Trajectory of Brain Structure and Function before and after the Onset of Psychosis: a Longitudinal Multicentre Study

Psychotic disorders such as schizophrenia are disabling and relatively common conditions. They usually begin in early adulthood, and are chronic illnesses. Treatment can partially reduce some of the symptoms, but is not curative. These disorders are preceded by an 'at risk mental state' (ARMS). People in this state have a 30% risk of developing psychosis. However, at present, it is not possible to predict which individual with an ARMS will go on to develop psychosis and which will not. The aim of my study is to test whether measuring the structure and function of the brain with Magnetic Resonance Imaging (MRI) scanning in people with an ARMS can help clinicians to predict who will later become ill. In order to do this, I will put together a large number of scans collected from several different research centres in Europe and Australia. People with an ARMS who take part in the study will be scanned when they first contact support clinics, and then again after 3 months, 12 months and 24 months. I will use the information in the scans to test whether the way the brain changes over time is different in people who will later develop psychosis compared to people who do not. There is evidence to suggest that specific regions of the brain called the hippocampus, insula and lateral ventricles change in volume close to the time-point that a psychotic illness starts. However it is not known exactly when these changes occur and if they happen before or after the beginning of psychosis. From the MRI scans I will be able to discover exactly when these changes occur. One popular theory of schizophrenia proposes that a region of the brain called the hippocampus becomes abnormally active and then causes problems in regulating another region of the brain called the striatum which is important in symptoms of psychosis. I will test this theory by using the MRI scans to measure the activity of the hippocampus compared to activity in the striatum.

The aim of this study is to characterise the trajectory of brain structure and function before and after the onset of psychosis. Specifically I will examine changes in the volume of the insula, hippocampus and lateral ventricles and test the hypothesis associated with the theory of Tony Grace that functional connectivity between these regions increases prior to psychosis. A secondary aim is to predict who will develop psychosis by applying machine learning techniques to the longitudinal data. The fellowship is to fund longitudinal MRI scans of subjects with an at risk mental state (ARMS), 25-30% of whom are expected to develop psychosis within 2 years of contacting medical services. The baseline MRI scans are already being collected as part of the EU-GEI project. However there were no plans in the original study to do subsequent scanning. My proposal aims to capitalise on this ARMS sample, which is an order of magnitude larger than any examined in previous studies. Scanning will be completed at 7 centres which each have ARMS clinics and a 3T MRI scanner. ARMS subjects will be scanned at the point of transition and 12 and 24 months after transition; those who do not transition to psychosis and healthy controls will be scanned 3, 12 and 24 months after the baseline scan. Analysis of the structural MRI data will be conducted with FreeSurfer and software developed by my collaborator Paul Thompson and resting state fMRI data will be analysed with FSL MELODIC. I will use support vector machines to predict psychosis transition from the imaging and clinical data which would have immediate translational benefits. Changes in brain structure and function before psychosis raises the intriguing possibility that future therapies may be able to arrest these changes. Charactering the trajectory of these changes would indicate which regions of the brain are first affected, and establish a reference point to compare the effects of preventative treatment strategies.

The key problem in the clinical management of people who present with prodromal signs of psychosis is that it is impossible to predict which individuals will go on to develop a psychotic disorder and which will not. As a result, potentially preventative clinical interventions have to be applied to all subjects, even though only about 30% will subsequently become psychotic. A successful algorithm which correctly identifies the latter subgroup would have an immediate translational benefit, as preventative treatment could then be targeted to those who need it most. This would permit a more efficient use of clinical resources and would be more ethically acceptable. The findings may thus significantly improve the clinical management of people at high risk of psychosis, particularly in terms of preventative interventions.

A less direct, but equally important impact will result from an improved understanding of the mechanisms underlying the onset of psychotic disorders. This is fundamental to the development of new treatments for psychosis, especially in its early phase. Our poor understanding of the pathophysiology of psychosis has been a key factor in the failure to produce new treatments.

Mental health charities and public health bodies would be able to use the results of this research to reduce the stigma of psychosis by publicising that physical changes in the brain are associated with the development of psychosis (similar to neurological illnesses). They could also use the findings to foster a culture of prevention in mental health, educating the public to recognise the symptoms of an at risk mental state and if necessary contact appropriate medical services.

The cost of schizophrenia to the economy in England has been estimated at 6.7 billion pounds a year (Mangalore & Knapp 2007). A significant portion of this cost is related to delays in diagnosis and if early diagnosis could be achieved by the findings in this study, the financial burden of schizophrenia and other psychosis disorders to the economy would be lessened. In addition, as the research is European wide the findings from this study are likely to impact on mental health policy of the wider European community rather than the UK alone.

References
Mangalore R, Knapp M (2007) Cost of schizophrenia in England.J Ment Health Policy Econ.10(1):23-41.