Pandemic-Friendly Speed Dating for Neuroscientists

BRAIN Initiative researchers developed a virtual format to foster data-sharing among experimental and computational projects.

When developing theory, models and analyses, researchers must test mathematical tools against real neurological data. Ideally, scientists aim to reuse previously collected experimental data, but identifying and interpreting those data has long been a bottleneck for theorists. As the BRAIN 2025 Report states, “True partnerships between theorists and experimentalists will yield large dividends for almost every conceptual and experimental problem to be tackled under The BRAIN Initiative®.”

This desire to bring together theorists and experimentalists has gained significant traction in recent years, in part because of the innovative work of the Simons Collaboration on the Global Brain (SCGB). In 2016, as part of the annual Cosyne conference, the SCGB pioneered a speed-dating format where experimentalists and theorists with mutual interests could be ‘matched’ as a way to foster collaboration. The event has remained a part of the conference in subsequent years — it was so popular in 2017 that registration needed to be capped.

But how can we, the neuroscience community, facilitate the interactions necessary to build partnerships in a pandemic-friendly virtual setting? Traditional collaborative opportunities rapidly declined with the COVID-19 pandemic, as many meetings went virtual, stifling the forums where scientists would meet each other and cross-fertilize their ideas. Though scientists have largely adjusted to this new format for presenting research, there is no substitute for the power of small, serendipitous, exploratory face-to-face conversations. (The virtual Neuromatch conference, which instituted its own version of virtual matchmaking, is an exception.)

To address this need, we formed a team to devise a new type of event that would enable this sort of collaboration entirely remotely. The goal was to promote collaboration between the recipients of two specific BRAIN Initiative grant types: U19s and Theories, Models and Methods (TMMs). The U19s are team-based projects where groups of labs join together to tackle large unanswered scientific questions. These projects generally have several data collection components. The TMM grants fund the development of analytical tools that would be broadly useful to the field, but these researchers generally do not collect data themselves. Our goal was to create conversations between U19 and TMM teams with shared interests to sow the seeds for collaborations. The team consisted of Grace Peng and Susan Wright from the BRAIN Initiative; Bill Lytton and Fidel Santamaria, who coordinate TMM grant recipients; Brent Doiron, a theoretical neuroscientist at the University of Chicago and an investigator with the SCGB; and me, Ben Dichter. Doiron and I coordinate data science for one of the U19 projects.

To start, we asked TMM and U19 groups to submit abstracts with the theme of data reuse. U19 abstracts described data that could have great value outside their own scientific questions and analyses. The theorists’ abstracts described what types of insights their tools could provide about data. The U19 groups then viewed the TMM abstracts and ranked the groups they would most like to meet, and the TMM groups did the same with the U19 abstracts. Using these rankings, we matched groups for a series of brief meetings where a single U19 group would be matched with a single TMM group. We used mixed-integer linear programming to determine an optimal schedule, maximizing the occurrence of groups meeting with their preferred counterparts.

The turnout for the event, held November 13, 2020, was far greater than expected. One hundred people across the 38 groups participated, showing that members have a strong desire to build connections in the field.

Participants gathered virtually to find ways to re-use data and share analysis tools. Grace Peng

Our event began with an introduction and a brief talk about BRAIN-funded tools such as NWB and DANDI that can facilitate data exchanges between groups. Meetings generally had two to four people in each room, an optimal size for a good conversation.

In our post-event discussion, participants expressed gratitude and even joy at the opportunity to forge new scientific connections in the digital landscape. “It exceeded my expectations,” said Srdjan Antic, a neuroscientist at the University of Connecticut. “Having the face and the camera enormously increases communication, rather than just emailing someone.”

Others agreed, reporting that they made more connections — often surprising ones — than they had anticipated. “It was super useful. There was a scientific match every time,” said Adrienne Fairhall, a computational neuroscientist at the University of Washington and a member of the SCGB’s executive committee. “People are working on things that I have data that would be relevant for, and people have methods that would be interesting to try on our data, so we definitely plan to follow up with everyone we talked to today.”

Our follow-up survey found that the vast majority of responders found the event useful — 86 percent said they would be interested in participating in similar events in the future.

Several participants suggested that we make the sessions longer. Though we plan to modify the meeting length to accommodate the desire for longer conversations, we encouraged participants to self-organize at the meeting’s close and to follow leads they discovered during the event.

One of the interesting sentiments to emerge from the discussion and feedback was a sense of genuine surprise, which was perhaps a side effect of the two-way ranking strategy we used to create the matches. We expected that the groups that put time and effort into the ranking process would benefit, but the surprise was that even groups that put in little effort had good matches. In fact, even participants that missed the ranking deadline benefited from the groups that chose them. This may have seemed like a happy coincidence to them, but in fact it was a consequence of the dual-matching design.

Our results revealed that we have found a way to foster collaboration among researchers who might not otherwise meet. This speed-dating format provides a convenient way for scientists to survey the field in a fun and social way. The digital nature of the event has several advantages: It is safer, more environmentally friendly, more cost- and time-effective, and more inclusive; those who would have had difficulty attending a traditional conference due to disability or geographical location could participate from the comfort of their homes.

Perhaps most importantly, this virtual format scales much better than in-person events. The in-person speed dating at Cosyne, for example, had to be capped at 30 participants. We’ve demonstrated that, using our ranking approach to find matches, a successful virtual event with 100 participants is easily achievable. With the right infrastructure and a few tweaks, we could see this type of event scale to thousands, unbound by the old limitations of physical buildings and rooms.

Inspired by the success of this event, we want to bring this style of meeting to more communities. We have started to build an app that would manage the abstract submission, ranking and video speed dating. The app will scale beyond the capabilities of Zoom breakout rooms to accommodate even larger communities. In the near future, we hope to connect with fellow organizers who are interested in using this sort of platform in their own communities. These efforts will be critical for building the tools to support a new generation of collaborative partnerships between scientists, across fields and around the world.

Ben Dichter is leader of the Ripple U19 data core and community liaison for Neurodata Without Borders.

Recent Articles