A response duration schedule: effects of training, extinction, and deprivation.

Researchers taught animals to press and hold a lever for 40 seconds. A click and food arrived only when the animal met the 40-second mark. They then stopped the click and food to see if the timing would vanish. Finally they brought the click and food back to test quick re-learning.

During training most holds clustered tightly around 40 seconds. When the click and food stopped the neat 40-second peak disappeared. Once the click and food returned the 40-second pattern came back fast.

KELLEHER et al. (1963) showed that more continuous reinforcement before extinction creates bigger bursts when reinforcement stops. G et al. found no burst, just a flat loss of the timed pattern. The difference is topography: T measured rate bursts; G measured duration precision.

Smith et al. (2022) later quantified resurgence, showing longer reinforcement history makes return of old behavior stronger. G’s rapid retraining aligns with this: the 40-second pattern bounced back quickly after a long reinforced history.

Staddon (1970), published the same year, showed longer food delivery stretches pauses on fixed-interval schedules. Both studies reveal that the length of the reinforcer itself can shape timing, but G focused on holding the response rather than pausing after it.

If you need a client to sustain an action for an exact time—like holding a toothbrush for two minutes—pair the end of the duration with a salient click or beep and immediate reinforcement. Remove that signal and the precise timing will fade. Bring the signal back and the skill returns quickly, saving retraining time.