The study of information processing can advance our understanding of psychiatric disorders (Keefe, 1995). Neuropsychological data may be used to predict prognosis or course of illness and can also be integrated into aetiological and treatment models (Southgate et al., 2005, Treasure et al., 2005, Cavedini et al., 2006 and Cavedini et al., 2006).
Review of the neuropsychological literature in Eating Disorders (ED) highlights that information processing styles mirror phenotypic characteristics (Southgate et al., 2006). Links are clearly seen between obsessive-compulsive personality traits (preoccupation with details, lists and order; perfectionism; inflexible, rigid and stubborn mental set), significant vulnerability factors in the complex aetiology of EDs (Jacobi et al., 2004), and deficits in set-shifting ability or ‘cognitive flexibility’ in anorexia nervosa (AN) (Holliday et al., 2005, Tchanturia et al., 2005 and Roberts et al., 2007). Heightened impulsivity, as reported in self-report personality measures and displayed in multi-impulsive behaviours, has been associated with binge/purge symptoms, with or without a diagnosis of AN (Lacey and Evans, 1986, Vitousek and Manke, 1994, Claes et al., 2002 and Fassino et al., 2004). Two studies employing the matching familiar figures test (MFFT) as a measure of information processing style have reported analogous findings (Toner et al., 1987 and Kaye et al., 1995).
Discrepancy in the performance seen across tests of visuo-spatial processing is suggestive of another information processing bias in AN that corresponds to clinical characteristics. Superior performance has been reported in tasks that require local information processing (e.g. embedded figures test) in comparison to those that require global information processing (e.g. object assembly, Rey-figure, Bender Gestalt test) (Jones et al., 1991, Gillberg et al., 1996, Kingston et al., 1996, Sherman et al., 2006, Tokley and Kemps, in press and Gillberg et al., 2007). This pattern of performance is consistent with the concept of weak central coherence, a cognitive style in which information remains fragmented as opposed to integrated, with processing occurring at the level of ‘detail’ as opposed to ‘whole’(Happe and Frith, 2006). The hypothesis that weak central coherence reflects a preferential processing style in AN, concords with the extensive attention to detail found to be an enduring personality trait amongst individuals with AN (Anderluh et al., 2003). Thus cognitive style may account for both optimal and impaired performance across different tasks, depending on their processing requirements.
The aim of the current study was to expand upon previous research. Following a hypothesis driven approach, the MFFT was employed to explore ‘the preferential mode of information processing a subject manifests in the majority of reasoning, learning and memorising experiences during an assigned task’ (Weijers et al., 2001). The MFFT is a difficult matching to sample task that involves high response uncertainty. Participants are required to identify a target object amongst seven very similar distracter objects. Outcome measures include the accuracy of the participant and the time taken to make their response. The use of a dimensional scoring procedure, as opposed to the traditional categorical approach, contributes novel information with regards to performance in this task in ED samples. By distinguishing between the skills/ability required (content) and the way in which participants go about solving this task (process), the dimensional scoring procedure can provide information with regards to the phenotypic characteristics discussed above; reflection–impulsivity, and local–global biases.
Reflection–impulsivity reflects the process employed in problem solving; “the tendency to gather and evaluate information before making a decision” (p. 515; Clark et al., 2006). As performance in the MFFT involves high response uncertainty (a prerequisite for the accurate measurement of reflection–impulsivity; Kagan, 1966), the trade-off between speed and accuracy in responding is taken into consideration in this dimensional score. Impulsivity relates to fast and inaccurate responding, (decisions are made and acted upon with little consideration with regards to the accuracy). Conversely, at the other end of the continuum, a reflective ‘perfectionist’ cognitive style is implicated by slow and accurate performance.
With regards to content, as a difficult matching to sample task with high demands on visual search skills (Clark et al., 2006), it is clear that performance would benefit from local processing, a focus on the detail of the figures as opposed to processing the stimuli at the level of the gestalt. Local processing bias is therefore measured here using a dimensional efficiency score (efficiency associated with fast and accurate responses).
The hypotheses for this investigation were determined based upon the current knowledge regarding clinical phenotypes in EDs. In comparison to the control group, it was thought that participants with EDs would display information processing styles indicative of elevated impulsivity and efficiency. Firstly, participants with BN would be associated with elevated impulsivity, in accordance with the finding of previous research. Secondly participants with AN would show greater efficiency in the task, in line with the literature that suggests a bias towards detail level processing in this clinical group.