Bringing crowd power to Alzheimer’s Research
Two years ago, Dr. Pietro Michelucci was a hammer looking for a nail. Michelucci, an artificial intelligence expert with a longstanding interest in distributed human-machine systems, had been a science advisor to DARPA, the research arm of the U.S. Department of Defense. After 10 years of supporting national security, he was ready to refocus his efforts onto humanitarian activities with global impact; he wanted to apply his expertise to building powerful problem-solving systems that combined the complementary abilities of humans and computers to achieve new capabilities that neither humans nor computers could achieve in isolation. And that’s exactly what he did.
At the same time, the Schaffer-Nishimura Lab had been studying brain blood flow in mouse models of Alzheimer’s disease. Using multiphoton microscopy, they found that a small fraction of capillaries were plugged and had no blood flow. Such stalled capillaries are also found in the brains of myeloproliferative disease mouse models. However, the image analysis for finding the stalled capillaries is extremely challenging. Professor Chris Schaffer explained that for each week’s worth of data collection, it could take a trained laboratory technician an entire year to analyze the data. Schaffer explained that, although their findings were promising, because of this analytic bottleneck it could take decades to run the series of studies needed to arrive at a treatment target.
“I was very excited to meet Chris,” says Michelucci. “Alzheimer’s disease is the only top 10 killer without an effective treatment, so it definitely met our criterion of being a high impact problem. The next question was: can we use human computation to help solve it?”
It turned out that Schaffer had already tried machine-based solutions, and the best one was only 85% accurate. But once Michelucci saw the specific visual tasks involved, he realized that those tasks mapped almost perfectly onto two existing and highly-successful citizen science platforms: stardust@home and EyeWire, which both use online games to crowdsource data analysis. Both platforms have had tens of thousands of volunteer participants and their crowd-based results have been published in top-tier journals.
Michelucci contacted the progenitors of both platforms, who recognized the good fit and graciously offered to help adapt their existing platforms to Schaffer’s data. With a high impact societal problem, a well-defined solution, and a dream team of collaborators in place, Michelucci was able to attract the funding from the BrightFocus Foundation.
Since the project kicked off at the beginning of 2016, the growing 15-member team has been heavily engaged in developing the citizen science platform for Schaffer’s research. Indeed, one of the two platforms is poised to launch in October 2016. Called “Stall Catchers,” this new software will allow members of the general public, ages eight and older, to play an online game that will analyze Schaffer’s Alzheimer’s data. With an active community of online participants, Michelucci estimates that the analysis time for each experiment will be reduced from one year to one week. “In practice, this means we could potentially get to an effective treatment target for blood flow deficits in Alzheimer’s disease in a few years instead of decades,” says Michelucci. “That’s the power of human computation.”
Michelucci, who published a seminal text and established the first scholarly journal on human computation, is also the founder of the Human Computation Institute (HCI), a distributed organization that now includes 28 external professors representing 15 related disciplines.
“Operationally, HCI conducts scientific research, but our agenda is driven by societal need,” says Michelucci. “So once organizational constructs were in place, I was looking for a big problem to solve—a way to help millions of people, and that’s when the stars began to align.”
Michelucci recently moved to Ithaca to work more closely with the Schaffer-Nishimura Lab as the project develops. He hopes to develop EyesOnALZ (formerly known as WeCureALZ) into a sustainable enterprise that could support researchers around the globe, enabling blood flow studies related to a variety of diseases, including diabetes, macular degeneration, and cancer tumors. To learn more about EyesOnALZ, and to pre-register to participate in an online activity that will directly contribute to Alzheimer’s research, visit eyesonalz.com.