For anyone old enough to remember SETI@home (guilty), the concept of mass-distributed science research has a long history. Now that crowdsourcing is a much better-known and more widely-applied idea, we’re starting to see some really cool innovations for getting the general public to help out with science and research.
One of our favourites is Stall Catchers, an online game with 10,000 players that’s helping speed up research into Alzheimer’s disease at Cornell University. It can do research in an hour that would take a week in the lab.
Funded by the BrightFocus Foundation and part of the EyesOnALZ project, Stall Catchers is a result of the Human Computation Institute, Cornell and some other citizen science organisations working together.
We caught up with the epically-named Eglė Ramanauskaitė from the HCI to find out how Stall Catchers came about and how you can get involved.
1. Hi Eglė, thanks for talking to us today. Could you introduce yourself for us?
My name is Eglė Marija Ramanauskaitė, I am the Citizen Science Coordinator at the Human Computation Institute, as well as the community and content manager, plus UI designer for Stall Catchers.
I have a background in Molecular & Cellular Biology and Education Science, and have been involved with science communication and citizen science (practice and research) since 2014.
Aside from that, I’m very involved with open science and the hacker movement, and I have a farm with goats in the midst of Lithuanian woods.
Eglė (on the left!)
2. Tell us a bit about Stall Catchers.
Stall Catchers is an online game where anyone can contribute to Alzheimer’s research. By looking at movies from a live mouse brain, players help classify blood vessels as flowing or stalled - exactly the same as it is done in the lab.
The game helps crowdsource promising Alzheimer’s research at Cornell University, concerned with stalls - clogged blood vessels in the brain. Stalls seem to have an important role in Alzheimer’s disease, and by learning more we could find the first ever treatment, but the analysis of these data is extremely time-consuming. We can’t wait for an Alzheimer’s cure for decades, so we’re asking everyone to help now.
The Stall Catchers game was developed by the Human Computation Institute as part of the EyesOnALZ project. Our project lead - Pietro Michelucci, had the idea. Through a mutual colleague, he met one of the lead scientists doing this research at Cornell University, Chris Schaffer, in 2014. Chris’ lab was facing an analytic bottleneck problem with the data they were getting - data from one experiment was taking up to a year to analyse. After seeing what kind of data it was, Pietro realised it’s very similar to that in stardust@home - one of the first “volunteer thinking” projects ever created. He reached out to the creators of stardust@home and they kindly agreed to “lend” their platform to Alzheimer’s research. We rebuilt the platform from scratch using the “virtual microscope” that the stardust@home team created, gamified it, and launched the game.
I have been involved with the project since its official start in January 2016. Pietro brought me on board as a community and content manager, due to my knowledge and skills regarding science communication and science education, and my interest in citizen science. I have since been involved in managing various aspects of the EyesOnALZ project, as well as designing the Stall Catchers game.
3. Why did the team choose an online game? What other approaches did you consider?
There are plenty of online citizen science projects that aren’t games, including those that involve classification tasks, such as many Zooniverse projects, for example. stardust@home is not a game either - their platform is fairly basic, focusing on the functionality of the “virtual microscope” rather than a fancy UI. They rely on a handful of dedicated “dusters” to do all the work.
“We need as big a crowd as we can get if we want to get to an Alzheimer’s cure fast enough”
While this approach works great for some projects, and we considered it as an option, we realised that, first and foremost, we need as big a crowd as we can get if we want to get to an Alzheimer’s cure fast enough. We need a way not just to attract people to try the task in the first place (and the vessel movies can look daunting at first!), but to stay and keep “catching,” get addicted, do it on their morning commute, tell their friends, enjoy the friendly competition with friends and other catchers…
This approach has been working out really well so far. Not only do we have a set of “super catchers” who enjoy keeping their top places in the daily, weekly and all-time leaderboards, we also have school and university teams who get involved largely because of the competitive aspect of the game.
One such example is a team of middle school students from Boise, Idaho, who got involved a year ago during our month-long #CrushALZ competition. Not only did they crush other teams in the end, they got so involved that the game literally went viral in their school and the team quickly grew from a handful of students to more than 250. Today their team “Middle School STEM” has over 400 members, and the students are urging us to run a national schools competition.
We awarded “Middle School STEM” with the first ever Citizen Science Trophy earlier this year to acknowledge their extraordinary achievement. Which wouldn’t have happened if not for the competitiveness of the students.
4. What have the results been so far? How do players react?
We have received an overwhelming amount of support from our community so far. A lot of our players are retirees, who get involved because they have seen their friends and family battle Alzheimer’s disease. Unfortunately, that’s a losing battle to this day. To use some of their own words, by discovering the game they realized that they can help fight this terrible disease and make a difference if not for their own, then for their children’s and grandchildren’s generations.
“Not only did the crowd meet our expectations, they surpassed even the lab experts in some cases!”
Within the first few months after the game went live we conducted our first “validation study” to see if the crowd answers could be trusted. Not only did the crowd meet our expectations, they surpassed even the lab experts in some cases! Since then we have also developed novel consensus algorithms to make sure to make the best use of each annotation, thus our catchers’ time, and have been working on machine learning methods to optimise the machine-human partnership even further and make the data analysis more effective.
We are in the process of analysing our second large “real” dataset from the Cornell lab, dealing with the role of a high fat diet in the formation of stalls. Previously the catchers helped us analyse a dataset looking at the relative locations of stalls and amyloid plaques, which are implicated in Alzheimer’s. We’re waiting for the final conclusions to these experiments to come from the lab, but crowd-generated results have already helped us gain a better understanding.
5. Where might the project go from here?
At the moment, we are very much focused on making the platform more efficient and inviting more people to join the crowd. There’s still lots to be done before we can get closer to a cure for Alzheimer’s - lots of research questions and datasets to be looked at. We are also passionate about the educational aspect of the project, and are preparing a school curriculum, as well as finding ways to recognise informal learning achievements of our players.
Stall Catching in progress
We are also continuously working to improve the link between the human and the machine parts of the platform. At the moment, existing machine learning algorithms don’t do well when it comes to finding stalls, but this might change in the future. Soon we might be able to open this part up to the crowd too and ask them to create catcher “bots” that catch alongside humans, leading to even better results.
Our platform, including the gamification elements, could be used to crowdsource similar data from other studies as well, such as sickle-cell anaemia. There are no concrete plans at the moment, but we’d be excited to reuse the platform for other purposes. In fact, the creators of stardust@home are thinking of taking our platform back to their data as well.
6. What would you say to young women who’d like to make a Stall Catchers of their own one day? Any words of encouragement or advice?
“I was brought on board for talents that others saw in me, even if I didn’t see them myself”
Believe in yourself, or surround yourself with people who believe in you! When I joined the project, I was actually going through a period of very low self-esteem and didn’t trust my own abilities. But I was brought on board for talents that others saw in me, even if I didn’t see them myself.
Being trusted with some of the most important aspects of the project allowed me to gain confidence, and slowly I began to remember one of the most fundamental “truths” I had formed in my mind when I was still a teenager: anything is possible if you firmly believe that it is (because you try hard enough to make it so).
Our enormous thanks to Eglė for taking the time to talk to us. You can follow her on Twitter here and of course, give Stall Catchers a try here.
