Eating disorders (ED) are important psychosomatic illnesses affecting young females in a critical phase of psychosexual development (Fairburn and Harrison, 2003). Core pathologic features of anorexia nervosa (AN) consist of drive for thinness, disturbed eating behaviour, malnutrition and hormonal dysregulation. There is an ongoing debate on subtyping the disorder, as patients are heterogeneous in terms of personality characteristics, comorbid symptoms and long-term outcome (Ward et al., 2003), the most common distinction comprising restrictive (AN-R) and binge-eating/purging (AN-B/P) subtypes.
Structural cerebral studies demonstrated a decrease of cerebral grey matter in AN, including the cingulate cortex (Mühlau et al., 2007, Castro-Fornieles et al., 2009 and Joos et al., 2010). Frontolimbic resting perfusion is reduced in AN, the anterior cingulate cortex (ACC) being particularly affected in AN-R (Naruo et al., 2001 and Takano et al., 2001). A first functional magnetic resonance imaging (fMRI) study demonstrated increased blood oxygen level dependent (BOLD) responses of the left amygdala, insula and ACC in a sample of six AN patients using pictures of high caloric drinks as a paradigm (Ellison et al., 1998). The AN subtype was not reported. A further fMRI study of the same research group used visual food stimuli and investigated a heterogeneous AN patient sample, i.e. AN-R and AN-B/P patients (Uher et al., 2004). Increased ACC and orbitofrontal cortex (OFC) BOLD responses were reported (Uher et al., 2004). A limitation of the study was that one-third of patients were on psychotropic medication. A further study of an adolescent AN-R sample did not show increased frontolimbic perfusion, however (Santel et al., 2006). Further fMRI studies of AN had used body image stimuli — paradigms and methodologies varying greatly (Seeger et al., 2002, Wagner et al., 2003, Uher et al., 2005, Sachdev et al., 2008, Fladung et al., 2009, Miyake et al., 2010a and Vocks et al., 2010a), the salience of body image paradigms being a critical aspect (Uher et al., 2005).
The present study tried to further clarify frontolimbic dysfunction in AN, as there had only been two studies in adults using a disease-specific nutritional paradigm and one in adolescent patients, and by avoiding methodogical problems of former investigations: A clearly defined sample of AN-R patients was studied, in order to avoid heterogeneous samples, which differ with regard to psychopathological features, and possibly with regard to neural reactivity. A paradigm that had previously been applied in ED was used in order to be able to compare findings with earlier results (Uher et al., 2003 and Uher et al., 2004). This aspect seems of importance, as many imaging studies investigate small numbers — and even in healthy controls there is variability of neural responses to food stimuli (Beaver et al., 2006). Stimuli variability was increased compared to previous studies (Uher et al., 2003 and Uher et al., 2004) in order to minimize habituation effects, as amygdala responsiveness in particular shows habituation when the same emotional stimuli are repeated, as shown in healthy subjects (LaBar et al., 2001, Phillips et al., 2003 and Ball et al., 2007). The instruction to the presentation of stimuli was held simple in order to reduce any cognitive interference with the emotional experience, as the latter might be influenced as well as neural activation patterns (Townsend et al., 2010). Similar to the study of Uher et al. (2004), we used a block design, which might be more effective with regard to amygdala response, compared to an event-related design (Sergerie et al., 2008).
