Collaboration: the mother of invention

Boston produces an outsized share of scientific breakthroughs. Researchers working side by side is a big reason why.

May 08, 2011 by Carolyn Y. Johnson, Globe Staff

Dominic Walsh was a junior scientist working in the same building at Brigham and Women’s Hospital as Dr. Dennis Selkoe, a noted Alzheimer’s disease researcher. The two would run into each other in the hallway and talk, leading Walsh to eventually join Selkoe’s laboratory.

Together, they did the research that resulted in a landmark paper published in the journal Nature, which pinpointed a particular protein as the likely cause of the memory loss of Alzheimer’s.

“It made a difference,’’ Selkoe said of Walsh’s decision to work in the lab. “His desk and bench were right outside my office, and we had a very close working professional relationship… . [Successful collaborations] can happen across continents and across oceans, but it almost always involves some face-to-face contact.’’

It’s long been thought that proximity fosters fruitful encounters among researchers, but a recent paper showed just how powerful it can be: An analysis of a decade of Harvard biomedical research collaborations — including Selkoe and Walsh’s — found that the closer the offices of key research partners, the more influential their joint papers were likely to be. It mattered whether collaborators were riding the same elevators in a building in Longwood, or working in labs on opposite banks of the Charles.

The proximity study is part of an effort to decipher the recipe for innovation, and what makes a city like Boston excel at it while other cities lag. Is it a certain density of universities, the volume of federal grants or venture capital, or a lifestyle that attracts creative people?

The first step for researchers is to identify which cities are truly exceptional innovation hubs, a task that is not as straightforward as it seems.

Boston researchers, for example, do not author the most papers in scientific journals. But Boston outranks other cities for research published in the most respected journals. Luis Bettencourt, a research scientist at Los Alamos National Laboratory, found that in 2009, 1999, and 1989 many more papers in the three top science journals were authored by researchers in the Boston metropolitan area than by scientists from any other city worldwide.

Other researchers at Elsevier, a leading publisher of scientific, technical, and medical information, analyzed a massive database of scholarly journal article references and ranked the top five US institutions across 1,700 topic areas by the number of publications they contributed. That analysis found that in 969 topic areas, the Boston area has an institution among the top five, well above Washington, D.C., New York, and the San Francisco Bay Area.

There are any number of ways to slice such rankings, however, so Bettencourt decided to compile large amounts of data on factors that conceivably could have a connection to innovation — everything from the length of electric power lines in a city to the number of patents. His goal was to define the typical attributes of a city of a given size.

That allowed him to drill down into the ways that cities buck trends. He found that the Boston metropolitan area invents things at a rate more like a place about twice its size, according to 2005 patent data. It’s wealthier than one would predict based on the size of the population — a factor that may be driven by high-paying jobs in the technology and life sciences. But it also acts like a smaller city in ways that may be important for attracting talent — for instance, the crime rate is relatively low, more like a city two-thirds its size.

What can’t be addressed by such broad datasets, Bettencourt says, are the particular behaviors that directly influence researchers and the quality of their output.

That’s what Dr. Isaac Kohane and Kyungjoon Lee at Harvard Medical School set out to do when they armed students with maps and GPS devices to measure the distances between collaborators’ offices.

Perhaps not surprisingly, their study, published in the journal PLoS ONE late last year, grew out of a chance encounter between the colleagues. Kohane had argued to a dean that his growing center needed to be consolidated in one spot so that people would run into one another when they were coming and going, and exchange ideas as they drank their coffee.

But it wasn’t until he was walking down the hallway one day and had a conversation with Lee that he began to talk about systematically researching the question.

Their study found that the shorter the distance between two key authors — the one listed first and the one listed last on a paper — the more likely the paper was to have been cited by other researchers in the field, one measure of a paper’s influence. The last author is typically a senior scientist who plays a leadership role and supports the work, and the first author carries out much of the research.

The continued importance of geography may seem counterintuitive in the era of Skype, iPhones, and other technologies that make it effortless and inexpensive to collaborate with people around the world. But location matters.

“We do certainly see that all this ability to communicate at a distance, [which] has helped us do many more things faster or that we couldn’t do before, hasn’t hit cities in the way people expect,’’ Bettencourt said. “Face-to-face contact seems to be more important as the noise goes up.’’

Robert Langer, an MIT professor, said that proximity has mattered throughout his career. When he worked at Children’s Hospital Boston in the 1980s, he often chatted on the elevator with a cardiologist who worked one floor below him. Those casual encounters led to a project focused on delivering drugs to replacement heart valves and a paper in a leading journal, Science.

After Walsh published the Alzheimer’s paper with Selkoe in 2002, he moved back to Ireland to start his own laboratory. He and Selkoe remained close collaborators, talking and e mailing every other day. But this year, Selkoe persuaded Walsh to move back to Boston — and they now work two floors apart.

“At this stage, I would like to collaborate and start doing something that would have a real outcome,’’ Walsh, 47, said. Boston, he said, is “one of the few places in the world where you can try to do something like that… . What makes a really good research environment as opposed to one that is mediocre or OK is the quality of the people present.’’

Carolyn Y. Johnson can be reached at This article can be found at Collaboration: the mother of invention.