Effects of endorphin blocking on conditioned SCR in humans

Behav. Res. Ther. Vol. 31, No. 8, pp. 775-779, 1993 Printed in Great Britain. All rights reserved

0005-7967/93 $6.00 + 0.00 Copyright 0 1993 PergamonPressLtd

CASE HISTORIES AND SHORTER COMMUNICATIONS Effects of endorphin blocking

on conditioned

SCR in humans

HARALD MERCKELBACH, ARNOUD ARNTZ, PETER DE JONG and ERIK SCHOUTEN Department of Mental Health Sciences, Limburg University, Maastricht, P.O. Box 616, 6200 MD, The Netherlands (Received I9 November 1992) Summary-In order to test the hypothesis that low levels of endogenous opioids (endorphins) predispose to strong conditioning effects, female Ss (N = 36) were assigned to a placebo group, a low-dose naltrexone group, or a high-dose naltrexone group and then underwent a classical conditioning procedure. This procedure consisted of an acquisition phase in which all Ss received 5 pairings of a CS+ (neutral picture) and a UCS (100 dB white noise). The CS- (neutral picture) was never followed by a UCS. During extinction, Ss received 4 unreinforced presentations of CS+ and CS-. Throughout the experiment, skin conductance responses (SCRs) to the CSs and UCSs were recorded. Acquisition was successful in that CS + slides elicited stronger SCRs than CS - slides. However, during acquisition, there was no interaction between drug and differential response (CS + vs CS -). During extinction, there was no overall remaining effect of conditioning. Again, no evidence was found to suggest that (remaining) effects of conditioning were stronger in the naltrexone treated Ss than in the placebo Ss. If anything, the opposite seemed to be true with especially high-dose naltrexone Ss showing relatively weak conditioning effects.

INTRODUCHON According to the classical conditioning account of phobias, fear of a stimulus (e.g. a dog) develops as soon as this stimulus (conditioned stimulus; CS) is paired with a painful aversive event (e.g. being bitten by a dog; unconditioned stimulus; UCS). Although this account has met considerable criticism, it is still one of the most influential theories on the etiology of phobias (Rachman, 1990, 1991; Eysenck, 1987; van den Hout & Merckelbach, 1991). One important point that has been raised against the classical conditioning view is that the experience of an aversive UCS in connection with a CS is not a sufficient condition for phobic fears to emerge (Rachman, 1977). For example, numerous people have been bitten by a dog and, nevertheless, do not develop a phobia of dogs (DiNardo, Guzy & Bak, 1988). Likewise, aversive experiences with dentists are frequently reported by persons who do not suffer from dental phobia (Lautch, 1971; see for similar conclusions with regard to animal phobias: Hekmat, 1987; Merckelbach, Arntz, Arrindell & de Jong, 1992). In an attempt to come to grips with this problem, Eysenck and Kelly (1987) and Kelley (1987) proposed that the phobogenic effects of CS-UCS pairings are modulated by the neurohormonal state of the organism. In Kellev’s fl987) words- “ neuroses = conditioning x neirohormones” (p. -403). More specifically, Kelley argued that high ~levkls of adrenocorticotrope hormone (ACTH) and vasopressine enhance the phobogenic effects of aversive classical conditioning, whereas high levels of endogenous opioids (i.e. endorphins) attenuate these effects. As for the latter category of neurohormones, there are indications from animal studies that injections of morphine causes abolition of conditioned responses (Mauk, Warren & Thompson, 1982). Conversely, naloxone-an antagonist that blocks the action of endorphins-has been found to retard the extinction of classically conditioned freezing responses (e.g. Davis & Hendersen, 1985; see also review by Messing, 1988). Consistent with these results is a human study in which extremely negative correlations were found between measures of state and trait anxiety and level of endorphines in the cerebrospinal fluid (Post, Pickar, Ballenger, Naber & Rubinow, 1984). Furthermore, anxiety symptoms are commonly found during withdrawal from opiate addiction (Redmond, 1981). Recent clinical studies convincingly demonstrate that phobic patients receiving naloxone or naltrexone (i.e. an opioid antagonist that can be taken orally and has longer lasting effects than naloxone) benefit less from systematic desensitization or exposure therapy than patients receiving a placebo (Egan, Carr, Hunt & Adamson, 1988; Merluzzi, Taylor, Boltwood & Gotestam, 1991). Taken together, the findings on endorphins and (classically conditioned) fear seem to imply that there is an inverse relationship between endorphin action and phobogenic effects of aversive classical conditioning (Kelley, 1987). This relationship has been explained in terms of the analgesic consequences of endorphins (i.e. endorphins attenuate the noxious quality of CSs and UCS; Kelley, 1987; Fanselow, 1991). However, some authors have underlined the attentional effects of endorphins: low levels of endorphins (induced by naloxone), indeed, seem to increase electrocortical signs of attention to stimuli and thus might promote the efficiency of learning (Arnsten, Neville, Hillyard, Janowsky & Segal, 1984). To the best of the authors’ knowledge, no human study concerning endorphin blocking and conditioning has been carried out so far (see review by Messing, 1988). Yet, such a study might provide a test of Eysenck and Kelley’s (1987) proposal. Thus, the present study examines the effects of naltrexone on classical conditioning in humans. Subjects were given naltrexone or placebo and then underwent a differential classical conditioning procedure in which electrodermal responses (SCRs) were conditioned to a neutral CS by means of a white noise UCS. The hypothesis tested was that naltrexone enhances, in a dose dependent manner, the acquisition of conditioned SCRs and/or retards extinction of conditioned SCRs. 775

716

CASE HISTORIES

AND

SHORTER

COMMUNICATIONS

METHOD

Subjects Subjects in the final sample were 36 female spider phobics who applied for treatment at the Limburg University Spider Project. Mean age was 28.5 yr (range: 17-57 yr). Subjects met DSM-III-r criteria for simple phobia. Fear of spiders was the central complaint of the Ss and other psychopathological symptoms were absent. Recent work by Davey (1992) shows that spider phobia is not associated with trait anxiety or an increased prevalence of other phobias. Subjects with liver or kidney dysfunctioning or disease, Ss who used opiate medication and drug- or alcohol-addicted Ss were excluded from the present study. Apparatus and stimulus materials Skin conductance levels (SCL) and SCRs were picked up using two Beckman Ag-AgCl electrodes (8 mm dia.) attached to the medial phalanges of the second and third finger of the S’s non-dominant hand. The electrodes were connected to a Beckman Skin Conductance Coupler (type 9844). SCL and SCR recordings were based on the constant voltage (0.5 V) technique. Respiration was measured with a Beckman respiration belt fastened around the S’s chest. The respiration belt was connected to a Beckman Pressure/Pulse/Voltage Coupler (type 9884). Skin conductance activity and respiration were recorded continuously on a Beckman R 711 polygraph. Slides of a triangle and a circle served as conditioned stimuli (CSs). They were presented with a Kodak Carousel. The slides were projected through a hole in the wall on a white screen, approx. 2 m in front of the S. The size of the projected image was approx. 75 x 110 cm. A 100 dB white noise served as unconditioned stimulus. It was recorded on tape and presented through an audioboxe located in the testing room. A PDP Mint 11 microcomputer controlled onset and offset of CSs and UCS. Procedure Upon arrival in the laboratory, the S was briefly informed about the drug, the experiment and the exposure treatment. Next, the S was examined by a physician in order to exclude persons at risk with the pharmacological intervention (see above). Subjects signed an informed consent form and some pre-treatment baseline measures were obtained. Subjects were then told that they might either receive a placebo or a drug. Following this, Ss swallowed a capsule in the presence of the examiner. The capsule contained placebo, 25 mg naltrexone (low dose) or 100 mg naltrexone (high dose). Subjects were randomly assigned to one of the three drug conditions, each condition containing 12 Ss. The examiner was blind as to the pharmacological condition of each S. After the study had been finished and responses had been scored, the hospital pharmacologist gave information concerning the drug status (placebo, low dose naltrexone, high dose naltrexone) of each S. Peak plasma concentration of naltrexone emerges within 1 hr following drug intake. A single dose of naltrexone may have opiate-antagonistic effects for as long as 24 hr (AHSF, 1989). Some 2 h after drug intake, the experiment proper started. Subjects were seated in a comfortable chair placed in a dimly-lit, sound-attenuated testing chamber. Recording apparatus and the Kodak Carousel were located in adjacent room. Electrodermal recording sites were cleaned with distilled water and the electrodes and respiration belt were attached. The experiment was described to the S as a study of physiological reactions to different sensory stimuli. No information about the contingency between CS and UCS was given. The experiment consisted of two parts. The first was an acquisition phase which involved 5 CS+ and 5 CStrials. Duration of CSs was 6 sec. White noise (UCS) was delivered at CS+ offset. Duration of the UCS was 1 sec. The second phase was an extinction phase which involved 4 CS + and 4 CS- presentations. No UCS occurred during extinction. The order of presentations of CS + and CS - during the two phases was quasi-random. Four different sequences of CS + and CS- were employed. These sequences were counterbalanced across Ss. Intertrial intervals varied between IO and 20 sec. the mean being 15 sec. The content (i.e. triangle or circle) of CS+ and CS- was counterbalanced across Ss. Following the experiment, Ss received a one-session exposure treatment which lasted for 2 h. Finally, post-treatment measures were obtained. The treatment results will be described elsewhere (Arntz, Merckelbach & de Jong, 1992). Data reduction and analysis As SCRs are affected by the basal conductance level (SCL; Martin & Rust, 1976) SCLs were measured to see whether there were differences between the two groups in this respect. SCLs were obtained on two occasions: at the beginning of the experiment and at the end of the extinction phase. SCR was defined as the maximal deflection occurring 14 set after CS or UCS onset. SCL and SCRs were measured in micromho. To normalize the distribution, both SCL and SCR were square-root transformed; this was done prior to any statistical analysis (Levey, 1980). Respiration was used as a control variable. Following the criteria of Stern, Ray and Davis (1980) SCRs due to respiratory irregularities were excluded from the statistical analysis. Such irregularities occurred on less than 1% of the trials. SCR values for these trials were estimated on the basis of the SCRs on adjacent trials. SCRs were analyzed as response magnitudes (with zero responses). SCLs were subjected to a 3 (groups/drug condition) x 2 (time: at the start vs at the end of the experiment) analysis of variance (ANOVA), with the last factor being a repeated measure. SCRs during acquisition and extinction were evaluated with separate 3 (drug condition) x 2 (CS + vs CS -) x trials ANOVAs, with the latter two factors being repeated measures. For the trial factor, Greenhouse-Geisser corrected probability levels were used.

RESULTS

One-way ANOVAs showed that the groups did not differ with regard to age [F(2,33) < I.01 or with regard to fear of spiders as measured by Klorman, Weerts, Hastings, Melamed and Lang’s (1974) Spider Phobia Questionnaire [F(2,33) d 1.01. A 3 (groups) x 2 (time) ANOVA performed on the SCL data yielded a main effect of time [F(2,33) = 48.9, P $ 0.011due to a general decrease in electrodermal level during the experiment. This et&t was not modulated by group (i.e. drug condition).

CASE HISTORIES AND SHORTER

COMMIJNlCATlONS

0.35 0.30 0.25

0.20

E

mo.15

57 0.10 0.05

High

0.00 Fig. 1. Skin conductance responses to reinforced (CS+) and non-reinforced (CS-) slides during group (middle), and the high-dose acquisition of the placebo group (left), the low-dose naltrexone I.. . naltrexone group (rtght). Figure 1 shows the SCRs to CS+ and CS- slides during the acquisition phase of the experiment. A 3 (groups) x 2 (reinforcement: CS+ vs CS-) x 5 (trials) ANOVA revealed that the conditioning procedure was successful: overall, responses to CS+ slides exceeded those to CS- slides [F(1,33) = 16.8, P Q 0.011. No other effects reached significance. Most importantly, there was no indication that conditioning effects were stronger in naltrexone treated Ss than in control Ss [i.e. no significant groups x reinforcement interaction occurred; F(2,33) = 1.8, P = 0.191. Using the r-statistic, mean differentiation between CS+ and CS- acquisition trials was evaluated for each group. It was found that placebo and low-dose naltrexone Ss responded significantly stronger to CS + than to CS - : r (11) = 2.6, P C 0.02, one-tailed; [(I 1) = 3.1, P Q 0.01, one-tailed, respectively. Yet, the CS+ vs CS - differentiation failed to reach the conventional level of significance in the high-dose naltrexone group [t(l 1) = 1.1, P = 0.15, one-tailed], indicating that conditioning was less successful in this group than in the placebo or low-dose naltrexone group. To examine whether the overall conditioning effect was caused by differential responding on the first acquisition trials, an ANOVA was carried from which the first CS+ and CS- trials were excluded. This ANOVA yielded, again, a significant conditioning effect [F(l,33) = 25.1, P < O.Ol], No other effects attained significance.

1.60 1.401.20I.00 gOJOu-l

5 0.600.40 t

0.20 t 0.001 trials Fig. 2. Skin conductance responses to white noise UCSs during acquisition of the placebo group, the low-dose naltrexone group, and the high-dose naltrexone group.

CASE HISTORIES

AND

SHORTER

COMMUNICATIONS

0.35 0.30

0.25 0.20 s ulo.15 5 0.10 0.05 0.00 Fig. 3. Skin conductance resonses to CS+ and CSextinction trials of the placebo group low-dose naltrexone group (middle) and the high-dose nahrexone group (right).

(left), the

Figure 2 shows the SCRs to the white noise UCSs during the acquisition phase. A 3 (groups) x 5 (trials) ANOVA yielded a highly significant trial effect [F(4,132) = 56.7, P < O.Ol], which was caused by a general decline in responses to successive white noise presentations. However, this effect was not affected by drug condition [i.e. no significant groups x trials interaction emerged: F(8,132) < 1.01. SCRs during the extinction phase of the experiment are depicted in Fig. 3. A 3 (groups) x 2 (reinforcement) x 4 (trials) ANOVA revealed no main or interaction effects. The reinforcement effect fell short of significance [F( 1.33) = 2.8, P = 0.1 I] and there was no significant interaction of reinforcement with groups [F(2,33) = 1.06, P = 0.361. Follow-up analyses with the I -statistic showed that in the placebo group, responding during CS + was only marginally higher than responding during CS- [r(ll) = 1.4, P = 0.09, one-taiied]. Remaining effects of conditioning were even less convincing in the low-dose and high-dose naltrexone groups: I (11)= 1.3, P = 0.12, one-tailed; t( 11) < 1.O, respectively.

DISCUSSION

The present study found no support for the hypothesis that lowered levels of endorphins enhance the effects of aversive classical conditioning. That is to say, Ss receiving a low-dose (25 mg) or high-dose (100 mg) of the endorphin antagonist naltrexone were not found to display a superior acquisition of a classically conditioned response to the CS + as compared to placebo Ss. Neither was there a retarded extinction of classically conditioned responses in naltrexone Ss as compared to placebo Ss. If anything, conditioning was less stable in naltrexone treated Ss, especially in Ss who had received a high dose. A number of technical limitations of the present study merit brief comment. Firstly, although one may safely assume that a dose of 100 mg naltrexone is potentially able to elicit psychological effects [e.g. Merluzzi et al. (1991) administered 50 mg naltrexone to phobic Ss and found that this dose undermined behavioral treatment results], it can not be ruled out that the conditioning procedure employed in the present study was not aversive enough. The fact that naltrexone treated Ss were not found to respond stronger to white noise UCSs than placebo Ss is certainly consistent with this possibility. Thus, it may well be that naltrexone treated Ss display enhanced acquisition and retarded extinction of conditioned responses if and only if they are exposed to strongly aversive UCSs, e.g. a painful heat stimulus (Pitman, van der Kolk, Orr & Greenberg, 1990). In that case, the suppression of endorphin activity by naltrexone might optimahze the impact of the UCS which, in turn, would affect the dynamics of classical conditioning (Fanselow, 1991). In a similar vein, Messing (1988) concludes in his review that nalaxone or naltrexone can only be expected to exert effects when external stimuli (e.g. UCSs) are intensive enough to give rise to endorphin release. Secondly, the present study relied exclusively on electrodermal indices of conditioned responding. Yet, it is possible that somatic indices (e.g. eyeblink startles) or behavioral indices (e.g. escape) of conditioned fear are more sensitive to the effects of naltrexone. In passing, it should be noted that most of the animal studies that reported evidence for a facilitatory effect of endorphin blocking on conditioning, indeed, monitored behavioral (e.g. Davis & Hendersen, 1985) or somatic responses (e.g. Hernandez & Powell, 1980). Meanwhile, it should not be overlooked that the failure to find a facilitatory effect of endorphin blocking on conditioning has its precedents in animal literature (see Messing, 1988). In sum, then, the results do not sustain Kelley’s (1987) and Eysenck and Kelley’s (1987) suggestion that low levels of endorphin intensify the impact of classical conditioning. However, the present findings were obtained in a context that might have been only mildly aversive to the Ss. Parametric studies in which aversiveness of the UCS and indices of conditioned responding are systematically varied are needed in order to examine the possibility that reduced endorphin action facilitates classical conditioning of fear responses in a strongly aversive context.

CASE HISTORIES AND SHORTER

COMMUNICATIONS

779

REFERENCES

AHSF (1989). American Hospifal Formulary Service Drug Information. Bethesda: American Society of Hospital Pharmacists. Amsten, A. T., Neville, H. J., Hillyard, S. A., Janowsky, D. S. & Segal, D. S. (1984). Naloxone increases electrophysiological measures of selective information processing in humans. Journal of Neuroscience, 4, 2912-2919. Arntz, A., Merckelbach, H. &de Jong, P. (1992). Opioid antagonist affects behavioral effects of exposure in viuo. Manuscript submitted for publication, Davey, G. C. L. (1992). Characteristics of individuals with fear of spiders. Anxiety Research, 4, 299-314. Davis, H. D. & Hendersen, R. W. (1985). Effects of conditioned fear on responsiveness to pain: long term retention and reversibility by naloxone. Behavioral Neuroscience, 99, 277-289. DiNardo, P. A., Guzy, L. T. & Bak, R. M. (1988). Anxiety response patterns in dog-fearful and non-fearful subjects. Behaviour Research and Therapy, 25, 245-252. Egan, K. J., Carr, J. E., Hunt, D. D. & Adamson, R. (1988). Endogenous opiate system and systematic desensitization. Journal of Consulting and Clinical Psychology, 56, 287-29 1. Eysenck, H. J. (1987). Behavior therapy. In Eysenck, H. J. & Martin, I. (Eds), Theoreticalfoundations of behavior therapy. New York: Plenum Press. Eysenck, H. J. & Kelley, M. J. (1987). The interaction of neurohormones with Pavlovian A and Pavlovian B conditioning in the causation of neuroses extinction and incubation of anxiety. In Davey, G. (Ed.), Cognifiue processes and Pavlovian conditioning in humans. Chichester: Wiley. Fanselow, M. S. (1991). Analgesia as a response to aversive Pavlovian conditional stimuli: cognitive and emotional mediators. In Ray Denny, M. (Ed.), Fear, avoidance, and phobias: A fundamental analysis. New Jersey: Erlbaum. Hekmat, H. (1987). Origins and development of human fear reaction. Journal of Anxiety Disorders, I, 197-218. Hernandez, L. L. and Powell, D. A. (1980). Effects of naloxone on Pavlovian conditioning of eyeblink and heart rate responses in rabbits, Li& Sciences, 27, 863-869. van den Hout, M. A. & Merckelbach, H. (1991). Classical conditioning: still going strong. Behauioural Psychotherapy, 19, 59-79. Kelley, M. J. (1987). Hormones and clinical anxiety: an imbalanced neuromodulation of attention. In Eysenck, H. J. & Martin, I. (Eds), Theoretical foundations of behavior therapy. New York: Plenum Press. Klorman, R., Weerts, T. C., Hastings, J. E., Melamed, B. G. & Lang, P. J. (1974). Psychometric description of some specific fear questionnaires. Behavior Therapy, 5, 401409. Lautch, H. (1971). Dental phobia. Bri%h Journal of Psychiatry, 119, 151-158. Levey, A. B. (1980). Measurement units in psvchophvsioloav. In Martin. I. & Venables. P. H. (Eds). ~ , Techniuues in psychophysiology. New York: Wiley. __ . . -. Martin, I. & Rust, J. (1976). Habituation and the structure of the electrodermal system. Psychophysiology, 13, 554-562. Mauk, M. D., Warren, J. T. & Thompson, R. F. (1982). Selective, naloxone-reversible morphine depression of learned behavioral and hippocampal responses. Science, 216, 434436. Merckelbach, H., Amtz, A., Arrindell, W. A. & de Jong, P. J. (1992). Pathways to spider phobia. Behaviour Research and Therapy, 30, 543-546. Merluzzi, T. V., Taylor, C. B., Boltwood, M. & Gotestam, K. G. (1991). Opioid antagonist impedes exposure. Journal of Consulting and Clinical Psychology, 59, 425430. Messing, R. B. (1988). Opioid modulation of learning and memory: multiple behavioral outcomes. In Rodgers, R. J. & Cooper, S. J. (Eds), Endorphins, opiates and behavioural processes. Chichester: Wiley. Pitman, R. K., van der Kolk, B. A., Orr, S. P. & Greenberg, M. S. (1990). Naloxone-reversible analgesic response to combat-related stimuli in posttraumatic stress disorder. Archives of General Psychiarry, 47, 541-544. Post, R. M., Pickar, D., Ballenger, J. C., Naber, D. & Rubinow, D. R. (1984). Endogenous opiates in cerebrospinal fluid: relationship to mood and anxiety. In Post, R. M. & Ballenger, J. C. (Eds), Neurobiology of mood disorders. London: Williams & Wilkens. Rachman, S. (1977). The conditioning theory of fear-acquisition: a critical examination. Behaviour Research and Therapy, IS, 375-381. Rachman, S. (1990). The determinants and treatment of simple phobias. Advances in Behaviour Research and Therapy, 12, l-30. Rachman, S. (1991). Neo-conditioning and the classical theory of fear acquisition. Clinical Psychology Review,, II, 155-173. Redmond, D. (1981). Clonidine and the primate locus coeruleus: evidence suggesting anxiolytic and anti-withdrawal effects, In Lal, H. & Fielding, S. (Eds.). Psychopharmacology of clonidine. New York: Alan Liss. Stern, R. M., Ray, W. J. & Davis, C. M. (1980). Psychophysiological Recording. New York: Oxford University Press.