Implicit learning during general anaesthesia may occur despite a lack of explicit recall, though findings across studies remain inconsistent and require further neuroimaging research.
What can surgical patients hear during the state of unconsciousness induced by general anaesthesia? Very few patients are able to recall details of their surgical procedure. Liu and colleagues interviewed 1000 patients one day after elective surgery under general anaesthesia and found that only two of them (0. 2%) had conscious recollections of intra-operative events 1. A few years earlier, the Rotterdam group reported that two out of 140 patients had post-operative recall of surgical events 2. Based on these studies, one might argue that learning during anaesthesia is a very rare phenomenon, occurring only as a result of inadequate anaesthesia 3. Over the past years, however, memory researchers have demonstrated that recall is a rather insensitive measure of learning. Graf and Schacter 4 made a distinction between explicit and implicit forms of memory. Tests of recall and recognition require conscious recollection of a learning episode and have been labelled explicit memory tasks. Implicit memory tasks, in contrast, do not refer to such conscious recollection but bring about a change or facilitation in task performance which can be attributed to a previous learning experience. A popular implicit-memory task is the wordstem completion test. Subjects are first presented with words (e. g. Window). After a delay they are given a number of word stems and asked to complete these stems with 'the first word that comes to mind'. Some of the word stems (e. g. Win____) belong to words presented during the acquisition phase, others do not. Implicit memory retention is evidenced when the critical stems are more often completed to the stimulus words (e. g. Window instead of Winter) than can be expected on the basis of chance. Implicit memory functions are more robust than explicit memory functions; patients with organic amnesia, e. g. those with Korsakoff's syndrome, perform very poorly on recall and recognition tests, but exhibit normal performance on implicit memory tasks 5. Some authors have drawn a parallel between patients with the amnesic syndrome and those under general anaesthesia 6. Both groups have impaired explicit learning capabilities but may exhibit preserved implicit memory functions. Is there any evidence of learning during anaesthesia without post-operative recall? Around 1960 the American gynaecologist Cheek studied a group of patients with a poor course of post-operative recovery 7. These women had no explicit memories of intra-operative events. When hypnotized they were allegedly able to remember 'negatively coloured' statements that had been made about them, whilst anaesthetized, by members of the surgical team. Some 5 years later, one of the first experimental studies on cognitive processing under anaesthesia was conducted by Levinson-an anaesthesiologist who changed his speciality to psychiatry 8. Ten volunteer patients who had general anaesthesia for minor surgery were subjected to a mock crisis, in which the anaesthesiologist had the operation stopped by exclaiming that the patient had turned blue and needed more oxygen. One month later the patients were hypnotized and age-regressed to the time of surgery. Four of them were able to repeat the ominous words of the anaesthesiologist almost exactly. Another four remembered having heard something, and some of them were able to identify the speaker as the anaesthesiologist. The latter group all showed marked anxiety. None of the patients in this experiment had any conscious recall of events occurring during surgery. It is difficult to evaluate the validity of these early investigations of implicit learning during anaesthesia. One problem concerns the use of hypnosis: recent work in experimental psychology has shown that 'false memories' can be created using hypnosis 9. In addition, Levinson failed to use a control group (and a double-blind procedure). In the 1960s and 1970s several other studies were published. These experiments either used explicit memory tests or intra-operative therapeutic suggestions. Some authors hypothesized that if positive messages presented during anaesthesia result in a reduction in post-operative morbidity and shorter duration of hospital stay, this would be indicative of cognitive processing under anaesthesia. In 1977, Trustman, Dubovsky, and Titley critically reviewed seven studies of intra-operative learning and concluded that all seven were methodologically flawed 10. A few years later cognitive processing in anaesthetized patients was demonstrated in studies which were without methodological problems. Using an implicit-memory measure, Millar and Watkinson were among the first to demonstrate convincingly the existence of intra-operative learning 11. Two years later, Bennett, Davis and Giannini showed that patients who, during anaesthesia, had been given suggestions to touch an ear post-operatively, pulled their ear more often than control patients who had not received any suggestions 12. About the same time, Bonke and associates found that older cholecystectomy patients who had been played therapeutic suggestions intra-operatively, left the hospital sooner than patients who had not heard any positive messages 13. Note that in these investigations there were no patients with explicit recall of surgical events. Unfortunately, the results of subsequent research on cognitive processing under anaesthesia were less clear-cut: some investigators found evidence of intra-operative learning in anaesthetized patients, whereas others were unable to demonstrate such learning. It is difficult to explain these discrepant findings. In a major review paper, Ghoneim and Block 14 stated that 'the studies in this area are mosaics of different methods and procedures, with different outcomes. It is impossible to determine if differences among studies with positive results are attributable to differences in the tests, the anaesthetic regimens, the number of repetitions, duration, and complexity of stimulus presentations, or other factors'. In June 1995 the Third International Symposium on Memory and Awareness in Anaesthesia was held in Rotterdam, The Netherlands. Several presentations at this conference dealt with implicit learning during anaesthesia, with about half of them showing cognitive processing in anaesthetized patients. A few of the papers presented in Rotterdam, however, may shed new light on the issue of learning during anaesthesia. Merikle and Daneman 15 conducted a meta-analysis on studies of cognitive processing under anaesthesia looking for consistent findings across experiments. They reported that studies employing therapeutic suggestions did not yield evidence of intra-anaesthetic learning. When memory studies were lumped together a different picture emerged: there was evidence of processing under anaesthesia if tests were administered shortly after the surgical procedure, but not with long study-test delays. According to Merikle and Daneman, there are two possible explanations why memory in anaesthesia can only be demonstrated using a short study-test interval. It could be that there is rapid decay of memory for stimuli presented during anaesthesia. Another explanation is that memory in anaesthesia is more likely to be detected when the test phase resembles the study phase-the so-called Encoding Specificity Principle. Shortly after the operation, e. g. in the recovery room, patients still have residual anaesthetic agents in their blood and they often feel sleepy and drowsy. Given the similarity between patients' state under anaesthesia and in the recovery room, it would be easier to find evidence of intra-operative learning with a short study-test interval than with longer delays. Another important paper was presented by Alkire and colleagues 16 who used Positron Emission Tomography (PET) -a modern neuroimaging technique-to investigate cognitive processing under propofol anaesthesia. They reported high glucose metabolism rates, during anaesthesia, in parts of the brain known to be involved in learning and memory (i. e. the frontal and medial-temporal cortices). In addition, there was a high correlation between glucose metabolism in these parts of the brain and post-operative implicit memory performance. It thus seems that the brain continues to process auditory information while anaesthetized with propofol. The meta-analytic and PET findings are indicative of learning during anesthesia. Nevertheless, processing under anaesthesia is not yet a well-established fact. One caveat of Merikle and Daneman's meta-analysis is that studies with different anaesthetic techniques were combined. Note that volatile anaesthetics are considered to suppress cognitive functions more than nitrous oxide with or without opioids 17. It could be the case that researchers who used short study-test delays also employed anaesthetic cocktails without volatile anaethetics. In a related vein, Alkire and colleagues only used one anaesthetic technique. Hence, generalizing their findings to other anaesthetic combinations may be problematic. What needs to be done is to use PET or another neuroimaging technique such as functional Magnetic Resonance Imaging (f-MRI) to study cognitive processing in patients anaesthetized with different anaesthetic drug combinations. In 1998 the Fourth International Symposium on Memory and Awareness in Anaesthesia will be held in Cambridge (UK). We expect that, by that time, a more definitive answer can be given to the question 'what can surgical patients hear during general anaesthesia? ' M. Jelicic B. Bonke Rotterdam
Jelícic et al. (1996) conducted a review in General anaesthesia. General anaesthesia was evaluated on Implicit and explicit memory or learning. Implicit learning during general anaesthesia may occur despite a lack of explicit recall, though findings across studies remain inconsistent and require further neuroimaging research.