Stimulus- and response-reinforcer contingencies in autoshaping, operant, classical, and omission training procedures in rats.

The team worked with rats in a lab. Each rat got one of five lever-training plans.

Plans were autoshaping, regular operant, classical pairing, omission, or random control. Later every rat faced the same omission phase where lever presses no longer produced food.

All training types quickly produced lever pressing. When food stopped, pressing dropped the same amount in every group.

Surprise: the rats still touched the lever even though it no longer paid off. The movement stuck around even when the payoff was gone.

Locurto et al. (1976) ran a similar rat study one year earlier. They showed that rats with a past history of omission training later learned new autoshaping more slowly. Atnip (1977) moved that history inside one experiment and still saw equal drops in pressing, proving that the learning history effect acts on new acquisition, not on later extinction within the same cues.

Farrant et al. (1998) later tweaked reinforcement rate and delay. They found that lean, brief delays can actually boost early response rates compared with continuous payoff. W’s equal drop across groups fits this: once the payoff ends, past richness no longer matters.

Keely et al. (2007) showed rats still track which lever earns delayed food even without signals. Their data help explain why W’s rats kept touching the lever; the topography was still under stimulus control even when the food rule changed.

For BCBAs the key point is topography versus contingency. A client may keep touching, flapping, or tapping because the physical form was strongly reinforced, even when the payoff is now zero. Before you pick an intervention, probe both the response form and the current payoff. If the form is strong but the payoff is gone, consider response-blocking or differential reinforcement of an alternative that produces the same sensory consequence. This move can break the topography loop without long extinction waits.