Human Brain Data Should Be Shared

Ueli Rutishauser, PhD, is professor of Neurosurgery, Neurology and Biomedical Sciences and director of Human Neurophysiology Research and the Center for Neural Science and Medicine at Cedars-Sinai. He studies the mechanisms behind learning, memory and decision-making, and his work would not be possible without data-sharing and collaboration.

The Rutishauser Lab collaborates with several leading universities and medical centers, including the California Institute of Technology, Johns Hopkins University, the University of Toronto, Boston Children’s Hospital and the University of Colorado at Denver. A key aspect of this collaboration is open data-sharing. Cedars-Sinai is part of the National Institutes of Health BRAIN Initiative Research Opportunities in Humans Consortium, representatives of which recently penned an article for the peer-reviewed journal Neuron on the benefits of data-sharing.

“Dr. Rutishauser’s work advances our understanding of the workings of the human brain and also connects Cedars-Sinai with top medical institutions across the country and beyond,” said Jeffrey Golden, MD, vice dean of Research and Research Education at Cedars-Sinai. “Work of this caliber simply isn’t possible if institutions guard their discoveries and data, encumbering collaboration.”

Rutishauser, who holds the Board of Governors Chair in Neurosciences, sat down with the Cedars-Sinai Newsroom to discuss the role of data-sharing and scientific collaboration in his work and the broader scientific community.

What type of data does your lab’s research generate, and how has it advanced our understanding of the human brain?

Our data is acquired from patients with epilepsy who are undergoing depth-electrode monitoring, which means they have tiny electrodes surgically inserted into the brain to monitor seizure activity.

We use these recordings of the electrical pulses sent between individual neurons within the brain to study how the brain records and recalls memories, how we make decisions, and how these processes go wrong under certain conditions. For example, we have discovered how we monitor our own behavior for errors using this approach.

The same data, which is rare and difficult to acquire, can also be used to study many other aspects of the human brain, but this is only possible if the data is made accessible to other investigators inside and outside of Cedars-Sinai.

What is the current state of data-sharing in human neuroscience?

Funding agencies and journals require that research teams make their data available to others. However, they require sharing of only the exact data needed to reproduce a given study, often only on request.

This requirement is often viewed as a burden by investigators. But in our recent editorial, we point out the many benefits of investigators sharing all of their data freely and in a standardized format, so it is easily accessible. We found that when we, as the data producer, released data in this way, we discovered new collaborators who used our data to explore questions we had never thought about. Within our own lab, using a standardized data format also facilitated reuse of that data. And schools and universities have even used the data for teaching purposes.

Why is cross-institutional collaboration so important to the future of science?

Collaborating and sharing data is a way for us to broaden our impact. One of the inherent difficulties of the work I do is that the experiments are challenging to perform and the number of patients available is limited. To increase the amount of data available, we collaborate with research groups at other institutions. This allows us to perform large, well-powered studies and increases confidence in our findings by replicating findings at other institutions.

What challenges need to be overcome to increase data-sharing in neuroscience?

There are three challenges. First, the field has to agree on a standard data format. While there are formats that could fill this requirement, there’s no universal agreement on which should be the standard format. The use of a single standard data format in the field of neuroimaging, for example, shows the immense benefits of the practice. Second, we need data archives where very large files can be uploaded, stored and made available. Third, and perhaps most challenging, investigators must be willing to openly share their data.

What is your advice for investigators who want to share their data?

I strongly advise using a standardized data format. We chose to use the Neural Data Without Borders (NWB) format. Develop an expectation in your lab that when a project is finished, your team will export the data in that format, document it and publicly release it upon publication.

Read more on the Cedars-Sinai Blog: Stop Signal Neurons