Learning the opportunity cost of time in a patch-foraging task.

Area of Science:

• Cognitive Neuroscience
• Behavioral Economics
• Decision Science

Background:

• Most decision research focuses on simultaneous choices, not serial encounters where individuals decide whether to exploit or search.
• Foraging tasks, involving sequential decisions, have been studied in animals, but human psychological processes and learning mechanisms remain unclear.

Purpose of the Study:

• To investigate how humans learn in a foraging task requiring decisions between harvesting a depleting resource and switching to a replenished one.
• To compare human learning strategies against the marginal value theorem (MVT) and temporal-difference (TD) learning models.

Main Methods:

• Two experiments were conducted where participants decided whether to harvest or search in a foraging task.
• Opportunity cost was varied across blocks, and participants' behavioral adjustments were analyzed.
• Choice adjustments were compared to MVT's threshold-learning rule and TD learning.

Main Results:

• Participants successfully adjusted their harvesting decisions to changes in opportunity cost.
• Trial-by-trial decisions were better explained by the MVT threshold-learning rule than by TD learning.
• This suggests a distinct computational mechanism for learning in switching tasks.

Conclusions:

• Human learning in foraging tasks relies on a distinct mechanism, aligning with the MVT's opportunity cost threshold estimation.
• Findings differentiate learning in sequential switching tasks from traditional decision tasks.
• This research connects foraging behavior to broader neuroscience research on average reward rates.