose of propofol to a target concentration of 5 µg ml1. After 10 min, a blood sample was taken for propofol assay. Patients in group 2 received the second list of Finnish words and patients in group 3 received the list of category examples. All patients then received vecuronium 5 mg and had their tracheas intubated. From the start of laryngoscopy, patients in group 2 received the category examples and patients in group 3 received the Finnish words. Anaesthesia then continued as for group 1. Only one list was presented during intubation because we suspected that the stimulation caused by intubation would not last for the time taken to present two lists.
Plasma analysis
Blood taken for propofol assay was stored at 4°C in oxalate and the assay was performed by high-performance liquid chromatography with fluorescence detection.13
Post-operative procedure
Patients attempted the memory tests in the recovery room as soon as possible after surgery (within 2 h of recovery). Post-operative analgesia was provided with i.v. boluses of morphine 2 mg as required. Before the patients attempted the tests, we ensured that they were comfortable and felt able to sit up in bed and perform them. Memory was tested as follows. First, all patients were asked to name the first three examples of vegetables (for practice), fruits and colours that came to mind. Then they were asked if they could remember anything about their operation or about events occurring immediately before their operation. They were also asked whether they remembered dreaming during this period. Questioning became progressively more specific: patients were asked whether they remembered listening to any words before or during their operation, hearing the names of any vegetables, fruits or colours, hearing any Finnish or other foreign words, and finally whether they recognized any of the exemplars they had just generated or examples from the phrase yellow banana green pear. Responses were written down by the experimenter.
The Finnish memory tests came next. For the implicit memory test, patients were asked to decide whether each word referred to something good or nice, or to something bad or nasty. As they did not understand Finnish, they were asked to guess, using the strategy of assuming that a word refers to something nice if it sounds nice and to something unpleasant if it sounds unpleasant. These instructions deliberately implied that there was a correct answer, even though patients were being asked to guess, to encourage diversity of responses.11 For the recognition task, patients were asked to decide which words they remembered hearing earlier in the procedure, regardless of whether they came before or during the propofol infusion. Patients wrote down their own responses for this phase of testing. Patients heard the test list twice, rating their preference or recognition of each word on a 4-point scale. Test order was counterbalanced, i.e. half the patients in each group rated recognition on the first presentation of the test list and rated preference on the second presentation; the other half rated preference first and then recognition. On the second presentation, patients were asked to ignore the fact that they had just heard the entire list once. Ideally, we would have tested preference for half the words and recognition for the other half, but this would have required more stimuli and it is not known how initial memory load affects subsequent explicit and implicit memory under these conditions.
Control group for the category generation test
A control group of a further 12 patients was needed for the category generation test of implicit memory. This group was selected from patients undergoing laparoscopy on the same surgical lists as those in the experimental groups, and fulfilling the same criteria, but whose scheduled operation time meant that they could not be anaesthetized for the extended period required by the experimental conditions.
Statistical methods
BIS readings were recorded at the start and end of each memory list, and a mean was calculated for each list for each patient. Two-way mixed measures analysis of variance was used to compare the mean BIS values for each list across the three groups, followed by planned t-tests. One-way independent measures analysis of variance was used to compare propofol blood concentrations across groups, again followed by planned t-tests.
Category generation test performance was scored as the number of hits (responses with one of the words yellow, banana, green or pear). The independent samples t-test was used to compare mean hits in the experimental groups with mean hits in the control group, and to compare across experimental groups. The preference rating and recognition tasks were scored from 3 for a response of definitely remember or very nice response to 0 for a response of definitely do not remember or very nasty response. Dependent samples t-tests were used to compare scores for presented words with those for distractor words, and to compare intraoperative learning, if found, across groups. To maximize the chance of detecting memory, we did not correct the t-test results to allow for multiple testing. Statistical significance was assessed with = 0.05 and all tests were two-tailed. All analyses were performed using StatView 5.0.
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
BIS was significantly reduced in all three experimental groups after starting the propofol infusion (P<0.01) (Table 1). BIS was significantly lower in the two groups receiving propofol 5 µg ml1 infusion than in the group receiving 2 µg ml1 infusion (P<0.001). In all patients, we noted an increase in heart rate after intubation. However, BIS scores were not significantly altered by intubation. Blood propofol concentrations were significantly higher in groups 2 and 3 than in group 1.
|
Recognition scores for the Finnish words showed explicit memory for the list presented before propofol infusion began (P<0.001), but no memory for the list presented during propofol infusion (Table 2).
|
|
To maximize power, we tested implicit and explicit memory for the whole set of Finnish words. However, the second presentation of the test list was possibly a less sensitive memory test than the first, because patients had to distinguish words they had heard earlier from the whole set of words that they had just heard. We therefore analysed the responses of each group to the first presentation of the test list, i.e. the preference rating task for half the patients and the recognition task for the other half in each group. Groups 1 and 2 showed no difference between ratings for the words presented during propofol and distractors. Group 3 had marginally higher ratings for the words presented during intubation with propofol infusion than the distractors (P=0.09).
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
We tested implicit and explicit memory for category examples and Finnish words presented during propofol infusion. Overall, there was no evidence of explicit or implicit memory for category examples presented during propofol infusion. There was a suggestion of implicit memory for the examples presented during intubation compared with anaesthesia without intubation, but this result must be treated with caution because (i) it was only marginally significant and would have been non-significant after correcting for multiple t-testing; (ii) the hit rate in the intubation condition was not significantly higher than for the control group. We observed explicit and implicit memory for the Finnish words presented before propofol infusion, despite the intervening anaesthesia and surgery. Overall, there was no memory for the Finnish words presented during propofol infusion, but again there was a weak trend towards memory for words presented during intubation. Further investigation of the effects of stimulation on learning may be worthwhile.
We obtained absolutely no evidence for learning during propofol infusion without stimulation. This study incorporated methodological improvements upon many previous studies. We used a single anaesthetic agent during stimulus presentation and tested patients scheduled for the same operation, thereby minimizing noise compared with opportunistic studies that tested patients receiving various drugs. Recall tests tend to be more difficult than recognition tests or implicit memory tests because they provide no information about the words to be remembered. We tried to make the explicit and implicit memory tests equally sensitive by using recognition for the explicit tests;14 therefore failure to observe explicit memory shows a genuine lack of explicit, conscious learning and not merely failure on a difficult test. We piloted the Finnish memory tests to confirm that the recognition test measured explicit memory and the preference rating test measured implicit memory. We manipulated depth of anaesthesia experimentally and measured it using the bispectral index.
A previous study has shown implicit memory for words presented to volunteers receiving propofol,15 but the infusion dose of propofol was considerably smaller (1.27 mg kg1 h1) than the 2 µg ml1 used in the present study. Volunteers failed to learn words presented during infusion of larger subanaesthetic doses of propofol.3 16 In contrast, studies with patients showed learning during surgery with sedative infusions of propofol,17 18 and general anaesthetic regimes that included propofol7 1822 (though some have shown no learning).2325
Volunteers may learn relatively little because they are not anxious and do not attend carefully to the stimuli while they are sedated (although lack of attention should have little effect on implicit memory). We assumed that the patients in our study would be more anxious and more attentive, yet they learned nothing even during conscious sedation. The evidence from other studies that patients can learn during surgery with propofol sedation or anaesthesia suggests that surgery may enable learning in anaesthetic conditions that would otherwise prevent learning. The present finding that patients showed marginally significant amounts of memory for words presented during intubation yet no memory for words presented during sedation, tentatively supports the hypothesis that physical stimulation facilitates learning. A stronger relationship may have been found if the period of stimulation had been longer. A recent study of learning during trauma surgery26 demonstrated that depth of anaesthesia (BIS) explained a modest portion of the variance in memory for intraoperative stimuli. We speculate that levels of physical stimulation would explain another portion of the memory variance, perhaps through increases in concentrations of circulating catecholamines.
Any conclusions must be considered in the light of the several weaknesses in the study. To avoid undue disruption to operating theatre schedules, we did not subject the control group for the category generation task to the full pre-operative sedation protocol. Rather, we selected patients for the control group whose scheduled operation time made them unsuitable for the full procedure. Although a fully randomized group would have been preferable, we are reassured by the fact that our control groups response rate (1.75) was similar to that reported by Jelicic and colleagues8 (1.84), though both differed from the baseline rate in a previous study7 of patients undergoing major surgery (0.79). Time constraints prevented us from testing anxiety levels but future research could investigate the effect of pre-operative anxiety on learning during anaesthesia. Fifteen patients received morphine before attempting the memory tests. Morphine may have affected their performance on the tests, as might the pain they would have experienced without morphine. Either way, it would have been unethical to withhold analgesia. We presented the full test list of Finnish words twice, once for the preference test and once for the recognition test. Ideally, we would have tested preference for half the words and recognition for the other half, but this would have required more stimuli and it is not known how initial memory load affects subsequent explicit and implicit memory under these conditions. Despite these weaknesses, we conclude that sedation and anaesthesia with propofol prevents learning. We observed a trend towards memory for stimuli presented during the brief stimulation caused by tracheal intubation. We therefore speculate that more prolonged stimulation may facilitate learning during anaesthesia.
![]() |
Appendix |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Preference rating and recognition tasks
Three lists of ten disyllabic Finnish words were recorded onto compact disk, one list per track. Each list was repeated 12 times at the rate of one word per 1.5 s hence a total list length of 3 min. A single test list, containing all 30 words in random order, was recorded onto a separate compact disk track with a 4 s interval between each word and the next. The response sheet for the preference rating task contained five columns, one listing the 30 words and the other four headed very nice/good, slightly nice/good, slightly nasty/bad and very nasty/bad. On the response sheet for the recognition task, columns 25 were headed definitely remember, seems familiar, seems unfamiliar and definitely do not remember. A score of 3 points was given for each definitely remember or very nice response, two points for each seems familiar or slightly nice response, 1 point for each seems unfamiliar or slightly nasty response, and 0 points for definitely do not remember or very nasty responses.
Patients received one list of Finnish words before propofol infusion and a second list during propofol infusion. The third list provided distractors for the memory tests; the use of lists before propofol, during propofol or distractor was fully counterbalanced. On recovery, patients rated their preference and recognition of all 30 words.
Pilot testing
The Finnish stimuli and tests were pilot-tested on 12 undergraduate volunteers who listened to one list (repeated 12 times) with instructions to attend fully and remember the sounds of the words, and to a second list (repeated 12 times) under conditions of divided attention, the second list being presented as background distraction to a demanding visual task. After a 10 min filled retention interval, volunteers listened to the test list and rated their preference for each word. They then listened again to the test list and rated their recognition of the words. Two-factor analysis of variance with repeated measures on test (preference, recognition) and condition (focused attention, divided attention) showed a significant interaction between test and condition (P<0.01). Divided attention reduced recognition test scores compared with focused attention (paired t-test, P<0.001) but did not affect preference ratings (P>0.10). Ratings in each condition on each test exceeded the relevant baseline ratings to distractor words (paired t-test, P<0.05 for each comparison). These results show that the recognition test is a good measure of explicit memory, because it was sensitive to the manipulation of attention, and that the preference rating task is a good measure of implicit memory, because it was insensitive to the manipulation of attention yet showed higher ratings for presented words than distractors.
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
2 Ghoneim MM, Block RI. Learning and memory during general anesthesia: an update. Anesthesiology 1997; 87: 387410[ISI][Medline]
3 Andrade J. Investigations of hypesthesia: using anesthetics to explore relationships between consciousness, learning, and memory. Conscious Cogn 1996; 5: 56280[ISI][Medline]
4 Glass PS, Bloom M, Kearse L, Rosow C, Sebel P, Manberg P. Bispectral analysis measures sedation and memory effects of propofol, midazolam, isoflurane, and alfentanil in healthy volunteers. Anesthesiology 1997; 86: 83647[ISI][Medline]
5 Gardner H, Boller F, Moreines J, Butters N. Retrieving information from Korsakoff patients: effects of categorical cues and reference to the task. Cortex 1973; 9: 16575[Medline]
6 Kunst-Wilson WR, Zajonc RB. Affective discrimination of stimuli that cannot be recognized. Science 1980; 207; 5578[ISI][Medline]
7 Roorda-Hrdlicková V, Wolters G, Bonke B, Phaf RH. Unconscious perception during general anaesthesia, demonstrated by an implicit memory task. In: Bonke B, Fitch W, Millar K, eds. Memory and Awareness in Anaesthesia. Lisse: Swets and Zeitlinger, 1990; 1505
8 Jelicic M, Bonke B, Wolters G, Phaf RH. Implicit memory for words presented during general anaesthesia. Eur J Cogn Psychol 1992; 4: 7180[ISI]
9 Block RI, Ghoneim MM, Pathak D, Kumar V, Hinrichs JV. Effects of a subanesthetic concentration of nitrous oxide on overt and covert assessments of memory and associative processes. Psychopharmacology 1988; 96: 32431[ISI][Medline]
10 Block RI, Ghoneim MM, Sum Ping ST, Ali MA. Human learning during general anaesthesia and surgery. Br J Anaesth 1991; 66: 1708[Abstract]
11 Murphy ST, Zajonc RB. Affect, cognition, and awareness: affective priming with optimal and suboptimal stimulus exposures. J Pers Soc Psychol 1993; 64: 72339[ISI][Medline]
12 Zajonc RB. On the primacy of affect. Am Psychol 1984; 39: 11724
13 Plummer G. Improved method for the determination of propofol in blood by high performance liquid chromatography. J Chromatogr 1987; 421: 1716[Medline]
14 Reingold EM, Merikle PM. Using direct and indirect measures to study perception without awareness. Perception Psychophys 1988; 44: 56375[ISI][Medline]
15 Polster MR, Gray PA, OSullivan G, McCarthy RA, Park GR. Comparison of the sedative and amnesic effects of midazolam and propofol. Br J Anaesth1993; 70: 6126[Abstract]
16 Chortkoff BS, Gonowski CT, Bennett HL, Levinson B, Crankshaw DP, Dutton RC, et al. Subanesthetic concentrations of desflurane and propofol suppress recall of emotionally charged information. Anesth Analg 1995; 81: 72836[Abstract]
17 Cork RC, Heaton JF, Campbell CE, Kihlstrom JF. Is there implicit memory after propofol sedation? Br J Anaesth 1996; 76: 4928
18 Stapleton C, Andrade J. An investigation of learning during propofol sedation and anaesthesia using the process dissociation procedure. Anesthesiology 2000; 93: 141825[ISI][Medline]
19 Villemure C, Plourde G, Lussier I, Normandin N. Auditory processing during isoflurane anaesthesia: a study with an implicit memory task and auditory evoked potentials. In: Sebel PS, Bonke B, Winograd E, eds. Memory and Awareness in Anesthesia. Englewood Cliffs, NJ: Prentice Hall, 1993; 99106
20 Bethune DW, Ghosh S, Gray B, Kerr L, Walker IA, Doolan LA, et al. Learning during general anaesthesia: implicit recall after methohexitone or propofol infusion. Br J Anaesth 1992; 69: 1979[Abstract]
21 Schwender D, Kaiser A, Klasing S, Peter K, Pöppel E. Explicit and implicit memory and mid-latency auditory evoked potentials during cardiac surgery. In: Sebel PS, Bonke B, Winograd E, eds. Memory and Awareness in Anesthesia. Englewood Cliffs, NJ: Prentice Hall, 1993; 8598
22 Munte S, Kobbe I, Demertzis A, Lullwitz E, Munte TF, Piepenbrock S, et al. Increased reading speed for stories presented during general anesthesia. Anesthesiology 1999; 90: 6629[ISI][Medline]
23 Charlton PFC, Wang M, Russell IF. Implicit and explicit memory for word stimuli presented during general anesthesia without neuromuscular blockade. In: Sebel PS, Bonke B, Winograd E, eds. Memory and Awareness in Anesthesia. Englewood Cliffs, NJ: Prentice Hall, 1993; 6473
24 Westmoreland C, Sebel PS, Winograd E, Goldman WP. Indirect memory during anaesthesia: effect of midazolam. Anesthesiology 1993; 78: 23741[ISI][Medline]
25 Macrae WJ, Thorp JM, Millar K. Category generation testing in the search for implicit memory during general anaesthesia. Br J Anaesth 1998; 80: 58893[ISI][Medline]
26 Lubke G, Kerssens C, Phaf H, Sebel PS. Dependence of explicit and implicit memory on hypnotic state in trauma patients. Anesthesiology 1999; 90: 67080[ISI][Medline]