Biomedical Engineering Reference
In-Depth Information
society before prematurely expiring. We all have our problems, but this seems a
bit severe. In reality, most organisms change their behavior during periods of
illness (2,3). These responses are often analyzed in terms of their benefits for the
afflicted, but they also have the potential to create a type of "social immune re-
sponse" that protects healthy individuals by altering patterns of interpersonal
contact. At the aggregate level, these changes amount to transient distortions in
the structure of social networks. Disease-reactive networks can be induced by a
centralized authority (e.g., quarantine), but they can also develop as an emergent
property of simple behavioral rules operating at the individual level (e.g., avoid
sick people). In fact, vertebrate biology appears to have evolved molecular sig-
naling pathways to generate disease-reactive behavior without any explicit rea-
soning by a host. Diagnosis and treatment also change the functional connec-
tivity of a disease transmission network, as do variations in host resistance or
pathogen virulence. In homogeneous social networks, individual disease-
reactive behavior aggregates into fairly simple nonlinear feedback at the popula-
tion level. Vertebrate social structures are actually quite heterogeneous, and dy-
namic linkage can produce highly complex behavior in such heavily structured
systems. The present studies seek to map those dynamics and understand their
significance for population survival. The findings that emerge suggest that the
neural substrates of disease-reactive behavior may have evolved in tandem with
biological immune responses to "strategically" alter host population structures
under the ecological press of socially transmitted disease.
2.
MODEL
The results presented here come from a series of epidemics simulated in
ActiveHost—an agent-based modeling system for analyzing interactions be-
tween biological and behavioral determinants of health. The general architecture
is summarized in Figure 1, and described in greater detail in the Appendix.
The basic modeled unit is an individual host—an "agent" that interacts with
other agents in some way that can transmit disease. Each agent maintains a list
of potential interaction partners and realizes those interactions according to
a specified and potentially variable probability during each time unit. The ag-
gregated set of "source-target" linkages constitutes the disease transmission
network. Additional linkage networks may also be provided to transmit disease-
blocking interventions, behavioral responses to illness, or changes in social con-
tact patterns. Each host maintains a submodel specifying pathogenetic processes
on the cellular and molecular level, and this model can delete the host from
the live population after a specified period of infection. The pathogenesis model
can include periods of latent infectiousness (host is not recognizably sick but
can transmit disease), overt illness (visibly sick and transmits disease), disease
remission/recovery (infectiousness ceases after a period), immunity (loss of
infectibility), and reactivation (infectiousness resumes after a period of recov-
