Bipolar affective disorder and schizophrenia are highly heritable psychiatric illnesses and the leading causes of worldwide disability. The orbitofrontal cortex (OFC) is a region of the frontal lobe with wide spread connectivity with other brain areas involved in reward, motivation and emotion. Evidence from various neuroimaging, genetic, post-mortem and brain lesion studies suggest that orbitofrontal cortex may play a role in pathophysiology of mental illnesses. This thesis sought to investigate the pathogenesis of major psychiatric illnesses through the investigation of orbitofrontal morphology in schizophrenia and bipolar disorder and through its associations with brain structure and function. Orbitofrontal morphology and its structural and functional associations were examined in healthy controls, patients with schizophrenia or bipolar affective disorder, and those at high genetic risk using functional and structural MRI. In the first study we found that the orbitofrontal type III is more frequent and the orbitofrontal type I is less common in the right hemisphere in patients with schizophrenia while in patients with bipolar disorder type III appears more often in both left and right hemispheres. We then sought to examine the relationship of orbitofrontal morphology to disease risk in a study of 146 people at high risk of developing schizophrenia and 110 people at high risk of developing bipolar disorder. We discovered that in the unaffected high risk groups the orbitofrontal type III predicted the development of later psychiatric illnesses, when combined with anterior cingulate morphology. Finally we showed, in a further study, that OFC morphology was associated with measures of schizotypy, brain structure, brain function and cognition. In conclusion, orbitofrontal morphology is linked to major psychiatric disorder and has significant structural and functional associations. As orbitofrontal sulcogyral patterns are formed in early life a fuller awareness of their relevance to brain function holds out the prospect that we could use such measures as an indicator of vulnerability to the development of illness later in life. This work points to the potential for the foundation of a theory of predictive associations between morphological patterns and the development of psychosis.