Skip to main content

We Need to Change the Trajectory of Mental Health Research

The first order of business: break down research silos and move toward open science

This article was published in Scientific American’s former blog network and reflects the views of the author, not necessarily those of Scientific American


After centuries of denial, the reality of the cost and the extent of mental health disorders is reaching global acceptance. The suffering of the 20 percent of children and adults affected by mental illness, as well as that of their families and loved ones, is a story in major media outlets. The staggering estimates of productivity losses and societal burden and the tragic outcomes of illness (substance use, suicide, school shootings) are the subject of political dialogue and debate. In response, our leaders turn to scientists for solutions.

Unfortunately, as a scientist and as a child and adolescent psychiatrist, while I am confident that we will one day get the answers we are looking for, I fear they may come at a pace too slow to meaningfully impact the current generation. Unless, that is, we find a way to break down research silos.

Concerns about the formation of research silos are nothing new. Traditionally, researchers have tended to organize first and foremost around their academic departments and institutions, then through a web of associations formed through interactions at scientific conferences and workshops. While the internet has undoubtedly brought researchers closer together and fostered new means of communication and interactions, it has not broken the silos down.


On supporting science journalism

If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.


Beyond tradition and inertia, the maintenance of research silos can be attributed to a range of factors. Logistically, there are the often prohibitive work demands and administrative hurdles to organizing and curating data and resources for sharing outside one’s laboratory or institution. Culturally, concerns about loss of competitive advantage are among the most cited. It is not uncommon to hear a data generator ask why they should go through all the effort to collect, organize and share their data just for a data analyst to make the big finding. Add the context of academic institutions that generally give little recognition for sharing in tenure decisions; funding agencies that encourage sharing but rarely factor it into grant decisions; and journals that rarely enforce pro-sharing guidelines, and one can see why silos persist.

How is this affecting mental health research? As in any area of research that remains wedded to a silo culture, progress is and will be slow. This is because silo culture produces findings that cannot be reproduced. Teams rely on only the data samples that they themselves can collect, which are relatively limited in size and thus prone to generating statistical errors. Proprietary analytic methods lead to dramatically different findings in the literature.

And because research silos don’t value standardization, it is difficult to put the pieces together across publications. At least one third of published findings are expected to be impossible to replicate—and we don’t even know which ones they are! It takes years and substantial costs to identify them; as a result, the pace of progress is dramatically slowed.

Research silos also stand directly in the face of some of the most promising new models for conducting mental health research (e.g., transdiagnostic research, precision medicine). These new models inherently require large-scale samples, data intensive computing and interdisciplinary collaboration. And silos inherently increase the cost of research. A recent publication found that a relatively small data sharing effort in the brain imaging community led to the generation of more than 1,000 publications; to publish the same body of papers using proprietary data would have cost between $1 billion and $2 billion.

With all the evidence against it, why do remain in a siloed culture? Scientists aren’t against the ideal of open science. Few if any argue that it is a bad idea. But the current terrain is not conducive to its easy implementation, and this is where the challenge arises. Weighing the costs versus benefits of the status quo, we ask ourselves: Are we willing to wait longer for solutions to illnesses that affect 20 percent or more of our child and adult populations, that are among the most costly disabilities at a societal level, and that directly account for 1.4 percent of all deaths worldwide via suicide? The answer is clearly no, and so it is time for academic institutions, granting agencies and commercial interests globally to take decisive actions to both increase the incentives for open science and remove the barriers.