Follow Datanami:
March 21, 2016

Psychiatry Looks to Data For Clinical Tools

Growing interest in harnessing big data and other numerical tools in the emerging field of “computational psychiatry” is reflected in the series of papers published this month in a science journal that together shed more light on how these tools can be used to treat mental illness.

The latest thinking on computational psychiatry was published in the March edition of the journal Nature Neuroscience, with a focus on neural computation and the clinical applications of computational psychiatry. Among the goals, practitioners note, is accelerating the transition from research to therapies.

“Translating advances in neuroscience into benefits for patients with mental illness presents enormous challenges because it involves both the most complex organ, the brain, and its interaction with a similarly complex environment,” note the authors of a paper on bridging the gap to clinical applications.

Computational psychiatry is moving along two complementary tracks, one theoretical and the other driven by big data techniques. “Data-driven approaches apply machine-learning methods to high-dimensional data to improve classification of disease, predict treatment outcomes or improve treatment selection,” according to researchers from University of Zurich, University of Lisbon, Portugal and Brown University.

Research into the application of computational psychiatry appears to face an increasingly common problem: huge datasets or, in the case of clinical researchers, “very high-dimensional datasets” consisting of neural, clinical, genetic and other large data volumes. Complicating matters is that data generated by a single brain scan, for example, includes a range of values that change over time.

This data “dimensionality” is viewed as both a blessing and a curse. Hence, researchers are attempting to apply machine-learning techniques to automate the analysis of these large datasets in search of meaningful patterns.

As the journal editors noted in an introduction to the series, machine-learning computational applications are increasingly being combined with theoretical “mechanistic models.” The goal is to leverage the “confluence” between big data tools and theoretical approaches to improve classification and treatment of mental illness.

“Making computation a core part of neurobiology curricula should be high on the agenda at every university,” the editors stressed. “The small benefit is that in five years’ time, we needn’t start off the computational focus by trying to justify its relevance for neuroscience. The great hope is that it leads to better biological understanding and treatment of disease.”

Those concerns reflect the fact that development of computational psychiatry tools remains in the early stages. “The next steps will be to validate these tools in longitudinal studies and examine how they could inform treatment decisions,” researcher Quentin Huys of the Swiss Federal Institute of Technology and the University of Zurich told Scientific American. “Then, their ability to improve outcomes will have to be tested in clinical trials.”

Still, combining data- and theory-driven approaches holds promise for transforming research into therapies, investigators stressed.

Recent items:

Unleashing Artificial Intelligence With Human-Assisted Machine Learning

Can Big Data Give Us Bionic Brains?

Datanami