Positional inference in rhesus macaques.

Area of Science:

• Cognitive Science
• Animal Behavior
• Neuroscience

Background:

• Understanding serial relationship learning is key to cognitive abilities.
• Transitive inference (TI) is a heuristic for serial learning studied across species.
• Previous research focused on TI in various serial learning tasks.

Purpose of the Study:

• To investigate how monkeys learn and generalize serial relationships using transitive inference.
• To explore the role of positional inference alongside transitive inference in decision-making.
• To test a spatial model explaining both transitive and positional inferences.

Main Methods:

• Monkeys learned the order of 7-item photographic stimulus lists via trial and error.
• Subjects were tested on novel 'derived' lists combining stimuli from different training lists.
• Behavioral responses to novel pairs were analyzed to assess inference strategies.

Main Results:

• Monkeys showed strong preferences for novel stimulus pairs, even across different training lists.
• Preferences aided performance when pair order was congruent with training ranks.
• Performance dropped below chance when pair order was incongruent with training ranks.

Conclusions:

• Monkeys utilize both transitive inference and positional inference for serial learning.
• Positional inference complements transitive inference in choices between novel stimuli.
• A spatial model accurately predicts behavioral responses and errors based on stimulus position and uncertainty.