Landmine-detection rats: an evaluation of reinforcement procedures under simulated operational conditions.

Five rats hunted for landmines in a sandbox. The trainers removed all real mines. They marked empty spots as if mines were there. Then they secretly soaked the soil with TNT scent and later cleaned it up.

When a rat dug at a marked spot, it got no food. When it dug where TNT scent had been, it got a pellet. This let the rats earn food even though no real explosives were present.

All five rats kept finding every fake mine. Their accuracy stayed high even though digging at the old mine spots no longer paid off.

The TNT-scent trick gave the rats a way to win food. Because food still showed up, the animals kept working hard.

Donahoe et al. (2000) showed that non-contingent food can hide extinction. The rat study uses the same idea. Food still appears, so the behavior keeps going.

Nighbor et al. (2018) found that keeping the same sights and sounds boosts resurgence. The rat team kept the same sand, smells, and lights. Yet resurgence did not happen because the TNT spots gave new rewards.

Hanley et al. (1997) removed the person to stop self-injury. The rat team removed the real mine but kept the reward. Both studies show that taking away the real cause while keeping some payoff prevents the behavior from dropping.

You can keep a skill alive even when the natural payoff is gone. Plant fake treats, praise, or tokens in places where the real reward used to be. This keeps clients working while you fade the original reinforcer. Try it when you need to maintain safety responses, academic accuracy, or compliance after the main prize is removed.