Neural Networking

  • Feature Story

Neural Networking


M.R. O'Connor
How casual connections in Building 46 yield innovative collaborations

Molecular biologist Elly Nedivi was looking for a microscope. Specifically, she wanted an instrument capable of taking high resolution images of entire neurons with all their dendrites and synapses in the brain of a living mouse, something no one had done before. Newly arrived at MIT as an assistant professor, Nedivi found herself at a talk given by Peter So, an associate professor of mechanical and biological engineering, and optics expert. So’s openness to new ideas struck Nedivi and she approached him afterwards. “It was clear to me that he was the person,” she said. So had never considered the applications of his research in microscopy to neuroscience before then. “I knew almost next to nothing about the brain,” he said. “When Elly comes up to ask if I’m interested in working together to do neuroimaging, I said, ‘I haven’t done any, why don’t we try something?’”

Since then, Nedivi, now the William R. (1964) and Linda R. Young Professor of Neuroscience, and So, now a professor of mechanical and biological engineering, have initiated a long-standing collaboration using multi-photon microscopy, a technique that can visualize fluorescent structures deep in living tissue, to track dendrites and neurons in mouse brains. Their early research supported the view that neurons in adult brains aren’t hard-wired, as previously believed, but are actually remarkably plastic and capable of remodeling structurally, even growing new branch tips. They have continued to expand imaging capabilities with a microscope that uses fluorophores that can be co-excited and easily separated. This two-color synaptic imaging enabled simultaneous visualization of inhibitory synapses and dendritic spine remodeling in vivo. They are now working to develop the first three-color imaging techniques. It’s been 20 years since their first conversation and Nedivi and So are still pioneering new instruments and insights into the plasticity of the brain, publishing some 12 papers together. Their labs are a four-minute walk apart, and their students and postdocs regularly cross the street to collaborate with one another. It’s been a remarkably smooth partnership, according to So. “After many years, there is a trust personally between Elly and me,” he says. As for Nedivi, she’s reminded of a similar partnership in rock ‘n’ roll. “Elton John and his songwriter, Bernie Taupin” she says, with a laugh.

Building 46, which houses the Department of Brain and Cognitive Sciences as well as the McGovern Institute for Brain Research, Picower Institute for Learning and Memory, and several other high-profile research centers, is home to 45 faculty researchers, all dedicated to working on fundamental questions about the brain and mind. In search of answers, many of them routinely reach outside the scope of their own disciplines to work with scientists in other fields of inquiry, often with colleagues within the building. Some of these collaborations have become partnerships lasting for decades, resulting in influential publications and new ideas. For the last 18 years, Walter A. Rosenblith Professor of Cognitive Science Nancy Kanwisher on the fourth floor has worked on and off with Jim DiCarlo, Peter de Florez Professor of Neuroscience and department head, on the second floor to develop and improve computational models of how the brain recognizes objects. Professor of Cognitive Science Laura Schulz on the second floor has partnered with Professor of Computational Cognitive Science Josh Tenenbaum on the fourth floor for 19 years to understand early childhood learning and the nature of memory. Since 2010, Associate Professor Mike Halassa on the fifth floor has worked with Guoping Feng, James W. (1963) and Patricia T. Poitras Professor of Neuroscience, on the third floor to understand how specific brain circuits relate to autism and schizophrenia. “We are lucky to have amazing colleagues here,” said Kanwisher. “Many of the people I want to collaborate with most in the world are in my department.”

When the complex at 43 Vassar Street was completed in 2005, it was touted as the largest neuroscience center in the world, built to bring together different disciplines in one place. The lead architect, Charles Correa (1930-2015), studied at MIT and designed the central atrium and surrounding balconies as common areas where people and ideas could commingle, creating opportunities for serendipitous discovery. Fifteen years later, Building 46 seems to have succeeded in fomenting collaborations. “The physical proximity is really important,” said DiCarlo. “I see Nancy all the time. You get to know someone and you speak their language and understand where they are coming from and you are not afraid to challenge them.”

For Mike Halassa, who also holds a Class of 1958 Career Development Professorship, Building 46 created the opportunity for an unexpected encounter that resulted in a rich partnership with Guoping Feng. In 2010, Halassa was a postdoc studying the thalamus and its cortical interactions. As it turned out, Feng had created a new transgenic mouse, the first that would give Halassa the ability to study the thalamic reticular nucleus and its role in generating cortical rhythms. “It was so unusual because he gave it to me before he published his own paper on generating the mouse,” said Halassa. This work resulted in a Nature Neuroscience paper published in 2011 where Halassa and Feng provided the first causal demonstration for the role of the thalamic reticular nucleus in cortical spindles, rhythms relevant to sleep function and known to be perturbed in both schizophrenia and Autism Spectrum Disorder. Three years later, Halassa was walking down a hallway when Feng saw him and said: “You wouldn’t believe this but I have this amazing finding that I think you would be interested in.” Feng had discovered a gene, Ptchd1, that seemed linked to altered sensory processing in Autism Spectrum Disorder and was specifically expressed in the thalamic reticular nucleus of the mouse during development. “What are the odds that there is this one brain region that I’ve established some expertise on and now one form of autism is linked to it?” said Halassa. The resulting study, co-authored by both researchers, was published in Nature in 2016.

Other collaborations at Building 46 originated in disagreements that evolved into ongoing debates that continue to inform the research and ideas of both collaborators. Two decades ago, Josh Tenenbaum was demonstrating an experiment that showed how humans might learn the relationship between a cause and effect by separately analyzing the structure or existence of a cause and its strength. Most of the people in the room agreed with his results, but Laura Schulz strongly dissented. “She exclaimed, ‘I don’t see it at all,’” said Tenenbaum. “We went back and changed the experiment based on her feedback. It was tough but she was right. It worked much better, it was much more interesting. And now it’s one of my most cited papers. It laid the foundation for much of my research.” In Schulz’s recollection, this initial disagreement transformed her understanding of the work she was doing in the field. “Learning about Josh’s ideas is like walking with a naturalist: you’re going up a mountain and someone points to a wildflower,” says Schulz, “and then points out that it’s connected to the entire geology of the landscape and to the Ice Age.” The ability to challenge each other works because, according to Schulz, they share a huge amount of common ground. “We bring different perspectives to what we are talking about but I would say there is unlimited respect,” she said. “It’s made my work much more far reaching and vigorous.”

Today, Tenanbaum and Schulz’s collaboration extends beyond co-authoring papers and collaborating on specific projects. Tenenbaum says some of their most significant work is co-mentoring students. “It’s a super additive benefit, being mentored by the two of us,” he says. “They get to learn from our creative, productive interplay.” One of their mentees was Julian Jara-Ettinger, whose doctoral dissertation won the Robert J. Glushko Prize in 2017. Now an assistant professor of psychology at Yale University, Jara- Ettinger studies how children use mental representations and computations to navigate the social and physical world. Tenenbaum proudly describes Jara- Ettinger as someone whose work “bears the stamp of what he learned and how he was trained, which was in the nature of this collaboration.” Jara-Ettinger said that he felt privileged to work with Schulz and Tenenbaum because their approaches complement one another so well. “Laura sees the field from a birds-eye view, identifying and solving problems long before others have even realized that a problem is looming over our theories,” he said. “And Josh has a way of taking foundational century-old problems and making their vagueness and intractability disappear.”

There is also a personal aspect when it comes to sustaining a partnership, a natural inclination towards friendship. “I feel like I couldn’t collaborate with someone if I didn’t like them and trust them, as a human being as well as a scientist,” says Kanwisher. Fifteen years ago, she describes meeting with DiCarlo so they could create a graduate seminar course and being struck by his ambition. “We would come up with a list of topics and then Jim would say, ‘I want to know how object recognition works.’ I would laugh and say, ‘Well, we’re not going to answer that.’ Now, 15 years later, I feel like he has done more to answer that question than pretty much anyone else.” DiCarlo describes Kanwisher as someone has he looked up to since the beginning of his career who has now also become his friend and colleague down the hall. “It’s not just her amazing scientific mind that I learn from, but how she mentors students, and asks hard questions without offending people,” he says. “It’s an art and those are skills that I try to model from her.”

Rather than exhausting themselves, long collaborations seem to generate productivity. “We’re in the same building and space, I see him maybe once a week and we have these brainstorm meetings once a month,” said Halassa of his relationship with Feng. “I feel like we are just getting started.” Indeed, there is a sense that the longer the collaboration, the more fruitful it becomes. “I would emphasize the long-term nature of it,” said Tenenbaum. “In order to nurture students, you have to have a really long-term collaborative relationship and build ideas together.” According to DiCarlo, working with Kanwisher “feels so much more scientifically productive at this moment. This is the most exciting work that we are doing together, and it just gets more exciting because we now have working computational models to pull our previously separate lines of work together and activate totally new questions.”

In recent months, these collaborations have been put to a new test. Nearly all research stopped when the Covid-19 pandemic arose, and the comeback has been slow; labs only returned to full occupancy in mid- October. Under the mantra “work that can be done remotely, should be,” many whose work focuses on computational modeling and machine learning continued to work remotely when wet-lab research ramped up.

Within weeks of the wind-down, however, it was evident the collaborations wouldn’t be stopped. Earl Miller, the Picower Professor of Neuroscience, and Professor Ila Fiete have moved ahead on a project to study how neural rhythms move around the cortex and what that means for brain function, holding a joint lab meeting on Zoom recently where graduate student Mikail Khona presented models of how timing relationships affect those waves. While this kind of an interaction between labs would normally have taken place in person in Building 46, neither lab let the pandemic get in the way.

The habit—the desire—the need—to forge new paths continues to be as strong as ever across the Building 46 community.

M. R. O’Connor is a journalist who writes about the politics and ethics of science, technology and conservation. Her work has appeared online in The Atavist, Slate, Foreign Policy, The New Yorker, Nautilus, UnDark and Harper’s. Her most recent book is Wayfinding: The Science and Mystery of How Humans Navigate the World.

David Orenstein, Director of Communications for the Picower Institute for Learning and Memory, also contributed to this article.