Extinction: It’s as Simple as Withholding the Functional Reinforcer . . . Or, Is It?

Photo by Gabby Orcutt on Unsplash

Harla B. Frank, M.S., BCBA

bSci21 Contributing Writer

The consequences of an act affect the probability of its occurring again. – B. F. Skinner

Extinction is the process of withholding a functional reinforcer of a target behavior in order to reduce or remove it completely; but it can also involve the interruption of the “response-reinforcer relation” (Castillo, Borrero, & Mendres-Smith, 2014, p. 617).  Extinction has been used to address a wide variety of behaviors, i.e., aggression; self-injury; inappropriate vocalizations; over-eating; gambling; drug and alcohol use/abuse; and many more.  While extinction is an effective procedure when implemented correctly, it is not without adverse side effects (Lerman & Iwata, 1995).

Most behavior analysts are familiar with the research, which spans decades, regarding extinction bursts and spontaneous recovery.  An extinction burst is the increase in frequency, intensity, or duration of the target behavior in the early stages of treatment (Lerman & Iwata, 1995).  If treatment integrity is maintained during these “bursts,” they are usually short lived.  Spontaneous recovery is the sudden reappearance of the target behavior after an extinction procedure has effectively reduced it – or removed it altogether (Lerman, Kelly, Van Camp, & Roane, 1999).  Again, if treatment integrity is maintained, this is a temporary situation.  Even though these adverse effects are usually short lived, consideration must be made regarding the use of extinction with aggression or self-injury due to the possibility of harm to the client or others (Lerman & Iwata, 1995).  While extinction bursts and spontaneous recovery are rather well known effects of extinction programs, there are other challenges that can arise, such as persistence; renewal; and resurgence.

Persistence is the phenomena that occurs when a target behavior resists extinction attempts due to a reinforcement history that involves high rates of reinforcement (MacDonald, Ahearn, Parry-Cruwys, Bancroft, & Dube, 2013).  Behavioral Momentum Theory postulates that higher rates of reinforcement increase a behavior’s resistance to extinction – or change (MacDonald et al., 2013; Podlesnik & Shahan, 2009).  MacDonald, Ahearn, Parry-Cruwys, Bancroft, and Dube (2013) found that dense reinforcement schedules characteristic of Functional Analyses (FA) increased resistance to extinction and increased extinction bursts.  These researchers suggest that switching from a dense continuous reinforcement (CRF) schedule to an intermittent (INT) schedule prior to treatment may mitigate the problem of resistance and extinction bursts (MacDonald et al., 2013).

Renewal is an intriguing phenomena in which a change in context (room size or location; color of walls or treatment materials; odor; sound; light; etc.) from the extinction context can bring about a re-emergence of the target behavior (Kelley, Liddon, Ribeiro, Greif, & Podlesnik, 2015; Podlesnik, Kelley, Jimenez-Gomez, & Bouton, 2017; Todd, Vurbic, & Bouton, 2014; Tsiang & Janak, 2006).  Testing for renewal usually consists of a “three-phase” process that includes training and reinforcing the target behavior in one context – say, at home; extinguishing the target behavior in the training context (at home) – or a novel context (such as, a clinic); and, while extinction is still in effect, returning the subject to the “original training context” (home) or a novel context (school) (Kelley et al., 2017, p. 675; Podlesnik et al., 2017; Todd et al., 2014).

One example of renewal is a student who’s attention-seeking and escape behaviors prompted his parents to withdraw him from school and continue his education at home.  During the period of home schooling, the behavior analyst – together with the parents, used an extinction program to remove the attention-seeking and escape behaviors.  Upon re-enrolling in public school in the fall, the target behaviors re-emerged. Re-entry into the context in which the target behaviors had been reinforced brought about the renewal of the behaviors.

Renewal can also occur as a result of the combination of context and cue (Tsiang & Janak, 2006).  Cues include the “visual, tactile, and olfactory” stimuli present in the training-reinforcement condition (Tsiang & Janak, 2006).  Tsiang and Janak (2006) conducted a study in which they taught mice to self-administer alcohol in a specific operant chamber.  Extinction was carried out in another chamber (Tsiang & Janak, 2006).  Reinstatement was tested by placing the mice back in the operant chamber in which they learned to lever-press for alcohol – without the cues that were present during training; placing the mice back in the extinction chamber with the cues used in the training-reinforcement chamber; and placing the mice back in the training-reinforcement chamber with the cues that were present during training (Tsiang & Janak, 2006).  Of the three reinstatement conditions, renewal of the target behavior was most pronounced in the treatment-reinforcement chamber with the cues (Tsiang & Janak, 2006).  The results of this study have important implications regarding the interaction of context and cues in the renewal of extinguished behavior.

Research into the renewal of extinguished behaviors have led to discoveries of possible approaches to mitigate this problem.  The first approaches hearken back to Stokes and Baer’s admonition to “program generalization” by training multiple exemplars; training loosely; using indiscriminable contingencies; and programming common stimuli (Stokes & Baer, 1977, p. 350).  Since renewal is specific to context – and context cues, one must train across many different contexts to promote generalization of treatment effects (Podlesnik et al., 2017).  Programming extinction cues could mitigate the problem of renewal by associating a specific stimuli with the extinction condition (Podlesnik et al., 2017).  For example, the subject could be provided with a green bracelet to wear during extinction (programming common stimuli).  When the subject leaves the extinction context and returns to the training-reinforcement context or a novel context, the bracelet will act as a cue that engaging in the target behavior while wearing the bracelet will not be reinforced (Podlesnik et al., 2017).  While the literature suggests having the therapists wear specific colored shirts or wristbands, it would be difficult to control the use of extinction cues in novel contexts in which the therapists are absent.  Having the subject wear the cue exercises more control over the novel contexts.  Another unique approach to mitigating renewal is the change from a specific, response-dependent reinforcer in the training-reinforcement condition to a response-independent, different reinforcer delivered on a time-based schedule during the extinction condition (Podlesnik et al., 2017).  Utilizing the response-independent reinforcer when the subject returns to the training-reinforcement context – while maintaining extinction, could lessen the potential for renewal (Podlesnik et al., 2017).

Resurgence refers to the emergence of a previously extinguished target behavior as a result of placing the alternative behavior on extinction (Lambert, Bloom, Samaha, Dayton, & Rodewald, 2015; Lieving, Hagopian, Long, & O’Connor, 2004).  Rarely are extinction programs used in isolation.  Typically, differential reinforcement of alternative behavior (DRA) is used to teach and reinforce an appropriate, alternative behavior that can serve the same – or similar – function as the target behavior.  The subject is reinforced for the expression of the alternative behavior with the same, or similar, reinforcer that maintained the target behavior (Lambert et al., 2015).  Variables that can increase the potential for resurgence include long reinforcement histories of, and high rates of reinforcement for, the target behavior (Lambert et al., 2015).  Interestingly, even brief periods of extinction of the alternative behavior can produce resurgence of the target behavior (Marsteller & St. Peter, 2014).  Let’s take a look at how resurgence might occur in an extinction plus DRA example.

Tiffany is 5-years-old and attends kindergarten at her local elementary school.  To gain access to desired tangibles (a doll in the play area), she screams loudly enough to be heard two classrooms away.  Her teacher would like to place the screaming behavior on extinction and utilize Functional Communication Training (FCT) to teach Tiffany how to appropriately request the desired toy.  The screaming is placed on extinction by withholding Tiffany’s access to the doll when she engages in screaming – the reinforcer maintaining the target behavior.  FCT is used to teach her an appropriate request, i.e., “May I play with the doll?”  Each time she uses the appropriate request, she is given access to the doll for seven minutes.  The frequency of screaming is reduced to zero within three days of implementing the extinction program and appropriate requesting is used each time she desires a specific tangible.  All is going well!  Then, the teacher comes down with the flu and a substitute teacher must take over for her in her absence.  The teacher failed to tell the substitute about the behavior intervention plan that she is implementing with Tiffany.  There were a few occasions when the substitute didn’t fulfill Tiffany’s request – in essence, appropriate requesting was placed on extinction.  Suddenly, the screaming re-emerged.  How can we mitigate the problem of resurgence in order to promote generalization and maintenance of treatment effects?  Research has given us some approaches that may help.

Lambert et al. (2015) tested a “serial DRA” approach in which multiple appropriate alternative behaviors were taught.  This provided the subjects with many behaviors that could be expressed if the first – or even the second – response was, inadvertently, unreinforced (Lambert et al., 2015).  The logic of this approach is that the subject may cycle through the appropriate responses to gain reinforcement before resorting to the target behavior (Lambert et al., 2015). Marsteller and St. Peter (2014) found that utilizing a response-independent, fixed-time (FT) reinforcement schedule that was linked to “reinforcement rates during and immediately preceding DRA” did not cause resurgence and maintained the alternative behaviors (pp.465-466).

While little research exists to support the prevention of resurgence by thinning the reinforcement delivery gradually during DRA, it may be possible to prevent resurgence by moving slowly from a continuous to an intermittent schedule (Ringdahl & St. Peter, 2017).  Alternative behaviors maintained on intermittent schedules of reinforcement may be more resistant to extinction than those maintained on continuous schedules (Cooper, Heron, & Heward, 2007).

Extinction programs are used extensively to reduce or eliminate unwanted behaviors.  One simply withholds the functional reinforcer maintaining the behavior and voila, the behavior is eliminated.  On the surface, extinction is deceptively simple.  However, failing to anticipate the problems of resistance, renewal, and resurgence can lead to serious setbacks in client progress – and perhaps expose the client to harm if the target behavior is self-injury.  Designing effective extinction programs that maintain over time and space requires building in solutions to problems you hope won’t happen.

By failing to prepare, you are preparing to fail. – Benjamin Franklin


Castillo, M. I., Borrero, J. C., & Mendres-Smith, A. E. (2014). Evaluating the presence versus absence of the reinforcer during extinction. Journal of Applied Behavior Analysis, 47, 617-622.

Cooper, J. O., Heron, T. E., & Heward, W. L. (2007). Applied behavior analysis (2nd ed.). Upper Saddle River, NJ: Pearson.

Kelley, M. E., Liddon, C. J., Ribeiro, A., Greif, A. E., & Podlesnik, C. A. (2015). Basic and translational evaluation of renewal of operant responding. Journal of Applied Behavior Analysis, 48, 390-401.

Lambert, J. M., Bloom, S. E., Samaha, A. L., Dayton, E., & Rodewald, A. M.  (2015). Serial alternative response training as intervention for target response resurgence. Journal of Applied Behavior Analysis, 48, 765-780.

Lerman, D. C., & Iwata, B. A.  (1995). Prevalence of the extinction burst and its attenuation during treatment. Journal of Applied Behavior Analysis, 28, 93-94.

Lerman, D. C., Kelley, M. E., Van Camp, C. M., & Roane, H. S. (1999). Effects of reinforcement magnitude on spontaneous recovery. Journal of Applied Behavior Analysis, 32, 197-200.

Lieving, G. A., Hagopian, L. P., Long, E. S., & O’Connor, J. (2004). Response-class hierarchies and resurgence of severe problem behavior. The Psychological Record, 54, 621-634.

MacDonald, J. M., Ahearn, W. H., Parry-Cruwys, D., Bancroft, S., & Dube, W. V. (2013). Persistence during extinction: Examining the effects of continuous and intermittent reinforcement on problem behavior. Journal of Applied Behavior Analysis, 46, 333-338.

Marsteller, T. M., & St. Peter, C. C. (2014). Effects of fixed-time reinforcement schedules on resurgence of problem behavior. Journal of Applied Behavior Analysis, 47, 455-469.

Podlesnik, C. A., Kelley, M. E., Jimenez-Gomez, C., & Bouton, M. E. (2017). Renewed behavior produced by context change and its implications for treatment maintenance: A review. Journal of Applied Behavior Analysis, 50, 675-697.

Podlesnik, C. A., & Shahan, T. A. (2009). Behavioral momentum and relapse of extinguished operant responding. Learning & Behavior, 37, 357-364.

Ringdahl, J. E., & St. Peter, C. (2017). Resurgence: The unintended maintenance of problem behavior. Education and Treatment of Children, 40, 7-26.

Stokes, T. F., & Baer, D. M. (1977). An implicit technology of generalization. Journal of Applied Behavior Analysis, 10, 349-367.

Todd, T. P., Vurbic, D., & Bouton, M. E. (2014). Mechanisms of renewal after the extinction of discriminated operant behavior. Journal of Experimental Psychology: Animal Learning and Cognition, 40, 355-368.

Tsiang, M. T., & Janak, P. H. (2006). Alcohol seeking in C57BL/6 mice induced by conditioned cues and contexts in the extinction-reinstatement model. Alcohol, 38, 81-88.

Harla Frank, M.S., BCBA earned her Master’s degree in Psychology, with an emphasis in Applied Behavior Analysis, from Florida State University.  Since receiving her certification as a Board Certified Behavior Analyst (BCBA) in 2007, she has worked primarily with children and young adults on the Autism Spectrum, but has also worked with adults with a variety of diagnoses and needs. She has served as an expert witness for Applied Behavior Analysis (ABA) in the Colorado court system and has had the privilege of providing “ABA approaches” training to foster care staff and families.

Since 2010, Harla has taught ABA course sequences, as well as general psychology courses, for Kaplan University.  Currently, she also contracts with a pediatric home healthcare company in Denver to provide ABA therapy to children with a variety of diagnoses. You can contact her at hfrank@kaplan.edu.

Print Friendly, PDF & Email

4 Comments on "Extinction: It’s as Simple as Withholding the Functional Reinforcer . . . Or, Is It?"

  1. Zuzana Maštenová | February 3, 2018 at 6:26 am | Reply

    Excellent article, very useful, but can you please explain following, because I’m obviously missing something here. Doesn’t this contradict – Behavioral Momentum Theory postulates that higher rates of reinforcement increase a behavior’s resistance to extinction – or change (MacDonald et al., 2013; Podlesnik & Shahan, 2009). Alternative behaviors maintained on intermittent schedules of reinforcement may be more resistant to extinction than those maintained on continuous schedules (Cooper, Heron, & Heward, 2007). Thank you!

    • Hi Zuzana,

      Thank you for your kind comments! You are definitely correct that higher rates of reinforcement increase a behavior’s resistance to extinction. When discussing the alternative, replacement behavior, we would begin with a continuous schedule of reinforcement and, once it reaches a desired and reliable state, we would want to slowly transition to an intermittent schedule of reinforcement. Basically, the understanding is that intermittent schedules are more resistant to extinction than continuous schedules due to the fact that it is almost immediately apparent to the individual whose behavior has been maintained on continuous reinforcement that the contingency has changed. Upon implementation of an extinction program for a behavior that has been maintained on continuous reinforcement, with the very next expression of the target behavior, the extinction program is felt. When behaviors are maintained on intermittent schedules, it may take a period of time before the individual realizes a change in contingency has occurred. Because the individual has expressed the target behavior a number of times without reinforcement, and only some of expressions of the behavior have obtained reinforcement, the extinction program, while officially “on,” may not actually influence the behavior the first time the reinforcer is withheld due to the seemingly random nature of the reinforcement.
      I apologize if I was a bit confusing. 🙂

  2. James Chastain, BCBA-D | February 5, 2018 at 10:21 am | Reply

    bSci21 has posted some pretty good reviews, but this one is pretty GREAT! Thanks for such a clear description of terms/processes, with potential clinical solutions.

    • Hi James,
      Thank you so much for your extremely kind and gracious comment! I appreciate you reading my piece!

Leave a comment

Your email address will not be published.