Late last year PARC received a new multi-year, multi-million dollar grant from the National Science Foundation (NSF) and the National Institutes of Health (NIH) for work on Fittle+. We are striving to transform how we address healthy behaviors, but behind all of this is intensive science and research.
The NSF is interested in promoting deep science, and a major component of the Fittle+ project is research on computational models capable of predicting how coaching interventions and group interactions influence individual changes in behavior. This will require integration of theories from individual health behavior, self-regulation, computational cognitive psychology, and computational models of online social interaction to develop fine-grained predictive models of behavior change.
Current theories of individual health behavior are not developed enough to support this kind of fine-grained prediction. For instance, the Theory of Planned Behavior (TPB) provides a conceptual framework for individual health behavior change cast at a level appropriate for traditional methods of persuasive messaging, education, and counseling. However, TPB needs to be greatly refined to meet the demands of fine-grained prediction and intervention with people going about their everyday lives.
PARC has had numerous projects developing computational neurocognitive models that are fine-grained, and predictive, but these models have not encompassed the motivational, social, and self-regulatory factors addressed by TPB. These predictive neurocognitive models simulate individuals, but behavior-change programs are often oriented to groups to enhance social processes that positively affect motivation and compliance. Agent-based models constrained by neurocognitive models need to be researched to understand the fine-grained dynamics of group interactions and predict the impact on individual behavior.
As part of the NSF-NIH Fittle+ project we are developing an integrated predictive model based on a combination of the ACT-R neurocognitive model, TPB, and advanced measurement models. The goal of this phase of analysis is to have a model capable of probabilistic predictions of individual behavior change with fine-grained parameters associated with elementary contexts of everyday life, attributes, beliefs, knowledge, etc. of the individuals themselves, challenge goals, and micro-intervention types and content. A second goal is the development of an agent-based model of social commitment and contribution rates in teams.
Scientific models are providing guidance on the automation of micro-targeted interventions and the social engineering of effective Fittle+ teams. These scientific foundations are also driving the development of Fittle+, and the deployment of Fittle+ at a major hospital will test and improve the underlying theories. We believe that a major driver of this science is “big data” – the more participants we engage, the more we will understand about what works for each individual and how supportive social engagement can be sustained over time. We are looking for partners, collaborators, and end-users at fittle.org, or you can email us at email@example.com.
Learn more about Fittle+