The study investigated the presence and characteristics of spontaneous imagery in women with bulimia nervosa (BN) and their links to childhood memories. Using a semi-structured interview, data was collected from three groups of female participants: BN participants (N=13), dieting (N=18) and non-dieting controls (N=20). BN participants reported more spontaneous images than non-dieting control participants. Their images were recurrent and significantly more negative and anxiety provoking than those of controls. They involved more sensory modalities than in dieting controls and were more vivid than in non-dieting controls. BN images typically involved the visual, organic and cutaneous modalities. They were linked to a specific childhood memory, similar in emotional tone and sensory modalities. Once depression was controlled, many of the between-group differences became non-significant. The results suggest that imagery may be a significant feature of BN, potentially distinguishing those with BN from controls, although further research into the link between mood, imagery and memory is needed. The findings have clinical implications, particularly for assessment and for the application of imagery rescripting in BN.
Cognitive theories of bulimia nervosa (BN) have focused primarily on the role of cognitions in the disorder. Beck (1976) however, advised that images, as well as verbal cognitions, may usefully provide access to the appraisals that people make and that these appraisals may need challenging for cognitive therapy to be successful. In BN relevant appraisals include underlying assumptions1 about eating, weight and shape and negative self or core beliefs2 (Cooper, 2005; Waller, Kennerley, & Ohanian, 2007).
It has been suggested that images may reveal idiosyncratic meanings not necessarily accessed through verbal cognitions (Hackmann, Surawy, & Clark, 1998), and provide a relatively quick route to deeper levels of meaning compared to verbal cognitions (Barnard & Teasdale, 1991). Treatment could benefit greatly from procedures that allow fast assessment and identification of relevant core beliefs (including that in BN). It might also usefully focus specifically on imagery modification or imagery rescripting if images are found to be important in maintaining the disorder.
Several studies have investigated imagery in anxiety disorders. These have highlighted imagery as a maintaining factor and as a means of rapid access to underlying assumptions and core beliefs (Holmes & Hackmann, 2004). Studies have found that images: (a) often reflect the idiosyncratic fears of the individual (Ottaviani & Beck, 1987; Wells & Hackmann, 1993); (b) tend to be recurrent, i.e. stable over time (Hackmann, Clark, & McManus, 2000); (c) often possess sensory qualities other than the visual, such as organic (internal body) and cutaneous (skin) modalities (Hackmann et al., 2000; Pratt, Cooper, & Hackmann, 2004); and (d) appear to include input from early memories, where a match is found in terms of their sensory and interpersonal content (for example, see Hackmann et al., 2000; Pratt et al., 2004). The latter might be expected given that early memories are commonly experienced retrospectively nonverbally (Pillemer, 1998) and that many early experiences occur at an age when material is more likely to be encoded in affective, visual or kinaesthetic forms rather than linguistically (Ohanian, 2002).
Of those studies that have employed appropriate control groups, some have found between-group differences in terms of the content and characteristics of images (Hackmann et al., 1998; Pratt et al., 2004). Others have found that it is not the presence of imagery that is crucial but various other characteristics. For example, a study in patients with body dysmorphic disorder (BDD) (Osman, Cooper, Hackmann, & Veale, 2004) found that patient and control groups only differed in degree of negative emotion associated with their images, rather than whether or not spontaneous images were experienced. Given that BDD and BN are both body–image related disorders, it is possible that the same may be true of BN. Equally, as many young women experience weight and shape concerns, and diet, it is also possible that these are normative experiences.
Little is known about imagery in BN, despite the fact that BN is often associated with considerable anxiety in relation to food, eating, weight and shape and in social situations (Hinrichsen, Wright, Waller, & Meyer, 2003). Only one study has investigated imagery in people with eating disorders (Cooper, Todd, & Wells, 1998). In this study, 12 women with BN and 12 women with anorexia nervosa were interviewed. Thirteen participants reported images, although these were not enquired about specifically. In nearly every case the images described how the patient felt she would look if she were to gain weight or become fat. Two studies report successful use of imagery rescripting in eating disorders including a case study in BN (Ohanian, 2002). A second study found that imagery modification resulted in significant changes in belief ratings for core beliefs compared to a control intervention (Cooper, Todd & Turner, 2007).
Given its potential importance, examination of imagery in BN warrants further investigation. The current study used a semi-structured interview to investigate the presence, content and characteristics of spontaneous imagery and links to childhood memories in women with BN and control participants. Dieting and non-dieting control groups were used. The hypotheses were as follows:
(1) The characteristics of spontaneous images will have a good test–retest reliability, a good inter-rater reliability and a good concurrent validity. Compared to dieting and non-dieting controls, women with BN will report (2) more negative and recurrent spontaneously occurring images when asked about their eating, shape or weight, (3) spontaneous images in a greater number of sensory modalities, experienced with greater intensity, and (4) more links between spontaneous images and childhood memories.
