NeuroPsyGrid: towards an ontology and multi-centre brain imaging for early psychosis.

This collaboration brings together two existing MRC-funded projects, to help develop national and international collaborations in psychosis research.

There are two main problems to be overcome in integrating multiple data sets for multi-centre studies. The first is to know what is in each data set and what it is intended for. This is annotation or meta data - data about data. The second is to specify precisely what is meant both by the data itself and the meta data - which is the business of ontologies .

We are working together towards a database, common metadata and a simple ontology for psychosis. We will then test this system as a method of combining data from existing studies to answer clinically relevant questions.

We are also studying structural and functional brain imaging reproducibility within and between five research centres.

These developments will put us and other mental health researchers in a better position to conduct large multi-centre studies of earlier detection methods, more effective treatments and possibly even preventative interventions in people at high risk of or in the early stages of psychosis.

This collaboration would enhance two existing MRC-funded Grid projects by working together to help develop national and international e-Science collaborations in psychosis research. Both projects are concerned with the collection of multi-centre data for psychosis research. Yet to link together and interpret these datasets requires a common consensus on the way the data objects are organised ? the metadata ? and on the termnology used to describe and annotate their contents ? an ontology. Thus, we aim to harmonise our joint work towards a common database, a shared metadata model and a simple ontology for psychosis. In particular, by adopting a novel focus on changes in clinical and biological variables over time, in the few years before and after diagnosis, we will be able to test the utility of these integrated clinical, behavioural and imaging data as an indexing and discovery aid to answering clinically relevant questions. We will also study structural and functional brain imaging reproducibility within and between three centres from PsyGrid and two centres from NeuroGrid. We will ensure that these data acquisition, storage and access arrangements are compatible with other major national and international initiatives. Developing this infrastructure and piloting these processes will put us and other mental health researchers in a better position to conduct large multi-centre clinical and neuro-imaging studies of earlier detection methods, more effective treatments and finally even preventative interventions in people at high risk of or in the early stages of psychosis.