Los Alamos National Laboratory

This proposal is to advance the theory and software for capturing realistic behavioral dynamics in an epidemiological model. The task is challenging because human behavior is hard to measure, and it varies due to myriad factors including age, gender, cultural background, and others. The dynamics of infectious disease spread are sensitive to individual and collective human behavior in the affected population. For example, negative attitudes towards childhood immunization programs have resulted in recent measles outbreaks. Behavior is intertwined with attitudes and objectives, and it can spread from person to person via social conduits such as in-person conversation, telephone, and social media. We recognize that the recent emergence of social media has introduced a fundamental shift in how information is spread among people: both the rapidity and breath of information transmission have qualitatively increased. Our proposed research will use social media data to identify behavior, and behavior changes, that directly impact the spread of an epidemic. We propose to dynamically mimic realistic changing behavior in an agent-based epidemiological model by integrating a cognitive model with our existing state-of-the-art simulation platform. Agents will participate in two simulated environments simultaneously: a traditional contact network-based epidemiological model, and a second cognitive simulation, which models the dynamics of attitudes and objectives. The latter will generate and select behaviors for agents based on attitudes (e.g., perceived severity) and objectives (e.g., avoid infection), informed by analysis of social media and other sources. Our goal is to deepen scientific understanding of (a) collective human behavior in response to infectious diseases and (b) the impact of emergent behavior on the accuracy and applicability of predictive models of disease spread. 

MIDAS Project Website:Los Alamos National Laboratory

Principal Investigator: Sara Y. Del Valle, Ph.D.

Investigator: James M. Hyman, Ph.D.

Collaborating Institutions: Tulane University