Meet the Semi-Finalists: Q&A with BiAffect

This post is part of a special “Meet the Semi-Finalists” series, featuring Q&As with the five semi-finalists of the Mood Challenge for ResearchKit.

Today’s “Meet the Semi-Finalists” post features BiAffect, a system for understanding mood and neurocognitive functioning in bipolar disorder using keystroke dynamics, such as typing speed and errors, to track and predict mood episodes. Alteration in communication is one of the main, problematic symptoms of bipolar disorder. This ResearchKit study will unobtrusively monitor non-verbal speech/behaviors to improve our understanding of mood disorders and provide a means of predicting future mood fluctuations.

BiAffect is one of five semi-finalists competing to become a finalist and receive $100,000 to develop their designs into prototypes to be piloted with iPhone users. Stay tuned for the finalist announcement in October!

Tell us about your team’s background.
The BiAffect team is composed of a group of neuroscientists, physicians, and computer scientists at the University of Illinois at Chicago, University of Illinois College of Medicine, and the University of Michigan.

Led by the Dean of College of Engineering Peter Nelson and mathematician-psychiatrist Dr. Alex Leow, an Associate Professor in Psychiatry and Bioengineering, our investigative team at UIC additionally boasts world-renowned expertise not only in data-mining (Philip S. Yu, UIC Distinguished Professor and Wexler Chair in Information Technology) but also the diagnosis and treatment of mood disorders (Drs. Scott Langenecker and Olu Ajilore, both Associate Professor in Psychiatry). Leading our collaborative team at the University of Michigan, Dr. Melvin McInnis is recognized in translational research and has led clinical research programs in bipolar disorder for 25 years. Kelly Ryan (Clinical Assistant Professor in Psychiatry), is a clinical neuropsychologist with expertise in neuropsychiatric assessment in relation to functional outcomes with a specific background in impulsivity measurement and mood assessment in bipolar disorder, as well as mobile-health methodologies.

Andrea Piscitello, who completed his MS in computer science at UIC, is a software engineer based out of Europe and has been involved with the project since his master’s thesis. Faraz Hussain obtained his MS in rehabilitation psychology from the Illinois Institute of Technology, and works as a medical researcher at Northwestern University in addition to developing the iOS companion app for the BiAffect keyboard.

Why is this Challenge important to you? What inspired your proposal for a ResearchKit study?

Our target audience is people affected by bipolar disorder, a major psychiatric disorder that has been deemed the most expensive behavioral health diagnosis with an estimated lifetime prevalence of nearly 4%. Currently, diagnosis and treatment of bipolar disorder rely on careful history-taking and mental status examination by an experienced clinician, at times aided by self-report or caretaker-informed questionnaires. In general, these reports have to be interpreted by providers in order to extract patterns that could indicate an imminent change in mood. Moreover they do not necessarily represent objective psycho-physiological markers.

For this reason we want to investigate if keyboard dynamics and sensor data from iPhone serve as more objective biomarkers. If successful, the proposed work will lead to a mobile technology that continuously and unobtrusively monitors mood and cognitive states in bipolar disorder, thus allowing doctors and relatives to promptly react before significant functional impairments occur.

What have been the biggest challenges and successes in developing your study thus far?
Our biggest success is that we have already implemented BiAffect on the Android platform and conducted a pilot feasibility study in which we recruited a group of 28 individuals who on average have used the BiAffect technology for 60 days and have completed daily assessments of their mood and cognition. Our biggest challenge relates to the technical issues we encountered in order to further develop a similar technology for iOS using ResearchKit on a much larger scale. Taking into account and balancing a multitude of factors including user friendliness, optimal user time commitment, backend data management and related entitlement and privacy issues has been challenging.

You’ve entered the Virtual Accelerator phase, which includes expert mentorship and participation in a live Boot Camp. What’s the biggest insight you’ve uncovered through this process so far?
Our biggest insight as we’ve gone through the Virtual Accelerator came when we realized that our BiAffect technology will provide a unique unobtrusive platform not only for us as researchers and scientists to learn about participants, but also for participants to learn about themselves and to gain better insight into the inner workings of their own brains.