Portal de Periódicos da CAPES

Comparing the effects of stimulation and propofol infusion rate on implicit and explicit memory formation

2001; Elsevier BV; Volume: 86; Issue: 2 Linguagem: Inglês

10.1093/bja/86.2.189

1471-6771

Jackie Andrade, L. F. ENGLERT, Courtney Harper, N. D. Edwards,

Memory Processes and Influences

Doubt remains about the conditions under which learning persists despite anaesthesia. This study investigated the relative importance of dose of anaesthetic and stimulation for learning during propofol infusion before surgery. Thirty-six patients were randomly assigned to three groups. Group 1 received two word lists (category examples and nonsense words) during infusion of propofol to a target concentration of 2 μg ml−1. Groups 2 and 3 received the word lists during infusion of propofol 5 μg ml−1. Group 2 received nonsense words before tracheal intubation and category examples during intubation; Group 3 heard category examples before and nonsense words during intubation. Bispectral index was recorded as a measure of depth of sedation/anaesthesia. We assessed explicit memory on recovery using a structured interview and a recognition test. We assessed implicit memory using a category generation test and a preference rating task. To establish baseline, a control group of 12 patients completed the category generation test without receiving the category examples during anaesthesia. Overall, there was no evidence for learning during propofol infusion, though the category generation task showed a trend towards more implicit memory for words presented during intubation than during anaesthesia. We conclude that learning does not occur during anaesthesia without surgery. Doubt remains about the conditions under which learning persists despite anaesthesia. This study investigated the relative importance of dose of anaesthetic and stimulation for learning during propofol infusion before surgery. Thirty-six patients were randomly assigned to three groups. Group 1 received two word lists (category examples and nonsense words) during infusion of propofol to a target concentration of 2 μg ml−1. Groups 2 and 3 received the word lists during infusion of propofol 5 μg ml−1. Group 2 received nonsense words before tracheal intubation and category examples during intubation; Group 3 heard category examples before and nonsense words during intubation. Bispectral index was recorded as a measure of depth of sedation/anaesthesia. We assessed explicit memory on recovery using a structured interview and a recognition test. We assessed implicit memory using a category generation test and a preference rating task. To establish baseline, a control group of 12 patients completed the category generation test without receiving the category examples during anaesthesia. Overall, there was no evidence for learning during propofol infusion, though the category generation task showed a trend towards more implicit memory for words presented during intubation than during anaesthesia. We conclude that learning does not occur during anaesthesia without surgery. Many studies have demonstrated learning during general anaesthesia.1Andrade J Learning during anaesthesia: a review.Br J Psychol. 1995; 86: 479-506Crossref PubMed Scopus (80) Google Scholar 2Ghoneim MM Block RI Learning and memory during general anesthesia: an update.Anesthesiology. 1997; 87: 387-410Crossref PubMed Scopus (140) Google Scholar In a typical study, patients are played a tape of words during surgery and then given memory tests for these words on recovery. The memory tests reveal unconscious or implicit memory for the words even though patients have no conscious or explicit memory of hearing anything during their operation. That is, patients respond differently on the implicit memory tests to the words that were presented during anaesthesia compared with similar distractor words, yet they neither recall hearing the words during anaesthesia nor recognize these words when they are presented on recovery. It is not known whether this learning occurs despite patients being unconscious or whether it is a sign that some patients are momentarily conscious during anaesthesia. The effects of surgery and other types of stimulation on learning are also unknown. Volunteer studies have tended to show little or no learning during conscious sedation, unless the dose of anaesthetic is very small and the stimuli to be learned are simple, such as single words.3Andrade J Investigations of hypesthesia: using anesthetics to explore relationships between consciousness, learning, and memory.Conscious Cogn. 1996; 5: 562-580Crossref PubMed Scopus (30) Google Scholar Volunteers differ from patients in two ways: absence of surgical stimulation and, presumably, lower levels of anxiety. Surgery and anxiety may facilitate learning by decreasing the depth of anaesthesia or by altering concentrations of circulating catecholamines independently from their effects on anaesthetic depth. 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.3Andrade J Investigations of hypesthesia: using anesthetics to explore relationships between consciousness, learning, and memory.Conscious Cogn. 1996; 5: 562-580Crossref PubMed Scopus (30) Google Scholar 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 index4Glass 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: 836-847Crossref PubMed Scopus (1105) Google Scholar 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.5Gardner H Boller F Moreines J Butters N Retrieving information from Korsakoff patients: effects of categorical cues and reference to the task.Cortex. 1973; 9: 165-175Crossref PubMed Scopus (61) Google Scholar 6Kunst-Wilson WR Zajonc RB Affective discrimination of stimuli that cannot be recognized.Science. 1980; 207: 557-558Crossref PubMed Scopus (735) Google Scholar Category generation was used for comparison with previous demonstrations of learning during general anaesthesia.7Roorda-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 Memory and Awareness in Anaesthesia. Swets and Zeitlinger, Lisse1990: 150-155Google Scholar 8Jelicic M Bonke B Wolters G Phaf RH Implicit memory for words presented during general anaesthesia.Eur J Cogn Psychol. 1992; 4: 71-80Crossref Scopus (32) Google Scholar Preference ratings have also been used to demonstrate learning during sedation9Block 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: 324-331Crossref PubMed Scopus (27) Google Scholar and anaesthesia,10Block RI Ghoneim MM Sum Ping ST Ali MA Human learning during general anaesthesia and surgery.Br J Anaesth. 1991; 66: 170-178Crossref PubMed Scopus (77) Google Scholar 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.11Murphy ST Zajonc RB Affect, cognition, and awareness: affective priming with optimal and suboptimal stimulus exposures.J Pers Soc Psychol. 1993; 64: 723-739Crossref PubMed Scopus (1149) Google Scholar 12Zajonc RB On the primacy of affect.Am Psychol. 1984; 39: 117-124Crossref Scopus (1141) Google Scholar Implicit memory was tested by a category generation task and by a preference rating task. Full details of the construction and pilot-testing of the memory tests are given in the Appendix. 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.7Roorda-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 Memory and Awareness in Anaesthesia. Swets and Zeitlinger, Lisse1990: 150-155Google Scholar 8Jelicic M Bonke B Wolters G Phaf RH Implicit memory for words presented during general anaesthesia.Eur J Cogn Psychol. 1992; 4: 71-80Crossref Scopus (32) Google Scholar 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';12Zajonc RB On the primacy of affect.Am Psychol. 1984; 39: 117-124Crossref Scopus (1141) Google Scholar 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. 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. 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 ml−1. 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 ml−1 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 ml−1 i.v. and anaesthesia was continued with propofol to a target concentration of 5 μg ml−1 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. Patients in groups 2 and 3 received a higher initial dose of propofol to a target concentration of 5 μg ml−1. 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. 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.13Plummer G Improved method for the determination of propofol in blood by high performance liquid chromatography.J Chromatogr. 1987; 421: 171-176Crossref PubMed Scopus (264) Google Scholar 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.11Murphy ST Zajonc RB Affect, cognition, and awareness: affective priming with optimal and suboptimal stimulus exposures.J Pers Soc Psychol. 1993; 64: 723-739Crossref PubMed Scopus (1149) Google Scholar 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. 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. 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 α = 0.05 and all tests were two-tailed. All analyses were performed using StatView 5.0. One patient in group 3 received additional propofol (50 mg bolus) because she moved just before intubation. Fifteen patients received morphine in the recovery area before memory testing (dose range 2–10 mg). Five of these patients were in group 1, four were in group 2 and six were in group 3. 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 ml−1 infusion than in the group receiving 2 μg ml−1 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.Table 1Bispectral index during stimulus presentation and mean blood propofol concentration just before presentation of list 2. Mean (range). Within-group comparisons: *P<0.01 compared with list 2 BIS values. Between-group comparisons: **P<0.001 compared with groups 2 and 3Bispectral indexBlood propofol concentration (μg ml−1)GroupList 1: before propofol infusionList 2: during sedation (group 1) or anaesthesia (groups 2, 3)List 3: during sedation (group 1) or intubation (groups 2, 3)192.0 (88.5–95.5)*85.4 (76.5–96.5)**82.4 (65.0–96.0)0.89 (0.19–2.36)**291.2 (82.5–96.5)*51.3 (29.5–73.0)51.9 (33.0–71.5)2.77 (0.73–5.40)390.5 (76.0–95.5)*55.8 (45.5–70.0)54.8 (44.5–62.0)3.13 (0.63–5.50) Open table in a new tab No patient reported recall of surgery. One patient in group 1 recalled hearing the category examples 'banana', 'pear' (both presented during infusion of propofol 2 μg ml−1) and 'orange' (not presented). One patient in group 2 recognized 'banana' (presented during intubation) when read the phrase 'yellow banana green pear'. All but two patients remembered hearing nonsense or Finnish words some time before their operation. One patient claimed to remember the entire period of sedation with propofol 2 μg ml−1 but could not recall any of the Finnish words. Four patients reported hearing a total of 13 Finnish words, six of which came from the list played before propofol infusion and seven of which were unidentifiable. No-one recalled Finnish words from lists played during propofol infusion. 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).Table 2Mean explicit and implicit memory scores for the Finnish words, by anaesthetic condition during word presentation. Mean (95% confidence interval). Memory scores are presented as priming index, i.e. score for presented list minus score for distractor list. *P<0.05; **P 0.10)Phrase during which phrase was presentedMean number of hitsSedation (group 1)1.83Anaesthesia (group 3)1.67Anaesthesia + intubation (group 2)2.08*Mean for the experimental groups1.86Mean for the control group1.75 Open table in a new tab Preference ratings for the Finnish words showed implicit memory for the words presented before propofol infusion began (P<0.05). For the words presented during propofol infusion, ratings did not differ from baseline, i.e. there was no implicit memory. 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). We experimentally manipulated depth of anaesthesia in two ways, by varying the propofol infusion rate and by giving the stimulus of tracheal intubation. To confirm this, we used BIS as an index of depth of sedation/anaesthesia. Our results confirmed that depth of anaesthesia differed between patients receiving propofol 2 μg ml−1 and those receiving 5 μg ml−1. However, BIS did not detect an effect of intubation in the patients receiving propofol 5 μg ml−1 (Table 1). Intubation is known to be a stimulating procedure, and there were clinical signs of lightened anaesthesia (increased heart rate) when patients were intubated in the present study. The failure of BIS to detect any effect of intubation may reflect its insensitivity to subtle changes in anaesthetic depth or suggest that BIS is a better monitor of depth of sedation than depth of anaesthesia. 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;14Reingold EM Merikle PM Using direct and indirect measures to study perception without awareness.Perception Psychophys. 1988; 44: 563-575Crossref PubMed Scopus (339) Google Scholar 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,15Polster MR Gray PA O'Sullivan G McCarthy RA Park GR Comparison of the sedative and amnesic effects of midazolam and propofol.Br J Anaesth. 1993; 70: 612-616Crossref PubMed Scopus (69) Google Scholar but the infusion dose of propofol was considerably smaller (1.27 mg kg−1 h−1) than the 2 μg ml−1 used in the present study. Volunteers failed to learn words presented during infusion of larger subanaesthetic doses of propofol.3Andrade J Investigations of hypesthesia: using anesthetics to explore relationships between consciousness, learning, and memory.Conscious Cogn. 1996; 5: 562-580Crossref PubMed Scopus (30) Google Scholar 16Chortkoff 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: 728-736PubMed Google Scholar In contrast, studies with patients showed learning during surgery with sedative infusions of propofol,17Cork RC Heaton JF Campbell CE Kihlstrom JF Is there implicit memory after propofol sedation?.Br J Anaesth. 1996; 76: 492-498Crossref PubMed Scopus (29) Google Scholar 18Stapleton C Andrade J An investigation of learning during propofol sedation and anaesthesia using the process dissociation procedure.Anesthesiology. 2000; 93: 1418-1425Crossref PubMed Scopus (27) Google Scholar and general anaesthetic regimes that included propofol7Roorda-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 Memory and Awareness in Anaesthesia. Swets and Zeitlinger, Lisse1990: 150-155Google Scholar 18Stapleton C Andrade J An investigation of learning during propofol sedation and anaesthesia using the process dissociation procedure.Anesthesiology. 2000; 93: 1418-1425Crossref PubMed Scopus (27) Google Scholar, 19Villemure 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 Memory and Awareness in Anesthesia. Prentice Hall, Englewood Cliffs, NJ1993: 99-106Google Scholar, 20Bethune 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: 197-199Crossref PubMed Scopus (33) Google Scholar, 21Schwender 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 Memory and Awareness in Anesthesia. Prentice Hall, Englewood Cliffs, NJ1993: 85-98Google Scholar, 22Munte 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: 662-669Crossref PubMed Scopus (31) Google Scholar (though some have shown no learning).23Charlton 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 Memory and Awareness in Anesthesia. Prentice Hall, Englewood Cliffs, NJ1993: 64-73Google Scholar, 24Westmoreland C Sebel PS Winograd E Goldman WP Indirect memory during anaesthesia: effect of midazolam.Anesthesiology. 1993; 78: 237-241Crossref PubMed Scopus (15) Google Scholar, 25Macrae WJ Thorp JM Millar K Category generation testing in the search for implicit memory during general anaesthesia.Br J Anaesth. 1998; 80: 588-593Crossref PubMed Scopus (3) Google Scholar 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 surgery26Lubke G Kerssens C Phaf H Sebel PS Dependence of explicit and implicit memory on hypnotic state in trauma patients.Anesthesiology. 1999; 90: 670-680Crossref PubMed Scopus (133) Google Scholar 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 group's response rate (1.75) was similar to that reported by Jelicic and colleagues8Jelicic M Bonke B Wolters G Phaf RH Implicit memory for words presented during general anaesthesia.Eur J Cogn Psychol. 1992; 4: 71-80Crossref Scopus (32) Google Scholar (1.84), though both differed from the baseline rate in a previous study7Roorda-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 Memory and Awareness in Anaesthesia. Swets and Zeitlinger, Lisse1990: 150-155Google Scholar 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. Thirty repetitions of the phrase 'yellow banana green pear' were recorded onto a compact disc track at the rate of one word per 1.5 s, making a total presentation time of 3 min.7Roorda-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 Memory and Awareness in Anaesthesia. Swets and Zeitlinger, Lisse1990: 150-155Google Scholar 8Jelicic M Bonke B Wolters G Phaf RH Implicit memory for words presented during general anaesthesia.Eur J Cogn Psychol. 1992; 4: 71-80Crossref Scopus (32) Google Scholar The phrase was played to the patients during propofol infusion, and on recovery they were asked to say the first three examples of fruits and colours that came to mind. 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 2–5 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. 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 0.10). 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.