Tracing neuroscience’s family tree to track its growth

The seed for Neurotree was planted during a lab meeting when I was a graduate student. We were discussing two papers promoting different ideas about the circuits that underlie complex cell function in the visual cortex, and my adviser, Jack Gallant, noted that the papers derived from two groups of researchers, each of which could trace its training back to two different labs. We found it illuminating to trace out on a scrap of paper who had trained with whom and see how the two schools of thought had formed—and to think about how experience and context shape how we as neuroscientists develop our experiments and conceptual models.

Soon after, my labmate Ben Hayden and I decided to broaden those efforts: We put the tangled network of visual neuroscience PIs and their graduate and postdoctoral trainees into a database, so that we could more easily browse the “family tree” of mentoring relationships. Tracing our own intellectual heritage back in time turned out to be a fascinating exercise: We discovered that we weren’t so distant from those whom we perceived as mythical founders of our field.

We quickly realized that the neuroscience network extended well beyond our immediate field, and that the database could benefit from the collective knowledge of our friends and colleagues. After a few weeks of not telling our adviser exactly how we were spending our time, we launched Neurotree, a home-brewed website with a simple interface to visualize and navigate the academic genealogy. Any user with a little extra time and knowledge was—and still is—invited to add themselves or their colleagues to the Neurotree site. Every contribution to the project increases its value as a resource.

Somewhat to our surprise, and thanks to the power of search engines, many researchers began stumbling on Neurotree, and the resource grew rapidly. The original website contained mentorship data for about 500 neuroscientists; 20 years later, it houses data for 150,000 scientists.

Researchers in other fields soon asked to establish their own genealogies, and Neurotree expanded into the Academic Family Tree (AFT), hosting genealogies for more than 800,000 researchers in more than 50 fields. We have also linked researchers to their publications and grants, so it’s possible not just to learn the identity of someone’s mentor, but also to get a sense of the ideas and expertise that the mentor provided to their trainees.

As Neurotree has grown, the project has transformed from a curiosity into a valuable resource, offering a unique source of data for studies of academic social networks. Studies using AFT data have explored questions around how factors such as lab size and publication rates influence trainees’ careers, for example. Neurotree data also makes it possible to track how various concepts and models evolve over time—a particularly useful analysis in a field where researchers from a range of disciplines, including biology, physics, math and psychology, all shape future academic generations.

T

o examine how different academic disciplines have influenced neuroscience over time, I traced the “distance”—number of mentor connections—between each neuroscientist in Neurotree and prominent historical figures in seven fields. The inverse distance to different fields provides a “fingerprint” of each researcher’s influences. Example fingerprints for some well-known neuroscientists, shown below, provide a simple picture of the broad influences shaping their work.

Because Neurotree tracks when researchers completed their training, it’s also possible to chart how the interdisciplinary influences captured by their fingerprints have shifted over time. I plotted the average fingerprint weights across Ph.D. recipients each year, revealing clear trends in how some fields, such as math and physics, have grown in influence, and others, including chemistry and philosophy, have decreased. The growing weights for math and physics may reflect the increasing role of quantitative methods and modeling within the field.

This quick study of the Neurotree database highlights the rich diversity of ideas and approaches that makes our field so dynamic. When viewing neuroscience as a family tree, it is impossible not to think about the evolutionary forces that shape it. The past few decades have seen a huge expansion of the field, and recent changes in the federal funding landscape raise poignant questions about whether this trend will continue. Will we be forced to narrow the range of ideas driving research? How will this affect the many promising but immature ideas for translating neuroscience to medicine? So many things are uncertain, but as a stable resource, Neurotree will help the community visualize and understand whatever changes the future brings.

Over the years that I have maintained Neurotree, I have prioritized its identity as a shared, collective project: The contents of the site are contributed by a community of volunteers, and the data are made available open access to the research community. I hope that others will use it to explore questions of their own and contribute ideas about how to make it more useful and more comprehensive.