This is part of a series on improving mental health research, diagnosis and treatment. Join the conversation on Twitter with the hashtag #OpenMinds
The next big thing in psychiatric research is "machine learning," an approach used in fields ranging from stock market analysis to robotics. Computer algorithms convert raw data into statistical models, which are then used to make predictions, such as whether a treatment will work.
In the CAN-BIND study, algorithms are sophisticated enough to find patterns both within – and between – the study's three research areas: brain imaging, blood biomarkers and clinical evaluations.
But this is only possible because the Ontario Brain Institute, one of CAN-BIND's major funders, has created a model for standardizing data sets.
In the past, experts in fields such as brain imaging could not pool their results because of inconsistencies in how data were collected and stored.
OBI worked with Indoc Research, an Ontario-based not-for-profit group, to come up with common data elements through consensus-based workshops attended by researchers in each field.
The approach extends beyond CAN-BIND: OBI's new funding model requires that researchers use common data elements for the study of conditions ranging from cerebral palsy to Parkinson's disease. Data from these studies are stored in a shared bank called Brain-CODE, accessible by researchers across the country.
OBI's model for collaborative research has drawn interest from groups including the Organisation for Economic Co-operation and Development, headquartered in Paris, France.
Gaining agreement across disciplines about data standards is challenging, said Dr. Donald Stuss, president and scientific director of OBI, but essential for gathering masses of information needed for advanced research into complex brain disorders.
The Centre for Addiction and Mental Health has purchased advertisements to accompany this series. The organization had no involvement in the creation or production of this, or any other, story in the series.