PROVIDENCE – Interviewing neuroscientists engaged in research is an art form. It requires great listening techniques, and the ability to recognize that they occupy a place in the scientific world that is both ethereal and concrete.

When they talk, they are often describing an innovative process that is often poetic, what the poet Robert Bly once likened to leaps in consciousness, with poets chasing what Bly called dragon smoke.

When it comes to the collaborative research effort known as the Flywire Consortium, which produced what is known as a “connectome” – the visualizaion of the neural structure of an adult fruit fly – it involved more than a dozen research labs sharing datasets.

ConvergenceRI recently interviewed Davi Bock, Ph.D., a neuroscience researcher who was deeply immersed in the Flywire initiative. Christopher Moore, Ph.D., associate director of the Carney Institute of Brain Science at Brown University, provided the connection. [Bock received his undergraduate degree from Brown.]

Bock currently works as an Associate Professor in the Department of Neurological Sciences at the University of Vermont, using high throughput electron microscopy. His research at the Janelia Institute in Virginia helped to develop the datasets that served as the foundation of much of the collaborative work in creating the connectome of an adult fruit fly’s brain.

Here is the ConvergenceRI interview with Bock.

ConvergenceRI: What is the nature of the research collaboration among neuroscientists? How does that work and why is that important?    
BOCK: That is a very interesting question. Is there any more to the question or would you like me to rip on it from there?

ConvergenceRI: Go from there and take it whichever way you want to go.    
BOCK: There is a software programmer named Eric Raymond who wrote a very influential essay called “The Cathedral and the Bazaar.” The idea was that there are two ways to build large software. The first way is for a monolithic organization to design the thing upfront, down to the last brick, execute according to the design, over a long time period. That’s how you build a cathedral. And that was [the model for] big companies building software.

And then, the open source version of that is: the “bazaar.” At that time, the Linux operating system was emerging. There is a long tradition of open source software and collaboration, with people just writing software to scratch their own itch and build something they wanted.

That mode is extremely productive, and Raymond’s argument was that the “bazaar” eventually overtakes the “cathedral” as a mode to generate value in software development. So, that is a very roundabout way of answering your question.

Basically, I think the same thing is true in my field of “connectomics,” the sub-area of neuroscience trying to pull out wiring diagrams for brain circuits. We have seen efforts that are essentially managed. I’m thinking now of the so-called “MICrONS” program, run by IARPA, the Intelligence Advanced Research Projects Agency, which is a sibling agency to DARPA in the federal government.

It had heavily managed milestones, with lots of process, and that was very successful. The Allen Brain Institute delivered a millimeter cube of calcium images of a mouse visual cortex from that initiative.

There are other approaches. And, in some ways, I think that the work that I did, on the Flywire Project, is much more in the “bazaar” mode.

I was actually [working on] the acquisition of the data and the processing of it was sort of an individual lab seeing an opportunity to contribute to a really cool project, and then joining in. After I left the Janelia Institute, or as I was leaving Janalia, I made the dataset available to the world.

And Sebastian Seung and Mala Murthy wrote a grant to realign and segment [the data] and proofread it. They didn’t say they were going to turn it into a connectome, but they made enough progress that it became clear that it would be possible.

A colleague of mine, Greg Jefford, and I had written a separate grant, in some coordination, but not much, so that in this “bazaar” mode to the BRAIN initiative at the National Institutes of Health. We basically diverted our resources and hopped on board the Flywire effort to help with the proofreading. We certainly helped a lot with the annotation of cell types.

Those are the modes, and they can sort of switch from one to the other. The key is that data should be open. That’s the key. You need to make your datasets open, especially if you want to operate in the bazaar mode. I don’t know if that was an overly long answer.

ConvergenceRI: No, it was quite good. I think it gets to the heart of what is sometimes referred to as the “innovation ecosystem,” which I think is often misunderstood.

I worked for years as a communications consultant for the John Adams Innovation Institute in Massachusetts, writing a lot of their materials. They were the people responsible for the $1 billion Life Sciences Initiative. I was intimately connected to that process and the efforts to try and bring universities together to collaborate rather than to compete. 

Which is not to say that competition isn’t good, but it seems to me what you are talking about the importance of having a conversation that is collaborative, in order to reach a better understanding of the work.    
BOCK: That is true. I think there is another idea in the mix, which is that of the gift economy.

In the gift economy, you gain status by giving away valuable and expensive things. You accrue social capital and credit. One of the incentives to make these datasets open is that by saying, “Hey, everyone can use this,” you sort of gain credibility in your community as somebody who can deliver high-value datasets. And that is worth something, right?

Because in our world, we review each other’s grants, unless we are close collaborators, in which case we are conflicted out. But if you gain a reputation as somebody who when they say they are going to do something, they do it, and that thing is to deliver valuable datasets to the world, then it helps you.

And that’s another dynamic that helps foster collaboration is if the funding environment is such that it promotes the gift economy.

ConvergenceRI: Can you talk about the connectome and how the images were produced? I interviewed with Christopher Moore around his use of bioluminescence optogenetics, or BLOG. I was wondering whether that played any role in terms of the imaging of how you illuminate what’s going on in the brain, and how you can see it? And how you can see the way that cells operate?    
BOCK: For the work that my lab did, no. We use electrons, and shine the electrons through thin sections of hundred-millimeter diameter of axons and dendrites of neurons.

However, before we do that, it can be very useful to do something called in vivo calcium imaging. Which is related to what you are talking about with Chris. What happens there is that you can load neurons with a calcium or a protein or a dye, depending on your approach, and when you shine long wavelengths on it, it fluoresces in shorter wavelengths, like green.

What you can actually do, with calcium, when a neuron fires, an action potential calcium rushes in, so the calcium signal, this calcium dependent fluorescence, is a proxy for the firing of the neuron.

And so, if you do that kind of imaging before you do the electron microscopy, you can actually register the activity of neurons and how they work, while the animal was still alive in the electron microscopy dataset, and then you can [compare] the wiring diagram of all those neurons and relate the circuit structure to the observed in vivo activity of the neurons and how those neurons respond to visual stimuli, or how they contribute to motor control or whatever it is your are studying.

I guess the point is that you can integrate across these imaging modalities to combine data.

ConvergenceRI: Can you talk about what your current work is?    
BOCK: My current work is [focused] on annotating this whole connectome, which has been a major focus for the last few years. We have built a bunch of back-end tools. This work is kind of bookend for me. And, I am thinking about new project ideas, some actually with Chris Moore, some with other colleagues, some just on my own. So, there are a lot of open possibilities right now.

I’m sort of figuring it out right now. There is a short list of strong possibilities. It is kind of nice. I haven’t been in this position in a while, because these projects are so long term.

While I was at Janelia, I managed a big operation, with large teams and coordinating with many other teams, inside and outside the Institute. But now, I live in rural Vermont and I work out of my house, most of the time.

ConvergenceRI: Within the context on how labs collaborate, is that a new thing? I get the sense that there is limited money, so there is always competition to get that money. At the same time, as you described the “bazaar,” the marketplace, labs are “learning,” if that is the right term for it, how to collaborate with each other and still preserve their identities.    
BOCK: That’s true. It is kind of a push-pull dynamic that is present at all times. The way that I’ve handled this in my work is to basically like, when we generated this dataset, which as described in a 2018 paper in Cell, there were many authors, and all of their contributions were actually essential. The idea was that, yes, I was the corresponding author on the paper, and my graduate student was the first author, and the [other authors] were kind of buried in the middle, you can look at it that way.

But the other thing was, the guy whose lab did the alignment and registration of the data can go out to his community and say, “Look at what I did, in this amazing application of my software.” And, if he wanted to, he could write up a specialty publication.

So, I try to structure things so that people can have their piece of the credit pie. It doesn’t always work perfectly. It works most of the time. And, it’s a nice way of doing things, rather than having people fighting over it.

But, you are right. There is definitely competition. You know, my world is one that is filled with very large egos – people whose entire identities are constructed around the science that they do.

Sometimes you get some people throwing elbows and the scientific equivalent of bullying. But this is just humans. Humans are not all sugar and spice and everything nice. Part of nature is red in tooth and claw.

All of the above happens. Competition, collaboration, et cetera. A lot of it depends on funding agencies, number one, having consistent funding, which isn’t always true. And number two, having funding mechanisms that encourage collaboration.