Contingency discrimination training and resurgence: Effects of reduced extinction session durations
Cut extinction time in half and you can wipe out resurgence.
01Research in Context
What this study did
Shahan et al. (2026) asked a simple question: can we spend less time in extinction and still stop problem behavior from coming back? They ran a lab study using contingency-discrimination training. Sessions alternated between reinforcement and extinction. The team kept the reinforcement half at 30 minutes and cut the extinction half to 15, 30, or even 5 minutes.
They tested different off-to-on ratios: 1:2 (15 min off, 30 min on), 1:1 (30 min each), and two shrinking schedules that started long and got shorter. The goal was to see which ratio wiped out resurgence.
What they found
Shorter extinction worked. A 1:2 ratio (half as much extinction) already slashed resurgence. A 1:1 ratio or the shrinking schedules erased it completely. Target responses stayed low and total extinction time dropped.
The best part: less extinction did not mean more problem behavior. Responses did not creep back up when the schedule tightened.
How this fits with other research
Briggs et al. (2018) and Muething et al. (2021) saw the opposite in clinics. They reported resurgence in 76% and 41% of FCT thinning steps. Shahan’s lab result now shows the swing factor: session ratio. Keep extinction brief and frequent and you can dodge the spike they measured.
Nist et al. (2023) tried giving rats five alternative responses instead of one and still got full resurgence. That looks like a clash, but the methods differ. Nist used expanded DRA; Shahan used timed CDT. The contradiction tells us it is not the number of choices that matters, it is how you schedule the breaks from extinction.
Trask et al. (2018) already showed that mixing in free reinforcers during tests cuts resurgence. Shahan extends that idea by shrinking the whole extinction block, giving us a second, easier lever to pull.
Why it matters
If you run FCT or any DRA with extinction, plan your thinning steps like Shahan did. Keep the off-session short. Try 15 minutes of extinction for every 30 minutes of reinforcement and watch the data. You may stop the post-thinning spike Briggs and Muething warn about without adding extra hours of therapy. Less extinction, same gain — a simple schedule tweak you can test next week.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Plot your next thinning step with a 1:2 off-to-on ratio and track resurgence for one week.
02At a glance
03Original abstract
Resurgence is an increase in a suppressed target behavior following a worsening of conditions for a more recently reinforced alternative behavior. Prior research shows that exposure to equal‐duration sessions of alternative reinforcement availability versus unavailability during treatment (i.e., contingency discrimination training; CDT) reduces resurgence. Clinically, minimizing exposure to extinction while maintaining the resurgence‐mitigating effects of CDT would be desirable. This experiment examined the effects of reduced off‐session durations by exposing groups of rats to different ratios of off:on session durations: All On (0 min: 30 min), CDT 1:1 (30 min: 30 min), CDT 1:2 (15 min: 30 min), CDT 1:6 (5 min: 30 min), and CDT escalate (i.e., [Esc] off‐session duration increased across sessions). Resurgence decreased exponentially with “off” session duration, with CDT 1:2 reducing resurgence and both CDT 1:1 and CDT Esc eliminating resurgence while generating control of alternative behavior by the prevailing reinforcement contingencies, without increasing the total number of target responses during treatment. Resurgence as choice in context theory described the data well with the assumption that the effect of the signaling properties of the reinforcement contingencies themselves increases linearly with the off:on duration ratio, as is true with the S−/S+ ratio in other discriminations.
Journal of the Experimental Analysis of Behavior, 2026 · doi:10.1002/jeab.70072