پردازش های انتخابی کلمات مربوط به وزن و شکل در بولیمیا: استفاده از آزمون استروپ کامپیوتری

A computerised Stroop colour-naming task was used to measure concerns about weight, shape, and eating in bulimia nervosa. Two versions of the computerised Stroop were compared, a voice-activated and a button-pressing-activated programme. Bulimia nervosa patients were significantly slower in colour naming shape- and weight-related words than their female age-matched controls. The button-pressing computerised Stroop was both more sensitive and more accurate at measuring colour-naming speeds than the voice-activated version. When the bulimia nervosa group were divided according to their Eating Attitudes Test (EAT) scores, those who showed extreme pathological attitudes to weight and shape were significantly slower in colour naming size words and in food disruption scores than those with a lesser degree of psychopathology. The computerised Stroop might be useful as a diagnostic tool and in the assessment of the effectiveness of therapy for the individual patient.

Dysfunctional attitudes to weight, shape, and eating play a central role in the psychopathology of anorexia and bulimia nervosa Fairburn et al., 1986 and Garner & Bemis, 1982. Patients see weight as being extremely important, equate their personal value with the shape of their bodies, and devote much of their time and energy to controlling what they eat (Vitousek & Hollon, 1990). Measurement of these concerns has traditionally relied on self-report questionnaires such as the Eating Attitudes Test (EAT) and Body Shape Questionnaire (BSQ), which have been able to pinpoint those aspects of shape and eating that are of particular concern for an individual. More recently, methods from cognitive science have been employed to measure not so much the attitudes that an individual has about weight, shape, and eating, but the mechanisms that serve to perpetuate those attitudes. Eating-disordered patients are thought to show selective attention to disorder-salient information, such that they are more likely to perceive and to remember information concerning body size and eating King et al., 1991, Schotte et al., 1990 and Vitousek & Hollon, 1990. This selective attention has been measured with a modified version of the Stroop colour-naming task (Stroop, 1935). Patients with anorexia and bulimia nervosa demonstrate a colour-naming interference of food- and shape-related words, which may be indicative of a ‘selective processing’ of information relevant to this domain Ben-Tovim et al., 1989, Channon et al., 1988, Fairburn et al., 1991, Jones Chesters et al., 1998 and Sackville et al., 1998. These results have led some authors to suggest that the Stroop test may be an objective, reliable, and specific measure of eating-disordered attitudes and thus used as a method of assessing the effectiveness of treatment or therapy Ben-Tovim & Walker, 1991, Fairburn et al., 1991 and Walker et al., 1992. However, it has also been shown that certain non-patient groups of individuals share in this colour-naming interference of words related to food and/or shape: those who are in a temporary state of food deprivation (Channon & Hayward, 1990), dieters and nondieters who have been given a high-calorie preload before testing Mahamedi & Heatherton, 1993 and Ogden & Greville, 1993, and restrained eaters Green & Rogers, 1993, Huon & Brown, 1996, Long et al., 1994 and Perpina et al., 1993. If normal individuals under particular circumstances can show a selective processing of food and shape words as demonstrated by the Stroop test, its specificity in the assessment of eating-disordered psychopathology is threatened. The content of experimental instructions is extremely important in determining the results that the modified Stroop test will produce. Mahamedi and Heatherton (1993) informed their subjects that ‘the experiment was designed to investigate the effects of personality on perceptual vividness’; Perpina et al. (1993) asked subjects to ‘help with a research project investigating perceptual abilities.’ In providing a cover story, these studies attempt to disguise the experimental hypothesis and to eliminate demand characteristics. Most of the other studies to date do not discuss the way in which the experiment was framed for its subjects. However, it is clear that asking subjects to complete questionnaires relating to eating disorders immediately prior to the Stroop task Cooper et al., 1992 and Cooper & Fairburn, 1992, or indeed asking subjects to fast (Channon & Hayward, 1990) or to drink high-calorie preloads before their tests Mahamedi & Heatherton, 1993, Ogden & Greville, 1993 and Overduin et al., 1995 may alert them to the purpose of the experiment and therefore alter their responses. In card-based Stroop tests, the simultaneous presentation of a card containing 100 words or so about eating and weight may also provide clues about the experimental hypothesis to the subject. In addition, it may lead to the build-up of significant anxiety in eating-disordered patients that will globally affect cognitive skills Cooper, 1995, Lovell et al., 1997 and Williams et al., 1988. The colour-naming interference observed would reflect concern about issues associated with weight and eating, but would not be indicative of a selective processing of these words. Another difficulty with card-based variants of the modified Stroop is that the experimenter measures colour-naming speeds with a stopwatch. As the experimenter is not blind to the identity of the subject, or to her placement in either the patient or control group, he is able to consciously or unconsciously alter the results of the study in the direction of the hypothesis. Several studies have attempted to overcome these difficulties by using a computerised version of the modified Stroop test Black et al., 1997, Green et al., 1994, Jones Chesters et al., 1998, Overduin et al., 1995 and Sackville et al., 1998. Experimenter expectancy effects are eliminated because the computer records the colour-naming speeds. Furthermore, the computerised Stroop measures colour-naming speeds of individual words. This allows words from the various categories to be presented to the subject in a random sequence, thus counterbalancing the presentation order of the different test and control categories. As the test words are mixed with control words, the demand characteristics of the experiment are disguised and the patient is more likely to be blind to the experimental hypothesis. Despite this obvious advantage, only two authors reported the use of a computerised Stroop programme that generated a random sequence of word presentation Jones Chesters et al., 1998 and Overduin et al., 1995. All other computerised Stroop studies presented words from each category in a ‘blocked’ fashion. This means that two major problems of experimental design, the blatant demand characteristics of the experiment, and the influence of anxiety in the successful completion of the task, are not eliminated. Both studies where words appeared on the computer screen in a random order used a voice-activated computer programme, rather than a programme that measured colour-naming speeds via a four-choice button press box Jones Chesters et al., 1998 and Overduin et al., 1995. As voice-activated programmes can be triggered by any sound, not just the correct colour spoken aloud, this approach may be less accurate than a button-pressing-activated programme, where one of the four colour choices must be made and mistakes can be accurately measured. The primary aim of this study is therefore to use a computerised Stroop approach, which encompasses a number of methodological refinements over previous studies, to determine whether patients with bulimia nervosa exhibit selective processing of food- and shape-related words. The second aim of the study is to compare the sensitivity and reliability of voice-activated and button-pressing-activated versions of the computer programme.