Neuroanatomical Bases of Eating Disorders
Rudolf Uher, Janet Treasure and Iain C. Campbell
There are several lines of evidence that may eventually converge to provide an integrated model of brain dysfunction in eating disorders (ED). In this section we have combined current knowledge and have constructed a neuroanatomically based explanation of ED. The picture is far from clear; and consequently, much of what is written contains suggestions and hypotheses rather than established explanations.
Diagnostic categories of ED span a spectrum of heterogeneous conditions, ranging from anorexia to obesity, in which maladaptive eating habits constitute a major symptom and which are thought to arise from primarily psychological disturbance. The inclusion of simple obesity in ED is controversial (Bruch, 1973; Treasure and Collier, 2001). Restrictive anorexia nervosa (RAN), one of the extreme conditions at the end of the ED spectrum, constitutes probably the most clear-cut syndrome, and is therefore a useful model for research. Much more knowledge has been collected for anorexia nervosa (AN) than for other ED. Therefore, this chapter will be substantially based on AN research and bulimia nervosa (BN) will be discussed to a lesser extent.
In the second part of this paper (Symptoms and Physiology) four groups of ED symptoms and their neural correlates are examined (eating and hunger/satiety perception, body image, pain perception and cognitive functions). A third part (Direct evidence of neural disturbance in ED) summarizes evidence from lesions, electrophysiological and neuroimaging studies. Finally, the fourth part provides a conclusion.
Given the heterogeneity of symptoms and syndromes within the spectrum of ED (Treasure and Collier, 2001), it is unlikely that a particular neuroanatomically defined disturbance will match on to DSM IV or ICD10 diagnostic categories, rather it may be helpful at this stage to examine brain function in relation to individual symptoms. However, a question may be raised: can these diverse symptoms be considered as separate disturbances or do they (all or some of them) represent several facets of a common psychopathological mechanism? For example, the apparent insensitivity of AN patients to hunger, pain and physical fatigue could be attributed to a general impairment or inhibition of signalling coming from the body. Moreover, as body image representation is dependent on interoceptive and exteroceptive sensory information, its distortion in ED may also be caused by primary or secondary sensory insensitivity (Smeets and Kosslyn, 2001). The concept of a 'proto-self (Damasio, 1999) as a representation of bodily awareness that constitutes a basic level of consciousness can be considered as a model which links these various symptoms. The proposed neuroanatomical substrate for this representation of body physical status is comprised of brainstem nuclei, hypothalamus, basal forebrain, insular, secondary sensory and medial parietal cortices.
Four groups of ED symptoms and their neural correlates are examined in this section: eating and hunger/satiety, body image, pain perception and cognitive functions. Although not clinically apparent, the insensitivity to pain is dealt with in detail, because considerable research has been carried out on this topic and it may provide a useful link in the pathophysiology of ED. Other constituents of ED symptomatology: affective and emotional disturbance, impulsivity and obsessive-compulsive symptoms are shared with other psychiatric disorders (affective disorders, personality disorders, obsessive-compulsive disorder) and their description and neural mechanisms can be found in other sections of this book.
Deleterious eating behaviour, the most obvious manifestation of ED, is associated with an abnormal perception of hunger and satiety. People with AN seem to lack the feeling of hunger, whereas those with BN have diminished feelings of satiety and still feel hungry after having eaten a large meal (Owen et al., 1985; Halmi, 1988; Halmi and Sunday, 1991; Hetherington and Rolls, 1991, 2001). The complementarity between hunger and satiety states seems to disappear in ED and people with AN of the bingeing-purging subtype (BPAN) may report feeling both full and hungry at the same time (Halmi and Sunday, 1991). In response to provocation by insulin or 2-deoxy-D-glucose the subjective rating of hunger by AN patients paradoxically decreases (Nakai and Koh, 2001). Despite this subjective absence of hunger, when AN patients were shown palatable food, they responded by increasing insulin secretion to an even greater extent than did healthy controls, but then chose not to eat the food (Broberg and Bernstein, 1989). This suggests that although their body is preparing for meal ingestion, people with AN do not follow this with appropriate behaviour (eating the food). In summary, hunger/satiety sensing is apparently disrupted in people with eating disorders and the subjective experience and behaviour are disconnected from the body's needs and autonomic functions. Several important questions remain unresolved: Do patients with AN not feel hunger or do they just not admit feeling it? Is this disturbance in hunger/satiety sensing primary (and possibly causative) or is it acquired during the course of the illness? Hetherington and Rolls (2001) in a recent review have argued in favour of a secondary, learned disturbance in ED.
Neural mechanisms involved in the response to food and its modulation by hunger and satiation have been extensively investigated with single neuron activity mapping in primates (Rolls, 1999). It has been established that taste, olfactory and visual pathways which process food-related stimuli converge to the amygdala and the caudal orbitofrontal cortex, where their