How do animals weigh conflicting information about reward sources over time? Comparing dynamic averaging models

Area of Science:

• Behavioral neuroscience
• Animal cognition

Background:

• Optimal foraging theory posits animals maximize food intake.
• Mechanisms for tracking behavioral alternatives and choice remain unclear.
• Spontaneous recovery of choice, where past reward distributions influence current behavior after a hiatus, is a key phenomenon.

Purpose of the Study:

• To investigate spontaneous recovery of choice in rats.
• To compare the predictive accuracy of different computational models for animal decision-making.
• To test how rats integrate past and recent reward information.

Main Methods:

• Three free-operant experiments using a serial reversal design in rats.
• Concurrent variable interval schedules were used to bait two responses (A and B).
• A delay period was introduced before testing choice preference without reward.

Main Results:

• Data strongly supported the occurrence of spontaneous recovery of choice.
• The observed form and extent of recovery were inconsistent with tested computational models.
• Rats showed a reversion to behavioral allocation based on more distant past reward distributions.

Conclusions:

• Spontaneous recovery of choice is a robust phenomenon in rats.
• Existing computational models, including exponentially weighted moving average and Temporal Weighting Rule, do not fully account for the observed recovery.
• Further research is needed to refine decision rules and valuation functions in foraging models.