International Journal of Neuroscience, 2014; 124(11): 806–811 Copyright © 2014 Informa Healthcare USA, Inc. ISSN: 0020-7454 print / 1543-5245 online DOI: 10.3109/00207454.2014.890619
RESEARCH ARTICLE
Working memory performance and exposure to pleasant and unpleasant ambient odor: Is spatial span special? G. N. Martin1 and A. Chaudry2 Human Olfaction Laboratory, Middlesex University, London, United Kingdom; 2 Associate Member, Business Psychology Group, UCL, London, United Kingdom
Int J Neurosci Downloaded from informahealthcare.com by The University of Manchester on 11/21/14 For personal use only.
1
Sensory influences on working memory (WM) performance were investigated in 86 healthy adults. Participants were exposed to an ambient pleasant odor (lemon), unpleasant odor (machine oil) or no odor during completion of three WM tests from the Wechsler Memory Scale-III: the letter–number sequencing, spatial span and digit span tests. A significant main effect of odor was found for spatial span but no other task: scores were significantly lower in the unpleasant odor condition than the pleasant odor condition. Significant odor × sex interactions were found for the spatial span, digit span and letter–number sequencing tasks: men’s spatial span scores were lower in the unpleasant odor condition than in the control condition, and women’s scores were significantly better in the pleasant odor condition than in the unpleasant odor condition. The results suggest that ambient odor may impair or facilitate specific types of WM depending on the task, sex of the participant and affective characteristics of the odor. KEYWORDS: odor, working memory, sex differences, spatial span, pleasantness
Introduction Odor has long been considered an effective memory cue in the fictional and scientific literature. The Proust phenomenon whereby an odor reinstates the context in which it was originally experienced (named after the author of A La recherch´e du Temps Perdu, the autobiographical novel provoked by this olfactory experience), has been well demonstrated experimentally [1–7]; although see [8]. Exposure to pleasant and unpleasant odors has been associated with the recall of positive and negative memories, respectively, [9] and verbal material is more effectively recalled when the same odor is presented at encoding and recall [6]. Studies also report that exposure to odor can influence cognition beyond contextreinstatement memory, to specific types of memory such as working memory (WM; e.g. [10,11]). In the original Baddeley and Hitch model [12], WM described the process whereby material was maintained and manipulated in short-term memory, and comprised two slave components – the articulatory loop, later re-named the phonological loop, and the Received 5 March 2013; revised 30 January 2014; accepted 30 January 2014 Correspondence: Dr. G. N. Martin, Department of Psychology, Middlesex University, London, United Kingdom. E-mail: [email protected]
806
visuospatial sketchpad – and a central controller or manipulator of the to-be-remembered material, the central executive. In the first systematic study of the effect of olfactory stimulation on a variant of WM, participants exposed to the odor of lavender but not rosemary showed significantly poorer WM performance and longer response times compared with the control condition [11]. In a follow-up study, exposure to the odor of peppermint was associated with higher scores on (combined) spatial and numeric WM tasks than was ylang ylang [13]. No significant difference between the odor conditions and the control conditions was found. A functional Magnetic Resonance Imaging (fMRI) study of WM and odor exposure [10] found that participants who responded negatively to the unpleasant odor of yeast (some did not) showed a reduction in 2-back task performance and reaction time was slower in this condition. The study did not include a pleasant odor as a comparison or as an internal control to account for the possibility of distraction/stimulation effects. The yeast odor was also rated more intense than the odorless air, which is a probable confound. When the effect of exposure to the odor of yeast on the 0-back/2-back continuous performance was investigated in a group of schizophrenic
Int J Neurosci Downloaded from informahealthcare.com by The University of Manchester on 11/21/14 For personal use only.
Odor and working memory
patients, exposure was associated with increased reaction time and fewer hits compared to healthy controls [14]. This effect of odor on WM is illuminating given that the primary, immediate response to olfactory stimulation is hedonic (see [15] for review). The effects of pleasant and unpleasant odors on WM performance might be predicted in the context of current theorizing concerning the operation of WM, specifically that WM resources are depleted in the context of distracting, unpleasant stimuli and less so during exposure to pleasant stimuli [16]. For example, negative emotion induction has been associated with impairment in spatial working memory (SWM) [17], and in visuospatial working memory [16,18]. The effect has been attributed to enhanced competition between attentional resources such that attention to a threat (a negative stimulus that leads to the experience of a negative emotion) competes with the attention required to concentrate on spatial WM performance [16]. If this explanation is correct, then an unpleasant odor should differentially affect spatial (and other types of) WM. However, if the WM system is competing for resources, then it is possible that any olfactory stimulus will create an impairment in the olfactory domain, as it provides additional attentional distraction and a drain on the cognitive resources required to permit WM to operate effectively and efficiently. Therefore, we hypothesize that exposure to ambient unpleasant odor will exert a more adverse effect on WM – specifically SWM – than would a pleasant odor or no odor. No study to date has examined or tested this hypothesis. Such a study would help clarify the role of extraneous, hedonic, sensory stimuli in WM performance, specifically whether certain types of odor disrupt WM differentially and whether the cause of WM disruption is limited to nonolfactory stimuli.
Methods Participants Eighty-six right-handed, nonsmoking undergraduates (45 women; 41 men; median age = 20 years; age range = 19–25) from a North London University participated. All reported being free of respiratory infection at the time of testing and none had history of head injury or anosmia. None were receiving prescription medication for illness. Participants were randomly assigned to one of three conditions: lemon odor (13 men; 15 women), machine oil (13 men; 15 women) or an unscented control condition (15 men; 15 women). As women appear to show superior olfactory sensitivity, discrimination and identification ability to men and are more susceptible to C
2014 Informa Healthcare USA, Inc.
807
suggestions concerning the effects of odor [15,19], sex was included as a variable in the current study.
Materials Odors and odor delivery Previous research has indicated that odors rated as pleasant and unpleasant are lemon and machine oil, respectively [20,21]. These odors were administered in the current experiment via an AromaCo Aromacube, a commercial and professional device designed to diffuse ambient odor in small- to medium-sized environments (see http://www.aromaco.co.uk). Odors were allowed to diffuse for 10 min prior to testing and allowed constant diffusion throughout the course of the experiment. The machine was placed in the corner of the room, was unobtrusive and was not visible to participants. All participants in the odor conditions reported being able to detect an odor in the testing room when asked to indicate so at the end of their session. Testing took place in a 5 m × 2 m × 2 m experimental room, purposebuilt for behavioral olfactory research, which was windowless and well ventilated. The room contained a table at which the participant sat and completed the tests. At the end of the experiment, participants were asked to indicate whether they could detect an odor in the room and those in the odor conditions indicated that the room was scented. They were then asked whether the scent was pleasant or unpleasant. Participants’ responses were consistent with the odor in the room.
Memory measures Hard copy measures of working or short-term memory were administered and comprised the letter–number sequencing, digit span and spatial span tests from the Wechsler Memory Scale-III. In the letter–number sequencing task, a WM task, the experimenter reads to the participants a string of numbers and letters, which the participants rearrange and repeat back either in numerical or alphabetical order, respectively. Initial trials begin with one letter and one number (e.g. L-2) and end with four letters and four numbers (e.g. 7-M-2-T-6-F-1-Z). Total score can range from 0 to 21. In the digit span task, participants are read a string of numbers and asked to repeat them in the same order and backward. The initial trials comprised two numbers (e.g. 1–7); the final trials, nine numbers (e.g. 2–7–5–8–6–2–5–8–4). Total score can range from 0 to 30. In the spatial span task, participants observe an experimenter touching a collection of blocks and then recall the order in which the blocks were touched,
Int J Neurosci Downloaded from informahealthcare.com by The University of Manchester on 11/21/14 For personal use only.
808
G. N. Martin & A. Chaudry
forward and backward. The initial trials comprised two numbers (e.g. 1–7); the final trials, nine numbers (e.g. 2–7–5–8–6–2–5–8–4). Total score can range from 0 to 32.
pleasant condition is associated with the poorest performance. The control condition falls between the scores of both.
Procedure Participants responded to advertisements placed around campus and were asked to arrive at a testing cubicle where they were met by the experimenter. They were informed that they were participating in an experiment investigating the effects of environmental variables on memory performance and were randomly assigned to one of the three odor conditions. Participants were seated at a table and the experiment was described to them. After providing written-informed consent, they completed a questionnaire requesting information about their sex, handedness, general health, anosmia, etc., followed by a range of cognitive and memory tests, including three measures of short-term and WM described above, and where appropriate, with the experimenter present (see above). The memory tests were presented in counterbalanced order. Each session, including the consent procedure and the full debriefing at the end of the experiment, lasted approximately 30–40 min. The study had received ethical approval from the Department of Psychology’s Ethics Committee.
Spatial span
Results The data from the three measures were subjected to a 2 (sex) × 3 (odor condition) independent groups analysis of variance (ANOVA). We made no predictions regarding the relative performance between tasks, only differences within tasks due to olfactory stimulation. Mean scores and standard deviations are presented in Table 1. As Table 1 suggests, the pattern of results indicates an effect of odor in that the pleasant condition is associated with better performance on all three tasks and the un-
The ANOVA revealed a significant main effect of odor for spatial span scores, F (2, 80) = 3.27, p < 0.05. Tukey’s HSD tests showed that participants in the lemon odor condition performed significantly better than did those in the machine oil condition (p < 0.05), but not the control condition. A main effect of sex was also found, F (1, 80) = 7.62, p < 0.01, with women significantly outscoring the men on this task. However, both main effects were tempered by a significant sex × odor interaction, F (2, 80) = 3.53, p = 0.04. Men in the control condition were significantly better at the task than were those in the unpleasant odor condition (p < 0.01).
Digit span A significant sex × odor interaction, F (2, 80) = 4.26, p < 0.05, was found for digit span. Tests of simple effects showed that women in the lemon odor condition performed significantly better than did women in the machine oil odor condition (p < 0.05). No main effects of sex, F (1, 80) = 3.86, p = 0.053, or odor, F (2, 80) = 0.82, p = 0.44, were found.
Letter–number sequencing No significant effect of odor, F (2, 80) = 1.58, p = 0.21, or sex, F (1, 80) = 0.90, p = 0.35, was found for letter–number sequencing but a significant sex × odor interaction did emerge, F (2, 80) = 3.50, p < 0.05. Tests of simple effects found that women performed significantly better in the lemon odor condition than did those in the machine oil odor condition (p < 0.05).
Table 1. Mean scores and standard deviations (in parentheses) for the letter–number sequencing, spatial span and digit span tasks, by sex and odor condition. Lemon
Letter–number sequencing Spatial span Digit span
Machine oil
Control
Male
Female
M and F
Male
Female
M and F
Male
Female
M and F
10.85 (1.86) 15.15 (3.29) 19.23 (2.49)
11.93 (2.49) 17.86 (4.03) 20.93 (4.89)
11.42 (2.25) 16.60 (3.89) 20.14 (3.99)
11.30 (1.25) 12.69 (1.65) 20.38 (1.98)
9.60 (2.13) 16.07 (3.24) 17.40 (3.52)
10.39 (1.95) 14.50 (3.10) 18.79 (3.24)
11.33 (1.76) 16.07 (3.06) 21.47 (4.32)
10.73 (2.07) 15.47 (2.45) 18.13 (3.36)
11.03 (1.92) 15.77 (2.74) 19.80 (4.16)
International Journal of Neuroscience
Odor and working memory
Int J Neurosci Downloaded from informahealthcare.com by The University of Manchester on 11/21/14 For personal use only.
Discussion In the current study, we sought to investigate whether certain types of olfactory stimulation would affect performance on WM tasks positively or adversely. Specifically, we investigated whether exposure to pleasant and unpleasant ambient odors affected WM performance in men and women. Based on previous research, we hypothesized that exposure to unpleasant odor would be associated with a significant impairment in WM performance compared to pleasant odor or no odor. We found that the performance of men and women on one of the WM tasks – spatial span – was significantly poorer in the unpleasant odor condition than the pleasant odor condition, thus partly replicating and supporting previous findings of a study administering a different SWM task to ours [14]. Furthermore, we found that women’s performance on the letter–number sequencing and digit span performance was significantly worse in the unpleasant odor condition than the pleasant odor condition, that women’s digit span performance was significantly better than men’s in the pleasant odor condition, and that men’s spatial span performance was significantly impaired in the unpleasant odor condition compared to the control condition. Across all tasks, exposure to the pleasant odor was associated with generally better performance than was exposure to the unpleasant odor, as Table 1 illustrates, although on only one of these tasks was the main effect of odor statistically significant. The unpleasant odor condition was associated with the poorest spatial span performance as well as the poorest performance on all tasks (although not significantly so on all tasks), suggesting that exposure to an unpleasant scent can lead to an impairment of aspects of WM. The effect of the unpleasant odor is consistent with previous research in which perception of negatively valenced photographs, and “threat-of-harm,” were associated with impaired SWM [16]. Our results are consistent with the explanation [16, 17], that the detrimental effect of negative emotional stimuli on WM may cause automatic distraction because of competition for resources and that this relationship may be mediated by the strength of the negative emotional stimulus [22]. We did not test the strength-of-stimulus hypothesis directly; we exposed participants to distinctive odors of opposing valence based on previous research and pilot work. One limitation of the current study is that the valence of the odor was not measured directly, but indirectly – all participants were asked whether they were able to detect an odor and whether they found it pleasant or unpleasant. The quality/strength-of-stimulus hypothesis could be tested more comprehensively via the administration of pleasant odors that are known to have demonstrable
C
2014 Informa Healthcare USA, Inc.
809
sedative/relaxing qualities, such as bergamot [23] or chocolate [24] and compare the response to these stimuli with that to arousing odors such as lemon or peppermint [25]. In similar studies, others have found that the odor of lavender and ylang ylang was associated with poorer WM performance [11,13]. Based on these findings, it is possible that pleasant and alerting stimuli/odors in general might facilitate performance on standard WM tests (as was partly found here) and pleasant and relaxing stimuli/odors might impair performance relative to the control condition because of the greater interference in/ loss of attention during WM task performance (and the shifting of cognitive resources). It is also noteworthy that although scores were lowest across all tests in the unpleasant odor condition, they were highest in the lemon odor condition, thus suggesting that additional extraneous olfactory stimuli in general do not draw on more WM resources (the control condition’s response fell between that of two odor conditions). The specific mechanism underpinning olfactoryrelated WM interference is unclear. It is possible that a WM memory task featuring odor-relevant content would lead to better performance because the two elements of the experiment are compatible and consistent with each other. Odor-irrelevant content would lead to more disrupted performance, or produce no effect on performance, because the context would be incompatible with, or irrelevant to, the content. There is some evidence that such a scenario might arise, as others [26] have demonstrated that chemosensory stimuli can be selectively attended to. The main effect of odor in the spatial span task was tempered by a sex × odor interaction whereby the men’s control group performance was significantly less impaired than performance of men in the unpleasant odor condition. There were no other interactions with sex. One of the most well-established sex differences in the WM literature is a male advantage for visuospatial WM and transformational visuospatial WM specifically [27], a finding which may have a neural basis [28]. The current findings suggest that the addition of an unpleasant stimulus appears to compromise this function more consistently in men, although it is notable that on this task in this study women generally performed better than men. The sex × odor interactions are noteworthy as sex is rarely considered a variable in relevant studies. One WM study, for example [17], recruited only men. Sex was included as a variable because women have been found to demonstrate superior olfactory sensitivity, consistently outperform men on tests of olfactory detection, identification and discrimination tasks are significantly influenced by odor stimulation (see [15] for
Int J Neurosci Downloaded from informahealthcare.com by The University of Manchester on 11/21/14 For personal use only.
810
G. N. Martin & A. Chaudry
review) and have better odor recognition memory [29]. In two tasks (digit span and letter–number sequencing), women performed better in the lemon odor condition than in the machine oil odor condition – or performed worse in the unpleasant odor condition. What is noteworthy about these particular interactions, however, is that they are different to that found for spatial span in that women were more adversely affected by the unpleasant stimulus in the digit span and letter–sequencing tasks, the two tasks with the less salient spatial content/weight, and compared to responses in the pleasant odor condition. When comparing responses to the two odors specifically, one might speculate that this reflects women’s greater sensitivity to odor, particularly pleasant odor [19,30,31] which, in turn, may make them more susceptible to its cognitive effects. But this does not explain their comparatively poorer performance during exposure to unpleasant odor. In conclusion, we found that exposure to unpleasant odor was associated with generally poorer WM performance compared to a pleasant odor or control, and significantly poorer spatial span performance. This finding extends those of other studies, which have demonstrated poorer spatial span and visuospatial performance during exposure to negative olfactory and nonolfactory emotional stimuli. The sex × odor interactions reported indicate the importance of sex as a variable in olfaction studies exploring memory and cognition and specifically suggest that men may be more detrimentally affected by exposure to unpleasant odor during spatial span WM, whereas women are more detrimentally affected by negative stimuli on other WM tasks when performance is compared with exposure to positive stimuli.
Acknowledgements The authors thank Kelly Lyons and two other reviewers for their helpful and constructive comments on an earlier draft of this manuscript.
Declaration of Interest The authors report no conflict of interest. The authors alone are responsible for the content and writing of this paper.
References 1. Aggleton JP, Waskett, L. The ability of odors to serve as statedependent cues for real world memories: Can Viking smells aid the recall of Viking experience? Br J Psychol 1999;90:1– 8.
2. Cann A, Ross DA. Olfactory stimuli as context cues in human memory. Am J Psycholo 1989;102:91–102. 3. Chu S, Downes J. Odor-evoked autobiographical memories: psychological investigations of Proustian phenomena. Chem Senses 2000;25:111–16. 4. Herz RS. The effects of cue-distinctiveness on odor-based context dependent memory. Mem Cogn 1997;25:375– 80. 5. Parker A, Ngu H, Cassaday HJ. Odor and Proustian memory: reduction of context-dependent forgetting and multiple forms of memory. Applied Cogn Psychol 2001;15:159–71. 6. Schab FR. Odors and the remembrance of things past. J Exp Psychol: Learning, Memory and Cognition 1990;16:648– 55. 7. Smith DG, Standing L, de Man A. Verbal memory elicited by ambient odor. Percept Mot Skills 1992;74:339–43. 8. Eich E. Fragrances as cues for remembering words. J Verbal Learn Verbal Behav 1977;17:103–11. 9. Ehrlichman H, Halpern JN. Affect and memory: effects of pleasant and unpleasant odors on retrieval of happy and unhappy memories. J Pers Soc Psychol 1988;55:769–79. 10. Habel U, Loch K, Pauly K, et al. The influence of olfactoryinduced negative emotion on verbal working memory: individual differences in neurobehavioural findings. Brain Res 2007;1152:158–70. 11. Moss M, Cook J, Wesnes K, Duckett P. Aromas of rosemary and lavender essential oils differentially affect cognition and mood in healthy adults. Int J Neurosci 2003;113:15–38. 12. Baddeley AD, Hitch GJ. Working memory. In: Bower G, ed. The Psychology of learning and motivation. New York: Academic Press; 1974:47–89. 13. Moss M, Hewitt S, Moss L, Wesnes K. Modulation of cognitive performance and mood by aromas of peppermint and ylangylang. Int J Neurosci 2008;118:59–77. 14. Schneider F, Koch K, Reske M, et al. Interaction of negative olfactory stimulation and working memory in schizophrenia patients: development and evaluation of a behavioural neuroimaging task. Psychiatry Res 2006;144:123–30. 15. Martin GN. The neuropsychology of smell and taste. Hove: The Psychology Press; 2013. 16. Lavric A, Rippon G, Gray JR. Threat-evoked anxiety disrupts spatial working memory performance: an attentional account. Cogn Ther Res 2003;10:489–504. 17. Li X, Chan RCK, Luo Y-J. Stage effects of negative emotion on spatial and verbal working memory. BMC Neurosci 2010;11:60. 18. Awh E, Jonides J. Overlapping mechanisms of attention and spatial working memory. Trends Cogn Sci 2001;5: 119–26. 19. Brand G, Millot J-L. Sex differences in human olfaction: between evidence and enigma. Q J Exp Psychol 2001;54B: 259–70. 20. Martin GN. The effect of exposure to odor on the perception of pain. Psychosom Med 2006;68:613–16. 21. Martin GN, Cooper JA. Adding zest to difficult journeys: odoreffects on simulated driving. BPS Annual Conference, University of York, York, 2007, 23 March. 22. Li X, Li X, Luo Y-J. Anxiety and attentional bias for threat: an event-related potential study. Neuroreport 2005;16:1501–5. 23. Gould A, Martin GN. A good odor to breathe? The effect of pleasant ambient odor on human visual vigilance. Appl Cogn Psychol 2001;15:225–32. 24. Martin GN. Human electroencephalographic (EEG) response to olfactory stimulation: two experiments using the aroma of food. Int J Psychophysiol 1998;30:287–302. 25. Ho C, Spence C. Olfactory facilitation of dual-task performance. Neurosci Lett 2005;389:35–40.
International Journal of Neuroscience
Odor and working memory
Int J Neurosci Downloaded from informahealthcare.com by The University of Manchester on 11/21/14 For personal use only.
26. Spence C, Kettenmann B, Kobal G, McGlone F. Selective attention to the chemosensory modality. Percept Psychophys 2000;62:1265–71. 27. Loring-Meier S, Halpern DF. Sex differences in visuospatial working memory: components of cognitive processing. Psychon Bull Rev 1999;6:464–71. 28. Speck O, Ernst T, Braun J, et al. Gender differences in the functional organization of the brain for working memory. Neuroreport 2000;11:2581–5.
C
2014 Informa Healthcare USA, Inc.
811
29. Larsson M, Lovden M, Nilsson L-G. Sex differences in recollective experience for olfactory and verbal information. Acta Psychol 2003;112:89–103. 30. Andersson L, Lundberg C, Astrom J, Nordin S. Chemosensory attention, habituation and detection in women and men. Int J Psychophysiol 2011;79:316–22. 31. Thuerauf N, Reulbach U, Lunkenheimer J, et al. Emotional reactivity to odors: olfactory sensitivity and the span of emotional evaluation separate the genders. Neurosci Lett 2009;456:74–79.
RESEARCH ARTICLE
Working memory performance and exposure to pleasant and unpleasant ambient odor: Is spatial span special? G. N. Martin1 and A. Chaudry2 Human Olfaction Laboratory, Middlesex University, London, United Kingdom; 2 Associate Member, Business Psychology Group, UCL, London, United Kingdom
Int J Neurosci Downloaded from informahealthcare.com by The University of Manchester on 11/21/14 For personal use only.
1
Sensory influences on working memory (WM) performance were investigated in 86 healthy adults. Participants were exposed to an ambient pleasant odor (lemon), unpleasant odor (machine oil) or no odor during completion of three WM tests from the Wechsler Memory Scale-III: the letter–number sequencing, spatial span and digit span tests. A significant main effect of odor was found for spatial span but no other task: scores were significantly lower in the unpleasant odor condition than the pleasant odor condition. Significant odor × sex interactions were found for the spatial span, digit span and letter–number sequencing tasks: men’s spatial span scores were lower in the unpleasant odor condition than in the control condition, and women’s scores were significantly better in the pleasant odor condition than in the unpleasant odor condition. The results suggest that ambient odor may impair or facilitate specific types of WM depending on the task, sex of the participant and affective characteristics of the odor. KEYWORDS: odor, working memory, sex differences, spatial span, pleasantness
Introduction Odor has long been considered an effective memory cue in the fictional and scientific literature. The Proust phenomenon whereby an odor reinstates the context in which it was originally experienced (named after the author of A La recherch´e du Temps Perdu, the autobiographical novel provoked by this olfactory experience), has been well demonstrated experimentally [1–7]; although see [8]. Exposure to pleasant and unpleasant odors has been associated with the recall of positive and negative memories, respectively, [9] and verbal material is more effectively recalled when the same odor is presented at encoding and recall [6]. Studies also report that exposure to odor can influence cognition beyond contextreinstatement memory, to specific types of memory such as working memory (WM; e.g. [10,11]). In the original Baddeley and Hitch model [12], WM described the process whereby material was maintained and manipulated in short-term memory, and comprised two slave components – the articulatory loop, later re-named the phonological loop, and the Received 5 March 2013; revised 30 January 2014; accepted 30 January 2014 Correspondence: Dr. G. N. Martin, Department of Psychology, Middlesex University, London, United Kingdom. E-mail: [email protected]
806
visuospatial sketchpad – and a central controller or manipulator of the to-be-remembered material, the central executive. In the first systematic study of the effect of olfactory stimulation on a variant of WM, participants exposed to the odor of lavender but not rosemary showed significantly poorer WM performance and longer response times compared with the control condition [11]. In a follow-up study, exposure to the odor of peppermint was associated with higher scores on (combined) spatial and numeric WM tasks than was ylang ylang [13]. No significant difference between the odor conditions and the control conditions was found. A functional Magnetic Resonance Imaging (fMRI) study of WM and odor exposure [10] found that participants who responded negatively to the unpleasant odor of yeast (some did not) showed a reduction in 2-back task performance and reaction time was slower in this condition. The study did not include a pleasant odor as a comparison or as an internal control to account for the possibility of distraction/stimulation effects. The yeast odor was also rated more intense than the odorless air, which is a probable confound. When the effect of exposure to the odor of yeast on the 0-back/2-back continuous performance was investigated in a group of schizophrenic
Int J Neurosci Downloaded from informahealthcare.com by The University of Manchester on 11/21/14 For personal use only.
Odor and working memory
patients, exposure was associated with increased reaction time and fewer hits compared to healthy controls [14]. This effect of odor on WM is illuminating given that the primary, immediate response to olfactory stimulation is hedonic (see [15] for review). The effects of pleasant and unpleasant odors on WM performance might be predicted in the context of current theorizing concerning the operation of WM, specifically that WM resources are depleted in the context of distracting, unpleasant stimuli and less so during exposure to pleasant stimuli [16]. For example, negative emotion induction has been associated with impairment in spatial working memory (SWM) [17], and in visuospatial working memory [16,18]. The effect has been attributed to enhanced competition between attentional resources such that attention to a threat (a negative stimulus that leads to the experience of a negative emotion) competes with the attention required to concentrate on spatial WM performance [16]. If this explanation is correct, then an unpleasant odor should differentially affect spatial (and other types of) WM. However, if the WM system is competing for resources, then it is possible that any olfactory stimulus will create an impairment in the olfactory domain, as it provides additional attentional distraction and a drain on the cognitive resources required to permit WM to operate effectively and efficiently. Therefore, we hypothesize that exposure to ambient unpleasant odor will exert a more adverse effect on WM – specifically SWM – than would a pleasant odor or no odor. No study to date has examined or tested this hypothesis. Such a study would help clarify the role of extraneous, hedonic, sensory stimuli in WM performance, specifically whether certain types of odor disrupt WM differentially and whether the cause of WM disruption is limited to nonolfactory stimuli.
Methods Participants Eighty-six right-handed, nonsmoking undergraduates (45 women; 41 men; median age = 20 years; age range = 19–25) from a North London University participated. All reported being free of respiratory infection at the time of testing and none had history of head injury or anosmia. None were receiving prescription medication for illness. Participants were randomly assigned to one of three conditions: lemon odor (13 men; 15 women), machine oil (13 men; 15 women) or an unscented control condition (15 men; 15 women). As women appear to show superior olfactory sensitivity, discrimination and identification ability to men and are more susceptible to C
2014 Informa Healthcare USA, Inc.
807
suggestions concerning the effects of odor [15,19], sex was included as a variable in the current study.
Materials Odors and odor delivery Previous research has indicated that odors rated as pleasant and unpleasant are lemon and machine oil, respectively [20,21]. These odors were administered in the current experiment via an AromaCo Aromacube, a commercial and professional device designed to diffuse ambient odor in small- to medium-sized environments (see http://www.aromaco.co.uk). Odors were allowed to diffuse for 10 min prior to testing and allowed constant diffusion throughout the course of the experiment. The machine was placed in the corner of the room, was unobtrusive and was not visible to participants. All participants in the odor conditions reported being able to detect an odor in the testing room when asked to indicate so at the end of their session. Testing took place in a 5 m × 2 m × 2 m experimental room, purposebuilt for behavioral olfactory research, which was windowless and well ventilated. The room contained a table at which the participant sat and completed the tests. At the end of the experiment, participants were asked to indicate whether they could detect an odor in the room and those in the odor conditions indicated that the room was scented. They were then asked whether the scent was pleasant or unpleasant. Participants’ responses were consistent with the odor in the room.
Memory measures Hard copy measures of working or short-term memory were administered and comprised the letter–number sequencing, digit span and spatial span tests from the Wechsler Memory Scale-III. In the letter–number sequencing task, a WM task, the experimenter reads to the participants a string of numbers and letters, which the participants rearrange and repeat back either in numerical or alphabetical order, respectively. Initial trials begin with one letter and one number (e.g. L-2) and end with four letters and four numbers (e.g. 7-M-2-T-6-F-1-Z). Total score can range from 0 to 21. In the digit span task, participants are read a string of numbers and asked to repeat them in the same order and backward. The initial trials comprised two numbers (e.g. 1–7); the final trials, nine numbers (e.g. 2–7–5–8–6–2–5–8–4). Total score can range from 0 to 30. In the spatial span task, participants observe an experimenter touching a collection of blocks and then recall the order in which the blocks were touched,
Int J Neurosci Downloaded from informahealthcare.com by The University of Manchester on 11/21/14 For personal use only.
808
G. N. Martin & A. Chaudry
forward and backward. The initial trials comprised two numbers (e.g. 1–7); the final trials, nine numbers (e.g. 2–7–5–8–6–2–5–8–4). Total score can range from 0 to 32.
pleasant condition is associated with the poorest performance. The control condition falls between the scores of both.
Procedure Participants responded to advertisements placed around campus and were asked to arrive at a testing cubicle where they were met by the experimenter. They were informed that they were participating in an experiment investigating the effects of environmental variables on memory performance and were randomly assigned to one of the three odor conditions. Participants were seated at a table and the experiment was described to them. After providing written-informed consent, they completed a questionnaire requesting information about their sex, handedness, general health, anosmia, etc., followed by a range of cognitive and memory tests, including three measures of short-term and WM described above, and where appropriate, with the experimenter present (see above). The memory tests were presented in counterbalanced order. Each session, including the consent procedure and the full debriefing at the end of the experiment, lasted approximately 30–40 min. The study had received ethical approval from the Department of Psychology’s Ethics Committee.
Spatial span
Results The data from the three measures were subjected to a 2 (sex) × 3 (odor condition) independent groups analysis of variance (ANOVA). We made no predictions regarding the relative performance between tasks, only differences within tasks due to olfactory stimulation. Mean scores and standard deviations are presented in Table 1. As Table 1 suggests, the pattern of results indicates an effect of odor in that the pleasant condition is associated with better performance on all three tasks and the un-
The ANOVA revealed a significant main effect of odor for spatial span scores, F (2, 80) = 3.27, p < 0.05. Tukey’s HSD tests showed that participants in the lemon odor condition performed significantly better than did those in the machine oil condition (p < 0.05), but not the control condition. A main effect of sex was also found, F (1, 80) = 7.62, p < 0.01, with women significantly outscoring the men on this task. However, both main effects were tempered by a significant sex × odor interaction, F (2, 80) = 3.53, p = 0.04. Men in the control condition were significantly better at the task than were those in the unpleasant odor condition (p < 0.01).
Digit span A significant sex × odor interaction, F (2, 80) = 4.26, p < 0.05, was found for digit span. Tests of simple effects showed that women in the lemon odor condition performed significantly better than did women in the machine oil odor condition (p < 0.05). No main effects of sex, F (1, 80) = 3.86, p = 0.053, or odor, F (2, 80) = 0.82, p = 0.44, were found.
Letter–number sequencing No significant effect of odor, F (2, 80) = 1.58, p = 0.21, or sex, F (1, 80) = 0.90, p = 0.35, was found for letter–number sequencing but a significant sex × odor interaction did emerge, F (2, 80) = 3.50, p < 0.05. Tests of simple effects found that women performed significantly better in the lemon odor condition than did those in the machine oil odor condition (p < 0.05).
Table 1. Mean scores and standard deviations (in parentheses) for the letter–number sequencing, spatial span and digit span tasks, by sex and odor condition. Lemon
Letter–number sequencing Spatial span Digit span
Machine oil
Control
Male
Female
M and F
Male
Female
M and F
Male
Female
M and F
10.85 (1.86) 15.15 (3.29) 19.23 (2.49)
11.93 (2.49) 17.86 (4.03) 20.93 (4.89)
11.42 (2.25) 16.60 (3.89) 20.14 (3.99)
11.30 (1.25) 12.69 (1.65) 20.38 (1.98)
9.60 (2.13) 16.07 (3.24) 17.40 (3.52)
10.39 (1.95) 14.50 (3.10) 18.79 (3.24)
11.33 (1.76) 16.07 (3.06) 21.47 (4.32)
10.73 (2.07) 15.47 (2.45) 18.13 (3.36)
11.03 (1.92) 15.77 (2.74) 19.80 (4.16)
International Journal of Neuroscience
Odor and working memory
Int J Neurosci Downloaded from informahealthcare.com by The University of Manchester on 11/21/14 For personal use only.
Discussion In the current study, we sought to investigate whether certain types of olfactory stimulation would affect performance on WM tasks positively or adversely. Specifically, we investigated whether exposure to pleasant and unpleasant ambient odors affected WM performance in men and women. Based on previous research, we hypothesized that exposure to unpleasant odor would be associated with a significant impairment in WM performance compared to pleasant odor or no odor. We found that the performance of men and women on one of the WM tasks – spatial span – was significantly poorer in the unpleasant odor condition than the pleasant odor condition, thus partly replicating and supporting previous findings of a study administering a different SWM task to ours [14]. Furthermore, we found that women’s performance on the letter–number sequencing and digit span performance was significantly worse in the unpleasant odor condition than the pleasant odor condition, that women’s digit span performance was significantly better than men’s in the pleasant odor condition, and that men’s spatial span performance was significantly impaired in the unpleasant odor condition compared to the control condition. Across all tasks, exposure to the pleasant odor was associated with generally better performance than was exposure to the unpleasant odor, as Table 1 illustrates, although on only one of these tasks was the main effect of odor statistically significant. The unpleasant odor condition was associated with the poorest spatial span performance as well as the poorest performance on all tasks (although not significantly so on all tasks), suggesting that exposure to an unpleasant scent can lead to an impairment of aspects of WM. The effect of the unpleasant odor is consistent with previous research in which perception of negatively valenced photographs, and “threat-of-harm,” were associated with impaired SWM [16]. Our results are consistent with the explanation [16, 17], that the detrimental effect of negative emotional stimuli on WM may cause automatic distraction because of competition for resources and that this relationship may be mediated by the strength of the negative emotional stimulus [22]. We did not test the strength-of-stimulus hypothesis directly; we exposed participants to distinctive odors of opposing valence based on previous research and pilot work. One limitation of the current study is that the valence of the odor was not measured directly, but indirectly – all participants were asked whether they were able to detect an odor and whether they found it pleasant or unpleasant. The quality/strength-of-stimulus hypothesis could be tested more comprehensively via the administration of pleasant odors that are known to have demonstrable
C
2014 Informa Healthcare USA, Inc.
809
sedative/relaxing qualities, such as bergamot [23] or chocolate [24] and compare the response to these stimuli with that to arousing odors such as lemon or peppermint [25]. In similar studies, others have found that the odor of lavender and ylang ylang was associated with poorer WM performance [11,13]. Based on these findings, it is possible that pleasant and alerting stimuli/odors in general might facilitate performance on standard WM tests (as was partly found here) and pleasant and relaxing stimuli/odors might impair performance relative to the control condition because of the greater interference in/ loss of attention during WM task performance (and the shifting of cognitive resources). It is also noteworthy that although scores were lowest across all tests in the unpleasant odor condition, they were highest in the lemon odor condition, thus suggesting that additional extraneous olfactory stimuli in general do not draw on more WM resources (the control condition’s response fell between that of two odor conditions). The specific mechanism underpinning olfactoryrelated WM interference is unclear. It is possible that a WM memory task featuring odor-relevant content would lead to better performance because the two elements of the experiment are compatible and consistent with each other. Odor-irrelevant content would lead to more disrupted performance, or produce no effect on performance, because the context would be incompatible with, or irrelevant to, the content. There is some evidence that such a scenario might arise, as others [26] have demonstrated that chemosensory stimuli can be selectively attended to. The main effect of odor in the spatial span task was tempered by a sex × odor interaction whereby the men’s control group performance was significantly less impaired than performance of men in the unpleasant odor condition. There were no other interactions with sex. One of the most well-established sex differences in the WM literature is a male advantage for visuospatial WM and transformational visuospatial WM specifically [27], a finding which may have a neural basis [28]. The current findings suggest that the addition of an unpleasant stimulus appears to compromise this function more consistently in men, although it is notable that on this task in this study women generally performed better than men. The sex × odor interactions are noteworthy as sex is rarely considered a variable in relevant studies. One WM study, for example [17], recruited only men. Sex was included as a variable because women have been found to demonstrate superior olfactory sensitivity, consistently outperform men on tests of olfactory detection, identification and discrimination tasks are significantly influenced by odor stimulation (see [15] for
Int J Neurosci Downloaded from informahealthcare.com by The University of Manchester on 11/21/14 For personal use only.
810
G. N. Martin & A. Chaudry
review) and have better odor recognition memory [29]. In two tasks (digit span and letter–number sequencing), women performed better in the lemon odor condition than in the machine oil odor condition – or performed worse in the unpleasant odor condition. What is noteworthy about these particular interactions, however, is that they are different to that found for spatial span in that women were more adversely affected by the unpleasant stimulus in the digit span and letter–sequencing tasks, the two tasks with the less salient spatial content/weight, and compared to responses in the pleasant odor condition. When comparing responses to the two odors specifically, one might speculate that this reflects women’s greater sensitivity to odor, particularly pleasant odor [19,30,31] which, in turn, may make them more susceptible to its cognitive effects. But this does not explain their comparatively poorer performance during exposure to unpleasant odor. In conclusion, we found that exposure to unpleasant odor was associated with generally poorer WM performance compared to a pleasant odor or control, and significantly poorer spatial span performance. This finding extends those of other studies, which have demonstrated poorer spatial span and visuospatial performance during exposure to negative olfactory and nonolfactory emotional stimuli. The sex × odor interactions reported indicate the importance of sex as a variable in olfaction studies exploring memory and cognition and specifically suggest that men may be more detrimentally affected by exposure to unpleasant odor during spatial span WM, whereas women are more detrimentally affected by negative stimuli on other WM tasks when performance is compared with exposure to positive stimuli.
Acknowledgements The authors thank Kelly Lyons and two other reviewers for their helpful and constructive comments on an earlier draft of this manuscript.
Declaration of Interest The authors report no conflict of interest. The authors alone are responsible for the content and writing of this paper.
References 1. Aggleton JP, Waskett, L. The ability of odors to serve as statedependent cues for real world memories: Can Viking smells aid the recall of Viking experience? Br J Psychol 1999;90:1– 8.
2. Cann A, Ross DA. Olfactory stimuli as context cues in human memory. Am J Psycholo 1989;102:91–102. 3. Chu S, Downes J. Odor-evoked autobiographical memories: psychological investigations of Proustian phenomena. Chem Senses 2000;25:111–16. 4. Herz RS. The effects of cue-distinctiveness on odor-based context dependent memory. Mem Cogn 1997;25:375– 80. 5. Parker A, Ngu H, Cassaday HJ. Odor and Proustian memory: reduction of context-dependent forgetting and multiple forms of memory. Applied Cogn Psychol 2001;15:159–71. 6. Schab FR. Odors and the remembrance of things past. J Exp Psychol: Learning, Memory and Cognition 1990;16:648– 55. 7. Smith DG, Standing L, de Man A. Verbal memory elicited by ambient odor. Percept Mot Skills 1992;74:339–43. 8. Eich E. Fragrances as cues for remembering words. J Verbal Learn Verbal Behav 1977;17:103–11. 9. Ehrlichman H, Halpern JN. Affect and memory: effects of pleasant and unpleasant odors on retrieval of happy and unhappy memories. J Pers Soc Psychol 1988;55:769–79. 10. Habel U, Loch K, Pauly K, et al. The influence of olfactoryinduced negative emotion on verbal working memory: individual differences in neurobehavioural findings. Brain Res 2007;1152:158–70. 11. Moss M, Cook J, Wesnes K, Duckett P. Aromas of rosemary and lavender essential oils differentially affect cognition and mood in healthy adults. Int J Neurosci 2003;113:15–38. 12. Baddeley AD, Hitch GJ. Working memory. In: Bower G, ed. The Psychology of learning and motivation. New York: Academic Press; 1974:47–89. 13. Moss M, Hewitt S, Moss L, Wesnes K. Modulation of cognitive performance and mood by aromas of peppermint and ylangylang. Int J Neurosci 2008;118:59–77. 14. Schneider F, Koch K, Reske M, et al. Interaction of negative olfactory stimulation and working memory in schizophrenia patients: development and evaluation of a behavioural neuroimaging task. Psychiatry Res 2006;144:123–30. 15. Martin GN. The neuropsychology of smell and taste. Hove: The Psychology Press; 2013. 16. Lavric A, Rippon G, Gray JR. Threat-evoked anxiety disrupts spatial working memory performance: an attentional account. Cogn Ther Res 2003;10:489–504. 17. Li X, Chan RCK, Luo Y-J. Stage effects of negative emotion on spatial and verbal working memory. BMC Neurosci 2010;11:60. 18. Awh E, Jonides J. Overlapping mechanisms of attention and spatial working memory. Trends Cogn Sci 2001;5: 119–26. 19. Brand G, Millot J-L. Sex differences in human olfaction: between evidence and enigma. Q J Exp Psychol 2001;54B: 259–70. 20. Martin GN. The effect of exposure to odor on the perception of pain. Psychosom Med 2006;68:613–16. 21. Martin GN, Cooper JA. Adding zest to difficult journeys: odoreffects on simulated driving. BPS Annual Conference, University of York, York, 2007, 23 March. 22. Li X, Li X, Luo Y-J. Anxiety and attentional bias for threat: an event-related potential study. Neuroreport 2005;16:1501–5. 23. Gould A, Martin GN. A good odor to breathe? The effect of pleasant ambient odor on human visual vigilance. Appl Cogn Psychol 2001;15:225–32. 24. Martin GN. Human electroencephalographic (EEG) response to olfactory stimulation: two experiments using the aroma of food. Int J Psychophysiol 1998;30:287–302. 25. Ho C, Spence C. Olfactory facilitation of dual-task performance. Neurosci Lett 2005;389:35–40.
International Journal of Neuroscience
Odor and working memory
Int J Neurosci Downloaded from informahealthcare.com by The University of Manchester on 11/21/14 For personal use only.
26. Spence C, Kettenmann B, Kobal G, McGlone F. Selective attention to the chemosensory modality. Percept Psychophys 2000;62:1265–71. 27. Loring-Meier S, Halpern DF. Sex differences in visuospatial working memory: components of cognitive processing. Psychon Bull Rev 1999;6:464–71. 28. Speck O, Ernst T, Braun J, et al. Gender differences in the functional organization of the brain for working memory. Neuroreport 2000;11:2581–5.
C
2014 Informa Healthcare USA, Inc.
811
29. Larsson M, Lovden M, Nilsson L-G. Sex differences in recollective experience for olfactory and verbal information. Acta Psychol 2003;112:89–103. 30. Andersson L, Lundberg C, Astrom J, Nordin S. Chemosensory attention, habituation and detection in women and men. Int J Psychophysiol 2011;79:316–22. 31. Thuerauf N, Reulbach U, Lunkenheimer J, et al. Emotional reactivity to odors: olfactory sensitivity and the span of emotional evaluation separate the genders. Neurosci Lett 2009;456:74–79.
