Effects of variable-interval value and amount of training on stimulus generalization

Area of Science:

• Behavioral Psychology
• Animal Cognition
• Operant Conditioning

Background:

• Stimulus generalization is a fundamental concept in learning theory.
• Previous research suggests generalization gradients may be influenced by training duration and reinforcement schedules.
• Understanding these factors is crucial for predicting behavioral responses to novel stimuli.

Purpose of the Study:

• To investigate the impact of variable-interval (VI) reinforcement schedules and training duration on stimulus generalization in pigeons.
• To determine if prolonged training alters generalization gradients differently depending on the reinforcement schedule.
• To examine the relationship between reinforcement rate and the shape of generalization gradients.

Main Methods:

• Pigeons were trained on a key-pecking task with food reinforcement under VI schedules (30s, 120s, 240s) and varying session numbers (10 sessions or until response stabilization).
• Generalization was tested using different light wavelengths or line orientations in extinction.
• Absolute and relative generalization gradients were analyzed to assess stimulus generalization patterns.

Main Results:

• Generalization gradients were similar across different VI schedules after 10 training sessions.
• For the VI 30-s schedule, generalization gradients were consistent regardless of training duration.
• For the VI 240-s schedule, prolonged training resulted in upward-shifted absolute gradients and flatter relative gradients compared to shorter training.

Conclusions:

• The shape of stimulus generalization gradients is not solely determined by short-term training.
• Prolonged training can alter stimulus generalization, but these effects are dependent on the specific reinforcement schedule.
• Reinforcement rate during training plays a critical role in shaping stimulus generalization under extended training conditions.