Elsevier

Biological Psychiatry

Volume 52, Issue 10, 15 November 2002, Pages 976-986
Biological Psychiatry

Review
Context, ambiguity, and unlearning: sources of relapse after behavioral extinction

https://doi.org/10.1016/S0006-3223(02)01546-9Get rights and content

Abstract

There is now ample evidence that extinction, the loss of learned performance that occurs when a Pavlovian signal or an instrumental action is repeatedly presented without its reinforcer, does not reflect a destruction of the original learning. This article summarizes the evidence and extends and updates earlier reviews. The main alternative to “unlearning” is the idea that extinction (as well as other retroactive interference processes, including counterconditioning) involves new learning that is stored along with the old. One consequence is that the Pavlovian signal or instrumental action has two available “meanings” and thus has the properties of an ambiguous word: its current meaning (and the resulting behavioral output) depends on what the current context retrieves. Contexts can be provided by a variety of background stimuli, including the physical environment, internal drug state, and time. The second thing learned (e.g., extinction, counterconditioning) seems especially dependent on the context for retrieval. A variety of evidence is consistent with this analysis, which highlights several important sources of relapse after extinction. The article concludes with several issues for future research, among them the question of how we can optimize extinction and other putative “unlearning” treatments so as to prevent the various forms of relapse discussed here.

Introduction

For at least the last decade, a consensus has been building that extinction, the loss of performance that occurs when a Pavlovian signal or an instrumental action is no longer paired with a reinforcer, does not destroy the original learning. The idea was actually present in Pavlov’s writing (Pavlov 1927) and in that of others who followed him (e.g., Konorski 1948, Pearce and Hall 1980, Wagner 1981. Nonetheless, the notion that extinction causes unlearning is so pervasive that it itself seems difficult to unlearn. Thus, it is second nature to think that exposure therapies are effective at reducing anxiety disorders because they destroy the learning that led to them. Similarly, it seems natural to accept certain assumptions in theories of learning and memory that imply new learning will destroy the old (e.g., McClelland and Rumelhart 1985, Rescorla and Wagner 1972 see McCloskey and Cohen 1989).

Evidence that extinction is not unlearning has been reviewed in other places (e.g., Bouton 1988, Bouton 2000, Bouton and Swartzentruber 1991, Falls 1998, Rescorla 2001. In this article, I briefly summarize and update that evidence in a way that illustrates what I think is the most reasonable alternative to the destruction hypothesis. Extinction does not destroy the first-learned information but instead reflects new learning. The result is that the signal (or instrumental action) acquires a second “meaning” that is available along with the first. In this sense, the current meaning of the signal or action is ambiguous Bouton 1984, Bouton and Bolles 1985. As with other ambiguous stimuli, such as ambiguous words, its current meaning is determined by the current context. Just as the word “Fire!” evokes different reactions in a movie theater and a shooting gallery, an extinguished Pavlovian signal will evoke different reactions in different contexts. The view rejects unlearning, accepts behavioral instability as a potential consequence of extinction, and highlights several possible sources of relapse.

Section snippets

Four mechanisms of relapse

Research on extinction has uncovered at least four phenomena indicating that it does not destroy the original learning. These are summarized in this section and in Table 1. All are potential mechanisms of relapse, and all indicate the importance of context in controlling performance after extinction has occurred. Most of the research has focused on extinction in Pavlovian conditioning, in which a conditional stimulus (CS, e.g., a tone) is first paired with a motivationally significant

Other forms of retroactive interference

The idea that extinction does not destroy the original learning is compatible with a great deal of other research. For example, extinction is just one example of several retroactive interference paradigms in which new learning is introduced to replace old learning (Bouton 1993). Interestingly, the ideas just described may generalize to all examples of interference (Bouton 1993). For example, in “counterconditioning,” a rat might first receive tone–shock pairings and then pairings of the tone

What kinds of cues are contexts?

When one translates animal research into clinical practice, it is natural to ask about the boundaries of the concept or definition of context. As I suggested earlier, it is useful to think that a variety of different background cues might play the role of context in extinction. In fact, research on context effects in both animal and human learning and memory suggests that a wide variety of stimuli might potentially function this way (Table 2). Drugs, for example, can provide salient internal

Conclusion

A large body of research suggests that extinction and retroactive interference do not cause unlearning. It may be tempting to conclude that therapeutic treatments are therefore doomed, but this is not the case. Although it is probably best to assume that the original information is always available, recognition of that fact will promote better, more realistic therapies. The ambiguity framework suggests that the trick will be to develop methods that promote ubiquitous retrieval of the second

Acknowledgements

Preparation of this article was supported by Grant No. RO1 MH64847–01 from the National Institute of Mental Health. The author thanks Laure Pain and Jay Sunsay for their comments.

Aspects of this work were presented at the conference, “Learning and Unlearning Fears: Preparedness, Neural Pathways, and Patients,” held March 21, 2002 in Austin, TX. The conference was supported by an unrestricted educational grant to the Anxiety Disorders Association of America (ADAA) from Wyeth Pharmaceuticals, and

References (103)

  • M. McCloskey et al.

    Catastrophic interference in connectionist networksThe sequential learning problem

  • JL Mystkowski et al.

    Treatment context and return of fear in spider phobia

    Behav Ther

    (2002)
  • S. Nakajima et al.

    Renewal of extinguished lever-press responses upon return to the training context

    Learn Motiv

    (2000)
  • J.B. Nelson

    Context specificity of excitation and inhibition in ambiguous stimuli

    Learn Motiv

    (2002)
  • C.A. Peck et al.

    Context and performance in aversive-to-appetitive and appetitive-to-aversive transfer

    Learn Motiv

    (1990)
  • S.J. Rachman

    The return of fearReview and prospect

    Clin Psychol Rev

    (1989)
  • B.I. Rodriguez et al.

    Context-specificity of relapseEffects of therapist and environmental context on return of fear

    Behav Res Ther

    (1999)
  • M.K. Rowe et al.

    Effects of an expanding-spaced vs. massed exposure schedule on fear reduction and return of fear

    Behav Res Ther

    (1998)
  • J.C.I. Tsao et al.

    Timing of treatment and return of fearEffects of massed uniform- and expanding-spaced exposure schedules

    Behav Ther

    (2000)
  • D.L. Walker et al.

    The role of amygdala glutamate receptors in fear learning fear-potentiated startle and extinction

    Pharmacol Biochem Behav

    (2002)
  • S.T. Ahlers et al.

    Administration of dexamethasone prior to training blocks ACTH-induced recovery of an extinguished avoidance response

    Behav Neurosci

    (1985)
  • A.G. Baker et al.

    Contextual conditioning and reinstatement of extinguished instrumental responding

    Q J Exper Psychol

    (1991)
  • M.E. Bouton

    Differential control by context in the inflation and reinstatement paradigms

    J Exper Psychol: Anim Behav Processes

    (1984)
  • M.E. Bouton

    Context time and memory retrieval in the interference paradigms of Pavlovian learning

    Psychol Bull

    (1993)
  • M.E. Bouton

    Conditioning remembering and forgetting

    J Exper Psychol: Anim Behav Processes

    (1994)
  • M.E. Bouton

    A learning theory perspective on lapse relapse and the maintenance of behavior change

    Health Psychol

    (2000)
  • M.E. Bouton et al.

    Role of conditioned contextual stimuli in reinstatement of extinguished fear

    J Exper Psychol: Anim Behav Processes

    (1979)
  • M.E. Bouton et al.

    Contexts event-memories and extinction

  • M.E. Bouton et al.

    Time and context effects on performance in a Pavlovian discrimination reversal

    J Exper Psychol: Anim Behav Processes

    (1993)
  • M.E. Bouton et al.

    State-dependent fear extinction with two benzodiazepine tranquilizers

    Behav Neurosci

    (1990)
  • M.E. Bouton et al.

    Contextual control of the extinction of conditioned fearTests for the associative value of the context

    J Exper Psychol: Anim Behav Processes

    (1983)
  • M.E. Bouton et al.

    Effect of context on performance to conditioned stimuli with mixed histories of reinforcement and nonreinforcement

    J Exper Psychol Anim Behav Processes

    (1986)
  • M.E. Bouton et al.

    Context-specificity of target versus feature inhibition in a feature-negative discrimination

    J Exper Psychol Anim Behav Processes

    (1994)
  • M.E. Bouton et al.

    The role of context in classical conditioningSome implications for cognitive behavior therapy

  • M.E. Bouton et al.

    Stimulus generalization context change and forgetting

    Psychol Bull

    (1999)
  • M.E. Bouton et al.

    Context effects on conditioning extinction and reinstatement in an appetitive conditioning preparation

    Anim Learn Behav

    (1989)
  • M.E. Bouton et al.

    Spontaneous recovery in cross-motivational transfer (counterconditioning)

    Anim Learning Behav

    (1992)
  • M.E. Bouton et al.

    Renewal of extinguished responding in a second context

    Anim Learn Behav

    (1994)
  • M.E. Bouton et al.

    Effects of contextual conditioning and unconditional stimulus presentation on performance in appetitive conditioning

    Q J Exper Psychol

    (1993)
  • M.E. Bouton et al.

    Slow reacquisition following extinctionContext encoding and retrieval mechanisms

    J Exper Psychol Anim Behav Processes

    (1989)
  • G.H. Bower

    Mood and memory

    Am Psychol

    (1981)
  • D.C. Brooks

    Recent and remote extinction cues reduce spontaneous recovery

    Q J Exper Psycho

    (2000)
  • D.C. Brooks et al.

    A retrieval cue for extinction attenuates spontaneous recovery

    J Exper Psychol Anim Behav Processes

    (1993)
  • D.C. Brooks et al.

    A retrieval cue for extinction attenuates response recovery (renewal) caused by a return to the conditioning context

    J Exper Psychol Anim Behav Processes

    (1994)
  • D.C. Brooks et al.

    Further evidence that conditioned inhibition is not the mechanism of an extinction cue’s effectA reinforced cue prevents spontaneous recovery

    Anim Learn Behav

    (2001)
  • D.C. Brooks et al.

    Reinstatement after counterconditioning

    Anim Learn Behav

    (1995)
  • D.C. Brooks et al.

    Extinction and spontaneous recovery of ataxic tolerance to ethanol in rats

    Psychopharmacology

    (2001)
  • D.C. Brooks et al.

    An extinction cue reduces spontaneous recovery of a conditioned taste aversion

    Anim Learn Behav

    (1999)
  • S. Campeau et al.

    Intra-amygdala infusion of the N-methyl-D-aspartate receptor antagonist AP5 blocks acquisition but not expression of fear-potentiated startle to an auditory conditioned stimulus

    Behav Neurosci

    (1992)
  • E.J. Capaldi

    The sequential viewFrom rapidly fading stimulus traces to the organization of memory and the abstract concept of number

    Psychonomic Bull Rev

    (1994)
  • Cited by (1320)

    View all citing articles on Scopus
    View full text