Abstract
The process by which genes and memes influence behaviour is poorly understood. Genes generally may have a strong influence as predispositions directing individuals towards certain behaviours; whereas memes may have a less direct influence as information inputs to cognitive processes determining behaviour. In certain areas of medical science, knowledge has progressed towards approximate quantification of genetic influences, while social psychology can provide models of mimetic influence as the spread of attitudes. This paper describes a computational model integration of genetic and mimetic influences in a healthcare domain. It models mimetic influences of advertising and health awareness messages in populations with genetic predispositions towards obesity; environmental variables influence both gene expression and mimetic force. Sensitivity analysis using the model with different population network structures is used to investigate the relative force of meme spread and influence.
Introduction
The "nature vs. nurture" debate in healthcare often treats genetic factors and social influences as separate silos. While we can quantify genetic predispositions for conditions like obesity, modeling how these interact with a constant barrage of information—ranging from government health advice to aggressive fast-food marketing—is incredibly complex. This paper addresses the lack of integrated, exploratory tools that can simulate these interactions, providing a "proof of concept" for medical researchers to test "what-if" scenarios: for instance, can a strong health campaign overcome a high genetic predisposition for obesity in a low-income neighborhood?
Methodology
The researchers employed a multi-layered modeling approach to bridge the gap between individual cognition and population-level trends.
Bayesian Networks (BN) and Agent-Based Modeling (ABM)
At the individual level, each "agent" is represented by a Bayesian Network. This allows the model to handle probabilistic rather than deterministic causal relationships. For example, a "Locus of Control" gene doesn't force a behavior; it changes the probability that an agent will follow health advice. These agents are then placed within an Agent-Based Model to observe emergent population behaviors.
Network Topologies
To test how information (memes) spreads, the researchers compared three network structures:
- Random: Simple diffusion across the population.
- Nearest Neighbor: Geographically constrained, spatial spread.
- Social Network: A "hub-and-spoke" model representing close friendships and weaker "friend-of-a-friend" ties.
The Integrated Obesity Model
The model's parameters and Conditional Probability Tables (CPTs) were synthesized from extensive literature reviews and expert consultations with medical schools. The core Bayesian Network structure links environmental factors (Socio-Economic Status), genetic traits (Locus of Control, Obesity gene), and competing memetic influences (Health Advice, Advertising).
Key Insights and Results
- Meme Competition and Decay: The model treats health advice and advertising as competing "memes." High-energy advertising memes tend to dominate health advice unless the environmental context is highly favorable. Furthermore, a decay factor reflects "campaign fatigue," where the influence of a message decreases as it becomes familiar.
- Environmental Dominance: In "worst-case" simulations (poor environment, high genetic risk), even strong health advice struggled to make an impact if advertising memes were present. The highest probability of positive health action occurred in "rich" environments with agents possessing a high Locus of Control and low genetic obesity risk.
- Network Sensitivity: The structure of the social network is a critical variable. Broadcast-style (random) diffusion was found to be more effective for spreading health advice than constrained social networks, where information can get "trapped" or diluted by the lack of strong ties between clusters.
Conclusion and Future Work
While the model uses generalized influences (High/Medium/Low) and requires further empirical validation, it demonstrates the potential for hybrid modeling in public health policy. Future work aims to incorporate real-world intervention data and simulate simultaneous broadcast and social spread to create a practical decision-support tool for policy makers.
Summary updated