Cognitive dysfunction such as body-image disturbance and undue influence of body weight on self-worth is a conspicuous feature of eating disorders. The cognitive problems are known to be extremely difficult to treat. Why and how, therefore, is cognitive behavioral therapy (CBT) recommended, with high quality evidence, in clinical guidelines such as the NICE guidelines in the UK?
 In reverse direction to the history of eating disorders, namely anorexia nervosa first and then bulimia, CBT was developed for bulimics first and then after its establishment, the skills were applied to anorexia nervosa. Anorexia treatment whether behavioral or familial, has tended to place patients in a passive mode. The CBT technique, on the other hand, invites patients to participate fully in the treatment, via formulation-making and symptom self-monitoring. This is particularly important because, unlike in the early days of adolescent anorexia 'epidemic', the number of adult patients has increased. Behavioral and family treatment is less applicable to adult patients who are expected to be more independent than early adolescent anorexics.
 CBT for bulimics consists of two parts. The first part, the normalization of eating pattern, is largely behavioral. In the enhanced CBT (CBT-E) by Fairburn, a standard CBT in the field of eating disorders research, patients are obliged to make two outpatient visits a week for the first four weeks in order to install a regular eating pattern. The cognitive work is added later on the basis that the patient has successfully achieved a regular meal schedule. This behavioral change through two sessions a week may be difficult in a Japanese clinical setting. Some modification such as a brief in-patient treatment may be considered. Also, the number of CBT therapists in Japan is lacking. Collaboration with clinical psychologists is necessary.
 The CBT for anorexia is a challenge. Fairburn has expanded the application of CBT to anorexia via his 'transdiagnostic' approach. Likewise, Pike et al started to use CBT-AN for relapse prevention for the patients who acquired sufficient weight through inpatient treatment. The research data is promising. In particular, Touyz et al show that CBT-AN had effects on severe and enduring AN (SE-AN), a category of AN which is often thought to be resistant to any type of treatment.
 It is of note that for both anorexics and bulimics, the effect of 'behavioral only' techniques expires early. By contrast, treatments which deal with psychological elements such as CBT and interpersonal psychotherapy (IPT) have a lasting effect. The time courses of CBT and IPT treatment effect seem sufficiently different that the matching of patient characteristics and the type of treatment should be investigated further. Another important aspect of cognitive dysfunction among eating disorder patients is 'denial of illness'. More research should be performed with regard to how patients, on improvement from eating disorders, look back on aspects of denial and whether a better understanding of these phenomena is helpful in relapse prevention.
 <Author's abstract>