Effects of schedules of reinforcement on pouched rats' performance in urban search-and-rescue training.

The team trained pouched rats to find a target in a mock disaster site.

Each rat had to sniff out the spot, then run back to the trainer for food.

They tested what happened when they removed food for finding the target, for returning, or for both.

No food for finding made the rats take longer to locate the target.

No food for returning made them sprint back faster.

You can speed up the part you reward, but the unrewarded part slows down.

Neuringer et al. (1968) showed pigeons divide time to match reward odds. The rat data fit: change the payoff point and time shifts with it.

Landon et al. (2002) found both quick and slow choice shifts after each reward. The rat study adds a real-world twist: you can pick which link—search or return—gets the quick fix.

Cohen et al. (1993) warned that reward rate only protects behavior in certain set-ups. Here, removing reward at different chain links proves the point: protection stays only where payoff stays.

In any chained task, decide which step must stay strong and put the reinforcer there. If you need a faster return, deliver the treat at the handler. If you need faster detection, pay at the target site. Match the schedule to the mission-critical link.