Schizophrenia and bipolar disorder are common, severe, and disabling psychotic disorders, which are difficult to research. We argue that the genetically determined neurodevelopmental disorder Prader-Willi syndrome (PWS), which is associated with a high risk of affective psychotic illness, can provide a window into genetic mechanisms and associated neural pathways. People with PWS can all show non-psychotic psychopathology and problem behaviours, but the prevalence of psychotic illness differs markedly by genetic subtype; people with PWS due to chromosome 15 maternal uniparental disomy have higher prevalence of psychotic illness compared with patients with PWS due to 15q11-13 deletions of paternal origin. On the basis of this observation and the neural differences between genetic subtypes, we hypothesise that the combined effects of the absent expression of specific maternally imprinted genes at 15q11-13, and excess maternally imprinted or paternally expressed genes on chromosome 15, affect the γ-aminobutyric acid-glutamatergic pathways and associated neural networks that underpin mood regulation and sensory processing, resulting in psychotic illness. We propose a model of potential mechanisms of psychosis in PWS, which might be relevant in the general population, and should inform future research.