"Machine Learning Approach" to Smartphone Data Garners $10,000 First Prize in The Michael J. Fox Foundation Parkinson's Data Challenge
The contest received an enthusiastic response from the scientific community -- the winning entry was chosen following more than 630 downloads of the dataset from teams in 21 countries.
The LIONsolver team's winning entry provided proof of concept for a "machine learning approach" that could unveil clues to PD onset and progression embedded in data collected on smartphones. LIONsolver's project proved the feasibility and value of gathering mobile data for monitoring PD, while laying the groundwork for further analysis of larger, and potentially more powerful, datasets using LIONsolver's machine learning platform.
Moving forward, this approach has the potential to contribute to technology-enabled strategies to improve feedback between doctors and Parkinson's patients, ultimately helping to increase patient well-being and lower long-term costs of care.
"Our team is concerned about the quality and affordability of healthcare," says LIONsolver founder and CEO Drake Pruitt. "As problem-solvers, we continually look for ways to attack these challenges. We believe we can offer a state-of-the-art platform for monitoring Parkinson's disease, and we're hopeful that this work could open new avenues for improving patients' lives."
Machine learning is a branch of artificial intelligence that allows computers to build powerful models and improve their own predictions by learning from data itself. Familiar examples of machine learning in everyday life include recommendations served up to users of online movie rental services or bookstores.
"We were amazed by the response to our competition from research teams around the world," said Maurizio Facheris, MD, MSc, associate director of research programs at The Michael J. Fox Foundation for Parkinson's Research (MJFF). "We look forward to working with LIONsolver to help translate cutting-edge computational analysis into state-of-the-art approaches to treating Parkinson's disease."
In addition to the winning project "Supervised and unsupervised machine learning for the detection monitoring and management of Parkinson's disease from passive mobile phone data," two teams' submissions were deemed worthy of special mention:
- "Remote Monitoring of Levodopa Response in Parkinson's Disease"-- James T.H. Teo and Parashkev Nachev, Charing Cross Hospital, London, UK
- "Identifying Parkinson's Disease from Passively Collected Acceleration Data"--Michelle Wang, MIT, Boston, MA
In conjunction with the Kinetics Foundation, MJFF will be awarding a member from both the winning and honorable mention teams to present their findings at the "Objective Measures in Parkinson's disease" dinner event, during the Movement Disorder Society's annual meeting taking place this June in Sydney, Australia.
About the Parkinson's Data Challenge
MJFF offered a $10,000 prize to the team with the best plan to leverage an existing data set collected from a group of Parkinson's patients and controls using a basic smartphone application. Hosted on Kaggle, the competition challenged research teams to develop the best way to improve diagnosis, treatment or therapeutic development in Parkinson's through analysis of these passively collected, objective data points. The development of the app and collection of the data were led by researchers, collaborating entrepreneurs and industry experts at Gecko Ventures and MIT.
The contest was judged by:
- Alexandra Carmichael -- Co-Founder of CureTogether and Director of Quantified Self
- Karl E. Case -- Professor of Economics Emeritus at Wellesley College and Co-Creator of the S&P Case Shiller Index
- Maurizio Facheris -- Associate Director, Research Programs at The Michael J. Fox Foundation for Parkinson's Research
- Ken Kubota -- Director, Kinetics Foundation
- Daniel Vannoni -- Entrepreneur and Managing Director, Gecko Ventures
To learn more about the winning entry, and potential practical applications moving forward, read a Q&A with Drake Pruitt at MJFF's FoxFeed blog.