A comparison of procedures for unpairing conditioned reflexive motivating operations

The team asked: how do you turn off a stimulus that now makes problem behavior spike? They compared two ways to break that learned link.

Six kids with different diagnoses watched a red light. The light used to mean "work now, no snacks." The researchers tried (1) plain extinction—no snacks ever—or (2) noncontingent unpairing—snacks given before the light, then the light still showed. They flipped the two methods across days to see which calmed the kids faster.

Giving snacks before the light (noncontingent unpairing) quickly dropped the kids' escape attempts and vocal protests. The effect stayed steady across days.

Plain extinction worked only half the time. On some days behavior fell, on others it bounced back or got worse. The team saw resurgence and side bursts.

Allison et al. (2012) ran a similar matchup with food selectivity. They also found noncontingent reinforcement beat an extinction package and parents liked it more. Together the two studies show the same pattern across very different problems—eating and CMO-R.

Rogers-Warren et al. (1976) saw the same edge four decades earlier in monkeys: adding any form of reinforcement during extinction gave faster suppression than extinction alone. Kettering's data simply update that lesson for human clinical settings.

Bensemann et al. (2015) warn that extinction packages like DRO can accidentally strengthen other odd behaviors. Kettering's spotty results with extinction echo that warning—without added reinforcement the procedure wobbles.

If you have a client who melts down the moment the demand card appears, the card may now be a CMO-R. Instead of just blocking escape, slip a bite of preferred food or a short video right before you flash the card. Do it every time for several sessions. The card loses its bite, problem behavior drops, and you skip the extinction burst. Keep data—if gains stall, thin the free snack schedule slowly.