An extinction burst is a temporary increase in the frequency or intensity of a behavior when the reward that previously maintained it is suddenly removed. It’s a well-documented phenomenon in behavioral psychology: before a behavior fades away, it often gets worse first. Understanding this pattern is essential for anyone trying to break a habit, change a child’s behavior, or simply make sense of why doing the right thing sometimes seems to backfire.
How Extinction Works
To understand the burst, you first need to understand extinction itself. In operant conditioning, behaviors persist because they’re reinforced. A child whines and gets a cookie. An employee complains and gets attention. A rat presses a lever and gets a food pellet. The behavior continues because it works.
Extinction is what happens when you stop delivering that reward. The reinforcing consequence that kept the behavior going is no longer provided, and over time, the behavior decreases. But “over time” is the key phrase. The decline isn’t immediate, and the path from reinforced behavior to no behavior passes through a predictable rough patch: the extinction burst.
What the Burst Looks Like
When a behavior suddenly stops working, most organisms don’t shrug and move on. They try harder. The behavior increases in frequency, intensity, or both. Think of it like pressing an elevator button. If the elevator doesn’t arrive, most people push the button several more times, often harder than before. Only after repeated failed attempts do they stop pressing and take the stairs.
A classic example involves a child who has learned that throwing a tantrum gets them out of brushing their teeth. If a parent suddenly stops giving in, the child doesn’t calmly accept the new rule. Instead, the tantrums initially get bigger, louder, and longer. The child is essentially testing whether the old strategy still works if applied with more force.
This same pattern shows up across contexts. A person who stress-eats may eat more intensely in the first days after committing to stop. Someone who checks their phone compulsively may feel an even stronger pull toward the screen when they first try to resist. The burst is the behavior’s last stand before it starts to weaken.
What Happens in the Brain
The neurological picture centers on dopamine, the brain chemical tied to reward prediction. When your brain expects a reward and gets it, dopamine neurons fire in a predictable pattern. When the expected reward doesn’t arrive, those neurons generate what researchers call a negative prediction error: a signal that something expected is missing.
Research published in The Journal of Neuroscience found that during extinction, dopamine neurons develop inhibitory responses, essentially learning to suppress the old expectation. This isn’t simply forgetting. It’s new learning layered on top of the original conditioning. The brain builds an inhibitory signal that actively suppresses the old reward association, rather than erasing it.
This distinction matters because the original learning and the new extinction learning decay at different rates. The extinction learning (the “stop expecting a reward” signal) fades faster than the original conditioning. This is why a behavior that seemed fully extinguished can sometimes reappear after a break, a phenomenon called spontaneous recovery.
How Long the Burst Lasts
The extinction burst is generally short-lived. Research in the Journal of the Experimental Analysis of Behavior found that when examining clinical cases, evidence of a burst appeared in roughly 36% to 62% of cases where extinction was used alone. Many studies define and measure the burst within the first session or the first few sessions of extinction, and the spike typically resolves quickly once the person or animal confirms the reward truly isn’t coming back.
The exact timeline varies depending on the behavior, how long it was reinforced, and how consistently the reward is now withheld. A habit reinforced for years may produce a more intense burst than one reinforced for weeks. But the core pattern is consistent: a sharp spike followed by a decline.
The Biggest Mistake: Giving In
Here’s where extinction bursts cause the most real-world damage. When the behavior escalates during the burst, the temptation to give in is enormous. The child is screaming louder than ever, the craving feels unbearable, the urge to check your phone becomes overwhelming. Giving in at this point doesn’t just reset the process. It makes the problem worse.
When you reinforce a behavior during a burst, you teach the organism that escalation works. The next time extinction is attempted, the behavior will likely escalate even faster and higher, because the escalation itself was rewarded. Research shows that this creates a pattern that may be difficult or even impossible to replicate in its original form. As behavioral researcher Murray Sidman noted, the initial transition from reinforcement to extinction may be “unrecoverable” once reinforcement is reintroduced and then removed again, meaning the second attempt at extinction can produce a very different and more challenging pattern of behavior.
There’s an additional wrinkle related to reinforcement schedules. Behaviors that were reinforced every single time (continuous reinforcement) actually produce more pronounced extinction bursts than behaviors that were reinforced only sometimes (intermittent reinforcement). This seems counterintuitive, but it makes sense: if a behavior always worked before, the sudden absence of reward is a more dramatic change. If the behavior only worked sometimes, the initial non-reward during extinction doesn’t feel as different from normal.
How To Manage an Extinction Burst
The most important strategy is simply knowing the burst is coming and committing to wait it out. Consistency is everything. If you’re trying to extinguish a behavior, whether in yourself, a child, or even a pet, the reward must stop completely and stay stopped through the burst phase.
A more structured approach used in clinical settings is called differential reinforcement of alternative behavior. The idea is straightforward: while you stop rewarding the unwanted behavior, you simultaneously teach and reward a replacement behavior that serves the same function. If a child throws tantrums to get out of a task, you stop rewarding the tantrum while teaching and reinforcing an appropriate way to ask for a break, like handing over a picture card or using a simple phrase. The replacement behavior gets the same reward the problem behavior used to get, giving the person a functional path to what they need.
This approach works better than extinction alone because it doesn’t just remove the old behavior; it channels the underlying need into something acceptable. Research shows that combining this strategy with full extinction of the problem behavior produces better outcomes than reinforcing alternatives while still occasionally giving in to the old behavior.
Spontaneous Recovery After Extinction
Even after a behavior has fully extinguished, it can reappear. Spontaneous recovery happens when enough time passes that the brain’s extinction learning has partially decayed, allowing the older, original conditioning to reassert itself. In research settings, behaviors that were fully extinguished showed significant recovery when tested after a seven-day break.
This doesn’t mean extinction failed. The recovered behavior is typically weaker than the original and extinguishes faster the second time. But it catches people off guard. A child who stopped throwing tantrums weeks ago might suddenly try one again. A person who kicked a craving might feel it resurface after a long gap. Recognizing this as a normal part of the process, not a sign of failure, makes it easier to stay consistent and let the behavior fade again.
The neurological explanation is elegant: because the brain’s extinction learning decays faster than the original conditioning, the suppression weakens before the underlying association does. The old behavior briefly peeks through the gap before the extinction learning re-establishes itself.