The phenomenology of food cravings: The role of mental imagery

Appetite 45 (2005) 305–313 www.elsevier.com/locate/appet

The phenomenology of food cravings: The role of mental imagery M. Tiggemann*, E. Kemps School of Psychology, Flinders University, GPO Box 2100, Adelaide 5001, Australia Received 20 September 2004; accepted 28 June 2005

Abstract The study aimed to investigate the role of mental imagery in the experience of food cravings. A sample of 130 undergraduate students was first asked to recall and rate a previous food craving experience. Second, they were asked to imagine themselves eating their favourite food and to rate the involvement of different sensory modalities. It was found that mental imagery was a key element in both the retrospective craving experience and the current food induction. In particular, craving intensity was related to the vividness of the food image (rZ.46). The specific sensory modalities most involved were visual (39.7%) and gustatory (30.6%), followed by olfactory (15.8%). There was little involvement of auditory or tactile modalities. It was concluded that cognitive experimental techniques aimed at reducing the vividness of visual, gustatory or olfactory imagery, might usefully be deployed to reduce unwanted food cravings. q 2005 Elsevier Ltd. All rights reserved. Keywords: Food craving; Craving intensity; Mental imagery; Sensory modalities

Introduction Food cravings refer to a motivational state whereby an individual feels compelled to seek and ingest a particular food (Baker, Morse, & Sherman, 1986). It is the intensity of this state and the specificity of the craved food, which distinguish food cravings from ordinary food choices or general hunger (Pelchat, 2002). Although such food cravings are normative everyday experiences (Hill & Heaton-Brown, 1994; Lafay et al., 2001), most research has addressed their potentially maladaptive nature. For example, food cravings have been associated with binge eating (Gendall, Joyce, Sullivan, & Bulik, 1998; McManus & Waller, 1995), which in turn contributes to both obesity (Wurtman & Wurtman, 1995) and eating disorders (Gendall et al., 1998; Waters, Hill, & Waller, 2000). Food cravings are also associated with guilt (Macdiarmid & Heatherington, 1995) and depression (Gendall et al., 1998), and have been shown to impair cognitive performance (Green, 2001). Most research interest has focused on the origins of food cravings, and it is now generally accepted that a

* Corresponding author. E-mail address: [email protected] (M. Tiggemann).

0195-6663/$ - see front matter q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.appet.2005.06.004

range of biological, affective and cognitive triggers all play a role. Although food cravings have commonly been associated with hunger (Hill, Weaver, & Blundell, 1991) and attributed to nutritional and calorific deficits (Wardle, 1987), it is clear that nutritional deprivation is not a necessary condition for food cravings (Pelchat & Schaefer, 2000). Other triggers include menstrual-related changes (Dye, Warner, & Bancroft, 1995), negative mood states (Hill & Heaton-Brown, 1994; Hill et al., 1991; Lafay et al., 2001), as well as expectations and cognitions (Weingarten & Elston, 1990). Experimentally, food cravings have been induced by the thought, smell or pictures of food (Alsene, Li, Chaverneff, & de Wit, 2003; Fedoroff, Polivy, & Herman, 2003; Green, Rogers, & Elliman, 2000). Broadly based surveys have established that food cravings are relatively common, that most people indulge their cravings on the majority of occasions, and that chocolate is the most commonly and intensely craved food in Western societies (Lafay et al., 2001; Weingarten & Elston, 1991). Less is known about the phenomenology of food cravings. Thus, we sought to study aspects of the actual experience of food craving episodes. In particular, a number of lines of evidence point to a role for mental imagery. For example, Green et al. (2000) reported a positive correlation between latency on a simple reaction time task following food-related imagery induction and

306

M. Tiggemann, E. Kemps / Appetite 45 (2005) 305–313

self-reported desire to eat. They concluded that food cravings reduce the resources available for cognitive performance and that such cravings can be induced by imagery instructions. Similarly, Harvey, Kemps, and Tiggemann (2005) found that craving intensity increased following instructions to imagine a food script, but not a non-food script. They also noted a positive correlation between self-reported vividness of the imagined food script and craving intensity. These observations illustrate the role of mental imagery in food cravings and corroborate anecdotal reports of food-related images in craving experiences (Salkovskis & Reynolds, 1994). The current study aimed to document the occurrence of such mental imagery in naturalistic food cravings. A specific objective was to further investigate the nature of any imagery processes involved in food cravings. In recent years, cognitive psychologists have mounted an argument that phenomenological experience is amenable to the methods of experimental cognitive psychology (Baddeley & Andrade, 2000). In particular, a working memory approach has been adopted to conceptualize the phenomenological experience of mental imagery (Holmes & Hackmann, 2004; Pearson, 2001). The most widely adopted account of working memory is the model postulated by Baddeley (2000) comprising three components: a central executive which operates as a supervisory system, and two limited-capacity slave systems, the visuo-spatial sketch pad and the phonological loop, which maintain visuo-spatial and verbal material, respectively. Harvey et al. (2005) used this model to interpret their finding that food craving ratings were lower following a visual but not auditory secondary imagery task as showing that the imagery processes underlying food cravings are visual, rather than verbal or auditory in nature. The present study aimed to extend these initial investigations by directly investigating the specific nature of the mental imagery in food craving experiences. Imagery can occur in other sensory modalities besides sight and hearing (Sheehan, 1967). Indeed, it is quite likely that images involving taste and smell will be particularly potent in food craving. In their study of the subjective aspects of everyday desires, May, Andrade, Pannabokke, and Kavanagh (2004) found that cravings (food, tobacco, soft-drink, and alcohol were combined) were generally attributed to spontaneous thoughts rather than to any identifiable cue in the environment, and that they involved primarily olfactory and visual imagery. As yet, there is no information available on the sensory modality of craving-related images for specifically food. Thus, we wished to investigate three research questions: (a) whether mental imagery plays a role in the experience of food craving, specifically whether people who experience strong cravings have more vivid food images; (b) whether people who have vivid food images have high imaging ability in general; and (c) the nature of the

specific sensory modalities involved in craving-related food images.

Method Participants The participants were 130 undergraduate students (56 men, 74 women) at Flinders University who took part for course requirements and credit. They ranged in age from 17 to 44 years. Procedure Participants completed a survey entitled ‘Food Craving Survey’ in small groups. The survey consisted of three separate questionnaires administered in turn. This served to prevent participants from flicking forward to future questions and thus reduced the potential for response biases due to demand effects. The first questionnaire entitled ‘Food Craving Experience’ concerned an actual previous craving experience. The second questionnaire ‘Favourite Food’ attempted to induce food craving in the laboratory. Both were aimed at assessing the nature of state food cravings. The final questionnaire contained trait measures of habitual food craving, imaging ability, and dietary restraint. Materials Food craving experience questionnaire After some background information including selfreported height and weight, the questionnaire instructed participants to think back to the last time they had a food craving (‘i.e. an intense urge to eat a particular food’). Participants were to recall the experience as clearly as they could, ‘as if it were happening again right now’, and to really put themselves into the situation. They were asked to write a short paragraph describing this food craving experience in as much detail as they could, focusing in particular on ‘what the actual craving was like’. They were provided with 14 lines to do this. Participants were then asked to rate the intensity of their craving at that time on a 10-point scale (1Z‘very slight’, 10Z‘overwhelming’) and to specify the food craved. They were then presented with a list of 14 potential triggers for the craving and asked to rate each one on a 5-point scale Likert scale (1Z‘not at all’, 5Z‘definitely’). This scale was expanded and adapted from that of May et al. (2004) to be specific to food (as opposed to nicotine or alcohol). The trigger statements were selected to cover a number of classes of antecedents, including biological (‘hunger’), affective (‘stress’), external cues (‘smelt the food’), intrusive thoughts (‘suddenly thought about the food’) and mental imagery (‘imagined the smell/taste of the food’).

M. Tiggemann, E. Kemps / Appetite 45 (2005) 305–313

307

greater imaging ability. Internal consistency in the present sample was high for the total scale (aZ.94), as well as for each subscale: vision (aZ.84), audition (aZ.86), touch, (aZ.82), gustation (aZ.85), olfaction (aZ.82).

Favourite food questionnaire The next questionnaire instructed participants to think about their favourite food and imagine they were eating it. Again they were to bring the experience to mind as vividly as possible, with as much detail as possible, ‘as if it were happening right now’. They were then asked to rate the vividness of their image (1Z‘no image at all’, 10Z‘image perfectly clear—as vivid as the actual experience’) and their current urge to eat the food (1Z‘no desire or urge’, 10Z ‘extremely strong desire or urge’). They were then asked to describe their experience by rating 10 descriptors on 5-point scales (1Z‘not at all’, 5Z‘definitely’). These descriptors were those of May et al. (2004), adapted for specifically food, designed to cover a number of general domains. Thus far, the questionnaire was deliberately constructed so that participants would be unaware of our particular interest in the mental imagery component of food cravings. At this point, however, participants were explicitly asked to rate sensory descriptors (addressing the five sensory modalities, e.g. ‘I saw the food’) on 10-point scales (1Z ‘not at all’, 10Z‘definitely’). Finally, they were asked to indicate the percentage of each sensory modality in their image of eating their favourite food (visual, auditory, olfactory (smell), gustatory (taste), tactile (touch)) to a total of 100%, and to specify the food they had imagined.

Results

Trait measures

Characteristics of the sample

Food craving. Habitual food craving was assessed by the Food Cravings Questionnaire—Trait (Cepeda-Benito, Gleaves, Williams, & Erath, 2000). This measure was designed to assess stable features of cravings within individuals across times and situations. Respondents rate how frequently each of 39 items (e.g. ‘I feel like I have food on my mind all of the time’) is true for them on a 6-point scale (1Z‘never’, 6Z‘always’). Cepeda-Benito et al. report good test–retest reliability over 3 weeks (rZ.88), as well as good discriminant and construct validity, in that the trait craving measure was correlated with the disinhibition and hunger factors, but not the cognitive restraint factor, of the Three-Factor Eating Questionnaire (Stunkard & Messik, 1985). In the present sample, internal consistency was high (aZ.94).

The sample consisted of 56 men and 74 women ranging in age from 17 to 44 years (mean age Z22.7 years, SDZ 6.8). Their mean BMI of 22.4 (SDZ3.9) fell within the normal range, as recommended by Garrow and Webster (1985). A few (10.0%) reported that they were currently on a diet to lose weight, while more (40.8%) reported that they watched what they eat in order to not put on weight. The majority of the sample (75.4%) had eaten 2–4 h previously. Current level of hunger was moderate (MZ5.6, SDZ2.4), as was urge to eat (MZ5.8, SDZ2.7).

Imaging ability. General imaging ability was measured by the Questionnaire Upon Mental Imagery—Brief Version (Sheehan, 1967) which assesses imaging ability in the five sensory modalities. Respondents are asked to think of pictures, sounds, feeling or touch, tastes and smells (five items each), and then to rate these in terms of vividness of the resulting image (1Z‘perfectly clear and as vivid as the actual experience’, to 7Z‘no image present at all, you only ‘know’ that you are thinking of the object’). Sheehan (1967) reports a strong correlation (rZ.92) between the shortened version and the original 135-item version (Betts, 1909). Here, items were reverse-scored so that high scores represent

Restraint. Dietary restraint was measured by the Revised Restraint Scale of Herman and Polivy (1980). This scale assesses dietary restraint as a relatively enduring habitual style of eating, likely encompassing both restrictive behaviours and periods of excessive eating (Heatherton, Herman, Polivy, King, & McGree, 1988). The scale consists of 10 items (e.g. ‘How often are you dieting?’) which are rated on 4- or 5-point Likert scales. Total scores range from 0 to 35 with high scores reflecting a high degree of dietary restraint. Factor analyses have shown that the scale gives rise to two subscales: Concern for Dieting, which reflects the attention paid to eating and dieting, and Weight Fluctuation, reflecting the extent of previously experienced weight gain and loss (Blanchard & Frost, 1983). Internal consistency was acceptable in the present sample for the total score (aZ.83) and two subscales (Concern for Dieting, aZ.84; Weight Fluctuation, aZ.65).

Qualitative analysis of craving experience Only two (out of 130) participants were unable to recall a craving experience. Written paragraphs were decomposed into four parts: settings (e.g. studying late at night, after a light breakfast), actual experience of the craving (feelings, thoughts), consequences (e.g. drove to the shop, food consumption), and incidental comments. On average, the paragraphs were 8.4 lines in length, of which 2.4 lines were devoted to setting, 3.4 to the actual experience, 1.2 to consequences and 1.4 to incidental aspects. Because we instructed participants to specifically focus on the actual experience of food craving, we subjected only these lines to content analysis. The most common theme to emerge from descriptions of the actual experience of the craving was its overwhelming nature. Over half (nZ68, 61.5%) of the respondents

308

M. Tiggemann, E. Kemps / Appetite 45 (2005) 305–313

Table 1 Mean ratings of triggers for craving experience Item

I suddenly thought about the food I felt hungry I imagined the smell/taste of it I pictured myself having it I had nothing else to do/I was bored I saw it I heard it I smelt it Other things I was thinking about reminded me of it I felt stressed/anxious/sad I always have it at that time/place I felt happy I was really busy I imagined the sound of myself having it

Mean

Correlation with craving intensity

Gender

Food category

Men

Women

Sweet

Savoury/ take-away

Meals

3.8 3.4 4.1 3.5 2.5 2.5 1.8 2.5 2.4

.22* .18* .32*** .27** .11 K.09 K.10 K.00 K.00

3.8 3.6 3.8 3.3 2.1 2.6 1.8 2.5 2.3

3.8 3.3 4.3 3.6 2.7* 2.5 1.7 2.5 2.5

4.0 2.8 4.1 3.5 2.8 2.4 1.5 2.2 2.5

3.7 3.8 4.3 3.4 2.4 2.6 1.7 2.4 2.4

3.5 3.9*** 3.8 3.4 1.9 2.6 2.2* 2.9 2.3

2.1 2.0 2.4 1.9 1.5

.10 K.04 .06 K.00 K.00

1.7 2.1 2.4 1.8 1.5

2.5* 1.9 2.4 2.0 1.6

2.5 2.1 2.5 1.9 1.5

1.8 1.9 2.3 2.0 1.5

1.9 2.0 2.4 1.7 1.6

*p!.05; **p!.01; ***p!.001.

described how they ‘couldn’t stop thinking about’ the food, how ‘it really played on (their) mind’, and how they ‘could not concentrate on anything else’. They characterized the craving as ‘an obsession’, ‘overpowering’ or ‘really intense’ (e.g. ‘I felt like I would die if I did not have the chocolate’). Most often (nZ42, 32.3%) this was accompanied by physical reactions like mouth watering and stomach grumbling, and occasionally (nZ6, 4.6%) by ambivalence or feelings of guilt. Although the physical reactions above implicate the imagining of food and/or eating, of particular relevance here, a number of responses (nZ39, 30.0%) also explicitly articulated imagery processes. Some of these responses were unspecified in modality (e.g. ‘imagining eating it’), but a number specified either taste (nZ24, 61.5%; e.g. ‘I can almost taste it in my mouth’; ‘I could almost taste and smell [the chocolate]’) or the visual modality (nZ11, 28.2%; e.g. ‘I could picture it in my mind, picture eating it’; ‘I began to visualize [spaghetti Bolognese]’). One particularly eloquent respondent described her craving for Laksa soup as follows: ‘I could taste the food on the top of my tongue and I was beginning to salivate. The food was very vivid in my mind and the colours of the food were bright and rich. I could even smell it and imagined how I would eat it and what I could eat first.. I end up feeling dizzy’. Quantitative analysis of craving experience By far the most commonly craved individual food was chocolate (24.2%), with chocolate foods, e.g. cake, cookies, constituting another 4.0%. Other sweet treats (e.g. ice cream) were craved by 12.1% of the sample, with savoury treats (e.g. nuts, chips) by 15.3%. Fully 22.6% craved takeaway foods, e.g. pizza, McDonalds, and 21.0% craved a meal (e.g. roast lamb and vegetables). Many more of

the women (42.0%) than men (11.1%) craved chocolate (or chocolate-containing foods), and more men than women craved a meal (33.3 vs 11.6%), c2Z18.11, p!.01. Table 1 provides the mean ratings for the 14 triggers, by gender and by food category. It can be seen that there were four highly endorsed items (rating O3): ‘I imagined the smell/taste of it’, ‘I suddenly thought about it’, ‘I pictured myself having it’ and ‘I felt hungry’. The first three relate to imagined and spontaneous triggers, while the fourth reflects a physiologically based trigger. A series of 3 (food category)!2 (gender) ANOVAs of the triggers revealed a significant main effect of food category for ‘I felt hungry’, F(2,117)Z10.54, p!.001, and ‘I heard it’, F(2,117)Z3.15, p!.05. Perhaps not surprisingly, the means indicate that cravings for meals were more triggered by hunger and hearing it (meal preparation) than were cravings for chocolate and other sweet treats. There was a significant gender effect for ‘I had nothing else to do/I was bored’, F(1,117)Z4.67, p!.05, and ‘I felt stressed/anxious/sad’, F(1,116)Z4.43, p!.05, whereby women rated these negative mood states more highly than did men. There were no significant food category!gender interactions. Imagery involvement in craving experience The mean intensity of the food craving experience was high (MZ8.2, SDZ1.6). This did not vary with gender, F!1, pO.05, nor type of food craved, F!1, pO.05. To assess the possible involvement of imagery, the triggers were correlated with intensity of the craving. As can be seen in Table 1, the first four triggers (suddenly thinking about the food, feeling hungry, imagining the smell or taste, and picturing oneself eating the food) were all positively related to rated intensity of the craving (although the first two had small effect sizes). Intensity was not significantly

M. Tiggemann, E. Kemps / Appetite 45 (2005) 305–313

309

Table 2 Mean ratings and correlations for descriptors of favourite food induction Item

Mean

Corr. with vividness

Corr. with craving

I want the food because I am hungry Eating it would feel very comforting right now Thinking how much better I would feel after I have eaten it I am imagining the taste of it I would feel more relaxed if I ate it I am visualizing it If I don’t think about it, my craving will go away I am trying to resist eating it I have it with me right now I can hear myself eating it

3.5 3.4 2.9 4.1 2.8 3.8 3.0 2.4 1.4 1.5

.12 .15 .13 .35*** .17 .43*** .11 .15 .10 .17

.34*** .37*** .37*** .43*** .37*** .39*** K.23** .11 .01 .14

*p!.05; **p!.01; ***p!.001.

related to the habitual craving trait measure, rZ.13, pO.05, nor overall imaging ability, rZ.07, pO.05, but was related to current level of hunger, rZ.24, p!.01. Favourite food induction The most common favourite foods imagined were meal foods (42.3%), followed by take-away foods (20.7%) and chocolate products (16.9%). Many more of the women (35.6%) than men (14.3%) nominated chocolate or other sweets as their favourite food, c2Z7.99, p!.05. Mean ratings for the 10 descriptors are provided in Table 2. The most highly ranked descriptor was ‘I am imagining the taste of it’, followed by ‘I am visualizing it’. A series of 3 (food category)!2 (gender) ANOVAs of the descriptors revealed a significant main effect of food category for ‘I am thinking of how much better I would feel after I have eaten it’, which was rated higher for meals (MZ3.4) than for the other food categories (M’sZ2.3, 2.5), F(2,123)Z7.53, p!.001. The only significant gender effect was for ‘I am trying to resist eating it’, which was rated more highly by women (MZ2.7) than men (MZ2.1), F(1,122)Z5.74, p! .05. There were no significant interactions. Imagery modalities in favourite food induction Table 3 presents the mean ratings for the sensory modalities. It can be seen that all modalities other than auditory (‘heard myself eating the food’) are moderately endorsed, led by seeing the food and then tasting the food. This was confirmed by the explicit percentages, where visual (39.7%) was the highest, followed by gustatory (30.6%), and then olfactory (15.8%). The tactile (9.5%) and auditory (4.4%) modalities were little used. These did not differ between men and women. Correlates of craving intensity and vividness in favourite food induction The mean intensity of craving (desire to eat) was 7.2 (SDZ2.3). Mean vividness of image was 7.9 (SDZ1.8). Neither the vividness of the image, nor the intensity of

the craving, was affected by gender or type of food imagined (Fs!1.5, pO.05). Importantly, craving intensity was correlated with the rated vividness of the image, rZ.46, p!.001. Tables 2 and 3 display the correlations between vividness of image and craving intensity with the general and sensory modality descriptors, respectively. Table 2 shows that the vividness of the image was associated with imagining the taste and visualizing the food. Craving intensity was related to many (7/10) of the descriptors. Table 3 shows that both vividness of the image and craving intensity were related to seeing the food, seeing oneself eating the food, and tasting the food. Additionally, vividness was also associated with smelling the food and feeling the texture in one’s mouth. Craving intensity of favourite food was related to habitual trait craving, rZ.22, p!.05, as well as to current level of hunger, rZ.51, p!.001. Vividness of the food image was significantly related to visual mental imagery ability, rZ.24, p!.01, and approached significance for overall imaging ability, rZ.16, pZ.06. Predictors of craving intensity in favourite food induction Hierarchical multiple regression analysis was used to examine potential predictors of craving intensity for the favourite food. The background variables of age, gender and Table 3 Mean ratings and correlations for sensory modality descriptors Item

Mean

Corr. with vividness

Corr. with craving

I saw the food I saw myself eating the food I heard myself eating the food I smelt the food I tasted the food I felt the texture of the food in my mouth

7.4 6.4

.33*** .37***

.21* .36***

2.8

.16

.15

5.2 7.2 6.5

.25** .42*** .35***

.13 .25** .17

*p!.05; **p!.01; ***p!.001.

310

M. Tiggemann, E. Kemps / Appetite 45 (2005) 305–313

Table 4 Standardized regression coefficients (betas) from hierarchical regression analyses to predict craving intensity for favourite food

equation, with a smaller contribution from seeing oneself eating the food, bZ.21, p!.05.

Variable Step 1 Age Gender BMI Step 2 Trait food craving Step 3 Current hunger Step 4 Visual imagery ability Olfactory imagery ability Gustatory imagery ability Step 5 ‘I saw the food’ ‘I saw myself eating the food’ ‘I smelt the food’ ‘I tasted the food’ Step 6 Image vividness

Craving intensity

Craving intensity

The role of dietary restraint

.07 K.13 K.06

.03 K.09 .02

Neither current dieting status nor dietary restraint (nor either of its subscales) was related to any of the state measures of intensity of food craving experience, vividness or craving intensity of favourite food (rs!.10, pO.05). Dietary restraint was, however, highly related to the trait measure of habitual craving, rZ.57, p!.001. This correlation was higher for the Concern for Dieting subscale, rZ.64, p!.001, than for Weight Fluctuation, rZ.34, p!.001. Restraint was not related to imagery ability in general, but was related to the sensory descriptors of ‘I saw myself eating the food’, rZ.19, p!.05, ‘I heard myself eating the food’, rZ.25, p!.01, and ‘I felt the texture of the food in my mouth’, rZ.21, p!.05.

.20*

.15

.45***

.43***

.01

K.08

.10

.06

K.02

.05

K.13 .28***

K.14 .21*

K.08 .25*

K.07 .12

–

.34***

*p!.05; **p!.01; ***p!.001.

BMI were entered on Step 1, trait craving on Step 2, current hunger level on Step 3, general imaging ability in the visual, olfactory and gustatory domains on Step 4, and the modality descriptors for sight, smell and taste of the imagined favourite food on Step 5. Overall these predictors accounted for 44% of the variance in craving intensity, RZ.66, F(12, 110)Z7.01, p!.001. As a set, the background variables (Step 1) did not offer significant prediction, F!1. However, trait craving offered additional unique prediction, R2change Z :08, Fchange(1,118)Z10.27, p!.01, as did current level of hunger, R2change Z :23, Fchange(1,117)Z38.74, p! .001. General imaging ability did not add significant prediction (pO.05). Finally, the modality descriptors offered significant additional prediction, R2change Z :10, Fchange(4,110)Z4.65, p!.01. The regression coefficients (beta values) for the final regression equation are presented in Table 4. It can be seen that trait food craving, bZ.20, p! .05, current hunger, bZ.45, p!.001, and the descriptors involving tasting, bZ.25, p!.05, and seeing oneself eating the food, bZ.28, p!.001, emerged as significant unique predictors. When vividness of the food imagery was added to the regression equation (Step 6), prediction significantly increased, R2change Z :07, Fchange(1,109)Z15.78, p!.001, to make a total of 51% of the variance accounted for, RZ .71, F(13,109)Z8.64, p!.001. Vividness of image, bZ.34, p!.001, and current hunger, bZ.43, p!.001, emerged as the most important predictors in the overall regression

Discussion The study has confirmed a number of previous findings. It appears that for Australian students, just like those in other Western countries, chocolate is the most commonly craved food (Hill & Heaton-Brown, 1994). This is especially the case for women, as also found by Weingarten and Elston (1991). Rogers and Smit (2000) have argued that this arises from hedonic effects paired with socially and culturally determined perceptions of appropriate intakes and uses. Thus, they argue that chocolate is the most highly craved food because it is highly palatable but should be eaten with restraint. Such a cultural basis for cravings is supported by the results of cross-cultural comparisons. The most commonly craved foods among Egyptian men and women, for example, are vegetable dishes (some stuffed or cooked with meat) (Parker, Kamel, & Zellner, 2003). The major aim of the present study, however, was to investigate the actual phenomenological experience of the craving episode in more detail than previously, with a particular focus on the involvement of imagery. This aim was realized in a number of different ways: by content analysis of past craving episodes, by the induction of craving via a food-related script, and by specific ratings on the involvement of particular sensory modalities. The general finding is clear. The role of mental imagery has been established as a key component in both naturalistic and induced craving experiences. Our findings demonstrate that mental imagery plays a number of different roles. First, the most highly rated triggers for the retrospective food craving experience were primarily thoughts or images as opposed to actual instances of the craved object, replicating May et al.’s (2004) finding for non-specific cravings. Second, the content analyses of the craving experience yielded a number of imagery

M. Tiggemann, E. Kemps / Appetite 45 (2005) 305–313

descriptors. Third, both the descriptors and explicit ratings of modality confirmed the involvement of mental imagery sensory domains. Finally, an important finding was that vividness of the image, over and above hunger level, was a significant unique predictor of craving intensity in the food induction scenario. Together these findings illustrate an important role for mental imagery in the phenomenology of food craving. Thus, the answer to our first research question is overwhelmingly affirmative. Our second research question as to whether people who have more vivid food images have high imaging ability in general was also answered in the affirmative. Vividness of the induced food image was correlated with imaging ability in the visual domain. This is a novel finding which warrants replication, but which suggests that food-related imagery is essentially no different from other types of imagery. Our third research question asked whether particular sensory modalities are involved in the mental imagery associated with food craving. Here, we have shown that visualizing and imagining the taste of the food are highly correlated with vividness of the imagery and craving intensity, both as descriptors and explicit modality statements. In specific percentages, the visual modality was assigned the highest percentage, followed by gustatory (taste), and then olfaction (smell). Finally, the multiple regressions identified ‘saw myself eating the food’ and ‘tasted the food’ as independent predictors of craving intensity. Taken together, the results identify the visual modality as the most important, followed by taste. Olfaction seems to play a (smaller) role, but the auditory and tactile modalities play almost no role. Thus, the research offers some insight into the food craving experience and thereby makes a contribution to our theoretical understanding of the nature of food cravings. Both physiological (level of hunger) and psychological (imagery) components were identified. In particular, the study details the imaginal component of food cravings. The practical significance is that such knowledge can, in turn, provide a basis for developing potential intervention tools to reduce the occurrence, intensity and duration of unwanted food craving episodes. This is important for two reasons. First, recurrent food craving episodes have been identified as a risk factor for the development of health problems, in particular obesity (Wurtman & Wurtman, 1995) and bulimia nervosa (Waters et al., 2000), and are associated with guilt and depression in non-clinical samples (Gendall et al., 1998; Macdiarmid & Heatherington, 1995). Second, the techniques currently available for curbing food cravings involve either the suppression of food-related thoughts (Johnston, Bulik, & Anstiss, 1999) or unreinforced exposure to food cues (Hetherington, 2001), and are very effortful and have not proved very successful. If, as we have shown, visual mental imagery is a key component of food craving, then concurrent visual activity may provide a useful technique for reducing food cravings. Analogously, Andrade, Kavanagh, and Baddeley (1997)

311

have been able to reduce the vivid and intrusive images symptomatic of post-traumatic stress disorder. They showed that a concurrent visuo-spatial task (eye movements) while imagining a traumatic event reduced the vividness and the emotional impact of the distressing image. According to working memory theory (Baddeley, 2000), concurrent visuo-spatial activity will have its craving reducing effect by diminishing the vividness of food-related images held in the limited-capacity visuo-spatial sketchpad. Thus, the increasing knowledge about memory and mental imagery in psychopathologies such as post-traumatic stress disorder, agoraphobia, body dysmorphic disorder and mood disorders (Hackmann & Holmes, 2004), can be usefully applied in the context of problematic food cravings. This theoretical framework is just beginning to give rise to attempts to reduce food cravings by using concurrent visuo-spatial tasks (Kemps, Tiggemann, Woods, & Soekov, 2004). Future research is needed to develop and evaluate such techniques for reducing unwanted food cravings. It might also usefully investigate whether similar techniques designed to interfere with the gustatory or olfactory imagery aspects of food craving might confer additional benefit. The role of dietary restraint is interesting. In the present sample, dietary restraint was related to habitual trait craving, but not to ‘state’ food craving. This may be due to the fact that dietary restraint (as conceptualized by the Revised Restraint scale) itself refers to a habitual and enduring mode of thinking about food (Heatherton et al., 1988). Here, we did not assess how frequent cravings were, but rather the intensity of two specific craving episodes. We conclude that the relationship between dietary restraint and craving will be a function of precisely how craving is operationalized. While a number of studies have found an association between dietary restraint and food craving (Fedoroff, Polivy, & Herman, 1997; Overduin & Jansen, 1996; Warren & Cooper, 1988), others have not (Harvey, Wing, & Mullen, 1993; Hill et al., 1991; Rodin, Mancuso, Granger, & Nelbach, 1991). Although there are many methodological differences between these studies, including the use of different measures to actually assess dietary restraint, on the basis of the present results we would tentatively suggest that dietary restraint will be related to global or aggregated measures of food craving, rather than to single episodes. Like all studies, the present study carries a number of limitations. First, the sample consisted of university students in a laboratory setting. Thus, the findings need to be replicated for naturally occurring cravings among more broadly based samples. Second, we did not experimentally control for levels of current hunger, which are liable to impact on how vividly one recalls a previous eating experience. Here, however, we were more interested in the phenomenological experience of craving, whether it be hunger-driven or by psychological factors. Nevertheless, future studies might usefully experimentally control or manipulate hunger by explicit instruction about eating prior

312

M. Tiggemann, E. Kemps / Appetite 45 (2005) 305–313

to testing. Third, some of the individual effect sizes were small, although the overall pattern of results was clear. Finally, all the measures in the present study were selfreport. This leaves open the possibility of the various reporting biases inherent in introspective data. On the other hand, Pelchat (2002) has recently suggested that self-ratings as we have used here are precisely the most appropriate way to measure the subjective experience that we are interested in. She argues that craving, as a psychological experience, is best measured by self-report. Objective measures such as consumption do not capture the essential subjective quality of what we mean by ‘craving’. Furthermore, our research has used not only a retrospective account, but also induced a current account to assess subjective feelings at the moment. Nevertheless, future research might usefully develop behavioural measures to investigate whether such ‘cravings’ do translate into unwanted eating behaviour. In sum, the study has made a start on a more detailed examination of the craving experience. In particular, it has clearly demonstrated that mental imagery is a central feature of food craving, and that this imagery is primarily visual in modality. Consequently, visual imagery techniques may prove useful therapeutic aids for reducing unwanted food craving in clinical and non-clinical populations.

Acknowledgements This research was supported by a Staff Research Project Grant from the Flinders University Research Budget. We are grateful to Kelly Marshall and Aja Franks for assistance with data collection.

References Alsene, K. M., Li, Y., Chaverneff, F., & de Wit, H. (2003). Role of abstinence and visual cues on food and smoking craving. Behavioural Pharmocology, 14, 145–151. Andrade, J., Kavanagh, D., & Baddeley, A. (1997). Eye-movements and visual imagery: A working memory approach to the treatment of posttraumatic stress disorder. British Journal of Clinical Psychology, 36, 209–223. Baddeley, A. D. (2000). The episodic buffer: A new component of working memory? Trends in Cognitive Sciences, 4, 417–423. Baddeley, A. D., & Andrade, J. (2000). Working memory and the vividness of imagery. Journal of Experimental Psychology: General, 129, 126–145. Baker, T. B., Morse, E., & Sherman, J. E. (1986). The motivation to use drugs: A psychobiological analysis of urges. In P. C. Rivers (Ed.), The Nebraska symposium on motivation: Alcohol use and abuse (pp. 257– 323). Lincoln: University of Nebraska Press. Betts, G. H. (1909). The distribution and functions of mental imagery. New York: Columbia University. Blanchard, F., & Frost, R. O. (1983). Two factors of restraint: Concern for dieting and weight fluctuations. Behavior Research and Therapy, 21, 259–267.

Cepeda-Benito, A., Gleaves, D. H., Williams, T. L., & Erath, S. A. (2000). The development and validation of the state and trait food-cravings questionnaires. Behavior Therapy, 31, 151–173. Dye, L., Warner, P., & Bancroft, J. (1995). Food cravings during the menstrual cycle and its relationship to stress, happiness of relationship and depression: A preliminary enquiry. Journal of Affective Disorders, 34, 157–164. Fedoroff, I. C., Polivy, J., & Herman, C. P. (1997). The effect of preexposure to food cues on the eating behavior of restrained and unrestrained eaters. Appetite, 28, 33–47. Fedoroff, I. C., Polivy, J., & Herman, C. P. (2003). The specificity of restrained versus unrestrained eaters’ responses to food cues: General desire to eat, or craving for the cued food? Appetite, 41, 7–13. Garrow, J. S., & Webster, B. S. (1985). Quetelet’s Index (W/H2) as a measure of fatness. International Journal of Obesity, 9, 147–153. Gendall, K. A., Joyce, P. R., Sullivan, P. F., & Bulik, C. M. (1998). Food cravers: Characteristics of those who binge. International Journal of Eating Disorders, 23, 353–360. Green, M. (2001). Dietary restraint and craving. In M. M. Hetherington (Ed.), Food cravings and addiction (pp. 521–548). London: Leatherhead. Green, M. W., Rogers, P. J., & Elliman, N. A. (2000). Dietary restraint and addictive behaviors: The generalizability of Tiffany’s cue reactivity model. International Journal of Eating Disorders, 27, 419–427. Hackmann, A., & Holmes, E. A. (2004). Reflecting on imagery: A clinical perspective and overview of the special issue of Memory on mental imagery and memory in psychopathology. Memory, 12, 389–402. Harvey, J., Wing, R. R., & Mullen, M. (1993). Effects on food cravings of a very low calorie diet or a balanced, low calorie diet. Appetite, 21, 105–115. Harvey, K., Kemps, E., & Tiggemann, M. (2005). The nature of imagery processes underlying food cravings. British Journal of Health Psychology, 10, 49–56. Heatherton, T. F., Herman, C. P., Polivy, J., King, G. A., & McGree, S. T. (1988). The (mis) measurement of restraint: An analysis of conceptual and psychometric issues. Journal of Abnormal Psychology, 97, 19–28. Herman, C. P., & Polivy, J. (1980). Restrained eating. In A. Stunkard (Ed.), Obesity (pp. 208–225). Philadelphia: W.B. Saunders. Hetherington, M. M. (2001). Food cravings and addiction. Surrey, UK: Leatherland. Hill, A. J., & Heaton-Brown, L. (1994). The experience of food craving: A prospective investigation in healthy women. Journal of Psychosomatic Research, 38, 801–814. Hill, A. J., Weaver, C. F., & Blundell, J. E. (1991). Food craving, dietary restraint and mood. Appetite, 17, 187–197. Holmes, E. A., & Hackmann, A. (2004). A healthy imagination? Editorial for the special issue of Memory: Mental imagery and memory in psychopathology. Memory, 12, 387–388. Johnston, L., Bulik, C. M., & Anstiss, V. (1999). Suppressing thoughts about chocolate. International Journal of Eating Disorders, 26, 21–27. Kemps, E., Tiggemann, M., Woods, D., & Soekov, B. (2004). Reduction of food cravings through concurrent visuospatial processing. International Journal of Eating Disorders, 36, 31–40. Lafay, L., Thomas, F., Mennen, L., Charles, M. A., Eschwege, E., Borys, J.M., et al. (2001). Gender differences in the relation between food crabings and mood in an adult community: Results from the Fleurbaix Laventie Ville Sante Study. International Journal of Eating Disorders, 29, 195–204. Macdiarmid, J. I., & Heatherington, M. M. (1995). Mood modulation by food: An exploration of affect and cravings in ‘chocolate addicts’. British Journal of Clinical Psychology, 34, 129–138. May, J., Andrade, J., Pannabokke, N., & Kavanagh, D. (2004). Images of desire: Cognitive models of craving. Memory, 12, 447–461. McManus, F., & Waller, G. (1995). A functional analysis of binge-eating. Clinical Psychology Review, 15, 845–863. Overduin, J., & Jansen, A. (1996). Food cue reactivity in fasting and nonfasting subjects. European Eating Disorders Review, 4, 249–259.

M. Tiggemann, E. Kemps / Appetite 45 (2005) 305–313 Parker, S., Kamel, N., & Zellner, D. (2003). Food craving patterns in Egypt: Comparisons with North America and Spain. Appetite, 40, 193–195. Pearson, D. G. (2001). Imagery and the visuo-spatial sketchpad. In J. Andrade (Ed.), Working memory in perspective (pp. 33–59). Hove, UK: Taylor and Francis. Pelchat, M. L. (2002). Of human bondage: Food craving, obsession, compulsion, and addiction. Physiology and Behavior, 76, 347–352. Pelchat, M. L., & Schaefer, S. (2000). Dietary monotony and food cravings in young and elderly adults. Physiology and Behavior, 68, 353–359. Rodin, J., Mancuso, J., Granger, J., & Nelbach, E. (1991). Food cravings in relation to body mass index, restraint and estradiol levels: A repeated measures study in health women. Appetite, 17, 177–185. Rogers, P. J., & Smit, H. J. (2000). Food craving and food ‘addiction’: A critical review of the evidence from a biopsychosocial perspective. Pharmacology Biochemistry and Behavior, 66, 3–14. Salkovskis, P. M., & Reynolds, M. (1994). Thought suppression and smoking cessation. Behaviour Research and Therapy, 32, 193–201.

313

Sheehan, P. W. (1967). A shortened form of betts’ questionnaire upon mental imagery. Journal of Clinical Psychology, 23, 386–389. Stunkard, A. J., & Messik, S. (1985). The three-factor eating questionnaire to measure dietary restraint, disinhibition and hunger. Journal of Psychosomatic Research, 29, 71–83. Wardle, J. (1987). Compulsive eating and dietary restraint. British Journal of Clinical Psychology, 26, 47–55. Warren, C., & Cooper, P. J. (1988). Psychological effects of dieting. British Journal of Clinical Psychology, 27, 269–270. Waters, A., Hill, A., & Waller, G. (2000). Bulimics’ responses to food cravings: Is binge-eating a product of hunger or emotional state? Behaviour Research and Therapy, 39, 877–886. Weingarten, H. P., & Elston, D. (1990). The phenomenology of food cravings. Appetite, 15, 231–246. Weingarten, H. P., & Elston, D. (1991). Food cravings in a college population. Appetite, 17, 167–175. Wurtman, R. J., & Wurtman, J. J. (1995). Brain serotonin, carbohydratecraving, obesity and depression. Obesity Research, 3, 477–480.