The Cognitive Behavioral Analysis System of Psychotherapy

I have been invited to write a clinical article on psychotherapeutic interventions for chronic depression. Rather than writing an article that only reviews the traditional cognitive behavioral analysis system of psychotherapy (CBASP) and provides outcome data, I focus primarily on an alternate approach.^1-3 I discuss CBASP, but I will also describe a strategy that can be added to CBASP practice that may improve clinical outcomes.

My discussion is informed in large part by contemporary learning theory (CLT) and, more specifically, by Prof Mark E. Bouton⁴ of the University of Vermont who published a splendid book in 2007 reviewing the synthesis that has taken place in the psychological learning laboratories. This is a learning theory revolution of which few psychiatrists and clinical psychologists are aware. Bouton describes how the cognitive revolution in psychology during the late 1960s, 1970s, and 1980s invaded the learning laboratories. The cognitive construct prompted researchers to study an expanded view of human and animal functioning when investigating behavior. The cognitive revolution subsequently thrust Pavlovian⁵ and Skinnerian6 learning into complementary positions.

One can no longer infer the presence of one type of learning (eg, Pavlovian learning) in the psychotherapeutic or pharmacotherapeutic context without implicating Skinnerian learning. For more than 7 decades, the 2 learning processes were dichotomized and conceptualized as mutually exclusive. The upshot of Bouton’s synthesizing view, which has significant ramifications for patient practice, is that metaphorically, Pavlov and Skinner are both sitting on the shoulders of all mental health practitioners who administer psychotherapy and/or pharmacotherapy.

The Bouton synthesis

Bouton’s synthesizing approach to Pavlovian and Skinnerian learning assumes that both models almost always interact to help organisms (1) learn about the stimulus properties of the interpersonal environment in which they live (Pavlov) and (2) learn to adapt to the interpersonal environment to obtain needed reinforcers (Skinner). Both learning processes must be concomitantly operative if humans (and animals) are to survive the environmental challenge.

Figure 1 illustrates Bouton’s synthesis of the learning processes in a basic paradigm.⁴

For the first time in the history of psychology and psychiatry, an adequate theory of learning is now available to practitioners to describe the following in CLT terminology:

• The etiology, diagnosis, and psychosocial functioning of chronically depressed patients.

• The learning goals of treatment and a way to assess the impact of learning on treatment outcome.

• What clinicians are trying to teach patients, defined as the “subject matter” of therapy.

• The prevention of the loss of extinction of the older pathological learning during the posttreatment period.

To date, the absence of an adequate learning theory has precluded clinicians from using learning theory to inform and describe the many activities that constitute psychological and psychiatric practice.

Any interpersonal stimulus situation, S:CS (conditional stimulus) that is followed by a biologically significant event, that is a reinforcer, S*:UCS (unconditional stimulus: pleasure, physical or psychological pain, fear, anxiety, hunger, thirst, and sexual satisfaction or deprivation) presents a learning moment in which the S*:UCS becomes associated with some behavior (response) and also with contextual cues or occasion setters.^4,7 The cues may be a place (eg, office, waiting room, workplace, home) or the context may involve particular individuals (eg, parent or caregiver, spouse, therapist, work colleagues, teacher, friend). In addition to being directly associated with S*:UCS, these originally neutral context cues can also become occasion setters that signal the availability of reinforcement if certain responses are made.

An example is seen in early childhood abuse when a parent (S:CS) repeatedly berates or physically abuses a child for breaking the rules. The child experiences fear/anxiety (because S:CS activates S*:UCS (fear/ anxiety) whenever he or she is around this significant other (who has now accrued conditional stimulus valence or excitor stimulus value). Active physical avoidance of contact with this parent, which is a response in the presence of the feared parent (S:CS), enables the child to decrease the fear/ anxiety (S*:UCS) and keeps it under control.

As shown in Figure 1, being in the presence of the abusing parent (S:CS) becomes an occasion setter and denotes that a response →[means: results in reinforcement] S*:UCS association has been acquired which subsequently signals that avoidance (response) will reduce the fear/anxiety. Bouton’s CLT synthesis is evident in that Pavlovian and Skinnerian learning work in concert-the child has exhibited both stimulus fear conditioning (Pavlov) and avoidance response learning (Skinner).

The Pavlovian variables that interest most clinicians are the patient’s emotional responses (past, present, and emotional reactions to anticipated future events), motivation to change, and the memory of previously learned associations (both short- and long-term memory). The Skinnerian variables we focus on in treatment are behavior and the reinforcers that behavior elicits.

CLT applied to clinical practice

CLT can add something that has been missing in our practice efforts. It will enable us to do more than just expose patients to psychological techniques and medicines and then observe the outcomes. Once patient learning concerns are added to practice, the clinician can address some important case management issues that may have been overlooked. For example, the following practice question moves to the forefront: what are patients learning from us psychologically?

This question implicates both psychotherapy as well as pharmacotherapy. CLT nudges us to ask the following specific questions: What are we trying to teach patients with our treatment? How much of the subject matter have patients learned? How does the learning affect treatment outcome and beyond?^7,8 Answering the first part of the last question is beyond the scope of this article; however, the other questions will be briefly addressed.

Some psychologists and psychiatrists may not see an immediate connection between CLT and pharma-cotherapeutic practice. However, pharmaco-therapy can provide a learning opportunity in which the drug (S:CS) evokes positive feelings (S*:UCS) in the patient. Biological treatment of patients with medication is pure Pavlovian conditioning accompanied with a strong Skinnerian component.^5,6 Think of it this way: I learn that taking (response) a particular antidepressant pill (S:CS) makes me feel better (S*UCS: Pavlov); the pill (S:CS) is itself associated with feeling better (S*:UCS). I also adhere to the medication regimen (response) because I have associated feeling better with taking a pill (response→S*:UCS association: Skinner).

In pharmacotherapeutic practice, compliance responses are critically important for patients. In addition, other neutral conditional stimuli (occasion setters, ie, place and person contextual cues such as the practitioner, his receptionist, the practitioner’s office, partner, children, friends) become associated with pill-taking to feel better (again, response→S*:UCS association). Furthermore, terminating the medication with the withdrawal of the associated S*:UCS at the end of treatment automatically places patients in an extinction trial/phase-many of the S:CSs are still present but are now encountered without the S*:UCS.

Without being aware of the dangers of the loss of extinction of the older psychopathological patterns (which are typically not erased in the brain), practitioners may have a serious relapse or recurrence problem during posttreatment. Relapse and recurrence are nonlearning labels that simply denote the loss of extinction of psychopathology following successful treatment. CLT alerts us to the importance of preserving the continued extinction of the older malevolent patterns.

Exposure-only practice

What is exposure-only practice? This is not the type of therapy where anxious/traumatized patients are exposed or required to confront the original trauma to overcome a debilitating

disorder or to acquire new coping mechanisms in the presence of the feared stimulus.^9,10 In our work, exposure-only practice reflects the type of training most of us received whereby we were taught to apply techniques to patients to make them feel better.

We may expose patients to several psychological or pharmacotherapy techniques; for example, we listen to their self-report, encourage them to talk and express how they feel, we empathize with their discomfort, we help them manage stress by teaching assertive behavior, we dispute their thinking patterns, and many of us expose patients to treatment by simply prescribing state-of-the-art medication.

By neglecting to pinpoint what it is we want to teach patients and by overlooking what they are actually learning from us, we are putting too much stock in what we can do using our techniques and medicines. What is being overlooked, particularly in the case of the chronically depressed patient, is what the individual requires to survive the posttreatment period. Put another way, if

patients don’t learn to do for themselves what you and I do to and for them, we leave them vulnerable to the ravages of psychiatric disorder-that is, to the dangers of relapse and recurrence.^4,7,11,12 This is the all-too-frequent consequence of exposure-only practice-patients leave treatment but remain vulnerable to relapse and recurrence because they have never really learned to take care of themselves. Exposure-based practice is only the first step. To inhibit a return of the disorder, we need to take one more step and CLT offers a way to do this. It removes practitioners from behind the wheel of treatment and puts patients into the driver’s seat, so to speak. The question becomes how do we take the second step?

The necessary second step

All CLT-based practitioners begin treatment administration by asking: what is it I want my patients to learn so that they can take care of themselves after treatment ends? How can I assess the learning that takes place?8 And finally, what is the likely impact of this learning on the outcome of treatment1,^8,11 and on the preservation of the extinction of the pathological patterns after treatment?

Case Vignette

Z is a 38-year-old secretary at a construction firm who has been married 3 times; she has no children. She reports being depressed since early adolescence. She was physically and sexually abused starting at age 13. On presentation, recurrent major depression with early-onset antecedent dysthymia (double depression) is diagnosed.

She had not been seen by a psychotherapist earlier but several antidepressants had been prescribed for her by her primary care physician. None of the antidepressants had been titrated to therapeutic levels and consequently had not relieved the depression.

Z’s early childhood memories include several malevolent significant others. Recalled information about her stepfather was elicited during the CBASP Significant Other History Exercise administered during session 2.^1,11 During this information gathering session it became evident that the extreme sexual abuse Z was subjected to as a child and adolescent contributed directly to her generalized interpersonal avoidance. Her interpersonal avoidance patterns were directly related to the maintenance of her chronic depression.

Toxic early developmental memories (Pavlovian) consisting of what Z learned about men at the hands of an abusive stepfather as well as other men automatically became associated with a well-meaning male clinician. In CLT terminology this is known as generalization and transfer of learning. Z’s dangerous early learning became the launching pad for a pervasive interpersonal avoidance strategy that generalized to people and especially to men. Her Pavlovian fear of people propelled her into a generalized detached interpersonal style resulting in keeping her distance from others.

Z’s interpersonal avoidance had 2 predictable effects on her life: (1) social isolation, which precluded Z from learning adaptive adolescent interpersonal skills; and a (2) generalized refractory Skinnerian avoidance of others, which maintained her Pavlovian fear of being abused by others (ie, avoidance of others reduced felt anxiety via negative reinforcement). However, interpersonal avoidance, while reducing anxiety, did not decrease the depression that had reached a full syndromal state. At screening, her Beck Depression Inventory–Second Edition (BDI-II) score was 48.

What does one teach a personally broken and fearful, chronically depressed patient who grew up in a learned helpless environment and who lives out an interpersonally detached lifestyle that would enable her to walk straight and obtain the positive reinforcers of relationship? Methodologically, Pavlov and Skinner lead the way with their stimulus learning and response learning foci, respectively.

A quick review of Figure 1 shows that Pavlovian learning emphasizes the acquisition of stimulus information about the psychotherapist (S:CS) as well as others (other S:CSs). New stimulus information about the psychotherapist was achieved through repeated administrations of the CBASP interpersonal discrimination exercise (IDE).^1,11 Situational analysis (SA), based on Skinnerian learning, was administered almost every session and taught Z how to make herself feel better (response→S*:UCS association) using newly acquired interpersonal skills (responses) to resolve interpersonal conflicts.^1,11 These novel skills were practiced with both the therapist and others and led to multiple reinforcing consequences.

CBASP

CBASP provided a 2-dimensional teaching program whereby Z progressively learned to discriminate the psychotherapist from “toxic” individuals and to respond emotionally to the psychotherapist as an inhibitor or safety signal. This resulted from an acquired ability to successfully discriminate the psychotherapist from malevolent significant others who had functioned as Pavlovian anxiety-fear excitors (eg, her stepfather, other men). Z was taught to take care of herself with others through situational analysis training in order to obtain positive interpersonal reinforcement.

Attaining interpersonal felt safety (S*:UCS) with the clinician (S:CS) was the crucial first step. Feeling safe sets the stage for the counter-conditioning (extinction) of interpersonal avoidance by replacing avoidance with interpersonal approach behavior. Shaping in new associations is facilitated when patients feel they are interpersonally safe. Figure 2 demonstrates what CBASP seeks to teach, measures the degree to which Z learned the tasks and presents the associated treatment outcome effects.

Treatment outcome targets 3 and 4 illustrate the impact that learning targets 1 and 2 may have exerted (Figure 2). The IDE (target 1) teaches interpersonal discrimination and is related to a quieting down of the original anxiety-fear elicitor stimulus value of the therapist (target 3); situa- tional analysis mastery (target 2) increased Z’s ability to identify the consequences of her behavior and was accompanied by decreasing BDI-II scores (target 4).

Extinction of the old learning

Chronically depressed patients must be strongly encouraged to protect the benefits of their psychotherapeutic and pharmacotherapeutic treatment for the remainder of their lives. Such encouragement is needed to prevent the reinstatement or reacquisition of the old pathological behaviors.^4,7 In short, chronically depressed patients are never done with treatment. This may sound surprising to some psychologists and psychiatrists who conclude that their work is completed once patients terminate treatment. It is not!

As noted earlier, treatment termination always puts patients on an extinction schedule for all the new learning they’ve acquired; since learning is rarely, if ever, erased in the brain, the old pathological floodwaters continually pound against the dam of the counter-conditioned training.^4,7 Failure to practice the new learning daily and to revisit the practitioner for buffer sessions increases the probability that the extinction dam will weaken and finally collapse-relapse and recurrence will not be far behind. This is the reason that exposure-only treatment would have been inadequate for Z. CLT sensitizes us to the dangers of the posttreatment period and for the need to continue medication for one’s lifetime as well as practice the skills learned in treatment.

Summary

I’ve tried to show that CLT may offer clinicians a way to increase response rates and reduce relapse and recurrence rates following treatment termination. Before such conclusions can be drawn, however, randomized clinical trials must be undertaken demonstrating that treatment outcomes are related to the degree of learning that occurs in practice; second, patients must be followed prospectively following treatment to determine whether posttreatment pre- paredness administered before termination protects the extinction of the older pathological learning (ie, produces lower relapse and recurrence rates). Preliminary data8 from a large clinical trial² are promising.

I’ve also tried to argue that, in light of CLT, simply exposing patients to psychological techniques or prescribing medicine is no longer a sufficient way to treat patients with chronic depression. The addition of CLT to practice will underscore the fact that patients are active learners. We sharpen up what they learn with us with a little forethought and planning concerning what we want to teach.

I feel strongly that treating patients as active learners, determining what it is they need to learn to overcome their psychiatric disorder, assessing the degree to which they learn the subject matter we teach and, finally, teaching patients to preserve the extinction of the pathological learning by daily practice of the counter-conditioned patterns we’ve taught will nudge psychiatric and psychological practice to a higher level of excellence.

 Outcome Assessment in Depression, by Waguih William IsHak, MD

Effects of Psychotherapy on Brain Function, by Gabriel S. Dichter, PhD, Jennifer N. Felder, and Moria Smoski, PhD

Disclosures:

Dr Mc Cullough is distinguished professor of psychology and psychiatry in the department of psychology at Virginia Commonwealth University in Richmond. He reports no conflicts of interest concerning the subject matter of this article.

References:

References

1. McCullough JP Jr. Treatment for Chronic Depression: CBASP. New York: Guilford; 2000.
2. Keller MB, McCullough JP Jr, Klein DN, et al. A comparison of nefazodone, the cognitive behavioral-analysis system of psychotherapy, and their combination for the treatment of chronic depression. N Engl J Med. 2000;342:1462-1470.
3. Klein DN, Santiago DJ, Vivian D, et al. Cognitive-behavioral analysis system of psychotherapy as a maintenance treatment for chronic depression. J Consult Clin Psychol. 2004;72:681-688.
4. Bouton ME. Learning and Behavior: A Contemporary Synthesis. Sunderland, MA: Sinauer Associates; 2007.
5. Pavlov I. Conditioned Reflexes. Anrep GV, trans. New York: Dover Publications; 1927/1960.
6. Skinner BF. Science and Human Behavior. New York: The Free Press; 1953.
7. Bouton ME. Context, ambiguity, and unlearning: sources of relapse after behavioral extinction. Biol Psychiatry. 2002;52:976-986.
8. Manber R, Arnow B, Blasey C, et al. Patient’s therapeutic skill acquisition and response to psychotherapy, alone or in combination with medication. Psychol Med. 2003;33:693-702.
9. Foa EB, Meadows EA. Psychosocial treatments for posttraumatic stress disorder: a critical review. Annu Rev Psychol. 1997;48:449-480.
10. Horowitz MJ. Stress Response Syndromes. London: Jason Aronson; 1986:chap 5.
11. McCullough JP Jr. Treating Chronic Depression With Disciplined Personal Involvement: CBASP. New York: Springer; 2006.
12. Thase ME. Long-term treatments of recurrent depressive disorders. J Clin Psychiatry. 1992;53 (suppl):32-44.