Approach and avoidance motivation in eating disorders

Psychiatry Research 188 (2011) 396–401 Contents lists available at ScienceDirect Psychiatry Research j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / p s yc h r e s Approach and avoidance motivation in eating disorders Amy Harrison a,⁎, Janet Treasure a, Luke D. Smillie b a b Kings College London, Institute of Psychiatry, Section of Eating Disorders, London, UK Department of Psychology, Goldsmiths, University of London, UK a r t i c l e i n f o Article history: Received 5 May 2010 Received in revised form 12 April 2011 Accepted 20 April 2011 Keywords: Eating disorders Anorexia nervosa Bulimia nervosa Recovery Reinforcement Sensitivity Theory BIS BAS Reward reactivity a b s t r a c t It has been proposed that approach and avoidance processes may be critically involved in the development and maintenance of eating disorders (EDs), including anorexia nervosa (AN) and bulimia nervosa (BN). The Behavioural Inhibition System and Behavioural Activation System Scales (BIS/BAS) and Appetitive Motivation Scale (AMS) questionnaires were administered to 286 participants: 91 healthy controls (HCs), 121 participants with a current ED, either AN (restrictive and binge purge subtypes), or BN and 74 participants recovered from an ED. Individuals with EDs had higher levels of sensitivity to punishment and lower levels of reward reactivity than controls. Individuals in recovery from an ED scored the same as those in the acute group, with the exception of BAS fun seeking, for which they scored significantly higher than those with restricting AN. Discriminant analysis revealed that HCs were maximally separated from those in the acute and recovered ED groups along a dimension reflecting high punishment sensitivity and low reward sensitivity. Classification analysis demonstrated that ED and HC group membership was predicted from reward and punishment sensitivity measures; however recovered participants tended to be misclassified as ED. This study suggests high punishment sensitivity and low reward reactivity/sensitivity might form a personality cluster associated with the risk of developing an ED. © 2011 Published by Elsevier Ireland Ltd. 1. Introduction Processes concerning approach and avoidance motivation are a prominent feature in the landscape of personality and psychopathology research, and it has been suggested that two brain-behaviour systems regulate approach of appetitive stimuli and avoidance of aversive stimuli (e.g. Cloninger, 1987; Gray, 1991; Fowles, 1993; Zuckerman, 2005; Depue, 2006; Elliot, 2008; Carver et al., 2009; Smillie et al., 2010). The approach system, the Behavioural Activation System (BAS), is theorised to underlie personality dispositions reflecting reward-sensitivity, including Extraversion or Impulsivity (see Pickering and Smillie, 2008). The avoidance system, the Behavioural Inhibition System (BIS), is thought to relate to personality dispositions reflecting punishment-sensitivity, including anxiety and Neuroticism (see Corr et al., 1997). In this article, we refer to punishment sensitivity more broadly, as the theoretical predictions and psychometric measures employed do not distinguish between these processes (fear/anxiety). In recent years, it has been proposed (Loxton and Dawe, 2001; Dawe and Loxton, 2004; Loxton and Dawe, 2006, 2007) that approach and avoidance processes may be critically involved in the development and maintenance of eating disorders (EDs). There are a range of reasons why ⁎ Corresponding author at: Kings College London, Institute of Psychiatry, Department of Psychological Medicine and Psychiatry, Eating Disorders Research Unit, 5th Floor, Bermondsey Wing, Guy's Hospital. St Thomas Street, London, SE1 9RT, UK. Tel.: + 44 207 188 0190; fax: + 44 207 188 0167. E-mail address: amy.harrison@kcl.ac.uk (A. Harrison). 0165-1781/$ – see front matter © 2011 Published by Elsevier Ireland Ltd. doi:10.1016/j.psychres.2011.04.022 individuals may engage in disordered eating behaviours such as bingeing and purging. For example, using confirmatory factor analysis, Wedig and Nock (2010) recently reported that people binge and purge to regulate their emotional state, specifically to decrease negative emotions and increase positive emotions, or for social reasons, namely to communicate needs to others, or to escape social interaction. Claes et al. (2010), using discriminant analysis found that individual differences in BIS motivation and top-down control independently predicted bingeing/purging behaviours. However, the proposed relationship between reinforcement sensitivity theory and EDs has been necessarily speculative (Dawe and Loxton, 2004, p. 7). Nevertheless, it has been suggested that the elevated anxiety frequently found in ED populations (e.g., Vitousek and Manke, 1994; Grau and Ortet, 1999) might have a basis in dispositional punishment sensitivity. For example, Ampollini et al. (1999) report that high levels of harm avoidance (a BIS-related construct) are associated with anxiety and depression, which are core features of EDs (Godart et al., 2003; Blinder et al., 2006; Pallister and Waller, 2008). Several authors have confirmed associations between punishment sensitivity and disordered eating (Loxton and Dawe, 2001; Kane et al., 2004; Nederkoorn et al., 2004; Claes et al., 2006). In a recent systematic review, Harrison et al. (2010) found that people with an ED had elevated scores on trait measures of punishment sensitivity (anxiety and harm avoidance measured using the Tridimensional Personality Questionnaire (TPQ) (Cloninger, 1987) and the Temperament and Character Inventory (TCI) (Cloninger, 1993) relative to healthy controls. Genetic data support the notion that such relationships reflect dispositional tendencies rather than transient states or symptoms. For instance, Wilksch and Wade A. Harrison et al. / Psychiatry Research 188 (2011) 396–401 (2009) report higher punishment sensitivity, as measured using the Sensitivity to Punishment and Sensitivity to Reward Questionnaire (Torrubia et al., 2001), in non-affected twin siblings of ED women, even after controlling for the temperament scores of the sister. This finding positively correlated with the importance of shape and weight (r=0.44). Dawe and Loxton (2004) suggest that increased reward sensitivity is associated specifically with vulnerability towards developing bingeeating behaviour. Their proposal has been supported using self-report and behavioural measures of reward sensitivity (Loxton and Dawe, 2001; Kane et al., 2004), although a later study failed to confirm these findings (Loxton and Dawe, 2007). These studies assessed reward sensitivity using a reaction-time paradigm called the Card Arranging Reward Reactivity Objective Test (CARROT; Powell et al., 1996). In order to provide an alternative assessment (i.e., not based on reaction time), Farmer et al. (2001) used the Taffle Task (Taffel, 1955), in which participants are rewarded financially each time they use certain pronouns during a sentence construction task. A higher frequency of binge eating episodes was associated with increased use of rewarded pronouns. In a student sample, Franken and Muris (2005) found that sensitivity to reward, measured using the Sensitivity to Punishment and Sensitivity to Reward Questionnaire (Torrubia et al., 2001) was significantly related to food craving and BMI, highlighting reward sensitivity as a potential vulnerability factor for developing an ED. Furthermore, Davis and Fox (2008) found that BMI was a statistically significant predictor of reward sensitivity, measured using the Sensitivity to Punishment and Sensitivity to Reward Questionnaire (SPSRQ) (Torrubia et al., 2001) for normal and overweight men. In a systematic review, Harrison et al. (2010) found that individuals with bulimia nervosa (BN) or anorexia nervosa (AN) binge/purge subtypes showed elevated scores on reward-sensitivity questionnaires (Tridimensional Personality Questionnaire (TPQ), Cloninger, 1987, and the Temperament and Character Inventory (TCI), Cloninger, 1993; Behavioural Inhibition System and Behavioural Activation System Scales (BIS/BAS scales, Carver and White, 1994) in comparison to controls. Conversely, individuals with AN restricting type were less sensitive to reward than matched controls. Since this review was published, Claes et al. (2010) have reported BIS BAS data for an additional 40 patients with EDs (AN (23 with restricting AN, 6 with binge purge AN and 11 with BN) and found no difference between those with a restricting ED in comparison to those with a binge purge ED. Wagner et al. (2007), using a monetary reward task, found that activity in the subgenual anterior cingulated cortex and its ventral striatal target were similar during positive and negative feedback, indicating that individuals with a history of AN may have difficulty discriminating between positive and negative feedback. Wagner et al. (2008) have also found that people with AN, in comparison to controls, have a reduced response to food stimuli in insular–striatal circuits, suggesting a difference in the way individuals with AN process taste stimuli. Kaye et al. (2009) have argued that in the absence of appropriate reward processing through ventral-striatal/DA paths, individuals who have recovered from AN might focus on a detailed strategy, rather than the global situation, which has important implications for informing clinicians regarding treatment. It is useful to look to recovered populations to understand more about the potential role of punishment and reward sensitivity in the aetiology of EDs, particularly considering the chronic nature of the illness. The systematic review by Harrison et al. (2010) found that novelty seeking did not alter in the recovered form of BN, although there was some evidence that the high levels of harm avoidance were less pronounced in the recovered from of AN, albeit with a high level of variability between studies. Using functional magnetic resonance imaging paradigms, Wagner et al. (2007) found altered reward processing in women who had recovered from AN with a monetary reward task, and Frank et al. (2006) found women recovered from BN had a reduced reward response to glucose in comparison to artificial saliva. Using positron emission tomography, Frank et al. (2005) also 397 found participants recovered from AN had reduced D2/D3 receptor binding in the ventral striatum, a region involved in the modulation of responses to reward stimuli. Wagner et al. (2006, p276), using latent profile analysis, identified “inhibited” and “disinhibited” personality clusters in individuals recovered from both AN and BN, and Klump et al. (2004, p. 1407) commented that higher levels of harm avoidance and lower self-directedness and cooperativeness found in individuals with EDs compared to controls “may be trait-related disturbances and contribute to the disorders' pathogenesis.” This finding is supported by the review of Kaye et al. (2009). Taken together, the findings suggest that reward and punishment sensitivity may play an important role in the development and maintenance of EDs and it would be beneficial to expand on these interesting data. 1.1. Aims This study therefore aimed to expand the available data on reward and punishment sensitivity measured using the BIS BAS scales in individuals with EDs and those who had recovered in comparison to healthy controls. 1.2. Hypotheses It was predicted that transdiagnostically, people with an ED would have elevated sensitivity to punishment, as measured by the BIS, relative to healthy controls. The second hypothesis was that people with BN and not those with restricting AN would have a higher sensitivity to reward, as measured by the BAS scale and AMS. Finally, we predicted similar levels of reward and punishment sensitivity in a recovered sample as in the acutely ill group. 2. Method 2.1. Participants Female participants with EDs and those in recovery from an ED were recruited from the Eating Disorder Unit, Institute of Psychiatry's volunteer database. The HC group was recruited from the local community and from Surrey and Loughborough Universities. Recovered individuals were recruited using posters and flyers and from the aforementioned volunteer database. Those with EDs were required to meet DSM-IV (American Psychiatric Association, 1994) criteria for either AN or BN. This diagnosis was made based on information provided regarding weight and height, number of binges per week, presence of compensatory behaviours such as self-induced vomiting and excessive exercise, duration of symptoms, food restriction and existence of undue influence of shape and weight on self-evaluation. This information was gathered using the Eating Disorder Diagnostic Scale (EDDS) (Stice et al., 2000), a 22-item self-report measure. Bardone-Cone et al. (2010) stated that a definition of recovery from an ED should have physical, behavioural, and psychological components. Therefore, recovered participants were required to have a body mass index N 18.5, restored menstruation for at least the past year, and an absence of ED behaviours such as restriction or binge-purge symptoms during this period. These data were self-reported by participants. The presence of recovered participants' past episode(s) of an ED was supported by the lifetime history component of the EDDS. HC participants were excluded if they reported a history of disordered eating, also assessed using the EDDS. They were also screened for the presence of substance abuse using the General Addiction Inventory (Greenberg et al., 1999), as were the ED and recovered groups. The research was conducted in accordance with institutional and international standards. Prior to participation, participants received written information about the study and written informed consent was collected from all participants. The study was approved by the Kings College London research ethics committee. 2.2. Measures and procedure 2.2.1. Behavioural Inhibition System and Behavioural Activation System Scales (BIS/BAS Scales; Carver and White, 1994) This self-report questionnaire comprises two scales measuring trait conceptualisations of reward and punishment sensitivity. The behavioural inhibition (BIS) scale items concern reactions to negative events (e.g., ‘Criticism or scolding hurts me quite a bit’). The three BAS scales (BAS-Drive (BAS-D), BAS fun seeking (BAS-FS), and BAS reward responsiveness (BAS-RR)) provide differing conceptualisations of sensitivity to rewarding stimuli (e.g., ‘I go out of my way to get things I want’). The BIS/BAS questionnaire has 20 398 A. Harrison et al. / Psychiatry Research 188 (2011) 396–401 items answered using a 4-point Likert scale ranging from 1 (‘very true for me’) to 4 (‘very false for me’). Within the present sample, Cronbach's alpha coefficient was 0.81 for the BIS, 0.67 for BAS reward responsiveness, 0.86 for BAS drive, and 0.81 for BAS fun seeking. 2.2.2. The Appetitive Motivation Scale (AMS; Jackson and Smillie, 2004) The AMS is an instrument designed to measure reward reactivity independently of low impulse control. Many BAS measures (e.g., the SPSRQ; Torrubia et al., 2001) reflect not only desire of and response to reward, but also a tendency to act rashly and without forethought, which diverges somewhat from the construct of approach motivation (Dawe and Loxton, 2004; Smillie et al., 2006b). The AMS was recently shown to be orthogonal to traits reflecting such rash impulsivity (Cooper et al., 2008). The AMS has 11 items (e.g., ‘I like to be rewarded’) and a Cronbach's alpha coefficient of 0.81 (Cooper et al., 2008). In this study, Cronbach's alpha coefficient was 0.88, and items were responded to using a 5-point Likert scale ranging from 1 (‘very true for me’) to 4 (‘very false for me’). recovered from BN regarding duration of illness, time in recovery and age. Those who had recovered from BN reported having 4.7 (S.D.= 2.4) objective binge episodes per week at the height of their illness, and the lowest BMI of the AN group was 14.11 (S.D.= 2.9). 3.2. Clinical and demographic characteristics Table 1 provides the clinical and demographic features of the sample. Participants did not differ in terms of age and ethnicity. None of the members of the control group were taking psychiatric medication, and their BMI fell into the normal range and was significantly higher than those in the ED group. 3.3. Transdiagnostic analysis 2.3. Analysis Preliminary data screening, which involved conducting normality tests (Kolmogorov– Smirnov tests) and visual inspections of histograms of continuous variables, indicated that the data were suitable for parametric analysis. Data were analysed using the Statistical Package for the Social Sciences Version 15 (SPSS). Effect sizes were calculated based on Cohen's D (Cohen, 1988), which describes an effect size of 0.2–0.3 as small, around 0.5 as medium, and 0.8 and above as large. The analysis plan was constructed so as to account for the transdiagnostic approach in the field, as well as to analyse data based on distinct diagnostic subgroups described in the DSM-IV (American Psychiatric Association, 1994). The aim is that this will permit comparison of the data with studies that have used mixed samples, as well as those who have used subtype analyses. Therefore, data analysis starts from a transdiagnostic perspective and subsequently explores differences between subgroups. Discriminant analysis was then used to predict group membership from the predictors (reward/punishment sensitivity). This analysis is the inverse of a one-way multivariate analysis of variance (MANOVA), such that the levels of the independent variable for the MANOVA become the categories of the dependent variable for discriminant analysis, and the dependent variables of the MANOVA become the predictors for discriminant analysis. Successive orthogonal discriminant functions are extracted (n functions=n groups−1) are constructed according to weighted combinations of the predictor variables which maximise group differences. In line with our predictions, we expected that (1) BIS scores would maximally separate participants with an ED from healthy controls, (2) BAS/AMS scores would maximally separate participants with BN from participants with AN, and (3) recovered participants would not be distinguishable from acutely ill participants in terms of these approach and avoidance dispositions. 3. Results 3.1. Sample The final sample consisted of 286 participants in total. Of these participants, 91 were in the healthy control group (HC), 121 participants had a current ED (there were 29 participants with restricting AN, 55 participants with binge-purge AN, and 34 participants with BN); 74 participants who had recovered from an ED also took part in the study. All participants were female. The mean duration of illness for the ED group was 6.3 years (S.D. = 2.5). Participants with BN reported a mean of 5.1 objective binge episodes per week (S.D.= 0.38). The mean time spent in recovery for the recovered group was 4.2 years (S.D. = 1.3) and recovered participants reported experiencing a length of illness of 3.8 years (S.D. = 2.9). There were no differences between those recovered from AN and those An ANOVA was employed to investigate group differences (HC, Recovered and ED) for self-reported levels of reward sensitivity (AMS/BAS scores) and punishment sensitivity (BIS score). Results are presented in Table 2. There was a main effect of group for the BIS, BAS-RR, BAS-FS and AMS scales, with small to medium effect sizes. There was no main effect of group for the BAS-D scale, although there was a small effect size of 0.2. Post hoc tests revealed that after application of a Bonferroni correction for multiple testing, there were significant differences between the ED group and the HC group across all remaining measures, with the ED group scoring significantly lower on the BAS scales and AMS, and significantly higher on the BIS. There were also significant differences between the recovered and HC groups for the BIS and AMS measures, with the recovered group scoring significantly higher on the BIS and significantly lower on the AMS than HCs. There were no significant differences between the ED and recovered groups. There was no significant difference for the BIS BAS or AMS between those in the ED groups who were taking anxiolytic medication in comparison to those who were not. 3.3.1. Correlations between the BIS BAS and AMS measures Table 3 provides a table of correlations between the BIS BAS and AMS measures for the entire sample. There were significant positive correlations between the AMS and the BAS-D, BAS-FS and BAS-RR measures, and there was a significant negative relationship between the AMS and BIS. 3.4. Subgroup analysis The ED group was divided into diagnostic subtypes according to DSM-IV criteria (American Psychiatric Association, 1994). There were 29 participants with restricting AN, 55 participants with binge-purge AN, and 34 participants with BN. The recovered group was considered as one complete group, as those who had recovered from AN (binge purge or restricting subtypes) or BN did not differ on any measures. There were no significant differences in age or ethnicity between the ED/recovered subgroups. Table 1 Demographic and clinical data for the eating disorder, recovered and healthy control groups. Age (mean and S.D.) BMI (mean and S.D.) Ethnicity (%) Medication (n who were currently taking prescribed medication) ED participants (n = 121) Recovered participants (n = 74) Healthy control participants (n=91) Test result 34.21 (10.55) 18.66 (5.53)a;b AN: 14.35 (2.56) BN 23.38 (2.68) White 91.90% Black 5.40% Asian 1.80% Other 0.90% Anxiety medication 84 Anti-depressants 98 35.25(10.39) 20.95 (3.73) 31.79 (12.16) 22.85 (3.68) F(2,281) = 2.19, p = 0.11 F(2,276) = 25.06, p = ≤0.001 White 92.60% Black 4.90% Asian 1.50% Other 1.00% Anxiety medication 0 Anti-depressants 0 White 92.70% Black 4.80% Asian 1.50% Other 1.00% Anxiety medication 0 Anti-depressants 0 χ2(2) = 11.89, p = 0.31 n = number of participants; S.D. = standard deviation; ED = eating disorder. For age and BMI, test results are based on an ANOVA; for ethnicity, a chi-square test was used. a = significantly less than healthy controls (p ≤ 0.001);b = significantly less than recovered group (p ≤ 0.001). 399 A. Harrison et al. / Psychiatry Research 188 (2011) 396–401 Table 2 A comparison of reward and punishment sensitivity between people with eating disorders, individuals who have recovered from eating disorders and healthy controls: Means and standard deviations. Measure ED group (n = 121) BIS total score Mean (S.D.) BAS fun seeking mean (S.D.) BAS reward responsiveness mean (S.D.) BAS drive mean (S.D.) AMS total mean (S.D.) score 25.16 9.04 15.38 9.60 17.28 (3.23)a (3.19)a (2.81)a (2.95) (6.97)a Recovered group (n = 74) 24.99 10.08 15.61 9.67 19.08 (3.28)a (3.08)a (2.62)a (3.18) (6.38)a HC Group (n = 91) 21.89 11.21 16.27 10.30 21.45 (3.78) b, c (2.66)b, c (2.07)b, c (2.78) (5.51)b, c Test statistic p value Effect size (D) ED vs HC F(2,275) = 26.04 F(2,275) = 13.18 F(2,275) = 3.31 F(2,275) = 1.82 F(2,275) = 10.87 b0.001 b0.001 0.04 0.16 b0.001 0.75 0.54 0.27 0.20 0.49 n = number of participants; S.D. = standard deviation; ED = eating disorder; BIS = Behavioural Inhibition System Scale; BAS = Behavioural Activation Scale; AMS = Appetitive Motivation Scale. Effect sizes are calculated based on Cohen's D (Cohen, 1988), which describes an effect size of 0.2–0.3 as small, around 0.5 as medium and 0.8 and above as large. a = different from healthy controls; b = different from recovered group; c = different from eating disorder group. An ANOVA was used to investigate subgroup differences (restricting AN, binge purge AN, BN and recovered) on self-reported levels of reward sensitivity (AMS/BAS scores) and punishment sensitivity (BIS score). Results are displayed in Table 4. The only measure on which the subgroups differed significantly was BAS-FS. Post hoc tests indicated that, after application of a Bonferroni correction for multiple testing, the only difference was between the acute restricting AN group and the recovered group. Those who were currently ill scored significantly lower for the BAS-FS than those who had recovered. 3.5. Discriminant analysis To further examine differences among groups (ED, HC and recovered groups) in terms of the reward and punishment sensitivity measures, a discriminant analysis was performed. As scores from the three BAS scales and the AMS showed near identical patterns of variation across groups, these scales were combined, giving two predictors (BAS/AM and BIS) of ED group membership. The full model comprising two discriminant functions was significant, χ2(4) = 62.52, p b 0.001, and the reduced model (i.e., after removal of the first, larger function) was marginally significant, χ2(1) =3.65, p = 0.05. The first function, however, explained most (94.7%) of the variance in the model (canonical correlation = 0.43), and thus only this function will be interpreted. The structure matrix indicated that this function reflected a combination of high BIS scores (pooled within group correlation = 0.90) and low BAS scores (pooled within-group correlation = −0.57). This means that the ED groups were maximally separated by a function reflecting high BIS and low BAS. Specifically, this function discriminated the HC group (function at group centroid = −0.70) from the ED group and recovered group (function at group centroid = 0.38 and 0.24, respectively). Classification data indicated that the model was reasonably accurate in classifying membership in the HC group (63% accurately classified), and highly accurate for the ED group (78% accurately classified). Classification accuracy for the recovered sample was very poor, as 70% were classified as EDs. This means that recovered individuals are not distinguishable from ED individuals on the basis of reward and punishment sensitivity measures. 4. Discussion This study aimed to expand previous data regarding BIS/BAS scale scores in individuals with EDs and also explored reward responsiveness Table 3 Correlations between the BIS BAS and AMS measures for the whole sample. BAS fun seeking BAS drive 0.518⁎⁎ BAS fun seeking 1 BAS reward responsiveness BIS score AMS score BAS reward responsiveness BIS score 0.586⁎⁎ 0.520⁎⁎ 1 0.073 0.336⁎⁎ − 0.104 ⁎⁎ Correlation is significant at the p = 0.05 level (2 tailed). 1 AMS score 0.537⁎⁎ 0.744⁎⁎ 0.510⁎⁎ − 0.318⁎⁎ 1 using the AMS. The first hypothesis, that individuals with an ED would have higher sensitivity to punishment than controls was supported. Individuals with an ED were found to have significantly elevated BIS scores when compared with the HC group. The second hypothesis, that people with BN and not those with restricting AN would have a higher sensitivity to reward was not supported. In fact, highly consistent findings were obtained using four separate trait conceptualisations of reward reactivity (BAS drive, reward responsiveness and fun seeking and the AMS), all of which were lower in the ED group overall, relative to the HC group. The third hypothesis, that individuals who had recovered from an ED would show similar levels of reward and punishment sensitivity as an acutely ill group, was supported. Scores on all measures employed did not vary between the ED and recovered groups with only one exception. Specifically, those with restricting AN scored significantly lower on the BAS-FS scale than those who had recovered. Discriminant analysis helped to consolidate these findings: HCs were maximally separated from those presently and formerly with an ED along a composite dimension reflecting high punishment sensitivity and low reward sensitivity. Finally, classification analysis demonstrated that while ED and HC group membership was relatively well predicted from reward and punishment sensitivity measures, there was a clear tendency for recovered participants to be misclassified as ED. 4.1. Comparisons with previous studies Three studies, including this one, Claes et al. (2006) and Kane et al. (2004), using the BIS, have now demonstrated that individuals with EDs consistently demonstrate high levels of punishment sensitivity compared to controls. The data for reward sensitivity are less clear, but in general, the findings of this study and Claes et al. (2006) support the existence of lower levels of reward sensitivity in people with EDs than HCs, although this is not replicated by Kane et al. (2004) who found the opposite. However, the present study is novel in reporting data for the AMS for an ED sample. This provides further evidence (i.e., beyond the BIS/BAS scales) that people with EDs show lower levels of appetitive motivation. Nevertheless, this finding raises difficult questions regarding the role of appetitive motivation in eating behaviour, which appears to vary from study to study, and therefore may interact with as yet unknown or unmeasured processes. The findings regarding high levels of punishment sensitivity and low levels of reward sensitivity in the recovered group, and the classification analysis in which recovered participants tended to be misclassified as being in the ED group provide support for models of EDs, such as that of Kaye (2008), which suggest that dysregulated reward systems are involved in the development of EDs. 4.2. Limitations The low internal reliability (0.67) of BAS reward responsiveness should be noted. The diagnosis of the participants was made based on self-reports, which may be inaccurate and it would have been preferable to have carried out semi-structured interviews to confirm diagnoses. However, participants are thoroughly screened when they enter the volunteer database and in addition provided a diagnostic category from 400 A. Harrison et al. / Psychiatry Research 188 (2011) 396–401 Table 4 A comparison of reward and punishment sensitivity within diagnostic subtypes (restricting anorexia, binge-purge anorexia and bulimia) of people with eating disorders and those recovered from an eating disorder: means and standard deviations. Measure Restricting AN group (n = 29) Binge purge AN group (n = 55) BN group (n = 34) Recovered group (n = 74) Test result p value Effect size (d) RAN vs BPAN vs BN vs recovered BIS total score mean (S.D.) BAS fun seeking mean (S.D.) BAS reward responsiveness mean (S.D.) BAS drive Mean (S.D.) AMS total score mean (S.D.) 25.79 (2.35) 7.97 (3.21) 15.10 (3.04) 25.21 (3.43) 8.96 (3.22) 15.09 (2.70) 24.32 (3.55) 10.06 (2.87) 16.00 (2.62) 24.99 (3.28) 10.08 (3.08) 15.61 (2.62) F(3,185) = 1.07 F(3,185) = 4.08 F(3,185) = 0.97 0.36 0.01 0.41 0.24 0.68 0.18 9.10 (2.98) 16.48 (7.07) 9.51 (2.96) 16.74 (6.90) 9.94 (2.80) 18.77 (6.65) 9.67 (3.18) 19.08 (6.38) F(3,185) = 0.413 F(3,185) = 1.929 0.74 0.13 0.18 0.29 n = number of participants; S.D. = standard deviation; AN = anorexia nervosa, BN = bulimia nervosa; BIS = Behavioural Inhibition System Scale; BAS = Behavioural Activation Scale; AMS = Appetitive Motivation Scale. Effect sizes are calculated based on the results of the ANOVA using Cohen's d (Cohen, 1988), which describes an effect size of 0.2–0.3 as small, around 0.5 as medium and 0.8 and above as large. NB: As the participants were screened for possible alcohol addiction using the General Addiction Inventory (Greenberg et al., 1999), it was possible to re-run the analysis with this as a covariate. All of the above main effects remained significant as above, except for the main effect of BIS (p = 0.3). their clinician where available. Due to power limitations, it was not ideal to further sub divide the BN group, but it would have been useful to split the BN group into purging and non purging subtypes as defined in the DSM-IV (American Psychiatric Association, 1994) and future work using semi-structured interviews for diagnosis would assist subtype discrimination. Power limitations may also explain why we did not observe differences between the subtypes of the recovered group. Finally, Bardone-Cone et al. (2010) recommend that a criterion of recovery is scoring within 1 standard deviation of age-matched community norms on all Eating Disorder Examination Questionnaire (EDE-Q; Fairburn and Beglin, 1994) subscales. The EDE-Q was not utilised in this study and as highlighted by Couturier and Lock (2006), using different definitions of recovery may limit comparison of this recovered group to other cohorts. Another relevant issue is the distinction which has recently been made between anxiety and fear as separate components of avoidance motivation (see Smillie et al., 2006a). It is not clear how this distinction might qualify the present results or results of other similar studies. Although it has been suggested that dividing Carver and White's (1994) BIS scale into two subscales can approximate the constructs of fear and anxiety (Heym et al., 2008), we did not do this in the present study as the resulting subscales showed unacceptably low reliability and high intercorrelations. This potentially reflects the fact that fear and anxiety are very similar at a descriptive level, even if they have separate biobehavioural bases. Behavioural or psychopharmacologic paradigms (e.g., Perkins et al., 2009) may be a more comparatively promising way to examine this issue in future ED research. The argument that high punishment sensitivity and low reward sensitivity are trait personality factors associated with a history of EDs is limited by the fact that the structure of these factors of the recovered group during the acute phase of their illness is unknown. Longitudinal studies are required to assess whether the high punishment sensitivity and low reward sensitivity observed in this study are risk factors or a scar of the illness. Based on Wilksch and Wade's (2009) findings that sensitivity to punishment was found at significantly elevated levels in non-affected twins, this could be a fruitful line of enquiry. 4.3. Future research Future studies could look at these self-report measures of reward and punishment sensitivity alongside behavioural measures, to compare self-reported reward/punishment sensitivity to experimental data. Prospective studies are also required to make stronger causal inferences. Future work may also involve exploring social reward, as Tiller et al. (1997) found individuals with EDs have poor social networks and are socially isolated and Kaye et al. (2009, p. 577) state that individuals with AN “find little in life that is rewarding aside from the pursuit of weight loss.” This statement is supported experimentally by Watson et al. (2010), who, using an econometric choice task (Hayden et al., 2007) and an eye-tracking paradigm, found women with AN did not find female faces rewarding and avoided looking at the face and eyes. 5. Conclusions This study provides further evidence for the association between high levels of punishment sensitivity and eating psychopathology. The study also suggests that people with EDs are characterised by lower levels of reward sensitivity. Those who had recovered from an ED did not differ from those who were acutely ill, suggesting high punishment sensitivity and low reward reactivity/sensitivity might form a personality cluster associated with the risk of developing an ED. These data add to the knowledge of how constructs from Gray's reinforcement sensitivity theory relate to the development and maintenance of disordered eating. Acknowledgements We thank Dr. Caroline Meyer at Loughborough University, who assisted with data collection, and Kimberley Davis at the University of Surrey, who assisted with both data collection and data input. We are also very grateful to the anonymous reviewers who contributed to the refinement of this work. References American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental Disorders: DSM-IV. American Psychiatric Press, Inc, Washington, DC. Ampollini, P., Marchesi, C., Signifredi, R., Ghinaglia, E., Scardovi, F., Codeluppi, S., Maggini, C., 1999. Temperament and personality features in patients with major depression, panic disorder and mixed conditions. Journal of Affective Disorders 52, 203–207. Bardone-Cone, A.M., Harney, M.B., Maldonado, C.R., Lawson, M.A., Robinson, D.P., Smith, R., Tosh, A., 2010. Defining recovery from an eating disorder: conceptualization, validation, and examination of psychosocial functioning and psychiatric comorbidity. Behaviour Research and Therapy 48 (3), 194–202. Blinder, B.J., Cumella, E.J., Sanathara, V.A., 2006. Psychiatric comorbidities of female in patients with eating disorders. Psychosomatic Medicine 68 (3), 454–462. Carver, C.S., White, T.L., 1994. Behavioral inhibition, behavioral activation, and affective responses to impending reward and punishment: the BIS/BAS scales. Journal of Personality and Social Psychology 67 (2), 319–333. Carver, C.S., Johnson, S.L., Joormann, J., 2009. Two-mode models of self-regulation as a tool for conceptualizing the role of serotonergic function in normal behavior and diverse disorders. Current Directions in Psychological Science 18, 195–199. Claes, L., Nederkoorn, C., Vandereycken, W., Guerrieri, R., Vertommen, H., 2006. Impulsiveness and lack of inhibitory control in eating disorders. Eating Behaviors 7 (3), 196–203. Claes, L., Robinson, M.D., Muehlenkamp, J.J., Vandereycken, W., Bijttebier, J., 2010. Differentiating bingeing/purging and restrictive eating disorder subtypes: the roles of temperament, effortful control, and cognitive control. Personality and Individual Differences 48, 166–170. Cloninger, C.R., 1987. A systematic method for clinical description and classification of personality variants: a proposal. Archives of General Psychiatry 44 (6), 573–588. Cloninger, C.R., 1993. A psychobiological model of temperament and character. Archives of General Psychiatry 44, 975–990. Cohen, J., 1988. Statistical Power Analysis for the Behavioural Sciences, 2nd edition. Academic Press, New York. Cooper, A.J., Smillie, L.D., Jackson, C.J., 2008. A trait conceptualisation of reward-reactivity: psychometric properties of the Appetitive Motivation Scale (AMS). Journal of Individual Differences 29, 168–180. Corr, P.J., Pickering, A.D., Gray, J.A., 1997. Personality, punishment and procedural learning: a test of J.A. Gray's anxiety theory. Journal of Personality and Social Psychology 73, 337–344. Couturier, J., Lock, J., 2006. What is recovery in adolescent anorexia nervosa? International Journal of";. International Journal of Eating Disorders 39 (7), 550–555. Davis, C., Fox, J., 2008. Sensitivity to reward and body mass index (BMI): evidence for a non-linear relationship. Appetite 50 (1), 43–49. A. Harrison et al. / Psychiatry Research 188 (2011) 396–401 Dawe, S., Loxton, N.J., 2004. The role of impulsivity in the development of substance use and eating disorders. Neuroscience and Biobehavioral Reviews 28 (3), 343–351. Depue, R.A., 2006. Neurobiological factors in personality and depression. European Journal of Personality 9 (5), 413–439. Elliot, A.J. (Ed.), 2008. Handbook of Approach and Avoidance Motivation. Taylor & Francis, New York. Fairburn, C.G., Beglin, S.J., 1994. Assessment of eating disorders: interview or self-report questionnaire? The International Journal of Eating Disorders 16, 363–370. Farmer, R.F., Nash, H.M., Field, C.E., 2001. Disordered eating behaviors and reward sensitivity. Journal of Behavior Therapy and Experimental Psychiatry 32 (4), 211–219. Fowles, D.C., 1993. Biological variables in psychopathology: a psychobiological perspective, In: Sutker, P.B., Adams, H.E. (Eds.), Comprehensive Handbook of Psychopathology, 2nd edition. Plenum Press, New York, pp. 57–82. Frank, G., Bailer, U.F., Henry, S.E., Drevets, W., Meltzer, C.C., Price, J.C., Mathis, C.A., Wagner, A., Hoge, J., Ziolko, S., Barbarich-Marsteller, N., Weissfeld, L., Kaye, W.H., 2005. Increased dopamine D2/D3 receptor binding after recovery from anorexia nervosa measured by positron emission tomography and [11C]raclopride. Biological Psychiatry 58, 908–912. Frank, G.K., Wagner, A., Achenbach, S., McConaha, C., Skovira, K., Aizenstein, H., Carter, C.S., Kaye, W.H., 2006. Altered brain activity in women recovered from bulimic-type eating disorders after a glucose challenge: a pilot study. The International Journal of Eating Disorders 39 (1), 76–79. Franken, I.H.A., Muris, P., 2005. Individual differences in decision-making. Personality and Individual Differences 39, 991–998. Godart, N.T., Flament, M.F., Curt, F., Perdereau, F., Jeammet, P., 2003. Comorbidity between eating disorders and anxiety disorders. The International Journal of Eating Disorders 23, 253–270. Grau, E., Ortet, G., 1999. Personality traits and alcohol consumption in a sample of nonalcoholic women. Personality and Individual Differences 27, 1057–1066. Gray, J.A., 1991. Neural systems of motivation, emotion and affect. In: Madden, J. (Ed.), Neurobiology of Learning, Emotion and Affect. Raven Press, New York, pp. 273–306. Greenberg, J.L., Lewis, S.E., Dodd, D.K., 1999. Overlapping addictions and self-esteem among college men and women. Addictive Behaviors 24 (4), 565–571. Harrison, A., O'Brien, N., Lopez, C., Treasure, J., 2010. Sensitivity to reward and punishment in eating disorders. Psychiatry Research 177 (1–2), 1–11. Hayden, B.Y., Parikh, P.C., Deaner, R.O., Platt, M.L., 2007. Economic principles motivating social attention in humans. Proceedings of the Royal Society of Biological Sciences 274, 1751–1756. Heym, N., Ferguson, E., Lawrence, C., 2008. An evaluation of the relationship between Gray’s revised RST and Eysenck’s PEN: distinguishing BIS and FFFS in Carver and White’s BIS/BAS scales. Personality and Individual Differences 45, 709–715. Jackson, C.J., Smillie, L.D., 2004. Appetitive motivation predicts the majority of personality and an ability measure: a comparison of BAS measures and a re-evaluation of the importance of RST. Personality and Individual Differences 36, 1627–1636. Kane, T.A., Loxton, N.J., Staigera, P.K., Daweb, S., 2004. Does the tendency to act impulsively underlie binge eating and alcohol use problems? An empirical investigation. Personality and Individual Differences 36, 83–94. Kaye, W., 2008. Neurobiology of anorexia and bulimia. Physiology and Behaviour 94 (1), 121–135. Kaye, W.H., Fudge, J.L., Paulus, M., 2009. New insights into symptoms and neurocircuit function of anorexia nervosa. Nature Reviews. Neuroscience 10, 573–584. Klump, K., Strober, M., Bulik, C.M., Thornton, L., Johnson, C., Devlin, B., Fichter, M.M., Halmi, K.A., Kaplan, A.S., Woodside, D.B., Crow, S., Mitchell, S., Rotondo, A., Keel, P.K., Berrettini, W.H., Plotnicov, K., Pollice, C., Lilenfeld, L.R., Kaye, W.H., 2004. Personality characteristics of women before and after recovery from an eating disorder. Psychological Medicine 34, 1407–1418. Loxton, N.J., Dawe, S., 2001. Alcohol abuse and dysfunctional eating in adolescent girls: the influence of individual differences in sensitivity to reward and punishment. The International Journal of Eating Disorders 29, 455–462. Loxton, N.J., Dawe, S., 2006. Reward and punishment sensitivity in dysfunctional eating and hazardous drinking women: associations with family risk. Appetite 47, 361–371. 401 Loxton, N.J., Dawe, S., 2007. How do dysfunctional eating and hazardous drinking women perform on behavioural measures of reward and punishment sensitivity? Personality and Individual Differences 42, 1163–1172. Nederkoorn, C., van Eijs, Y., Jansen, A., 2004. Restrained eaters act on impulse. Personality and Individual Differences 37, 1651–1658. Pallister, E., Waller, G., 2008. Anxiety and eating disorders: understanding the overlap. Clinical Psychology Review 28, 366–386. Perkins, A.M., Ettinger, U., Davis, R., Foster, R., Williams, S.C.R., Corr, P.J., 2009. Effects of lorazepam and citalopram on human defensive reactions: ethopharmacological differentiation of fear and anxiety. The Journal of Neuroscience 29, 12617–12624. Pickering, A.D., Smillie, L.D., 2008. The behavioural activation system: challenges and opportunities. In: Corr, P.J. (Ed.), The Reinforcement Sensitivity Theory of Personality. Cambridge University Press, Cambridge, pp. 120–154. Powell, J.H., Al-Adawi, S., Morgan, J., Greenwood, R.J., 1996. Motivational deficits after brain injury: effects of bromocriptine in 11 patients. Journal of Neurology, Neurosurgery, and Psychiatry 60, 416–421. Smillie, L.D., Pickering, A.D., Jackson, C.J., 2006a. The new Reinforcement Sensitivity Theory: implications for psychometric measurement. Personality and Social Psychology Review 10, 320–335. Smillie, L.D., Jackson, C.J., Dalgleish, L.I., 2006b. Conceptual distinctions among Carver and White's (1994) BAS scales: a reward-reactivity versus trait impulsivity perspective. Personality and Individual Differences 40, 1039–1050. Smillie, L.D., Loxton, N.J., Avery, R.E., 2010. Reinforcement sensitivity theory, research, applications and future. In: Chamorro-Premuzic, T., Furnham, A.F., von Stumm, S. (Eds.), Handbook of Individual Differences. Wiley, Blackwell, London. Stice, E., Telch, C.F., Rizvi, S.L., 2000. Development and validation of the Eating Disorder Diagnostic Scale: a brief self-report measure of anorexia, bulimia, and binge-eating disorder. Psychological Assessment 12 (2), 123–131. Taffel, C., 1955. Anxiety and the conditioning of verbal behavior. Journal of Abnormal and Social Psychology 51, 496–501. Tiller, J.M., Sloane, G., Schmidt, U., Troop, N., Power, M., Treasure, J., 1997. Social support in patients with anorexia nervosa and bulimia nervosa. The International Journal of Eating Disorders 21, 31–38. Torrubia, R., Avila, C., Moltó, J., Caseras, X., 2001. The sensitivity to punishment and sensitivity to reward questionnaire (SPSRQ) as a measure of Gray's anxiety and impulsivity dimensions. Personality and Individual Differences 31, 837–862. Vitousek, K., Manke, F., 1994. Personality variables and disorders in anorexia nervosa and bulimia nervosa. Journal of Abnormal Psychology 103, 137–147. Wagner, A., Barbarich-Marsteller, N.C., Frank, G.K., Bailer, U.F., Wonderlich, S.A., Crosby, R.D., Henry, S.E., Vogel, V., Plotnicov, K., McConaha, C., Kaye, W.H., 2006. Personality traits after recovery from eating disorders: do subtypes differ? The International Journal of Eating Disorders 39 (4), 276–284. Wagner, A., Aizenstein, H., Venkatraman, V.K., Fudge, J., May, J.C., Mazurkewicz, L., Frank, G.K., Bailer, U.F., Fischer, L., Nguyen, V., Carter, C., Putnam, K., Kaye, W.H., 2007. Altered reward processing in women recovered from anorexia nervosa. The American Journal of Psychiatry 16, 1842–1849. Wagner, A., Aizenstein, H., Mazurkewicz, L., Fudge, J., Frank, G.K., Putnam, K., Bailer, U.F., Fischer, L., Kaye, W.H., 2008. Altered insula response to a taste stimulus in individuals recovered from restricting-type anorexia nervosa. Neuropsychopharmacology 33, 513–523. Watson, K.K., Werling, D.M., Zucker, N.L., Platt, M.L., 2010. Altered social reward and attention in anorexia nervosa. Frontiers in Psychology 1 (36). doi:10.3389/fpsyg.2010.00036. Wedig, M.M., Nock, M.K., 2010. The functional assessment of maladaptive behaviors: a preliminary evaluation of binge eating and purging among women. Psychiatry Research 178 (3), 518–524. Wilksch, S.M., Wade, T.D., 2009. An investigation of temperament endophenotype candidates for early emergence of the core cognitive component of eating disorders. Psychological Medicine 39, 811–821. Zuckerman, M., 2005. Psychobiology of Personality, Second edition. Cambridge University Press, New York.