A Markov model description of changeover probabilities on concurrent variable-interval schedules

Area of Science:

• Behavioral science
• Animal behavior studies
• Operant conditioning research

Background:

• Concurrent schedules of reinforcement are used to study choice behavior.
• A common hypothesis suggests animals choose options with the highest immediate probability of reinforcement.
• This implies sequential dependencies in response patterns.

Purpose of the Study:

• To investigate sequential dependencies in pigeon choice behavior under concurrent variable-interval schedules.
• To determine if momentary choice probabilities align with reinforcement probabilities.
• To assess the applicability of Markov chain models to describe molecular-level performance.

Main Methods:

• Collected peck-by-peck sequential data from four pigeons.
• Utilized concurrent variable-interval schedules.
• Analyzed response sequences to identify dependencies.
• Modeled performance using Markov chains.

Main Results:

• No significant sequential dependencies were found in response-by-response performance.
• Simple Markov chain models effectively described the molecular-level performance.
• Momentary changeover probabilities were proportional to overall relative reinforcement frequencies.
• Relative response frequencies closely approximated relative reinforcement frequencies, even without a changeover-delay procedure.

Conclusions:

• The hypothesis of momentary maximization of reinforcement probability was not supported by the data.
• Pigeon choice behavior in this context can be adequately modeled by Markov chains, indicating independence from immediate past responses.
• Reinforcement history, rather than momentary probability, may be a more significant factor in overall choice patterns.