Alcohol Abuse and Dysfunctional Eating in Adolescent Girls: The Influence of Individual Differences in Sensitivity to Reward and Punishment Natalie J. Loxton and Sharon Dawe* School of Applied Psychology, Griffith University, Brisbane, Queensland, Australia Accepted 18 July 2000 Abstract: Objective: An unusually high comorbidity of eating disorders and alcohol abuse has been found in clinical and community samples of young women. This paper proposes that individual differences in sensitivity to reward and punishment may influence the propensity of young women to engage in dysfunctional eating and drinking behaviour. Method: The Drive for Thinness scale, the Alcohol Use Disorders Identification Test and the BIS/BAS scales were administered to 232 high school girls. Results: Heightened sensitivity to reward was the better predictor of alcohol misuse while heightened sensitivity to both reward and punishment was predictive of dysfunctional eating. When categorised by group, alcohol abusing, dysfunctional eating, and comorbid girls reported greater sensitivity to reward than non-disordered girls. Girls with dysfunctional eating with and without comorbid alcohol abuse reported greater sensitivity to punishment than alcohol abusing only girls. Discussion: These findings suggest that girls who abuse alcohol and have dysfunctional eating may share a vulnerability to heightened sensitivity to reward, yet be differentiated by sensitivity to punishment. © 2001 by John Wiley & Sons, Inc. Int J Eat Disord 29: 455–462, 2001. Key words: reward; punishment; alcohol; girls; personality, bulimia INTRODUCTION An unusually high comorbidity between eating disorders and alcohol use disorders has been well documented in recent years, in both clinical (Holderness, Brooks-Gunn & Warren, 1994; Krahn, 1991) and subclinical populations (Krahn, Kurth, Demitrack, & Drewnowski, 1992; Lundholm, 1989). Researchers have proposed two possible mechanisms: one in which an innate biological vulnerability to disordered eating and drinking exists (Krahn, 1991), a proposal supported by evidence of high rates of drug and alcohol problems in first degree relatives of eating disordered women (Strober, Salkin, Burroughs, *Correspondence to: Sharon Dawe, School of Applied Psychology, Griffith University, Mt. Gravatt, 4111, QLD. Australia. E-mail: s.dawe@mailbox.gu.edu.au © 2001 by John Wiley & Sons, Inc. 456 Loxton and Dawe & Morrell, 1982). Alternatively, the co-occurrence of both disorders may be influenced by personality traits, such as anxiety and impulsivity, frequently observed in women with disordered eating and drinking behaviour (Grau & Ortet, 1999; Vitousek & Manke, 1994). Whilst these proposed mechanisms of biological and psychological vulnerability have typically been investigated as opposing explanations, we suggest that both these mechanisms may be accommodated within a biologically based theory of personality. Gray (1970) posits the personality traits of impulsivity and anxiety as influenced by two independent, biologically based, motivational systems: the Behavioural Approach System (BAS), which regulates response to rewarding stimuli, and the Behavioural Inhibition System (BIS), which regulates response to conditioned cues of punishment. Impulsivity is thought to reflect an individual’s sensitivity to cues in the environment, which come to be associated, via conditioning processes, with potential reward. Individuals with greater BAS sensitivity should therefore exhibit a greater propensity to engage in approach (impulsive) behaviour and experience positive affect in situations cued with reward (Carver & White, 1994). On the other hand, BIS is involved in the inhibition of behaviour by increasing arousal and attention in response to conditioned cues of punishment and is implicated in the expression of negative affective states, such as fear and anxiety in response to such cues. Individuals with high BIS sensitivity experience increased behavioural inhibition and negative affect in situations cued for potential punishment (Gray, 1982). As these are distinct neural systems, individuals may vary in their degree of sensitivity to punishment and reward. Accordingly, an individual may have any combination of high and low sensitivity to reward and punishment. This theory may prove a useful framework for studying comorbidity in women with disordered eating and alcohol abuse. Previous studies investigating personality traits report high levels of trait impulsivity (novelty seeking; sensation seeking) and trait anxiety (harm avoidance) in eating disordered women. Bulimic women report higher trait impulsivity and trait anxiety than normal controls while anorexic women report similar levels of impulsivity to normal controls but greater trait anxiety than bulimic women (Vitousek & Manke, 1994; Battaglia, Przybeck, Bellodi & Cloninger, 1996; Bulik, Sullivan, Weltzin & Kaye, 1995). In terms of Gray’s theory, these findings suggest that eating disordered women may have greater sensitivity to punishment than non-disordered women, with anorexic women particularly sensitive to punishment, resulting in inhibited behaviour such as restrictive eating, and bulimic women being more sensitive to cues of reward resulting in approach behaviour such as binge eating. Research into personality characteristics of alcohol abusers suggest that personality characteristics of impulsivity and sensation seeking may predispose individuals to engage in problematic drinking, and that individuals who eventually abuse alcohol appear to be more sensitive to the reinforcing effects of alcohol (Cox, 1987). Recent research has found that young women who abuse alcohol score higher on measures of trait impulsivity and sensation seeking than either normal or psychiatric controls (Battaglia, et al., 1996; Grau & Ortet, 1999). A recent community study found a positive correlation between scores on a self-report measure of BAS and self-reported alcohol use (Jorm et al., 1999). These findings tentatively suggest a relationship between behavioural approach and the initiation and maintenance of alcohol use and related problems. The association between anxiety and alcohol misuse is somewhat equivocal though, with some studies finding a positive relationship (Brady, Grice, Dunstan, & Randall, 1993), and others a negative correlation (Battaglia et al., 1996), whilst others find no relationship (Grau & Ortet, 1999). Few studies though have compared eating disordered women to alcohol abusing Alcohol Abuse and Eating Disorders 457 women. Battaglia et al. (1996) though, found both eating disordered women and substance abusing women scored higher on novelty seeking than controls, and eating disordered women scoring higher than substance abusing women and controls on harm avoidance, thus indicating that alcohol misusing women, like bulimic women, may have a heightened sensitivity to reward, but may be less sensitive to cues of punishment than eating disordered women. Finally, women with comorbid bulimia and alcohol use disorders report higher trait anxiety and impulsivity than population norms (Bulik, Sullivan, Carter, & Joyce, 1997; Bulik, Sullivan, McKee, Weltzin & Kaye, 1994). Comorbid women, however, report greater trait impulsivity than women with bulimia only. Indeed, Wonderlich & Mitchell (1997) suggest that impulsivity may be one aspect that differentiates bulimic women with no substance misuse from those women who engage in both disordered behaviours. Comorbid women, therefore, may have a greater sensitivity to reward than women with bulimia only, in addition to heightened sensitivity to punishment. Research to date has tended to compare either eating disordered only or alcohol misusing only women to normal controls, or comorbid women to eating disordered only women. No single study has made comparisons in personality between all four groups. Furthermore, previous studies have used women with diagnosed disorders when assessing the influence of personality traits on psychopathology, which is problematic as behaviour and personality are often affected by the clinical disorder itself (Vitousek & Manke, 1994). To control for the confounding influence of psychopathology on reported personality, the current study will use non-clinical adolescent girls rather than women with diagnosed disorders. Further, as relatively few girls in this population meet diagnostic criteria for bulimia or anorexia this study will focus on dysfunctional eating behaviour and attitudes, such as dieting and fear of gaining weight, which are strong predictors of subsequent eating disorders (Joiner, Heatherton, & Keel, 1997). In sum, this paper proposes that bulimic and alcohol abusing women may share a propensity towards the (mis)use of naturally occurring rewards (such as food and psychoactive substances) due to an inherently heightened sensitivity to reward and subsequent approach behaviour (BAS). Furthermore, women with disordered eating may be prone to experience anxiety due to a heightened sensitivity to punishment (BIS). In essence, whereas reward sensitivity may link disordered eating (primarily bulimia) and alcohol abuse, heightened sensitivity to punishment may differentiate disordered eating and alcohol abuse. Accordingly, the aims of the present study are to: 1) investigate the predictability of scores on measures of dysfunctional eating and drinking from scores on a self-report measure of BIS/BAS and 2) explore the differences in self-reported sensitivity to reward and punishment between those girls who are in the clinical range on measures of dysfunctional eating and/or alcohol use and those do not score in the clinical range on either measure. METHOD Participants Two hundred and thirty-two senior high school girls (age range 16–18 years) were recruited from two schools in a major metropolitan centre in Australia. The sample was primarily Caucasian (74.9%) with smaller proportions of Indigenous Australians (4.3%), Asians (16.5%) and others (4.3%). 458 Loxton and Dawe Measures BIS/BAS Reactivity. The BIS/BAS scales were used to measure Gray’s behavioural inhibition (BIS) and behavioural approach (BAS; Carver & White, 1994). The instrument is a 20-item self-report measure to derive a global BIS score and three BAS related subscale scores: Drive (DRV), Fun Seeking (FS) and Reward Responsiveness (RR). Dysfunctional Eating. The Drive for Thinness scale (DT) of the EDI-2 (Garner, 1991) assesses respondents’ concerns regarding weight gain and preoccupation with weight and dieting, and was used to measure severity of dysfunctional eating. DT scores have been found to predict the development of bulimic behaviour over a ten-year period (Joiner et al., 1997). Alcohol Abuse. The Alcohol Use Disorders Identification Test (AUDIT; Saunders, Aasland, Babor, de le Fuente, & Grant, 1993), a measure consisting of 10 items that assess frequency and quantity of alcohol use, drinking behaviour and problems resulting from drinking was used. Procedure Upon receipt of parental consent, girls were permitted to take part in the study. Questionnaires were administered during class time where the participants were instructed to complete the questionnaire booklet alone and anonymously. RESULTS Descriptive Statistics Seventy-two percent of the sample reported drinking alcohol and 32% engaged in hazardous drinking behaviour (AUDIT score $8). Twelve percent of the sample scored in the clinical range of the DT scale ($14). The mean score on this scale in the current sample (M = 4.76, SD = 5.69) was somewhat lower than that found in American high school girls (M = 7.1, SD = 5.9; Shore & Porter, 1990). Mean scores on the BIS/BAS scales (BIS, M = 21.28, SD = 3.06; BAS-DRV, M = 10.40, SD = 2.00; BAS-FS, M = 11.38, SD = 1.90; BAS-RR, M = 16.29, SD = 1.86) were similar to those found in an Australian sample of young women (Jorm et al., 1999). However, while the BAS DRV and FS subscales correlated with each other (r = .42, p < .001) and were uncorrelated with BIS (r = .12, ns. & .00, ns. respectively) as expected, the BAS-RR subscale was as highly correlated with BIS (r = .36, p < .001) as with the BAS-DRV (r = .35, p < .001) and FS subscales (r = .22, p < .01). As such, BAS-RR was not used in subsequent analyses. Association Between Sensitivity to Reward and Punishment and Dysfunctional Eating and Alcohol Misuse Hierarchical regressions were employed to investigate the relationship between sensitivity to reward (BAS) and punishment (BIS), and alcohol misuse and dysfunctional eating. In the first analysis AUDIT score was the dependent variable and BIS was entered Alcohol Abuse and Eating Disorders 459 in the first step to account for any interaction BIS may have with BAS on this variable. As shown in Table 1 BIS score accounted for 2% of the variance, F(1, 230) = 3.91, p < .05. BAS scores accounted for an additional 16% of the variance, which was a significant improvement in the prediction of alcohol abuse over BIS alone, F(2, 228) = 21.89, p < .001. Together, BIS and BAS sensitivity accounted for 18% of the variance in alcohol use, R = .42, F(3, 228) = 16.14, p < .001. Examination of the b values suggest that BAS was the strongest predictor of alcohol abuse in adolescent girls with higher FS and DRV scores associated with greater alcohol use. BIS was markedly less predictive of alcohol misuse, with lower BIS scores associated with higher AUDIT scores. In the second analysis DT score was the dependent variable and BAS was entered into the first step. As shown in Table 2, BAS scores accounted for 7% of the variance, F(2, 229) = 8.07, p < .001. BIS score accounted for an additional 5% of the variance, which was a significant improvement in the prediction of dysfunctional eating over BAS alone, F(1, 228) = 12.15, p < .001. Together, BIS and BAS sensitivity accounted for 11% of the variance in dysfunctional eating, R = .34, F(3, 228) = 9.69, p < .001. Examination of the b values suggest that BIS and BAS FS scores were the strongest predictors of dysfunctional eating with higher BIS and FS scores associated with more severe dysfunctional eating. Differences Between Dysfunctional Eating Girls, Alcohol Misusing Girls, Comorbid Girls and Controls in Sensitivity to Reward and Punishment Cut-offs of $13 on the DT and $8 on the AUDIT were utilised to categorize girls as high risk for disordered eating and alcohol misuse, respectively. Based on these cut-offs girls were classified into one of the following four groups: Dysfunctional eating only (DE; n = 17); Alcohol misusing only (ALC; n = 58); Comorbid (CM; n = 18) and a control group of non-disordered girls (C; n = 138). As this resulted in much larger ALC and C groups than DE and CM groups, an ALC group comprising girls who scored the highest on the AUDIT whilst scoring low (# 2) on DT (n = 24) and a C group comprising girls with scores of zero on both the AUDIT and DT (n = 28) were formed. To assess differences in reported sensitivity to reward between groups, a one way between groups MANOVA was performed on two dependent variables: BAS-DRV and BAS-FS. Using Wilk’s criteria, the combined DVs were significantly affected by group, F(6, 164) = 5.91, p < .001. Univariate statistics revealed each subscale of BAS was significantly Table 1. Hierarchical multiple regression of BIS and BAS on alcohol misuse (AUDIT; n = 232) Variables Step 1 BIS Step 2 BAS–DRV BAS–FS Intercept = −.4.431 B −.340 .793 .810 b Partial r2 −.158* .02* .240** .233** R2 = .18 Adjusted R2 = .16 R = .42** .16** Note: BIS = Behavioral Inhibition System; BAS = Behavioral Approach System; DRV = Drive; FS = Fun Seeking; AUDIT = Alcohol Use Disorders Identification Test. *p < .05. **p < .001. 460 Loxton and Dawe Table 2. Hierarchical multiple regression of BIS and BAS on disordered eating (Drive for Thinness; n = 232) Variables Step 1 BAS–DRV BAS–FS Step 1 BIS Intercept = −13.09 B b Partial r2 .364 .474 .128 .158* .07** .408 .219** R2 = .11 Adjusted R2 = .10 R = .34** .05** Note: BIS = Behavioral Inhibition System; BAS = Behavioral Approach System; DRV = Drive; FS = Fun Seeking. *p < .05. **p < .001. affected by group, BAS DRV F(3, 83) = 5.74, p < .001; BAS FS F(3, 83) = 10.26, p < .001. Post-hoc analyses using Tukey’s HSD procedure were performed on each subscale to determine which groups were differentiated by BAS subscale scores. As shown in table 3 all disordered groups (ALC, DE, CM) scored significantly higher than C on the BAS-DRV scale. On the BAS-FS scale, the ALC and CM girls scored higher BAS scores than C. There was a trend for DE girls to score higher on this subscale than C (p = .064). Disordered girls did not differ from each other on either BAS scale. To assess differences in reported sensitivity to punishment between the four groups a one way between-groups ANOVA was performed on BIS scores. BIS scores were significantly affected by group, F(3, 83) = 5.12, p < .01. Tukey’s HSD post hoc tests revealed no significant differences between the disordered groups and C. CM and DE girls, though, scored significantly higher than ALC girls. Overall, disordered girls were distinguished from controls by heightened sensitivity to reward. Further, sensitivity to punishment distinguished girls who reported engaging in dysfunctional eating, with and without alcohol abuse, from alcohol abusing only girls. Table 3. Mean BIS/BAS scores across group n BAS–DRV Control Dysfunctional eating Alcohol misuse Comorbid BAS–FS Control Dysfunctional eating Alcohol misuse Comorbid BIS Control Dysfunctional eating Alcohol misuse Comorbid M SD 28 17 24 18 9.39 11.00* 11.16* 11.44* 1.88 2.00 1.73 2.17 28 17 (p = .064) 24 18 10.11 11.53 12.29* 12.88* 1.79 1.94 1.83 1.81 28 17 24 18 20.79 23.17** 19.79 22.56** 3.97 1.78 3.23 2.38 Note: BAS = Behavioral Approach System; BIS = Behavioral Inhibition System; DRV = Drive; FS = Fun Seeking. *Significantly higher than control (p < .05). **Significantly higher than alcohol misuse (p < .05). Alcohol Abuse and Eating Disorders 461 DISCUSSION The purpose of the current study was to examine whether heightened sensitivity to reward (BAS) and punishment (BIS) may be implicated in dysfunctional eating and alcohol misuse in young women. Alcohol misuse was indeed predicted by BAS reactivity. High scores on the BAS-DRV subscale reflect an individual’s persistence in obtaining desired goals, and high scores on BAS-FS reflect a willingness to seek out and spontaneously approach potentially rewarding experiences. Thus, in this study, it would appear that adolescent girls who engage in high levels of alcohol misuse may have a heightened sensitivity to cues associated with reward. This finding adds to earlier studies in which young alcohol using women have scored higher on measures of impulsivity and sensation seeking than controls (Battaglia et al., 1996; Grau & Ortet, 1999). Moreover, it supports suggestions that individuals prone to engage in problematic drinking may find the pleasurable effects of drinking more rewarding than non-problematic drinkers (Cox, 1987). Furthermore, as hypothesised, scores on both BIS and BAS measures were predictive of dysfunctional eating. The contribution of both BAS, as a measure conceptually linked to impulsivity, and BIS, a measure conceptually linked to anxiety, is in line with other research findings in which bulimic women report both higher impulsivity and trait anxiety than normal controls (Vitousek & Manke, 1994; Battaglia et al., 1996). While the current sample was not differentiated by those girls who engage in anorexic type behaviour and those who engage in bulimic-type behaviour, it is likely that such a heterogeneous sample would tend to have girls who would engage in bulimic rather than anorectic tendencies, as a) the prevalence of bulimia is considerably higher than anorexia and b) the current measure of disordered eating has shown predictive validity in the later development of bulimic symptoms rather than anorexic symptoms (Joiner et al., 1997). Thus, it is not surprising that the results obtained follow the personality patterns evidenced by bulimic rather than anorexic samples in previous research. Therefore, like alcohol misusers, girls with dysfunctional eating may be both predisposed to find the consumption of food more rewarding and more sensitive to negative appraisal of women with body shapes that do not conform to the thin ideal. When further categorized into four groups: DE, ALC, CM & C we found that greater BAS reactivity, particularly in regards to seeking out and approaching novel stimuli, appears to predispose some adolescent girls to engage in disordered eating and drinking behaviour. However, unlike earlier findings (Bulik et al., 1997; Bulik et al., 1994), in which comorbid women scored significantly higher than bulimic women on a measure of trait impulsivity, there were no significant differences between disordered groups on BAS subscales. Overall, it appears that a heightened sensitivity to reward is associated with dysfunctional eating and drinking behaviour, although differences in BAS reactivity may not be enough to determine which specific dysfunctional behaviour will be expressed. Sensitivity to cues of punishment though appears to differentiate DE girls with and without alcohol abuse from ALC only girls, a finding consistent with previous studies in which anorexic and bulimic individuals show higher trait anxiety than substance use disordered women (Battaglia et al., 1996; Vitousek and Manke, 1994). Furthermore, like previous research (Bulik et al., 1997), DE girls did not differ significantly from CM girls in terms of sensitivity to punishment. The difference in BIS scores between CM/DE girls and ALC girls suggests that girls who engage in dysfunctional eating with or without alcohol abuse may be more sensitive to potential punishment than non-eating disordered girls with alcohol use problems. These findings suggest that girls with inherently high sensitivity to cues of both reward 462 Loxton and Dawe and punishment may be more vulnerable to engage in either or both disordered eating and alcohol abuse. The exact nature of this vulnerability and interaction with other psychological and environmental factors though has yet to be explicated. REFERENCES Battaglia, M., Przybeck, T.R., Bellodi, L., & Cloninger, C.R. (1996). Temperament dimensions explain the comorbidity of psychiatric disorders. Comprehensive Psychiatry, 37, 292–298. Brady, K.T., Grice, D.E., Dunstan., L., & Randall. (1993). Gender differences in substance use disorders. American Journal of Psychiatry, 150, 1707–1711. Bulik, C.M., Sullivan, P.F., Carter, F.A., & Joyce, P.R. (1997). Lifetime comorbidity of alcohol dependence in women with bulimia nervosa. Addictive Behaviours, 22, 437–446. Bulik, C.M., Sullivan, P.F., McKee, M., Weltzin, T.E., & Kaye, W.H. (1994). Characteristics of bulimic women with and without alcohol abuse. American Journal of Drug and Alcohol Abuse, 20, 273–283. Bulik, C.M., Sullivan, P.F., Weltzin, T.E., & Kaye, W.H. (1995). Temperament in eating disorders. International Journal of Eating Disorders, 17, 251–261. Carver, S.C., & White, T.L. (1994). Behavioural inhibition, behavioural activation and affective responses to impending reward and punishment: The BIS/BAS scales. Journal of Personality and Social Psychology, 67, 319–333. Cox, W.M. (1987). Personality theory and research. In H.T. Blane & K.E. Leonard (Eds.), Psychological theories of drinking and alcoholism. New York: Guilford Press. Garner, D. (1991). Eating Disorders Inventory-2. Professional manual (pp. 55–89). Odessa, FL: Psychological Assessment Resources. Grau, E., & Ortet, G. (1999). Personality traits and alcohol consumption in a sample of non-alcoholic women. Personality and Individual Differences, 27, 1057–1066. Gray, J.A. (1970). The psychophysiological basis of introversion-extraversion. Behaviour Research and Therapy, 8, 249–266. Gray, J.A. (1982).The neuropsychology of anxiety: An enquiry into the functions of the septo-hippocampal system. Oxford: Clarendon Press. Holderness, C.C., Brooks-Gunn, J., & Warren, M.P. (1994). Comorbidity of eating disorders and substance abuse: Review of the literature. International Journal of Eating Disorders, 16, 1–34. Joiner, T.E., Heatherton, T.F., & Keel, P.K. (1997). Ten-year stability and predictive validity of five bulimiarelated indicators. American Journal of Psychiatry, 154, 1133–1138. Jorm, A.F., Christensen, H., Henderson, A.S., Jacomb, P.A., Korten, A.E., & Rodgers, B. (1999). Using the BIS/BAS scales to measure behavioural inhibition and behavioural activation: Factor structure, validity, and norms in a large community sample. Personality and Individual Differences, 26, 49–58. Krahn, D.D. (1991). The relationship of eating disorders and substance abuse. Journal of Substance Abuse, 3, 239–253. Krahn, D.D, Kurth, C., Demitrack, M., & Drewnowski, A. (1992). The relationship of dieting severity and bulimic behaviors to alcohol and other drug use in young women. Journal of Substance Abuse, 4, 341–353. Lundholm, J.K. (1989). Alcohol use among university females: Relationship to eating disordered behaviour. Addictive Behaviours, 14, 181–185. Saunders, J.B., Aasland, O.G., Babor, T.F., de le Fuente, J.R., & Grant, M. (1993). Development of the alcohol use disorders identification test (AUDIT). WHO collaborative project on early detection of persons with harmful alcohol consumption-II. Addiction, 88, 791–804. Shore, R.A., & Porter, J.E. (1990). Normative and reliability data for 11 to 18 year olds on the Eating Disorders Inventory. International Journal of Eating Disorders, 9, 201–207. Strober, M., Salkin, B., Burroughs, J., & Morrell, W. (1982). Validity of the bulimia-restrictor distinction and family psychiatric morbidity. Journal of Nervous and Mental Disease, 170, 345–351. Vitousek, K., & Manke, F. (1994). Personality variables and disorders in anorexia nervosa and bulimia nervosa. Journal of Abnormal Psychology, 103, 137–147. Wonderlich, S.A., & Mitchell, J.E. (1997). Eating disorders and comorbidity: Empirical, conceptual, and clinical implications. Psychopharmacology Bulletin, 33, 381–390.