Response-bout analysis of interresponse times in variable-ratio and variable-interval schedules

Area of Science:

• Behavioral Neuroscience
• Animal Behavior
• Reinforcement Learning

Background:

• Variable-ratio (VR) and variable-interval (VI) reinforcement schedules are fundamental concepts in operant conditioning.
• Understanding response patterns under different schedules is crucial for behavioral analysis.
• Interresponse time (IRT) distributions offer detailed insights into the temporal structure of behavior.

Purpose of the Study:

• To quantitatively analyze the differences in response patterns between VR and VI reinforcement schedules.
• To model and compare interresponse time distributions generated by rats' lever pressing behavior under VR and VI schedules.

Main Methods:

• Eight rats' lever pressing data were collected under yoked VR and VI reinforcement schedules.
• A comprehensive model comparison was conducted, evaluating 30 different distribution models (single and multi-component) for IRT data.
• The Akaike Information Criterion (AIC) and visual inspection were used to determine the best-fitting model.

Main Results:

• All rats exhibited higher overall response rates under the VR 30 schedule compared to the yoked VI schedule.
• A four-component Log-Normal distribution model (L4) provided the best fit to the interresponse time data.
• Parameter estimates from the L4 model indicated that the VR-VI response rate difference is attributable to variations in short IRTs, reflecting differences in within-bout response patterns.

Conclusions:

• The observed difference in response rates between VR and VI schedules is not due to an altered overall motivation to respond.
• Instead, the discrepancy arises from the distinct response patterns elicited by each schedule, specifically concerning the patterning of responses within bouts.
• This study highlights the utility of detailed IRT distribution modeling for discerning subtle differences in behavioral control by reinforcement schedules.