A Control Theory Model of Smoking

Abstract

We present a heuristic control theory model that describes smoking under restricted and unrestricted access to cigarettes. The model is based on the allostasis theory and uses a formal representation of a multiscale opponent process. The model simulates smoking behavior of an individual and produces both short-term (“loading up” after not smoking for a while) and long-term smoking patterns (e.g., gradual transition from a few cigarettes to one pack a day). By introducing a formal representation of withdrawal- and craving-like processes, the model produces gradual increases over time in withdrawal- and craving-like signals associated with abstinence and shows that after 3 months of abstinence, craving disappears. The model was programmed as a computer application allowing users to select simulation scenarios. The application links images of brain regions that are activated during the binge/intoxication, withdrawal, or craving with corresponding simulated states. The model was calibrated to represent smoking patterns described in peer-reviewed literature; however, it is generic enough to be adapted to other drugs, including cocaine and opioids. Although the model does not mechanistically describe specific neurobiological processes, it can be useful in prevention and treatment practices as an illustration of drug-using behaviors and expected dynamics of withdrawal and craving during abstinence.

Towards a Smart Smoking Cessation App: A 1D-CNN Model Predicting Smoking Events

“…Using the reported smoking events, nicotine concentration was calculated using the control theory model of smoking [16] as shown in Figure 6. Each peak in the figure represents a smoking event, followed by a gradual decrease in the nicotine level until the next smoking event.…”

“…Bobashev et al [16] modelled the behaviour of smokers and employed the opponent process scheme of control theory. The model did not present any complex neurobiological process, only providing a mathematical model with a cascading feedback loop, aimed at presenting the scientific narrative of the opponent process as shown in Figure 1.…”

“…Smoking frequency; each peak represents a smoking event using smoking events reported from a randomly-selected day from our collected data of one of our participants [2]. (a) a simulated smoking behaviour generated by the control theory model [16] , and (b) real smoking behaviour.…”

“…Following this hypothesis, this research proposes a deep-learning model, which when combined with a control theory model of smoking, will be able to adapt to the smoker's unique behaviour and predict future smoking events. The Bobashev et al [16] model was chosen due to its ability to capture the nicotine effect using the pharmacokinetic equation. The model can be later employed to develop a smart mobile app that will send automated interventions.…”

Decision Tree Model of Smoking Behaviour

“…Bobashev et al [7] modelled the behaviour of smokers and employed the opponent process scheme of control theory. The model did not present any complex neurobiological process, only providing a mathematical model with a cascading feedback loop, aimed at presenting the scientific narrative of the opponent process as shown in Fig.…”