A model of changeover behavior in two-alternative choice

Area of Science:

• Behavioral science
• Animal behavior
• Mathematical psychology

Background:

• Organismal dwell time on variable-interval schedules often follows an exponential distribution.
• This suggests an innate behavioral pattern, not influenced by the duration of the current visit.
• Concurrent-schedule performance can be modeled using continuous-time Markov chains.

Purpose of the Study:

• To characterize changeover behavior in concurrent schedules.
• To derive expressions for leaving rates from each alternative.
• To assess the performance of a novel model against existing ones.

Main Methods:

• Modeling concurrent-schedule performance using continuous-time Markov chains.
• Deriving leaving rate expressions from the matching law and principles of reinforcement or Baum's laws.
• Assessing model fit using a large dataset from a study manipulating reinforcement rates.

Main Results:

• Dwell time on concurrent variable-interval schedules follows an exponential distribution.
• A novel model based on the matching law accurately predicts changeover behavior.
• The proposed model demonstrates superior performance compared to competing models.

Conclusions:

• Organismal dwell time on concurrent schedules reflects an innate pattern, accurately modeled by continuous-time Markov chains.
• The matching law, combined with reinforcement principles, provides a robust framework for predicting changeover behavior.
• This research offers a superior, mathematically grounded model for understanding complex behavioral allocation.