Distinct Functional and Structural Cerebral Abnormalities in Eating Disorders in the Light of Diagnostic Classification Systems

prised the cingulate cortex as well as the parietal, temporal, insular, supplementary motor and subcortical (caudate nucleus) regions [7] . A photon emission computed tomography study reported almost reverse perfusion patterns of AN and BN when subjects were compared in the resting condition and after food intake [8] . Furthermore, using magnetic resonance spectroscopy of the anterior cingulate cortex, BN patients showed a positive correlation of glutamate with ‘drive for thinness’, whereas restrictive AN subjects did not [9] . With respect to ventromedial dysfunction (including the anterior cingulate cortex), it should however be kept With DSM-5 advancing there is a debate with respect to the classification of eating disorders (EDs). In principle, the categorical approach of DSM-IV will be kept, with anorexia nervosa (AN) and bulimia nervosa (BN) representing the main ED entities. However, there are also rationales to conceptualize ED as a continuum: Most ED patients – independent of diagnostic category – are characterized by an overevaluation of shape and weight, a strong dissatisfaction with their body and an engagement in some form of weight control behavior. They often switch diagnostic categories over time, and there is genetic cross-transmission. Furthermore, about half of ED cases do not fulfill all criteria of AN or BN. Here we raise the question whether the increasing number of neurobiological studies, in particular imaging data, might contribute to this debate. One of the first functional magnetic resonance imaging (fMRI) studies was suggestive of functional cerebral substrates common to various EDs, demonstrating increased reactivity of the medial prefrontal cortex [1] . Recent fMRI studies have shown different results. One study that used words concerning body image demonstrated an increased response of the right amygdala in AN, while ventromedial abnormalities characterized BN [2] . Investigations using visual body images yielded amygdalar activation in AN, but not in BN [3, 4] . Our group used a food paradigm, which demonstrated an increased response of the right amygdala and decreased signals of the midcingulate cortex in restrictive AN [5] , while BN patients had decreased frontocingular and temporal signals [6] . Contrasting these samples directly, BN was characterized by decreased frontocingular activation, as shown in figure 1 . A further recently published fMRI investigation used the active instruction to imagine eating the food shown in the images and compared AN with BN patients [7] . Again, differences in blood oxygen level-dependent signals were demonstrated, which comReceived: November 10, 2011 Accepted after revision: March 2, 2012 Published online: September 9, 2012