Measurement and Extinction of Conditioned Withdrawal-Like Responses in Opiate-Dependent Patients


As O’Brien has reviewed elsewhere in this volume (), there has been much experimental work on opiates and Pavlovian conditioning processes since Wikler’s original observations of withdrawal-like responses in drug-free patients (). Several studies have found evidence of conditioned withdrawal-like and opiate-like responses in rats, monkeys, and humans (). Addict patients viewing slides or videotapes of drug-related stimuli () or handling drug objects in a preparation ritual () experience subjective craving and withdrawal-like changes in physiological measures of skin temperature, heart rate, pupillary dilation, etc. Research from our own laboratory has demonstrated that opiate withdrawallike responses in humans can be conditioned to an arbitrary conditioned stimulus ().

These studies leave little doubt that conditioned withdrawal-like phenomena exist and can be both reliably elicited and measured. They do not, however, address the clinical significance of these responses. Though Wikler (1948) proposed conditioned withdrawal as the primary cause of relapse in drug-free patients, this link has not been clinically tested and is still controversial. Based on interviews with Baltimore street addicts, McAuliffe () had recently suggested that conditioned withdrawal-like phenomena are relatively infrequent and rarely trigger opiate use.


Though interesting, the McAuliffe interview data cannot substitute for research to empirically determine 1) the actual incidence of conditioned withdrawal-like phenomena, and 2) their role in clinical outcome, including relapse. We are currently conducting a large-scale treatment-outcome study which directly addresses both these issues. In this study, each patient’s conditioned with-drawal-like responses are first measured in the laboratory. Following measurement, one group of patients is given repeated, non-reinforced exposure to drug-related stimuli in an attempt to extinguish the conditioned withdrawal-like responses. If conditioned withdrawal-like responses trigger drug use/relapse, then reducing or removing these responses through extinction should have a beneficial effect on clinical outcome.

The present paper will present the methodology and early results of our attempts to measure and to extinguish conditioned withdrawal-like responses in opiate-dependent patients. Clinical outcome data concerning the possible benefits of extinction are currently being collected and will be presented in future papers.


Subjects – The subjects for this ongoing study are male veteran methadone patients from the Drug Dependence Treatment Unit of the Philadelphia VA Medical Center. Patient volunteers are recruited through direct contact or referral from their drug counselor. All patients are clinically screened to rule out diagnoses of major thought disorders (schizophrenia) or organic brain syndrome.

Design Considerations – In an earlier protocol () we attempted extinction of conditioned withdrawal-like responses by asking patients maintained on an opiate antagonist (naltrexone) to undergo double-blind cook-up and unreinforced self-injection rituals. In this procedure, opiate administration was either omitted (saline trials) or pharmacologically blocked due to the antagonist treatment. After a few initial trials, most patients experienced such strong dysphoria, withdrawal and craving that they refused to participate in further extinction sessions. Though no subject completed extinction, there was some suggestion that patients who completed more trials had somewhat better outcomes at six-month follow-up than other non-extinction naltrexone patients ().

In the current study, the extinction procedure was modified in two ways designed to increase patient comfort and compliance: 1) patients were given early trials with a graded hierarchy or drug-related stimuli as a prelude to the highly-evocative cook- up/ self-injection ritual; 2) each extinction trial was followed by 15-20 minutes of deep relaxation training to allow the patient to “wind down” from any discomfort or craving stirred by exposure to the drug-related stimuli.

As a final consideration, we recognized that an extinction procedure – even if well-tolerated by patients – would address only the conditioned factors of their disorder. If the significant psychological, social and vocational components of the addiction were left untreated, the possible clinical benefits of the extinction could be overshadowed and perhaps not even measurable. With this in mind, we decided to integrate our laboratory-derived extinction procedure with professional psychotherapy, a clinical treatment which had previously produced pervasive therapeutic benefits for our clinical population ().

Procedure – Patients eligible for the study are randomly assigned to one of three treatment groups. The clinical outcome of patients receiving cognitive-behavioral (CB) psychotherapy, extinction, and relaxation will be compared against two control groups: one group receiving CB therapy and relaxation (but no extinction) and a standard treatment control group which receives extra drug counseling and educational/control materials. Professional attention, session length, and small payments contingent upon session attendance are equivalent for all treatment groups.

Measurement of Conditioned Withdrawal-Like Responses – Prior to treatment, and at the end of treatment, and at 1 and 6 month follow-up points; each patient’s conditioned withdrawal-like responses are assessed in laboratory measurement sessions. All laboratory sessions are conducted in an environmentally controlled, electrically shielded recording chamber. Physiological measures include skin temperature, galvanic skin resistance (GSR. a general arousal index), heart rate; respiration and blood pressure. These physiological measures (except blood pressure) are continuously recorded on a polygraph and then converted to computer storage for later analysis.

In addition to the physiological measures, patients are asked to rate the degree of subjective high, carving or withdrawal they experience in response to test stimuli.

Both physiological and subjective responses are measured under two types of stimulus conditions: Neutral and Drug-Related. Each patient experiences both conditions, acting as his own control. For either stimulus condition, the following sequence obtains, lasting approximately one hour: 1) Resting Baseline; 2) Videotape (Neutral or Drug-Related); 3) Baseline; 4) Activity (Neutral or Drug-Related); and 5) Baseline.

The neutral videotape features a travelogue; the neutral (non-drug-related) activity allows patients to play a computerized “pong” game. The drug-related videotape features a cook-up-shoot-up ritual; the drug-related activity requires patients to go through a mock cook-up and tie-off, with optional self-injection of saline. Previous research in our center has shown pre-injection (drug preparation and cook-up) stimuli to be powerful elicitors of conditioned withdrawal-like responses ().

Extinction – Each hour-long treatment session for patients in the extinction group begins with 30 minutes of psychotherapy. followed by approximately 10 minutes of exposure to extinction stimuli. Each session ends with 15-20 minutes of relaxation, guided by audio cassette. Extinction stimuli include self- produced verbal imagery (“drug stories”), audiotapes of drug talk, color slides of cook-up-shoot-up rituals, videotape of drug purchase, cook-up and injection, and finally, handling of drug objects in a mock cook-up/tie-off procedure. Saline self-injection, the final member of the extinction series, is encouraged but optional. For each patient; the ordering of extinction stimuli across sessions is the same, and we now employ a fixed trials procedure which determines the number of exposures to each stimulus category.

Data for the extinction trials is currently based on the Within-Session Rating Scale (1982), a quantified subjective report listing 24 withdrawal-like and 24 high-like symptoms. The WSRS is administered before and immediately after exposure to the extinction stimuli. We have also recently begun to record GSR and skin temperature during treatment sessions, allowing us to track the course of the extinction across sessions and to compare subjective with physiological responses.

Extinction sessions for outpatient subjects are conducted three times weekly, with 35 sessions comprising a complete course of treatment. We have recently initiated the same study with in-patients undergoing gradual methadone detoxification over a four-week period. For these inpatients, extinction trials are conducted five times weekly, for a total of 22 treatment sessions.

Daily methadone is administered immediately after measurement or extinction sessions so that its onset effects not interfere with physiological or subjective measures.


Pretreatment Measurement of Conditioned Withdrawal – We have now obtained laboratory measurement of responses to drug-related stimuli for more than 35 patients. In the measurement sessions, patients respond to drug-related stimuli with a variety of physiological responses, including an increase in arousal (a decrease in GSR) and transient charges in heart rate and respiratory patterns. Of all the physiological measures recorded, a time-linked decrease in skin temperature has usually provided the most reliable and specific index of a conditioned withdrawal-like response.

Thus far, 35 to 40% of the patients tested exhibit a withdrawal like decrease in skin temperature which is specific to drug-related stimuli. Temperature change data from a representative ‘responder’ is presented in Figure 1. In this figure, the vertical bars represent temperature difference scores obtained by subtracting the mean skin temperature (°F) for a 4-minute baseline period (immediately preceding the neutral or drug-related video) from the mean skin temperature for an equivalent stimulus period (video, activity, post-activity). As shown in the right half of the graph, a decrease in skin temperature begins to develop during the drug video presentation and becomes quite pronounced during the cook-up ritual, persisting into the post-cook-up interval. Temperature response to the neutral stimuli () is negligible or shows no consistent pattern.

The average decrease in skin temperature to drug-related stimuli for clear-cut ‘responders’ is nearly 7°F. Some responders have experienced drops in skin temperature exceeding 12°F in a 15-20 minute period. Recovery time of skin temperature back to baseline level is usually roughly proportionate to the degree of temperature decrease, and usually occurs within 10 to 15 minutes of its nadir.

As many as one-third of the patients tested can be characterized as non-responders – they show no withdrawal-like temperature response to the drug-related stimuli. The remaining patients are more difficult to characterize, but several fall into the category of ‘nonspecific arousers’, showing mild arousal patterns to both neutral and drug-related activities, but no differential response in skin temperature or the other physiological measures.

More than a third of the patients report increases in subjective craving and withdrawal following exposure to the drug-related test stimuli. Interestingly, though subjective intensity of craving/ withdrawal is roughly correlated with the degree of physiological response, patients sometimes show a withdrawal-like decrease in skin temperature without reporting an increase in subjective withdrawal or craving. The reverse situation, in which a patient reports increased subjective craving but exhibits no profound change in physiological response also occurs, but is less common.

Extinction – Within extinction sessions, from one-half to two-thirds of the patients report increases in subjective craving and/ or withdrawal following exposure to the drug-related stimuli: In the early extinction trials, a few patients have also reported an increase in high-like symptoms, but these responses usually fade quickly and are replaced by relatively persistent withdrawal-like symptoms. Withdrawal symptoms elicited by the drug-related stimuli usually subside by the end of the 15-minute relaxation period.

Of patients who reliably respond to extinction stimuli with an increase in withdrawal-like symptoms, over half show a reduction in subjective response across the 35 sessions, suggestive of extinction. For at least two pilot patients the withdrawal-like responses persisted beyond 50 extinction trials.

Post-Treatment Measurement of Conditioned Withdrawal – Though the number of ‘responder’ patients who have completed this measurement phase is still relatively small, we do have early encouragement that extinction trials may reduce conditioned withdrawal-like responses. In general physiological ‘responders’ who undergo extinction tend to show a diminution of the temperature response to drug-related stimuli in post-treatment testing. ‘Responder’ patients in the non-extinction groups (therapy or extra counseling) do not exhibit this trend; and the temperature response is often similar to that at the outset of treatment.


Data from laboratory measurement sessions indicate that a substantial proportion – at least 40% – of opiate-dependent patients show physiological evidence of conditioned withdrawal in response to drug-related stimuli. The index response, a time-linked decrease in skin temperature, is often accompanied by increases in subjective craving and withdrawal. In extinction sessions, up to two-thirds of the patients tested respond with an increase in subjective craving and/or withdrawal to drug-related stimuli. Preliminary extinction data suggest that conditioned withdrawal-like responses, though relatively persistent, may be reduced with sufficient trials. Post-treatment laboratory measurements indicate that the extinction procedure may attenuate the withdrawal-like reduction in skin temperature.

Up to one-third of our patient population can be characterized as ‘non-responders ‘ – they fail to show physiological and/or subjective withdrawal-like responses to the drug-related stimuli used in our procedure. This lack of response is particularly intriguing since many of these ‘non-responders’ have had extensive drug use histories which should have allowed ample opportunity for conditioning to occur. We are currently exploring the possibility that certain emotional states (anxiety, etc.) may have become an integral part of the conditioned stimulus complex which elicits craving/withdrawal (), such that drug-related stimuli alone-unaccompanied by the mood state – are insufficient to elicit the conditioned response.

Although the clinical impact of our integrated treatment package awaits determination from outcome data, the presence of conditioned withdrawal-like responses in a significant proportion of our patient population suggests its possible benefit. The modified extinction procedures have been successfully integrated with psycho-therapeutic techniques to produce a treatment with retention rates approaching 70%. We are optimistic that this combined methodology will finally allow evaluation of the role of conditioned factors in opiate use and relapse.


Selections from the book: “Problems of Drug Dependence, 1983. Proceedings of the 45th Annual Scientific Meeting, the Committee on Problems of Drug Dependence, Inc.” Louis S. Harris, Ph.D., ed. A collection of papers which together record a year’s advances in drug abuse research; also includes reports on tests of new compounds for efficacy and dependence liability. National Institute on Drug Abuse Research Monograph 49, March 1984.