DescriptionIn operant conditioning, "extinction" refers to a procedure in which reinforcement is consistently withheld after conditioned responding. The "extinction effect" is observed when learned responding declines to its baserate, or extinguishes. Evidence suggests that the original association is preserved through extinction because conditioned responding can be restored. A hallmark of extinction is that it dissipates with time, and as subjects again exhibit the conditioned response. This phenomenon, spontaneous recovery, led Pavlov (1927) to conclude that learning is permanent. Extinction manipulations have been used in research with infants to eliminate undesirable behavior, to study emotion, and as test periods in instrumental learning preparations. However, it is unknown whether the properties of extinction are the same for human infants as for human adults and nonhuman animals.
In order to systematically characterize the extinction process early in ontogeny, 3-month-olds were first trained using the mobile conjugate reinforcement paradigm to kick to move an overhead mobile. Once the response was acquired, the extinction was presented and spontaneous recovery was assessed over the course of the normal retention interval for the task. The duration and temporal placement of the extinction phase were manipulated.
Infants did not reduce ongoing responding during the extinction manipulation, but the extinction effect was evident during subsequent testing. More than three minutes of nonreinforcement immediately following acquisition was effective at decreasing conditioned responding during subsequent long-term retention test. Paradoxically, when the extinction session was separated from acquisition by at least one day, 3 min was sufficient to cause a reduction in conditioned responding, while 6 min enhanced retention. No evidence of spontaneous recovery was observed in this study.