A cognitive process model captures near-optimal confidence-guided waiting in rats

Area of Science:

• Behavioral neuroscience
• Cognitive psychology
• Decision science

Background:

• Decision-makers allocate more time to pursue rewards with higher confidence.
• Willingness to wait for delayed rewards is often used as a proxy for decision confidence.
• Previous research lacked clarity on how environmental statistics and internal variability affect optimal waiting behavior.

Purpose of the Study:

• To investigate how environmental statistics, including travel time, influence optimal waiting strategies.
• To model the impact of internal variability sources on decision-making during confidence-guided waiting.
• To develop a process model for decision-making in confidence-guided waiting tasks.

Main Methods:

• Trained rats on a confidence-guided waiting task.
• Derived mathematical expressions for optimal waiting incorporating environmental statistics.
• Developed an accumulation-to-bound process model to analyze internal variability.

Main Results:

• Rats exhibited longer wait times than theoretically optimal agents.
• Rat behavior closely matched optimal agents when travel times were constrained.
• Mean wait times increased with confidence, but variability did not, contrary to scalar invariant timing.
• Variability was best explained by fluctuations in the stopping bound.

Conclusions:

• The study provides a tractable process model for confidence-guided waiting.
• Environmental statistics and internal variability significantly influence decision processes.
• Internal variability's effect on waiting time is primarily mediated by the stopping bound, not scalar timing.