Simplified Dynamical Systems Modeling of Daily Mood in Adolescent Cannabis Use to Guide Development of Individually Tailored Preventive Interventions

Introduction: This study used an intensive longitudinal time series methodology and dynamical systems modeling to identify regulatory dynamics of mood in adolescents using cannabis to facilitate development of individually tailored mHealth preventive interventions for cannabis use disorder.

Methods: Adolescents (n=6, aged 15-18) that regularly used cannabis completed EMAs 5 times/day for 14 days, including reports on mood (valance, energetic arousal, and calmness) and cannabis use. Simplified dynamical systems change-as-outcome multilevel modeling (group analysis) and regression (individual analysis) was employed to model the daily mood regulatory system. These models identify if and how cannabis use functioned as a system control parameter modifying the homeostatic set-point and/or stability of daily mood.

Results: For group analyses, on average there was greater stability in mood valence (b = -.30, p = .02) and calmness (b = -.41, p = .04) in the context of cannabis use. Yet, when examining individual teens we found unique dynamics in mood regulatory system functioning in the context of cannabis use. From a behavioral perspective, these individual differences indicate that between teens the reinforcing functions of cannabis use on mood varied.

Conclusions: A personalized, rather than a group, approach was needed to model how cannabis use modified mood dynamics for each individual. That personalized assessment can then be translated into individualized tailoring of mHealth interventions targeting prevention of cannabis use disorder. Finally, simplified dynamical systems analyses may be broadly useful in small n intensive longitudinal designs to identify key features of daily intraindividual variation in health behaviors to facilitate tailored mHealth intervention.