Computational Modeling and Systems Biology of Drug Abuse

United States

NIDA Organizer: Jane B. Acri, Ph.D.

Meeting Purpose and Intent:

Mathematical modeling is a relatively new but rapidly developing area of substance abuse research, providing researchers with additional dynamical dimension in both epidemiological and biological areas. In epidemiology, modeling allows scientists to simulate the consequences of various treatment and prevention scenarios, while in biology, modeling allows us to describe and simulate complex neurophysiologic processes.

The purpose of the meeting was to showcase the spectrum of modeling approaches and to discuss their application to solve real-world problems in drug abuse. The symposium was aimed at fostering dialogue and collaborations between NIDA and computational scientists working in both epidemiological and biological areas of substance abuse and addiction. During the symposium several modeling examples were presented, including models describing Injecting Drug Users (IDU) networks, ethnography of drug dealer?drug user interaction, neurocomputational models that relate decision making to neurotransmitter interaction, control theory models of addiction, reinforcement learning models of the transition from experimental use to addiction, and applications to clinical trials.

Brief Discussion of Meeting Outcome:

Participants were asked to consider several points of discussion:

1. What kind of scientific knowledge related to substance abuse could benefit from computational modeling?
2. What biological systems are perturbed by substance abuse, and is there data from these systems that can be modeled?
3. Can treatment response be predicted through analysis and modeling of genes and other environmental or behavioral factors of substance abusers that are in treatment? Is it possible to differentiate the groups and/or ultimately to tailor treatment according to genotype or phenotype?
4. How can epidemiological models of distribution and use patterns in adolescents be amenable to prevention efforts?
5. Can existing theories in substance abuse (opponent process theories, self-medication theories, etc.) be tested by modeling methods?

Two related program announcements were also brought to the attention of the attendees:

• PAR-08-023 Predictive Multiscale Models of the Physiome in Health and Disease
  http://grants.nih.gov/grants/guide/pa-files/PAR-08-023.html
• NOT-NS-08-008 Collaborative Research in Computational Neuroscience (CRCNS)
  http://grants.nih.gov/grants/guide/notice-files/NOT-NS-08-008.html

General Discussion - The meeting concluded with a broad discussion of the most pressing questions facing researchers. The major challenge of this type of research is the question of how to bring together researchers who can generate data in particular systems with those who have mathematical expertise in modeling. Another challenge is the translation of modeling results into applicable conclusions. Meetings are a first step in introducing the field to the concepts and advantages of modeling. Suggestions for future meetings included both: focused and broad content. Broad meetings are needed to synchronize terminology and concepts, and to facilitate cross-discipline collaboration. Tightly focused meetings are needed to concentrate on particular systems, or have breakout groups to discuss modeling in 1) biological systems, 2) behavioral paradigms, 3) genetic analysis, 4) risk factors, 5) prevention, 6) treatment, or other areas pertinent to substance abuse. A public resource for available computational models and data in addiction would be a welcomed development.

Brief Description of Resulting Publications:

Meeting organizers are discussing the possibility of engaging a journal editor to produce a special edition on modeling, and to include the presentations from this meeting.