Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
This article was downloaded by: [Cardiff University] On: 10 July 2010 Access details: Access Details: [subscription number 917204045] Publisher Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 3741 Mortimer Street, London W1T 3JH, UK Ergonomics Publication details, including instructions for authors and subscription information: http://www.informaworld.com/smpp/title~content=t713701117 State-dependent memory produced by aerobic exercise Christopher Miles Elinor Hardman To cite this Article Hardman, Christopher Miles Elinor(1998) 'State-dependent memory produced by aerobic exercise', Ergonomics, 41: 1, 20 — 28 To link to this Article: DOI: 10.1080/001401398187297 URL: http://dx.doi.org/10.1080/001401398187297 PLEASE SCROLL DOWN FOR ARTICLE Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf This article may be used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material. ERGONOMICS , 1998, VOL. 41, NO . 1, 20 ± 28 State-dependent memory produced by aerobic exercise CHRISTOPHER MILES and ELINOR HARDMAN School of Psychology, University of Wales College of CardiŒ, CardiŒCF1 3YG, UK Downloaded By: [Cardiff University] At: 09:37 10 July 2010 Keywords: Aerobic exercise; Cue-dependent memory; Recall levels; State-dependent memory. In a free recall experiment, participants learned lists of words in two physiological states: at rest and while exercising aerobically on a bicycle ergometer. Recall of the words was required in either the state consistent with learning or in the alternative state. Word lists learned during aerobic exercise were recalled best during aerobic exercise and vice versa. Greater changes in heart rate in the changed state conditions were associated with greater retrieval decrements. Recall levels for words both learned and recalled at exercise were equivalent to those for words both learned and recalled at rest. This ® nding rules out the possibility that exercise per se interfered with the original learning. The study is consistent with the view that state-dependent memory should be viewed as a particular form of cue-dependent memory. 1. Introduction The functioning of the human memory system may be characterized as involving the permanent retention of material with remembering and forgetting conceptualized respectively as successful recall and a temporary failure to recall. During learning the participant processes not only the to-be-learned material, but in addition associates that material with a variety of internal and external contextual cues. The participant therefore processes a multidimensional complex of stimuli. Forgetting will occur, at least at the behavioural level, if the contextual cues at test do not su ciently discriminate between the desired memory and some other competing memories (Capaldi and Neath 1995). By this account, a change in context necessarily results in altered stimulus conditions and changing the stimuli present from study to test can lead to retrieval failure (Tulving 1983). Context can aŒect the accessibility of items by reducing the eŒectiveness of potential retrieval cues to discriminate between the target memory and other distracter memories (Eich 1989). Perhaps the most widely known experiment examining the eŒect on recall of a context switch between study and test is that reported by Godden and Baddeley (1975). Experienced scuba divers studied a list of 36 common and unrelated words in one of two environments: either on land or underwater. Participants later retrieved the list items either in the original learning environment or in the alternate context. Even though the location of the dives and time of day were not controlled for, there was a pronounced context-dependen t retrieval ® nding using free recall: recall averaged 35% in the consistent learning and retrieval conditions but was markedly reduced to 24% in the changed conditions. A second experiment required one-half of the participants to learn a list of words and to recall them on land. The other half also learned and recalled on land but they were required to enter the pool, swim a 0014± 0139/98 $12× 00 Ó 1998 Taylor & Francis Ltd Downloaded By: [Cardiff University] At: 09:37 10 July 2010 State-dependen t memory produced by aerobic exercise 21 short distance, dive to a depth of 20 ft, and then return to land prior to recall. Recall performance of both groups was equivalent, thus ruling out the possibility that disruption between learning and test was the cause of poorer recall in the ® rst experiment. A similar context manipulation study with greater ecological validity (Martin and Aggleton 1993), which used novice divers who were required to learn and recall decompression tables, has shown an equally powerful context-dependen t retrieval eŒect. Each of these studies represents an example of context-dependen t retrieval with external manipulation of context. However, context may also be manipulated internally. When context is provided by the pharmacological state of the participant for instance, the phenomenon is known as state-dependent memory. Such statedependent learning in humans has been demonstrated by Goodwin et al. (1969). Participants who were given alcohol in training and saline at test performed worse on a variety of tasks compared to the same state groups. This ® nding does not appear to be attributabl e to anything speci® c to the drugged state. For example, Weingartner and Faillace (1971) conducted a state-dependent memory experiment with two diŒerent groups of participants. One group comprised chronic alcoholics and the other comprised an equal number of non-alcoholics closely matched to the alcoholic participants with respect to a variety of demographic variables. Participants were either intoxicated or sober at list learning and intoxicated or sober at test. Both groups showed state-dependent memory eŒects. Recall was higher when the test state matched the learning state in comparison to the changed state condition, thus ruling out any objection to a separate role of alcohol. Other state manipulations demonstrating this ® nding include mood congruency (Bartlett and Santrock 1979) and odour congruency (Schab 1990). In the Bartlett and Santrock study the mood or aŒective state of participants was altered so that they were either in a happy or neutral mood at study and a happy or neutral mood at test. Performance was better for participants in the same-mood conditions. Schab (1990) ® lled a small room with the odour of chocolate either at study or test or both. After a 24-h delay performance on a surprise memory test was better when the odours at learning and test matched. Thus a change in context (whether it be external or internal) may be conceptualized as in¯ uencing memory performance by removing or altering potential retrieval cues. For both state- and context-dependen t memory eŒects the disruption to performance is greatest when the contexts are more discriminable because of the concomitant change in retrieval cue availability. Eich (1980) has argued that the ® nding of state-dependent retrieval is determined to a large extent by the nature of the retrieval task employed. He argues that retrieval paradigms that provide an explicit cue to the participant with respect to the target item (e.g. paired associate learning) are not sensitive to state-dependent retrieval because the probing of memory via an explicit cue provides the participant with a direct link to the target item in memory. The existence of a direct cue allows the retrieval process to bypass state-dependent cues that might otherwise be activated (otherwise known as `the outshining hypothesis’). The deleterious eŒect of changing state is therefore negated. In those memory paradigms where no explicit retrieval cue is provided for the participant (e.g. free recall) then the state-dependent retrieval eŒects are reliable and robust. In this instance, Eich (1980) suggests that `invisible’ cues guide the process of retrieval. The available data therefore force the conclusion that state-dependent retrieval should be viewed as a cue-dependent phenomenon. State-dependent retrieval eŒects depend critically on the precise nature of the Downloaded By: [Cardiff University] At: 09:37 10 July 2010 22 C. Miles and E. Hardman information or cues that are available to the participants at retrieval. If the cues at test match those at learning then recall is facilitated. The present study was concerned with extending the ® ndings of state-dependent retrieval beyond these conditions where internal state is typically manipulated by the administration of a particular dose of a drug such as alcohol. Participant state was manipulated by the requirement to pedal a bicycle ergometer at a rate such that heart rate was elevated to approximately twice the normal resting level. In a repeated measures design participants’ heart rates were manipulated such that they learned and retrieved information either at rest or when working at a prede® ned level of aerobic exercise. Manipulation of heart rate in this way allowed an assessment of the association between the porportional change in heart rate and the associated change in recall levels for the changed state combinations. The present authors employed the same class of memory materials as reported by Godden and Baddeley (1975) in their study of context-dependen t retrieval. Superior recall is predicted in those conditions where the heart rate of the participant is consistent at both learning and retrieval in comparison to the changed state conditions. In the latter conditions fewer of those state cues present at the original encoding of the word lists were present at recall. In addition, the extent to which the porportional change in heart rate between rest and exercise conditions predicted change in retrieval levels was assessed. In particular, the authors aimed to determine the extent to which greater changes in heart rate (and therefore retrieval cues) between learning and retrieval are associated with greater de® cits in retrieval of information. 2. Method 2.1. Participants Twenty-four undergraduat e participants (6 males and 18 females) from the University of Wales College of CardiŒtook part in the study. Their age range was 18 ± 22 years with a mean of 20 years 3 months. Each participant received a small payment. 2.2. Assessment of exercise intensity This study investigated the eŒect of a change in heart rate between learning and retrieval on free recall performance. It therefore considered the intensity of exercise necessary to induce a change in cardiovascular state. Aerobic work commences at a level of exercise that requires the heart to beat at 60% of an individual’s maximum heart rate (McArdle et al. 1994). An accepted estimate of maximum heart rate (in beats per minute, bpm) is `220 Ð the participant’ s age’. Although participants of the same age will not necessarily have equivalent maximum heart rates, individual diŒerences should not be of any practical signi® cance in this regard (McArdle et al. 1994). In the current study the participants’ mean age was 20 years 3 months and the estimated maximum heart rate employed was 200 bpm. Therefore, the minimum heart rate giving rise to a change in heart rate (as de® ned by aerobic work) is 200 0× 6 = 120 bpm. To ensure that aerobic work levels were maintained during the exercise conditions, the points at which anaerobic work starts and aerobic ® nishes required de® nition. McArdle et al. (1994) describe that point at which maximal aerobic and minimal anaerobic respiration occurs as the Maximal Aerobic Function (MAF). MAF is calculated by subtracting chronological age from 180 and adding to this ® gure a ® tness-dependent number (range: Ð 10 to + 5). In the current study, a ´ State-dependen t memory produced by aerobic exercise 23 Downloaded By: [Cardiff University] At: 09:37 10 July 2010 minimum level of ® tness was assumed for all participants and an adjustment to MAF of Ð 10 was therefore employed. Thus, for a participant population with a mean age of 20 years 3 months a MAF of 150 bpm was employed. The aerobic zone for the exercise condition was hence calculated as a minimum of 120 bpm and a maximum of 150 bpm. 2.3. Materials Four stimulus lists were prepared. Each list comprised 36 trisyllabic words all with Kucera and Francis (1967) frequency ratings of 1 per million. The lists were spoken in monotone by a female voice at a rate of 1 word per 2-s and recorded directly on to an audio tape. Each list was recorded twice with a 10-s interval between lists. The word lists were presented to the participants over headphones in order to minimize external noise. At recall, the participants spoke their responses directly into a microphone that was connected to a second audio cassette recorder. The exercise condition was undertaken on a Seca bicycle ergometer with an adjustable work load (resistance) and a meter that provided a visual indication of cadence (pedalling rate). Each participant wore a Polar Edge heart rate monitor throughout the experiment. This provides a second-by-second digital record of heart rate. 2.4. Design A 2-factor repeated measures design was adopted. The ® rst factor refers to learning condition (exercise versus rest) and the second refers to retrieval condition (exercise versus rest). Allocation of participants to conditions was counterbalanced such that each of the 24 participants completed the four experimental combinations in a diŒerent order. The participants were presented with a diŒerent word list pair at learning in each experimental combination and each word list pair was used a total of six times in each of these. Participants attended the laboratory on four consecutive days at approximately the same time of day and completed a diŒerent experimental combination on each occasion. 2.5. Procedure Each participant was tested individually in a soundproofed laboratory. Upon entering the laboratory the experiment was described to the participants and all were provided with written task instructions. For all participants the experiment comprised a 3 min learning phase, a 5 min consolidation phase, and a 2 min retrieval phase. The participant was made comfortable on the bicycle ergometer and then put on the heart rate monitor, a pair of headphones and had a microphone positioned as close to the mouth as possible. During the learning phase of the experiment, participants heard two presentations of the same to-be-recalled 36-word list 10 s apart. Following Godden and Baddeley (1975) the second word list was followed by a spoken sequence of ® ve digits, presented at the rate of one digit per second for immediate recall. This task was included to limit the extent to which free recall of the lists was a direct product of subvocal rehearsal strategies. Following this task, there was a 5-min consolidation interval during which the participant remained on the bicycle ergometer and was not distracted. There then followed a 2-min free recall period. The four conditions in which each participant was tested are detailed below. (1) Rest-Rest(RR): The participant performed the complete task while seated on, but not pedalling, the bicycle ergometer. C. Miles and E. Hardman Downloaded By: [Cardiff University] At: 09:37 10 July 2010 24 (2) Rest-Exercise (RE): The participant sat at rest on the bicycle during the list presentation (learning) phase. During the 5-min (consolidation) interval the participant remained sitting quietly for the ® rst 3-min. For the remaining 2min, the participant was required to pedal the bicycle at a candence of 60 revolutions per minute (rpm). The resistance of the ergometer was adjusted such that the participant’s heart rate was elevated to a rate between 120 and 150 bpm. This 2-min period allowed the participant to adapt to the cardiovascular demands of exercise prior to the retrieval phase. During the 2min retrieval period the participant continued to pedal at the same rate. (3) Exercise-Rest (ER): Two minutes prior to the learning phase of the experiment the participant began pedalling at a cadence of 60 rpm. The resistance of the ergometer was adjusted such that the participant’ s heart rate was elevated to between 120 and 150 bpm. The participant continued to pedal at this rate throughout the 2-min learning phase. The participants rested during the 5-min consolidation phase. The participant then recalled the words during the 2-min rest period. (4) Exercise-Exercise(EE): Two minutes prior to the learning phase of the experiment the participant began pedalling at a cadence of 60 rpm. The resistance of the ergometer was adjusted such that the participant’ s heart rate was elevated to between 120 and 150 bpm. The participant continued to pedal at this rate throughout the learning phase. The participant rested for the ® rst 3-min of the consolidation phase. For the remaining 2-min the participant pedalled at the same load and cadence as in the presentation phase. The participant continued to pedal at the same load and cadence throughout the 2-min recall phase. The participant’s heart rate was noted at the beginning and end of each of the learning, consolidation and retrieval phases of the experiment. 3. Results 3.1. Heart Rate The mean heart rate associated with all combinations of learning and retrieval is shown in table 1. A related t-test showed no signi® cant diŒerence (t < 1) between the mean heart rate data for either the learning and retrieval at rest or learning and retrieval at exercise combinations. The heart rate data were converted to a proportional change from resting heart rate data for each participation prior to analysis, because of individual diŒerences in cardiovascular e ciency. For each participant the resting heart rate was taken as the mean heart rate in the learning at rest condition. For every other condition heart rate Table 1. Mean heart rate (BPM) at learning and retrieval for each of the Rest-Rest (RR); Rest-Exercise (RE); Exercise-Rest (ER) and Exercise-Exercise (EE) combinations. Condition Mean heart rate (BPM) Learning Retrieval RR RE ER EE 80× 1 80× 0 83× 6 131× 1 134 × 2 85× 5 136 × 3 134 × 1 State-dependen t memory produced by aerobic exercise 25 was calculated as a proportional change from that resting rate. These data were subjected to a 2-factor (2 2) repeated measures ANOVA with factors participant state at list learning (exercise versus rest) and participant state at list retrieval (exercise versus rest). Neither the main eŒect of participant state at learning nor that of participant state at recall was signi® cant (both Fs < 1) but their interaction was highly signi® cant (F(1,23)= 247 × 1, p < 0× 001) and the means are shown in table 2. Further analysis (Newman-Keuls) showed the proportional change in the RE and ER combinations to be signi® cantly greater (p < 0× 001) in comparison to both the RR and EE combinations. No other comparisons were signi® cant. ´ Downloaded By: [Cardiff University] At: 09:37 10 July 2010 3.2. Correct free recall The total number of words (max. = 36) recalled by each participant in each combination was calculated and the data submitted to the same model 2-factor Table 2. Proportional change in mean heart rate between learning and retrieval for each of the Rest-Rest (RR); Rest-Exercise (RE); Exercise-Rest (ER) and ExerciseExercise (EE) combinations. Condition RR 0× 008 Figure 1. RE 0× 6 ER 0× 6 EE 0× 02 Mean number of words recalled for each of the Rest-Rest (RR); Rest-Exercise (RE); Exercise-Rest (ER) and Exercise-Exercise (EE) combinations. 26 C. Miles and E. Hardman Downloaded By: [Cardiff University] At: 09:37 10 July 2010 ANOVA as described above. Neither the main eŒect of participant state at list learning nor at list retrieval was signi® cant (both Fs < 1) but their interaction was highly signi® cant (F(1,23) = 11× 83, p < 0× 005) and is represented in ® gure 1. The pattern of the interaction is consistent with predictions. Further analysis of the interaction (Newman-Keuls) showed that both same-state learning-retrieval conditions produced signi® cantly (p < 0× 05) higher levels of recall than either of the changed-state learning-retrieval conditions. 3.3. False alarms The data were further analysed to check that the state-dependent eŒect was not a function of a guessing strategy. There was an examination of the possibility that participants produced more guess words in the same-state learning retrieval conditions, which would result in a higher probability of producing a previously presented word. The same ANOVA model as described above produced no main eŒects of participant state at either learning or retrieval (both Fs < 1), nor an interaction between the factors (F < 1). 3.4. Correlation The extent to which porportional change from the resting heart rate level was associated with change in recall level for the changing-stat e learning and retrieval combinations was assessed. The mean proportional change in heart rate across both changing-stat e conditions was calculated for each participant (ignoring the sign of the change for the exercise-rest combination). These scores were correlated (Pearson’s r) with the average change in recall levels across both changing-stat e conditions (r = 0× 38, df = 22, p = 0× 06). 4. Discussion The results of the study demonstrate that a change in cardiovascular activity between learning and retrieval can exert a powerful state eŒect on retrieval of word lists. The pattern of data mirrors that shown for shifts in context from above or below water shown previously (Godden and Baddeley 1975, Martin and Aggleton 1993). On average the changing state conditions produced a 20% drop in recall performance in comparison to the same state conditions. It is unlikely that the eŒect of changing state can be attributed to greater disruption for the changing state combinations (Strand 1970) during the consolidation phase because the exercise-exercise combination also required participants to both stop and start pedalling during the consolidation phase. In addition, the pattern of the interaction as shown in ® gure 1 demonstrates that the eŒect of changing state on recall worked in both directions: that is going from rest to exercise and exercise to rest. Further analysis showed that the state of learning per se (exercise versus rest) had no in¯ uence on later recall. This ® nding is consistent with the work of both Tomporowski et al. (1987) who found no impairment in immediate free recall of 15 word lists for participants on a treadmill working at approximatel y 80% maximum heart rate and Sjoberg (1980) who showed that exercise on a bicycle ergometer had no eŒect on paired-associate learning at a range of loadings up to 75% maximum heart rate. The comparison of rest and exercise conditions is important because it eliminates the possibility that the interaction was due to either poor encoding of the stimuli or an inability to retrieve the words while in the exercise condition. The Downloaded By: [Cardiff University] At: 09:37 10 July 2010 State-dependen t memory produced by aerobic exercise 27 data indicate a clear eŒect of changing state on the retrieval of word lists. Material learned either at rest or while exercising aerobically is better retrieved by participants in the same as opposed to the alternate state. The environmental context was consistent across all learning-retrieval combinations and therefore the poorer retrieval in the changed state conditions may be attributed to a mismatch between the incidental cues encoded at original presentation of the material and the cues present at retrieval. The association between the porportional change in heart rate for the changing state conditions and the number of words retrieved, although only marginally signi® cant, does suggest that greater changes in cardiovascular activity are associated with greater reductions in retrieval. This ® nding is consistent with the view that statedependent retrieval is a cue-dependent phenomenon. Those incidental cues assimilated by the participant at learning become increasingly less eŒective as retrieval cues as the change in heart rate increases because of the associated increase in potential distracter cues. Finally, the act of pedalling the bicycle ergometer will have brought about a number of changes in physiological state in addition to that of increasing the heart rate (e.g. rate of oxygen uptake, increase in body temperature). It is likely that a change in a number of physiological parameters would together have acted to bring about a change of state in, and therefore associated retrieval cues available to, the participant. This also applies to the Godden and Baddeley (1975) and the Martin and Aggleton (1993) studies where the change in context from above and below water and vice versa was undoubtedly associated with a number of changes in the physiological state of the participant. The authors measured only change in heart rate but it is possible that changes in other such physiological parameters may better predict the performance change. These data may have implications for sports people who need to memorize information beyond the capacity of their immediate memory span and who need to recall it while competing. Initial learning of the information should take place when the physiological state of the competitor most closely matches that in which recall is required during competition. Acknowledgements The authors wish to thank John Aggleton and an anonymous referee for their comments on an earlier draft of this paper. References BARTLETT, J. C. and SANTROCK, J. W. 1979, AŒect-dependent episodic memory in young children, Child Development, 50, 513 ± 518. CAPALDI. E. J. and NEATH, I. 1995, Remembering and forgetting as context discrimination, Learning & Memory, 2, 107 ± 132. EICH, E. 1980, The cue-dependent nature of state-dependent retrieval, Memory and Cognition, 8, 157 ± 173. EICH, E. 1989, Theoretical issues in state dependent memory, in H. L. Roediger III and F. I. M. Craik (eds), Varieties of Memory and Consciousness: Essays in Honour of Endel Tulving (Erlbaum, Hillsdale, NJ), 351 ± 354. GODDEN , D. R. and BADDELEY , A. D. 1975, Context dependent memory in two natural environments: on land and under water, British Journal of Psychology, 66, 325 ± 332 GOODWIN, D. W., POWELL, B., BREMER, H., HOINE and STERN, J. 1969, Alcohol and recall: state dependent eŒects in man, Science, 163, 1358. KUCERA, H. J. and FRANCIS, W. N. 1967, A Computational Analysis of Present-day American English (Brown University Press, Providence, RI). 28 C. Miles and E. Hardman Downloaded By: [Cardiff University] At: 09:37 10 July 2010 MARTIN, K. M. and AGGLETON, J. P. 1993, Contextual eŒects on the ability of divers to use decompression tables, Applied Cognitive Psychology, 7, 311 ± 316. MCARDLE, W. D., KATCH, F. I. and KATCH, V. L. 1994, Essentials of Exercise Physiology (Lea and Febiger, London). SCHAB, F. R. 1990, Odors and the remembrance of things past, Journal of Experimental Psychology: Learning, Memory and Cognition, 16, 648 ± 655. SJOBERG, H. 1980, Physical ® tness and mental performance during and after work, Ergonomics, 23, 977 ± 985. STRAND, B. Z. 1970, Change of context and retroactive inhibition, Journal of Verbal Learning and Verbal Behavior, 9, 202 ± 206. TOMPOROWSKI, P. D., ELLIS, N. R. and STEPHENS, R. 1987, The immediate eŒects of strenuous exercise on free-recall memory, Ergonomics, 30, 121 ± 129. TULVING, E. 1983, Elements of Episodic Memory (Oxford University Press, New York). WEINGARTNER , H. and FAILLACE, L. A. 1971, Alcohol state-dependent learning in man, Journal of Nervous and Mental Disorders, 153, 395 ± 406.