The longitudinal course of psychiatric illness and response to medication in people with PWS is fully described. Further research is needed regarding the effect of psychotropic medications, particularly mood-stabilizing medication. These data will enable informed decisions to be made regarding management options and provide information on the possible long-term outcome of illness.

Psychiatric illness in PWS is predominately affective with atypical features. The prevalence and possibly the severity of illness are greater in those with mUPD. We present a 'two-hit' hypothesis, involving imprinted genes on chromosome 15, for the development of affective psychosis in people with PWS, regardless of genetic subtype.

Objective: To investigate the neural basis of the abnormal eating behaviour in Prader-Willi syndrome (PWS), using brain imaging. We predicted that the satiety response in those with PWS would be delayed and insensitive to food intake. Design and participants: The design of this study was based on a previous investigation of the neural activation associated with conditions of fasting and food intake in a nonobese, non-PWS group. The findings were used to generate specific hypotheses regarding brain regions of interest for the current study, in which 13 adults with PWS took part (mean7s.d. age ¼ 2976; BMI ¼ 31.575.1; IQ ¼ 7178, six were female). Measurements: Regional cerebral blood flow was measured using positron emission tomography in three sessions: one following an overnight fast and two following disguised energy controlled meals of similar volume and appearance -one of 1674 kJ (400 kcal) and another of 5021 kJ (1200 kcal). Subjective ratings of hunger, fullness and desire to eat, and blood plasma levels of glucose, insulin, leptin, ghrelin and PYY were measured before and after each imaging session. Results: The neural representation of hunger, after an overnight fast, was similar to that found in nonobese individuals in the control study. In contrast, after food intake, the patterns of neural activation previously associated with satiety were not found, even after the higher-energy load. Lateral and medial orbitofrontal cortical activation was associated with consumption of the 400-and 1200-kcal meals, respectively. The medial orbitofrontal activation, however, was only found in those who had shown a large percentage change in fullness ratings following the higher-energy meal. Conclusion: We conclude that there is a dysfunction in the satiety system in those with PWS. These findings suggest that brain regions associated with satiety are insensitive even to high-energy food intake in those with the syndrome. This may be the neural basis of the hyperphagia seen in PWS.

It would appear that incentive motivation alone plays a less powerful role in individuals with PWS than in those without the syndrome. This is likely to be due to the overriding intrinsic drive to eat because of a lack of satiety in those with PWS, and the impact of this on activity in the incentive processing regions of the brain. Activity in such reward areas may not then function to guide behaviour selectively towards the consumption of high preference foods.

The two main causes of Prader-Willi syndrome (PWS) are a paternally derived deletion in the maternally imprinted 15q11-q13 region or UPD(15)mat. Both mechanisms result in a loss of the active paternal contribution to the region. The affective psychosis associated with PWS has been found to be mainly confined to the propositi with UPD(15)mat rather than to those with a deletion. This suggests that the psychosis may be related to the presence of two copies rather than a single copy of a gene or genes located in the distal half of the region which is paternally imprinted, but maternally active, and whose loss results in Angelman syndrome (AS). A large population-based study of PWS allowed the identification of 12 people with a 15q11-q13 deletion who had suffered psychotic episodes and four adults with UPD(15)mat who so far had not. When these people were investigated using microsatellite markers, the 12 with a deletion were found to have two maternally derived copies of a narrow region between D15S975 and D15S661 making them effectively disomic for these loci. Thus all of the people with psychosis had two active copies of any imprinted genes in the region while all non-psychotic people (including controls) had only one. Quantitative RT-PCR studies suggest that a lack of expression of FLJ33332, either as a result of or resulting in gene dysregulation, may be associated with psychosis in PWS.