- Email: [email protected]
The repeated confrontation with videotapes of spiders in multiple contexts attenuates renewal of fear in spider-anxious students
ARTICLE IN PRESS
Behaviour Research and Therapy 45 (2007) 1169–1179 www.elsevier.com/locate/brat
The repeated confrontation with videotapes of spiders in multiple contexts attenuates renewal of fear in spider-anxious students Debora Vansteenwegen,1, Bram Vervliet, Carlos Iberico, Frank Baeyens, Omer Van den Bergh, Dirk Hermans Department of Psychology, University of Leuven, Leuven, Belgium Received 16 January 2006; received in revised form 23 August 2006; accepted 28 August 2006
Abstract In a treatment-analogue experiment, extinction of fear of spiders was investigated in a group of spider-anxious students. Two groups were created: in the single extinction group the extinction trials consisted of repeated presentations of a videotaped spider in one specific location of a house, whereas in the multiple extinction group the trials consisted of videotapes of the same spider in three different locations of a house. Also a control group was included that was exposed to videotapes of the location but without the spider. As reflected in skin conductance responses and self-report data, fear of spiders was significantly reduced in the two extinction groups compared to the control group. Moreover, when the extinction groups were confronted with the videotape of the spider in a new location, the single extinction group did not show generalisation of extinction, whereas the multiple extinction group did. These results corroborate the existing evidence for context dependence of extinction of fear and provide new evidence that the use of multiple contexts during extinction might improve the generalisability of extinction in humans. Implications for exposure therapy are discussed. r 2006 Elsevier Ltd. All rights reserved. Keywords: Extinction; Exposure; Renewal; Context; Electrodermal responding; Phobia
Introduction Exposure, the repeated confrontation with the object of fear, is a core element of many treatments for phobia, anxiety and anxiety-related disorders. It is assumed that exposure therapy involves processes analogous to extinction: repeatedly presenting the conditioned stimulus (CS) without the unconditioned stimulus (US) after acquisition most often leads to a decrease in conditioned responding (Eelen, Hermans, & Baeyens, 2001).
Corresponding author. Tel.: +32 16 32 61 34; fax: +32 16 32 59 24. 1
E-mail address: [email protected] (D. Vansteenwegen). Debora Vansteenwegen is postdoctoral researcher of the Fund for Scientific Research, FWO-Vlaanderen.
0005-7967/$ - see front matter r 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.brat.2006.08.023
ARTICLE IN PRESS 1170
D. Vansteenwegen et al. / Behaviour Research and Therapy 45 (2007) 1169–1179
In the Pavlovian conditioning literature, the last decade, a consensus has grown that the observed decrease in responding during extinction not necessarily reflects erasure of the original learning. Instead evidence is provided that during extinction new learning takes place that masks or inhibits the original learning (Myers & Davis, 2002; Rescorla, 2001). This approach clearly has consequences for exposure therapy as it indicates that the original fear associations are not destroyed and that extinguished fear can return after successful exposure. As such, the study of the causes and mechanisms of this return can be indicative for the understanding and prevention of clinical relapse (Rachman, 1989). Demonstrations of return of conditioned responding after extinction are crucial in this endeavour. In the animal conditioning literature, several post-extinction manipulations have been demonstrated to result in a return of conditioned responding. For example, Bouton and King (1983) demonstrated that executing extinction in a context different from the acquisition context and afterwards going back to the acquisition context resulted in a return of conditioned suppression in rats (ABA renewal). Bouton & Ricker (1994) demonstrated another type of renewal (return of conditioned responding due to a post-extinction context-change): acquisition and extinction took place in the same context but the test was performed in a new context (AAB renewal). This demonstration provides strong evidence for the fact that retrieval of extinction is more context-dependent than retrieval of acquisition and it points to a factor that might play an important role in the return of fear after successful exposure: leaving the therapeutic context can be sufficient to evoke a return of fear after successful exposure. In addition to a context-change, also the mere passage of time (spontaneous recovery, for a recent review see Rescorla, 2005) or the presentation of unpredicted USs (reinstatement, Bouton & King, 1983; Rescorla & Heth, 1975) have been shown to lead to a return after extinction. Recently, some of these phenomena of return of conditioned responding after extinction were replicated in human conditioning paradigms. In the fear conditioning paradigm, renewal of fear after extinction was demonstrated using electrodermal responses and verbal ratings as indices of learning (Vansteenwegen et al., 2005; Vervliet, Vansteenwegen, & Eelen, 2004). Using a computerised version of a kind of human-conditioned suppression paradigm (see Arcediano, Ortega, & Matute, 1996 for details), Havermans, Keuker, Lataster, and Jansen (2005) showed ABA renewal when changing the background context on the computer screen and Neumann (2006) demonstrated ABA and ABC renewal when changing between physical contexts in which participants executed this task. Several contingency learning studies have similarly provided evidence for renewal in humans (Garcı´ a-Gutie´rrez & Rosas, 2003a, b; Vila & Rosas, 2001). Finally in a more clinical context, there is analogous evidence for the context-specificity of exposure. In four studies spider-fearful individuals were tested at follow-up in either the treatment or a non-treatment context. More return was observed in the non-treatment context than in the treatment context. Rodriguez, Craske, Mineka, and Hladek (1999) observed more return in heart rate level but not in the self-reported fear at 2-week follow-up. Mineka, Mystkowski, Hladek, and Rodriguez (1999) found a small but significant return of self-reported fear during a behavioural approach test (BAT), while the effect was not evident in the spider phobia beliefs questionnaire (SBPQ). Mystkowski, Craske, and Echiverri (2002) replicated the return of self-reported fear at 1-week followup but the measurement of the heart rate and the SPBQ failed to generate the predicted context effect. Finally, Mystkowski, Mineka, Vernon, and Zinbarg (2003) demonstrated context-dependency using shifts in caffeine state as an internal context. Participants showed more return of fear when they received caffeine during exposure and not during follow-up (or vice versa) than when they received caffeine or placebo during both exposure and follow-up. Given that there is ample evidence that context changes can result in a return of fear, the subsequent question is how this return can be prevented. Against this background, several methods have been suggested in the animal conditioning literature. One method is to present the CS in different contexts during extinction. Gunther, Denniston, and Miller (1998) showed that rats that received extinction in three different contexts showed less return of conditioned suppression in yet a new context when compared to rats that received extinction in only one context. Chelonis, Calton, Hart, and Schachtman (1999) demonstrated in a conditioned taste aversion procedure that after exposure in three different contexts, test in the original acquisition context produced less renewal than when exposure took place in one and the same context. Inspired by the memory literature, Brooks and Bouton (1993, 1994) investigated an alternative method to increase the generalisability of extinction to other contexts. They showed that cues that were also present during extinction, could reduce
ARTICLE IN PRESS D. Vansteenwegen et al. / Behaviour Research and Therapy 45 (2007) 1169–1179
1171
the renewal-effect when they were presented at test. The cue would activate the memory representation of the extinction context and in that way would reduce the impact of the context-change. Demonstrations of the effectiveness of these prevention methods in humans are still scarce. Vansteenwegen Vervliet, Hermans, Beckers, Baeyens and Eelen (2006) showed in a human fear conditioning ABA-renewal design larger return of fear when, during test, cues were given that were also present during acquisition than when cues were given that were also present during extinction. A very nice illustration of the effectiveness of the retrieval cue method is the clinical-analogue study of Collins and Brandon (2002). They demonstrated that the return of alcohol cue reactivity after extinction due to a context-change could be reduced by the use of retrieval cues that were present during extinction. Furthermore, Neumann (2006) demonstrated using computer-conditioned suppression task that extinction in multiple contexts can reduce ABA and ABC renewal. Finally, in a clinical sample, Rowe and Craske (1998) showed return of fear towards a new spider exemplar after exposure with one and the same spider. This return was reduced when multiple stimulus examples were used during exposure. Using multiple stimulus, samples can be considered as an analogue of using multiple contexts as it is expected that both methods will increase the generalisation of what is learned during therapy. The present study aimed to replicate the finding that extinction in multiple contexts can attenuate renewal in human fear learning. A clinical-analogue study was conducted in which spider-anxious students were repeatedly confronted with short videotapes of a spider. By the use of blue screen montage the spider shot was placed against different contexts. Different locations/places in a house (bathroom, kitchen, etc.) were used as contexts. Three groups of students were tested: the single extinction group was confronted with the spider presented in one and the same location, the multiple extinction group saw the same spider shot but presented in three different locations and a control group saw a videotape with shots of contexts but no spiders. At the end, all students were confronted with the spider in a new context. It was hypothesised that the single and the multiple extinction groups would show a decrease in their fear for the spider in comparison with the control group. Moreover, when confronted with the spider in a new context, renewal should be observed for the single extinction group but not, or less so, for the multiple extinction group, illustrating the attenuating effect of extinction in multiple contexts on renewal. Methods Participants Student volunteers were asked to fill in a Dutch translation of the spider phobia questionnaire (SPQ, Muris & Merckelbach, 1996). When they met the criterion2 of a score of 16 or more, they were invited to participate in the experiment. Fifty-four participants (two men and 52 women) were randomly assigned to one of three conditions (all n ¼ 18). Their age ranged between 18 and 28 years with a mean of 19.2. All gave informed consent. One participant of the multiple group and one of the control group did not complete the experiment as they were too frightened to watch the spider video. They were excluded from all the analyses. Instruments Two questionaires were used in order to test fear of spiders. The SPQ was originally developed by Klorman, Weerts, Hastings, Melamed, and Lang (1974). It consists of a series of true–false items and is able to differentiate between spider phobic and non-phobic persons. The fear of spiders questionaire (FSQ) was developed by Szymanski and O’Donohue (1995) and translated in Dutch by Muris and Merckelbach (1996). It was designed to complement the information from the SPQ. The FSQ consists of items that are evaluated by means of an 7-point scale. In a direct comparison between the two questionnaires, Muris and Merckelbach (1996) demonstrated that although less often used, the FSQ is superior to the SPQ in measuring fear in the non-phobic range. 2
This criterion corresponds with the lowerbound of the 95%reliability interval of the mean of a comparable group of 45 Dutch female spider phobics who applied for treatment at the University Spider Phobia Project in the study of Muris and Merckelbach (1996).
ARTICLE IN PRESS 1172
D. Vansteenwegen et al. / Behaviour Research and Therapy 45 (2007) 1169–1179
Materials Four clearly distinguishable rooms in a house were selected (kitchen, basement, living room and bathroom). A stable image of each of these rooms served as contexts in the experiment. As an example a black-and-white picture of the kitchen context with a spider on the tableinfront is presented in Fig. 1. Coloured video montages of 60 s duration were created depicting one of the rooms in the background. On part of these video montages, after 15 s, a tarantula became visible who was sitting at one side of a small table. During the next 30 s, the spider was moving from one side of the table to the other side. The room was clearly visible in the background and the spider shot was placed on top of the room image using blue screen montage. The spider shot was identical for all context-plus-spider videos. Electrodermal activity was recorded with Fukuda standard Ag/AgCl electrodes (1 cm diameter) filled with a Unibase electrolyte and attached to the hypothenar palm of the left hand, which was cleaned with tap water. The inter-electrode distance was 2.5 cm. The Coulbourn skin conductance coupler (V71-23) provided a constant 0.5 V across electrodes. The analogue signal was passed through a 12-bit AD-converter and digitised at 10 Hz during the full 60 s of the presentation of each video fragment. Procedure Before the experiment started, all participants filled in the SPQ and the FSQ. Then, electrodes were attached. An overview of the course of the experiment as well as a scheme of the course of one trial is presented in Fig. 2. All participants received 14 presentations of a 60 s video shot with intertrial intervals of 60 s (black screen). Participants were instructed to carefully watch the videotapes. They were asked to imagine that they were in the place/room that was shown on the tape and to focus on the spider and not to suppress or avoid their fear. They were told that their fear would decline if they followed the instructions. During the ITI, participants had to indicate on three 11-point graphic rating scales how they felt during the past video-presentation. For the fear scale, they had to indicate how afraid they were from 10 (anchored: not afraid at all) to +10 (anchored: very afraid). Because Merckelbach, De Jong, Arntz, and Schouten (1993) suggested that evaluative conditioning (changes in valence due to the repeated paired presentation of a neutral stimulus with a highly valence stimulus) as well as disgust sensitivity play a role in the acquisition of fear, we
Fig. 1. An example of the type of contexts that were used in the experiment. This black-and-white picture represents the kitchen context with the tarantula on the table in front.
ARTICLE IN PRESS D. Vansteenwegen et al. / Behaviour Research and Therapy 45 (2007) 1169–1179
Post test trial
New test trial
Spider in context A
Spider in context A
Spider in context D
Spider in context C
Spider in context A
Spider in context A
Spider in context D
Context A without spider
Context A without spider
Spider in context A
Spider in context D
Pre test trial
Exposure trials
Single
Spider in context A
Spider in context A
Spider in context A
Spider in context A
Spider in context A
Spider in context A
Spider in context A
Spider in context A
Spider in context A
Spider in context A
Spider in context A
Multiple
Spider in context A
Spider in context C
Spider in context C
Spider in context A
Spider in context B
Spider in context C
Spider in context A
Spider in context B
Spider in context B
Spider in context B
Control
Spider in context A
Context A without spider
Context A without spider
Context A without spider
Context A without spider
Context A without spider
Context A without spider
Context A without spider
Context A without spider
Context A without spider
Group
Screen
Measures
Videomontage 60 s : context with or without spider
Registration of skin conductance
1173
Black Screen 60 s Filling in rating scale: Fear Valence Disgust
Fig. 2. The upper panel represents the course of the different trials during the experiment for the three groups: single, multiple and control. In this figure only the groups that were tested first in the new context were shown. The lower panel presents the course of one trial.
also asked the participants to indicate changes in their feelings of disgust and valence. They had to indicate how much disgust they felt on a scale ranging from 10 (not disgusting at all) to +10 (very disgusting). For the valence scale, they indicated how pleasant they felt ranging from 10 (very negative/not pleasant at all) to +10 (very positive/very pleasant). Three different randomisations of the order of the scales were used and equally divided over the participants. For each participant, the scales were always presented in the same order. For all participants, the first presentation included a spider in context A (pretest trial). Then, the participants of the control group received 11 presentations of context A without spider. Participants of the single group received 11 presentations of context A with spider. Finally, participants of the multiple group received three presentations of the spider in context A, four spiders in context B and four spiders in context C. Presentation order was equal for all subjects (CCABCABBBCA). In the test phase, responses of all participants were tested for the spider in context A and in context D. The order of the posttest trial (A) and the new test trial (D) was counterbalanced. For half of the participants of each group A was the basement-context and D the bathroom. For the other half of the participants this assignment was reversed. In the multiple group context B was the kitchen and context C the living room. Results The mean SPQ-score for the participants was 20.04 (SD ¼ 3.28). For the single group (n ¼ 18) the mean SPQ-score was 20.11, for the multiple group (n ¼ 17) M ¼ 19.94, and for the control group (n ¼ 17) M ¼ 20.06. Pairwise comparisons between the groups did not reveal statistical differences in SPQ-score, all Fo1. Skin conductance Skin conductance responses were visually inspected and corrected for artefacts before they were analysed statistically. The mean response during the first 15 s of the video fragment (context without the spider) was considered as a baseline and was subtracted from the largest response during the first 20 s of the spider presentation to reveal the skin conductance index. Mean responses were range corrected using the largest response observed during all spider presentations (Lykken & Venables, 1971) as the maximum range for each
ARTICLE IN PRESS D. Vansteenwegen et al. / Behaviour Research and Therapy 45 (2007) 1169–1179
1174
individual. The corrected responses were subjected to a square root transformation in order to normalise the distribution prior to statistical analysis. Three subjects were excluded from the skin conductance analysis due to technical problems. The data of the 49 remaining subjects were analysed using a 3 (Group) 3 (Context) analysis of variance. As the counterbalancing for type of contexts as well as for order of testing did not affect the crucial interactions, these variables were not included in the presented analysis. For the single context group (n ¼ 18), the multiple context group (n ¼ 15) and the control group (n ¼ 16) responses to the spider were registered during the first presentation in the extinction context (pre-A), the test-presentation in the extinction context (post-A) and the test-presentation in the new context (new-D). Means are presented in Fig. 3. The analysis revealed a significant main-effect of Group, F (2,46) ¼ 6.99, po0.01, MSE ¼ 0.10 and context, F(2,92) ¼ 21.49, po0.001, MSE ¼ 0.10 as well as a significant interaction, F(4,92) ¼ 11.10, po0.001, MSE ¼ 0.10. In order to test the effect of extinction, within this analysis planned comparisons were executed that compared responses to the spider at the beginning (pre-A) and at the end (post A) of the extinction trials. A significant decrease in responding was observed in the single extinction group, F(1,17) ¼ 22.13, po0.001, MSE ¼ 0.11 as well as in the multiple extinction group, F(1,14) ¼ 24.36, po0.001, MSE ¼ 0.15. As expected no decrease was obtained in the control group, F(1,15) ¼ 1.49. The decrease in the single as well as in the multiple group differed significantly from the control group (single: F(1,32) ¼ 15.89, po0.001, MSE ¼ 0.13; multiple: F(1,29) ¼ 19.50, po0.001, MSE ¼ 0.14). No difference in extinction-effect was obtained between the single and the multiple groups, Fo1. As a test for renewal, responses to the spider in the old context were compared to responses in the new context. In the single extinction group, larger responses were obtained for the new context than for the old context, F(1,17) ¼ 7.07, p ¼ 0.017, MSE ¼ 0.03 indicating a renewal effect. In contrast, the presentation of a spider in the new context in the control group did not reveal an increase, Fo1, and comparing this effect with the increase in the single extinction group revealed an almost significant interaction, F(1,32) ¼ 3.92, p ¼ 0.056, MSE ¼ 0.05. As predicted, in the multiple extinction group no significant renewal effect was obtained: there was no significant difference between the responses to the spider in the old and the new context, F(1,14) ¼ 1.19, p ¼ 0.290, MSE ¼ 0.05. In addition, no significant difference with the control group was obtained, Fo1. When comparing the increase from the old to the new context in the single and the multiple extinction groups directly, a clear interaction was found between the groups, F(1,31) ¼ 5.90, p ¼ 0.021, MSE ¼ 0.04. This interaction revealed strong evidence for our hypothesis that extinction in multiple contexts attenuates renewal.
1.4 1.2 SINGLE MULT IPLE CONT ROL
SCR-amplitude
1 0.8 0.6 0.4 0.2 0 PRE
POST
NEW
Fig. 3. Mean amplitudes of the skin conductance responses to the spider before extinction (pre), at the end of extinction (post) and in a new context (new) separately for the control group, the single extinction group and the multiple extinction group. Standard deviations are presented in the error bars.
ARTICLE IN PRESS D. Vansteenwegen et al. / Behaviour Research and Therapy 45 (2007) 1169–1179
1175
Apart from testing whether a renewal effect occurred, one can also test whether the renewal was complete, or in other words whether there was a residual effect of the extinction trials when testing in a new context. To test this, we compared the responses to the spider in the new context with the responses to the spider at the beginning of the session. In the multiple extinction group, it was not surprising that a significant decrease was obtained, F(1,14) ¼ 64, po0.001, MSE ¼ 0.07 because no renewal was observed. However also in the single extinction group, a significant decrease was observed F(1,17) ¼ 8.37, p ¼ 0.008, MSE ¼ 0.15 despite the significant renewal in this group. This indicates that the renewal-effect was not complete and that there was a residual effect of extinction. As predicted, however, the impact of the extinction trials was significantly stronger in the multiple extinction group than in the single extinction group, F(1,31) ¼ 5.49, p ¼ 0.026, MSE ¼ 0.12, indicating that extinction in multiple contexts effectively facilitates generalisation of extinction to new contexts. Self-report data For 52 participants valence, fear and disgust ratings were analysed in three separate 3 (Context) 3 (Group) univariate analyses of variance. As the counterbalancing for type of contexts as well as for order of testing did not affect the crucial interactions, these variables were not included in the presented analysis. Mean valence, disgust and fear ratings for three rating moments in the single, multiple and control groups are presented in Table 1. Valence: The analysis revealed a significant main effect of Group, F(2,49) ¼ 10.29, po0.001, MSE ¼ 19.56 and of Context, F(4,98) ¼ 7.22, p ¼ 0.001, MSE ¼ 2.84. Also the interaction was significant, F(4,98) ¼ 2.51, p ¼ 0.047, MSE ¼ 2.84. Within this analysis planned comparisons were executed. The tests revealed that both extinction groups show a significant decrease in responding from pre to post (single: F(1,17) ¼ 13.67, p ¼ 0.002, MSE ¼ 2.93; multiple: F(1,16) ¼ 8.91, po0.009, MSE ¼ 2.97), whereas no decrease was obtained in the control group, Fo1. The decrease in the single as well as in the multiple group differed significantly from the control group (single: F(1,33) ¼ 8.16, p ¼ 0.007, MSE ¼ 2.66; multiple: F(1,33) ¼ 5.64, p ¼ 0.024, MSE ¼ 2.67). No difference in extinction-effect was obtained between the single and the multiple groups, Fo1. To test the renewal effect, test ratings for the spider in the old and the new contexts were compared. In the single extinction group, a significant difference was obtained between the old and the new contexts, F(1,17) ¼ 16.66, p ¼ 0.001, MSE ¼ 1.30 whereas in the control group no differences were obtained between
Table 1 Self-reported valence, fear and disgust for the spiderclip at the beginning of extinction (pre), at the end of extinction (post) and in a new context (new) separately for the control group, the single extinction group and the multiple extinction group Moment
Group
Rating
Single
Multiple
Control
Valence Pre Post New
5.89 (2.70) 3.78 (2.73) 5.33 (2.74)
5.18(3.00) 3.41(3.99) 4.00(3.46)
7.88 (2.39) 8.00 (2.35) 8.00 (2.35)
Fear Pre Post New
6.83 (1.76) 4.83 (2.60) 6.39 (2.30)
6.35 (1.97) 4.59 (2.92) 5.47 (2.83)
7.53 (2.29) 7.82 (1.74) 8.24 (1.64)
Disgust Pre Post New
8.22 (1.86) 6.11 (2.03) 7.17 (1.50)
8.12 (2.00) 6.94 (2.19) 7.41 (1.87)
8.47 (1.62) 8.94 (1.14) 8.88 (1.27)
Standard deviations are presented between brackets.
ARTICLE IN PRESS 1176
D. Vansteenwegen et al. / Behaviour Research and Therapy 45 (2007) 1169–1179
the spiders in the old and the new contexts, Fo1. Moreover, the increase in the single extinction group differed significantly from the effect that was observed in the control group, F(1,33) ¼ 6.75, p ¼ 0.013, MSE ¼ 1.52. As expected, in the multiple extinction group no significant renewal effect was obtained: there was no significant difference between the response to the spider in the old and the new contexts, Fo1 and no difference was obtained with the effect in the control group, Fo1. Testing the difference in renewal between the single and the multiple extinction groups directly, revealed no significant between-group interaction; however, F(1,33) ¼ 1, 51, p ¼ 0.23, MSE ¼ 2.71. When comparing the test responses in the new context with the responses at the beginning of the extinction trials, the return in the single extinction group seemed to be complete as no difference was obtained anymore between the response to the spider in the new context and the response to the first presentation of the spider at the beginning of the extinction, F(1,17) ¼ 1.34, p ¼ 0.260, MSE ¼ 2.07. However in the multiple extinction group no significant difference was presented either, F(1,16) ¼ 2.13, p ¼ 0.163, MSE ¼ 5.51 despite a clearly significant extinction effect in this group and a non-significant return. The lack of this effect was probably due to the fact that responding to the spider in the new context was slightly larger than in the old context, but not significantly so. When comparing the strength of this effect in the two groups no significant interaction was obtained, F(1,31)o1. Disgust: The analysis revealed a significant main effect of Group, F(2,49) ¼ 6.78, p ¼ 0.003, MSE ¼ 5.46 and of Context, F(2,98) ¼ 6.05, p ¼ 0.003, MSE ¼ 1.89. Also the interaction was significant F(4,98) ¼ 3.99, p ¼ 0.005, MSE ¼ 1.89. Planned comparisons executed within this analysis revealed that the single extinction group showed a significant decrease in disgust ratings from pre to post, F(1,17) ¼ 22.81, po0.001, MSE ¼ 1.76. In contrast to the predictions, in the multiple extinction group the decrease was not significant, F(1,16) ¼ 2.64, p ¼ 0.124, MSE ¼ 4.45. However, no significant difference in decrease was obtained between the single and the multiple groups, F(1,33) ¼ 1.24, p ¼ 0.270, MSE ¼ 3.06. In the control group no decrease between pre and post was obtained, Fo1. The decrease in the single extinction group differed significantly from the control group, F(1,33) ¼ 17.48, po0.001, MSE ¼ 1.67 but the multiple extinction group differed not significantly from the control group, F(1,33) ¼ 3.83, p ¼ 0.059, MSE ¼ 3.01. When testing for renewal in the single extinction group, a significant increase in disgust was obtained from the old to the new context, F(1,17) ¼ 13.67, p ¼ 0.002, MSE ¼ 0.73 whereas no difference between the old and the new contexts was obtained in the control group, Fo1. Moreover, this increase in the single extinction group was significantly larger than the increase in the control group, F(1,33) ¼ 7.48, po0.001, MSE ¼ 0.73 indicating a renewal effect. In the multiple extinction group, there was no significant difference between the disgust rating for the spider in the old and the new contexts, F(1,16) ¼ 1.43, p ¼ 0.249, MSE ¼ 1.32 and no difference with the ratings in the control group, F(1,32) ¼ 2.07, p ¼ 0.169, MSE ¼ 1.02. The renewal-effect was not significantly stronger in the single group than in the multiple group; however, F(1,33) ¼ 1,46, p ¼ 0.23, MSE ¼ 1.02. In the single extinction group, the renewal-effect did not seem to be complete because a significant difference was obtained between the ratings for the spider in the new context and the ratings for the first presentation of the spider at the beginning of extinction, F(1,17) ¼ 6.69, p ¼ 0.019, MSE ¼ 1.50. For the multiple extinction group, on the other hand, no difference was obtained between the ratings at the beginning of extinction and the ratings for the spider in the new context, F(1,16) ¼ 1.06, p ¼ 0.320, MSE ¼ 3.99. This is not surprising because there was also no significant extinction-effect in this group. However when comparing the two groups directly, the decrease in disgust due to the extinction-trials when tested in the new context did not differ significantly, F(1,33)o1. Fear: The analysis revealed a significant main effect of Group, F(2,49) ¼ 6.49, p ¼ 0.003, MSE ¼ 12.37 and of Context, F(2,98) ¼ 12.68, po0.003, MSE ¼ 1.56. Also the interaction was significant, F(4,98) ¼ 4.77, p ¼ 0.001, MSE ¼ 1.56. Planned comparisons revealed that the single extinction group as well as the multiple extinction group showed a significant decrease in responding from pre to post (single: F(1,17) ¼ 21.86, po0.001, MSE ¼ 1.66; multiple F(1,16) ¼ 8.06, p ¼ 0.012, MSE ¼ 3.28) and no difference in decrease was obtained between the single and the multiple groups, Fo1. In the control group no decrease between pre and post was obtained, Fo1. The decrease in the single group as well as in the multiple group differed significantly from the control group, single: F(1,33) ¼ 17.35, po0.001, MSE ¼ 1.33; multiple: F(1,32) ¼ 8.44, p ¼ 0.007, MSE ¼ 2.13.
ARTICLE IN PRESS D. Vansteenwegen et al. / Behaviour Research and Therapy 45 (2007) 1169–1179
1177
When testing the spider in the single extinction group a significant difference was found between the old and the new contexts, F(1,17) ¼ 22.22, po0.001, MSE ¼ 0.95, whereas no differences were obtained in the control group, Fo1. Moreover, the increase in the single extinction group was larger than the increase observed in the control group, F(1,33) ¼ 7.77, po0.01, MSE ¼ 0.74, indicating a renewal effect. In the multiple extinction group, there was in contrast with the predictions, a significant difference between the response to the spider in the old and the new contexts, F(1,16) ¼ 8.22, p ¼ 0.011, MSE ¼ 0.81; however this increase was not significantly larger than the increase in the control group, F(1,32) ¼ 1.44, p ¼ 0.24, MSE ¼ 0.65. When comparing the single and the multiple extinction groups directly, no significant interaction was obtained, F(1,33) ¼ 2.25, p ¼ 0.143, MSE ¼ 0.88. When looking at the impact of the extinction trials on extinction of fear in a new context, no significant difference was obtained in the single extinction group indicating that the renewal in fear ratings was complete, F(1,17) ¼ 1.24, p ¼ 0.279, MSE ¼ 1.42. Also for the multiple extinction group no significant difference was obtained between the first rating at the beginning of extinction and the rating to the spider in a new context, F(1,17) ¼ 1.89, p ¼ 0.187, MSE ¼ 3.49. The between-group interaction was also not significant, F(1,33)o1.
Discussion This study was designed to evaluate the effect of extinction in multiple contexts on renewal of fear for spiders. First of all, our procedure was effective in reducing fear of spiders in spider-anxious students when comparing beginning and end of the session. In the extinction groups electrodermal responding, fear and valence ratings decreased over extinction trials whereas no such decrease was observed in the control group that did not receive the repeated confrontation with a spider in the videoclips. Only for the disgust ratings in the multiple exposure group no significant decrease was obtained. Hence, we can conclude that the repeated presentation of the spider videoclips effectively produced a within-session reduction. In contrast with the clear decreases in electrodermal responding, the rather small decreases (usually 1–2 points on an 11-point scale) in the ratings might be surprising. However, one cannot expect that spider-anxious students are released from their fear of spiders after a single session of simply watching video clips of a spider accompanied with only minimal instructions. In this study, we wanted to create an analogue extinction procedure that clearly had an impact on naturally occurring fear for spiders and that allowed us to test renewal and methods to attenuate renewal. It was not our intention to provide the students with a real treatment for their fear. Hence, our first attempt was successful. Second, the context-manipulation in this procedure seemed to be effective as well and produced the predicted renewal-effect. In the single extinction group, fear as indexed by electrodermal responding, fear ratings, valence ratings and disgust ratings returned when a videotape was presented showing the spider in a new context compared with the spider in the old context. When the control group received a similar contextchange, no increase in fear was obtained. Hence, we were able to demonstrate that changing a context after extinction can indeed cause return of fear in a group of spider-anxious students. This observation is in line with experimental studies on renewal in animals (e.g., Bouton & King, 1983) and humans (e.g., Vansteenwegen et al., 2005) as well as with some clinical studies investigating renewal of fear among spider phobics (e.g., Rodriguez et al., 1999). The present data contribute to the existing literature on renewal of naturally occurring fear in humans with regard to the following points: (1) an extinction control group was included, (2) a strictly controlled extinction procedure was used instead of an exposure treatment and (3) clear evidence was obtained not only in the verbal reports but also in the psychophysiological index, electrodermal responding. Finally and most importantly, the present experiment provided evidence for the fact that renewal can be attenuated by conducting extinction in multiple contexts. Spider-anxious students that received videotapes with the spider presented in different locations of a house showed less return of fear when confronted with the spider in a new context than students that saw similar tapes but with the spider presented in one and the same context. Hence, these data corroborate the findings in the animal conditioning literature (Chelonis et al., 1999; Gunther et al., 1998) and provide new evidence for the effectiveness of this method in order to reduce renewal of naturally occurring fears in humans.
ARTICLE IN PRESS 1178
D. Vansteenwegen et al. / Behaviour Research and Therapy 45 (2007) 1169–1179
Based on failures to replicate the multiple context extinction effect in their lab, Bouton, Moody, Garcia-Gutierrez, and Zilski (2006) warn for presenting this method as the ‘magic bullet’ to prevent relapse. They argue that several factors that are possibly influencing the effect are not yet known. One of these factors is whether or not acquisition took place in different contexts. Gunther et al. (1998) showed in their second experiment that extinction in three different contexts does not prevent renewal when acquisition took also place in three different contexts. Based on this finding one can question the clinical relevance of the multiple extinction method as the acquisition of fears treated in therapy is often unknown. However, in the present study, we worked with a naturally occurring fear and nevertheless demonstrated the impact of extinction in multiple contexts on the return of fear in a new context. This finding underlines the possible usefulness of this method for clinical practice. However, we agree with Bouton et al. (2006) that more research is needed to further unravel the preconditions for this phenomenon and the mechanisms that can explain it. Also, other methods to prevent return of fear after successful treatment, such as for example the use of retrieval cues (Collins & Brandon, 2002) should be further explored. Some limitations of the present study should be highlighted. First, the data of the verbal indices were not as convincing as the data of the electrodermal responses. Although as expected, renewal was only obtained in the single extinction group and not in the multiple extinction group for the valence and the disgust ratings, for the fear ratings in the multiple extinction group a return was nevertheless obtained. Additionally, the between-group interactions that compared the ratings for the single and the multiple extinction groups were not significant. The return in the fear ratings can easily be explained by the fact that only for this rating, the control group also showed a small increase when comparing the spider in the old and the new contexts. In fact when taking this baseline-increase into consideration, no additional increase for the multiple extinction group was left over. The lack of between-group interactions in the ratings is more difficult to explain and might possibly be due to the moderate impact of the extinction presentations on the verbal indices in general. As the data of the electrodermal responses do show all necessary interactions and as this index is considered to be less controllable and less vulnerable to demand effects, the observation that the data pattern of the ratings is highly similar to the data pattern of the electrodermal responses can be considered satisfying. Second, in comparing the single and the multiple extinction groups, we did not control for equal exposure to all the contexts in the two groups. We did however control for novelty in two other ways. As skin conductance responses, that are often regarded as orienting responses, are extremely sensitive to novelty and changes of all types, we controlled for the different reactions to the context itself. On every trial the context was first presented for 15 s without the spider, and then the spider occurred. These 15 s were used as a baseline that was subtracted from the largest response to the spider of that specific trial. Hence it is expected that most reactions to the novelty of the context as such were habituated before the spider occurred. Secondly, we always compared the extinction groups with the control group for which the new context was as new as for the single extinction group. However, future studies can include an even better control and equate the number of exposures to the different contexts for all groups. Third, there is the analogue nature of this study. All participants were to some extent afraid of spiders but their fear was not disturbing other aspects of life and they were not seeking treatment. Additionally, there was only a minimal extinction intervention restricted to watching videotapes. Future studies could use real spider phobics and include other aspects of the actual treatment for spider phobics, in order to guarantee that this method will useful in a real clinical context. Based on the data in this study, we can conclude that extinction in multiple contexts can prevent the return of naturally occurring fear in humans caused by the confrontation with the fear-eliciting object in a new context. This finding is of clinical importance as it suggests that conducting treatment in multiple contexts can improve the generalisability of exposure-based therapies. Acknowledgements The authors would like to thank An Schrauwen, An Dylst and Tine Lauwers for their assistance with running the experiments.
ARTICLE IN PRESS D. Vansteenwegen et al. / Behaviour Research and Therapy 45 (2007) 1169–1179
1179
References Arcediano, F., Ortega, N., & Matute, H. (1996). A behavioural preparation for the study of human pavlovian conditioning. Quarterly Journal of Experimental Psychology: Comparative and Physiological Psychology, 49, 270–283. Bouton, M. E., Moody, E. W., Garcia-Gutierrez, A., & Zilski, J. (2006). Extinction in multiple contexts does not necessarily make extinction less vulnerable to relapse. Behaviour Research and Therapy, 44(7), 983–994. Bouton, M. E., & King, D. A. (1983). Contextual control of the extinction of conditioned fear: Tests for the associative value of the context. Journal of Experimental Psychology: Animal Behavior Processes, 9, 248–265. Bouton, M. E., & Ricker, S. T. (1994). Renewal of extinguished responding in a 2nd context. Animal Learning and Behavior, 22, 317–324. Brooks, D. C., & Bouton, M. E. (1994). A retrieval cue for extinction attenuates response recovery (renewal) caused by a return to the conditioning context. Journal of Experimental Psychology: Animal Behaviour Processes, 20, 366–379. Chelonis, J. J., Calton, J. L., Hart, J. A., & Schachtman, T. R. (1999). Attenuation of the renewal effect by extinction in multiple contexts. Learning and Motivation, 30, 1–14. Collins, B. N., & Brandon (2002). Effects of extinction context and retrieval cues on alcohol cue reactivity among nonalcoholic drinkers. Journal of Consulting and Clinical Psychology, 70(2), 390–397. Eelen, P., Hermans, D., & Baeyens, F. (2001). Learning perspectives on anxiety disorders. In E. J. L. Griez, C. Faravelli, D. Nutt, & J. Zohar (Eds.), Anxiety disorders: An introduction to clinical management and research (pp. 249–264). London: Wiley. Garcı´ a-Gutie´rrez, A., & Rosas, J. M. (2003a). Empirical and theoretical implications of additivity between reinstatement and renewal after interference in causal learning. Behavioral Processes, 63, 21–31. Garcı´ a-Gutie´rrez, A., & Rosas, J. M. (2003b). Context change as the mechanism of reinstatement in causal learning. Journal of Experimental Psychology: Animal Behavior Processes, 29(4), 292–310. Gunther, L. M., Denniston, J. C., & Miller, R. R. (1998). Conducting exposure treatment in multiple contexts can prevent relapse. Behaviour Research and Therapy, 36, 75–91. Havermans, R. C., Keuker, J., Lataster, T., & Jansen, A. (2005). Contextual control of extinguished conditioned performance in humans. Learning and Motivation, 36(1), 1–19. Klorman, R., Weerts, T. C., Hastings, J. E., Melamed, B. G., & lang, P. J. (1974). Psychometric descriptions of some specific fear questionnaires. Behavior Therapy, 5, 401–409. Lykken, D., & Venables, P. H. (1971). Direct measurement of skin conduction: A proposal for standardization. Psychophysiology, 8, 656–672. Merckelbach, H., De Jong, P. J., Arntz, A., & Schouten, E. (1993). The role of evaluative conditioning and disgust sensitivity in the etiology and treatment of spider phobia? Advances in Behaviour Research and Therapy, 15(4), 243–255. Mineka, S., Mystkowski, J. L., Hladek, D., & Rodriguez, B. I. (1999). The effects of changing contexts on return of fear following exposure treatment for spider fear. Journal of Consulting and Clinical Psychology, 67, 599–604. Muris, P., & Merckelbach, H. (1996). A comparison of two spider fear questionnaires. Journal of Behaviour Therapy and Experimental Psychiatry, 27(3), 241–244. Myers, K. M., & Davis, M. (2002). Behavioral and neural analysis of extinction. Neuron, 36, 567–584. Mystkowski, J. L., Craske, M. G., & Echiverri, A. M. (2002). Treatment context and return of fear in spider phobia. Behaviour Therapy, 33, 399–416. Mystkowski, J. L., Mineka, S., Vernon, L. L., & Zinbarg, R. E. (2003). Changes in caffeine states enhance return of fear in spider phobia. Journal of Consulting and Clinical Psychology, 71(2), 243–250. Neumann, D. L. (2006). The effects of physical context changes and multiple extinction contexts on two forms of renewal in a conditioned suppression task with humans. Learning and Motivation, 37(2), 149–175. Rachman, S. (1989). The return of fear: Review and prospect. Clinical Psychology Review, 9, 147–168. Rescorla, R. A. (2001). Experimental extinction. In R. R. Mowrer, & S. B. Klein (Eds.), Handbook of contemporary learning theories (pp. 119–154). Mahwah, NJ: Lawrence Erlbaum Associates. Rescorla, R. A. (2005). Spontaneous recovery of excitation but not inhibition. Journal of Experimental Psychology: Animal Behavior Processes, 31(3), 277–288. Rescorla, R. A., & Heth, C. D. (1975). Reinstatement of fear to an extinguished conditioned stimulus. Journal of Experimental Psychology: Animal Behavior Processes, 104(1), 88–96. Rodriguez, B. I., Craske, M. G., Mineka, S., & Hladek, D. (1999). Context-specificity of relapse: Effects of therapist and environmental context on return of fear. Behaviour Research and Therapy, 37, 845–862. Rowe, M. K., & Craske, M. G. (1998). Effects of varied-stimulus exposure training on fear reduction and return of fear. Behaviour Research and Therapy, 36, 719–734. Szymanski, J., & O’Donohoe, W. (1995). Fear of spiders questionnaire. Journal of Behaviour Therapy and Experimental Psychiatry, 26(1), 31–34. Vansteenwegen, D., Hermans, D., Vervliet, B., Francken, G., Beckers, T., et al. (2005). Return of fear in a human differential conditioning paradigm caused by a return to the original acquisition context. Behaviour Research and Therapy, 43(3), 323–336. Vansteenwegen, D., Vervliet, B., Hermans, D., Beckers, T., Baeyens, F., & Eelen, P. (2006). Stronger renewal in human fear conditioning when tested with an acquisition retrieval cue than with an extinction retrieval cue. Behaviour Research and Therapy, 53, 131–142. Vervliet, B., Vansteenwegen, D., & Eelen, P. (2004). Generalization of extinguished skin conductance responding in human fear conditioning. Learning and Memory, 11(5), 555–558. Vila, N. J., & Rosas, J. M. (2001). Reinstatement of acquisition performance by the presentation of the outcome after extinction in causality judgements. Behavioral Processes, 56, 147–154.
Behaviour Research and Therapy 45 (2007) 1169–1179 www.elsevier.com/locate/brat
The repeated confrontation with videotapes of spiders in multiple contexts attenuates renewal of fear in spider-anxious students Debora Vansteenwegen,1, Bram Vervliet, Carlos Iberico, Frank Baeyens, Omer Van den Bergh, Dirk Hermans Department of Psychology, University of Leuven, Leuven, Belgium Received 16 January 2006; received in revised form 23 August 2006; accepted 28 August 2006
Abstract In a treatment-analogue experiment, extinction of fear of spiders was investigated in a group of spider-anxious students. Two groups were created: in the single extinction group the extinction trials consisted of repeated presentations of a videotaped spider in one specific location of a house, whereas in the multiple extinction group the trials consisted of videotapes of the same spider in three different locations of a house. Also a control group was included that was exposed to videotapes of the location but without the spider. As reflected in skin conductance responses and self-report data, fear of spiders was significantly reduced in the two extinction groups compared to the control group. Moreover, when the extinction groups were confronted with the videotape of the spider in a new location, the single extinction group did not show generalisation of extinction, whereas the multiple extinction group did. These results corroborate the existing evidence for context dependence of extinction of fear and provide new evidence that the use of multiple contexts during extinction might improve the generalisability of extinction in humans. Implications for exposure therapy are discussed. r 2006 Elsevier Ltd. All rights reserved. Keywords: Extinction; Exposure; Renewal; Context; Electrodermal responding; Phobia
Introduction Exposure, the repeated confrontation with the object of fear, is a core element of many treatments for phobia, anxiety and anxiety-related disorders. It is assumed that exposure therapy involves processes analogous to extinction: repeatedly presenting the conditioned stimulus (CS) without the unconditioned stimulus (US) after acquisition most often leads to a decrease in conditioned responding (Eelen, Hermans, & Baeyens, 2001).
Corresponding author. Tel.: +32 16 32 61 34; fax: +32 16 32 59 24. 1
E-mail address: [email protected] (D. Vansteenwegen). Debora Vansteenwegen is postdoctoral researcher of the Fund for Scientific Research, FWO-Vlaanderen.
0005-7967/$ - see front matter r 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.brat.2006.08.023
ARTICLE IN PRESS 1170
D. Vansteenwegen et al. / Behaviour Research and Therapy 45 (2007) 1169–1179
In the Pavlovian conditioning literature, the last decade, a consensus has grown that the observed decrease in responding during extinction not necessarily reflects erasure of the original learning. Instead evidence is provided that during extinction new learning takes place that masks or inhibits the original learning (Myers & Davis, 2002; Rescorla, 2001). This approach clearly has consequences for exposure therapy as it indicates that the original fear associations are not destroyed and that extinguished fear can return after successful exposure. As such, the study of the causes and mechanisms of this return can be indicative for the understanding and prevention of clinical relapse (Rachman, 1989). Demonstrations of return of conditioned responding after extinction are crucial in this endeavour. In the animal conditioning literature, several post-extinction manipulations have been demonstrated to result in a return of conditioned responding. For example, Bouton and King (1983) demonstrated that executing extinction in a context different from the acquisition context and afterwards going back to the acquisition context resulted in a return of conditioned suppression in rats (ABA renewal). Bouton & Ricker (1994) demonstrated another type of renewal (return of conditioned responding due to a post-extinction context-change): acquisition and extinction took place in the same context but the test was performed in a new context (AAB renewal). This demonstration provides strong evidence for the fact that retrieval of extinction is more context-dependent than retrieval of acquisition and it points to a factor that might play an important role in the return of fear after successful exposure: leaving the therapeutic context can be sufficient to evoke a return of fear after successful exposure. In addition to a context-change, also the mere passage of time (spontaneous recovery, for a recent review see Rescorla, 2005) or the presentation of unpredicted USs (reinstatement, Bouton & King, 1983; Rescorla & Heth, 1975) have been shown to lead to a return after extinction. Recently, some of these phenomena of return of conditioned responding after extinction were replicated in human conditioning paradigms. In the fear conditioning paradigm, renewal of fear after extinction was demonstrated using electrodermal responses and verbal ratings as indices of learning (Vansteenwegen et al., 2005; Vervliet, Vansteenwegen, & Eelen, 2004). Using a computerised version of a kind of human-conditioned suppression paradigm (see Arcediano, Ortega, & Matute, 1996 for details), Havermans, Keuker, Lataster, and Jansen (2005) showed ABA renewal when changing the background context on the computer screen and Neumann (2006) demonstrated ABA and ABC renewal when changing between physical contexts in which participants executed this task. Several contingency learning studies have similarly provided evidence for renewal in humans (Garcı´ a-Gutie´rrez & Rosas, 2003a, b; Vila & Rosas, 2001). Finally in a more clinical context, there is analogous evidence for the context-specificity of exposure. In four studies spider-fearful individuals were tested at follow-up in either the treatment or a non-treatment context. More return was observed in the non-treatment context than in the treatment context. Rodriguez, Craske, Mineka, and Hladek (1999) observed more return in heart rate level but not in the self-reported fear at 2-week follow-up. Mineka, Mystkowski, Hladek, and Rodriguez (1999) found a small but significant return of self-reported fear during a behavioural approach test (BAT), while the effect was not evident in the spider phobia beliefs questionnaire (SBPQ). Mystkowski, Craske, and Echiverri (2002) replicated the return of self-reported fear at 1-week followup but the measurement of the heart rate and the SPBQ failed to generate the predicted context effect. Finally, Mystkowski, Mineka, Vernon, and Zinbarg (2003) demonstrated context-dependency using shifts in caffeine state as an internal context. Participants showed more return of fear when they received caffeine during exposure and not during follow-up (or vice versa) than when they received caffeine or placebo during both exposure and follow-up. Given that there is ample evidence that context changes can result in a return of fear, the subsequent question is how this return can be prevented. Against this background, several methods have been suggested in the animal conditioning literature. One method is to present the CS in different contexts during extinction. Gunther, Denniston, and Miller (1998) showed that rats that received extinction in three different contexts showed less return of conditioned suppression in yet a new context when compared to rats that received extinction in only one context. Chelonis, Calton, Hart, and Schachtman (1999) demonstrated in a conditioned taste aversion procedure that after exposure in three different contexts, test in the original acquisition context produced less renewal than when exposure took place in one and the same context. Inspired by the memory literature, Brooks and Bouton (1993, 1994) investigated an alternative method to increase the generalisability of extinction to other contexts. They showed that cues that were also present during extinction, could reduce
ARTICLE IN PRESS D. Vansteenwegen et al. / Behaviour Research and Therapy 45 (2007) 1169–1179
1171
the renewal-effect when they were presented at test. The cue would activate the memory representation of the extinction context and in that way would reduce the impact of the context-change. Demonstrations of the effectiveness of these prevention methods in humans are still scarce. Vansteenwegen Vervliet, Hermans, Beckers, Baeyens and Eelen (2006) showed in a human fear conditioning ABA-renewal design larger return of fear when, during test, cues were given that were also present during acquisition than when cues were given that were also present during extinction. A very nice illustration of the effectiveness of the retrieval cue method is the clinical-analogue study of Collins and Brandon (2002). They demonstrated that the return of alcohol cue reactivity after extinction due to a context-change could be reduced by the use of retrieval cues that were present during extinction. Furthermore, Neumann (2006) demonstrated using computer-conditioned suppression task that extinction in multiple contexts can reduce ABA and ABC renewal. Finally, in a clinical sample, Rowe and Craske (1998) showed return of fear towards a new spider exemplar after exposure with one and the same spider. This return was reduced when multiple stimulus examples were used during exposure. Using multiple stimulus, samples can be considered as an analogue of using multiple contexts as it is expected that both methods will increase the generalisation of what is learned during therapy. The present study aimed to replicate the finding that extinction in multiple contexts can attenuate renewal in human fear learning. A clinical-analogue study was conducted in which spider-anxious students were repeatedly confronted with short videotapes of a spider. By the use of blue screen montage the spider shot was placed against different contexts. Different locations/places in a house (bathroom, kitchen, etc.) were used as contexts. Three groups of students were tested: the single extinction group was confronted with the spider presented in one and the same location, the multiple extinction group saw the same spider shot but presented in three different locations and a control group saw a videotape with shots of contexts but no spiders. At the end, all students were confronted with the spider in a new context. It was hypothesised that the single and the multiple extinction groups would show a decrease in their fear for the spider in comparison with the control group. Moreover, when confronted with the spider in a new context, renewal should be observed for the single extinction group but not, or less so, for the multiple extinction group, illustrating the attenuating effect of extinction in multiple contexts on renewal. Methods Participants Student volunteers were asked to fill in a Dutch translation of the spider phobia questionnaire (SPQ, Muris & Merckelbach, 1996). When they met the criterion2 of a score of 16 or more, they were invited to participate in the experiment. Fifty-four participants (two men and 52 women) were randomly assigned to one of three conditions (all n ¼ 18). Their age ranged between 18 and 28 years with a mean of 19.2. All gave informed consent. One participant of the multiple group and one of the control group did not complete the experiment as they were too frightened to watch the spider video. They were excluded from all the analyses. Instruments Two questionaires were used in order to test fear of spiders. The SPQ was originally developed by Klorman, Weerts, Hastings, Melamed, and Lang (1974). It consists of a series of true–false items and is able to differentiate between spider phobic and non-phobic persons. The fear of spiders questionaire (FSQ) was developed by Szymanski and O’Donohue (1995) and translated in Dutch by Muris and Merckelbach (1996). It was designed to complement the information from the SPQ. The FSQ consists of items that are evaluated by means of an 7-point scale. In a direct comparison between the two questionnaires, Muris and Merckelbach (1996) demonstrated that although less often used, the FSQ is superior to the SPQ in measuring fear in the non-phobic range. 2
This criterion corresponds with the lowerbound of the 95%reliability interval of the mean of a comparable group of 45 Dutch female spider phobics who applied for treatment at the University Spider Phobia Project in the study of Muris and Merckelbach (1996).
ARTICLE IN PRESS 1172
D. Vansteenwegen et al. / Behaviour Research and Therapy 45 (2007) 1169–1179
Materials Four clearly distinguishable rooms in a house were selected (kitchen, basement, living room and bathroom). A stable image of each of these rooms served as contexts in the experiment. As an example a black-and-white picture of the kitchen context with a spider on the tableinfront is presented in Fig. 1. Coloured video montages of 60 s duration were created depicting one of the rooms in the background. On part of these video montages, after 15 s, a tarantula became visible who was sitting at one side of a small table. During the next 30 s, the spider was moving from one side of the table to the other side. The room was clearly visible in the background and the spider shot was placed on top of the room image using blue screen montage. The spider shot was identical for all context-plus-spider videos. Electrodermal activity was recorded with Fukuda standard Ag/AgCl electrodes (1 cm diameter) filled with a Unibase electrolyte and attached to the hypothenar palm of the left hand, which was cleaned with tap water. The inter-electrode distance was 2.5 cm. The Coulbourn skin conductance coupler (V71-23) provided a constant 0.5 V across electrodes. The analogue signal was passed through a 12-bit AD-converter and digitised at 10 Hz during the full 60 s of the presentation of each video fragment. Procedure Before the experiment started, all participants filled in the SPQ and the FSQ. Then, electrodes were attached. An overview of the course of the experiment as well as a scheme of the course of one trial is presented in Fig. 2. All participants received 14 presentations of a 60 s video shot with intertrial intervals of 60 s (black screen). Participants were instructed to carefully watch the videotapes. They were asked to imagine that they were in the place/room that was shown on the tape and to focus on the spider and not to suppress or avoid their fear. They were told that their fear would decline if they followed the instructions. During the ITI, participants had to indicate on three 11-point graphic rating scales how they felt during the past video-presentation. For the fear scale, they had to indicate how afraid they were from 10 (anchored: not afraid at all) to +10 (anchored: very afraid). Because Merckelbach, De Jong, Arntz, and Schouten (1993) suggested that evaluative conditioning (changes in valence due to the repeated paired presentation of a neutral stimulus with a highly valence stimulus) as well as disgust sensitivity play a role in the acquisition of fear, we
Fig. 1. An example of the type of contexts that were used in the experiment. This black-and-white picture represents the kitchen context with the tarantula on the table in front.
ARTICLE IN PRESS D. Vansteenwegen et al. / Behaviour Research and Therapy 45 (2007) 1169–1179
Post test trial
New test trial
Spider in context A
Spider in context A
Spider in context D
Spider in context C
Spider in context A
Spider in context A
Spider in context D
Context A without spider
Context A without spider
Spider in context A
Spider in context D
Pre test trial
Exposure trials
Single
Spider in context A
Spider in context A
Spider in context A
Spider in context A
Spider in context A
Spider in context A
Spider in context A
Spider in context A
Spider in context A
Spider in context A
Spider in context A
Multiple
Spider in context A
Spider in context C
Spider in context C
Spider in context A
Spider in context B
Spider in context C
Spider in context A
Spider in context B
Spider in context B
Spider in context B
Control
Spider in context A
Context A without spider
Context A without spider
Context A without spider
Context A without spider
Context A without spider
Context A without spider
Context A without spider
Context A without spider
Context A without spider
Group
Screen
Measures
Videomontage 60 s : context with or without spider
Registration of skin conductance
1173
Black Screen 60 s Filling in rating scale: Fear Valence Disgust
Fig. 2. The upper panel represents the course of the different trials during the experiment for the three groups: single, multiple and control. In this figure only the groups that were tested first in the new context were shown. The lower panel presents the course of one trial.
also asked the participants to indicate changes in their feelings of disgust and valence. They had to indicate how much disgust they felt on a scale ranging from 10 (not disgusting at all) to +10 (very disgusting). For the valence scale, they indicated how pleasant they felt ranging from 10 (very negative/not pleasant at all) to +10 (very positive/very pleasant). Three different randomisations of the order of the scales were used and equally divided over the participants. For each participant, the scales were always presented in the same order. For all participants, the first presentation included a spider in context A (pretest trial). Then, the participants of the control group received 11 presentations of context A without spider. Participants of the single group received 11 presentations of context A with spider. Finally, participants of the multiple group received three presentations of the spider in context A, four spiders in context B and four spiders in context C. Presentation order was equal for all subjects (CCABCABBBCA). In the test phase, responses of all participants were tested for the spider in context A and in context D. The order of the posttest trial (A) and the new test trial (D) was counterbalanced. For half of the participants of each group A was the basement-context and D the bathroom. For the other half of the participants this assignment was reversed. In the multiple group context B was the kitchen and context C the living room. Results The mean SPQ-score for the participants was 20.04 (SD ¼ 3.28). For the single group (n ¼ 18) the mean SPQ-score was 20.11, for the multiple group (n ¼ 17) M ¼ 19.94, and for the control group (n ¼ 17) M ¼ 20.06. Pairwise comparisons between the groups did not reveal statistical differences in SPQ-score, all Fo1. Skin conductance Skin conductance responses were visually inspected and corrected for artefacts before they were analysed statistically. The mean response during the first 15 s of the video fragment (context without the spider) was considered as a baseline and was subtracted from the largest response during the first 20 s of the spider presentation to reveal the skin conductance index. Mean responses were range corrected using the largest response observed during all spider presentations (Lykken & Venables, 1971) as the maximum range for each
ARTICLE IN PRESS D. Vansteenwegen et al. / Behaviour Research and Therapy 45 (2007) 1169–1179
1174
individual. The corrected responses were subjected to a square root transformation in order to normalise the distribution prior to statistical analysis. Three subjects were excluded from the skin conductance analysis due to technical problems. The data of the 49 remaining subjects were analysed using a 3 (Group) 3 (Context) analysis of variance. As the counterbalancing for type of contexts as well as for order of testing did not affect the crucial interactions, these variables were not included in the presented analysis. For the single context group (n ¼ 18), the multiple context group (n ¼ 15) and the control group (n ¼ 16) responses to the spider were registered during the first presentation in the extinction context (pre-A), the test-presentation in the extinction context (post-A) and the test-presentation in the new context (new-D). Means are presented in Fig. 3. The analysis revealed a significant main-effect of Group, F (2,46) ¼ 6.99, po0.01, MSE ¼ 0.10 and context, F(2,92) ¼ 21.49, po0.001, MSE ¼ 0.10 as well as a significant interaction, F(4,92) ¼ 11.10, po0.001, MSE ¼ 0.10. In order to test the effect of extinction, within this analysis planned comparisons were executed that compared responses to the spider at the beginning (pre-A) and at the end (post A) of the extinction trials. A significant decrease in responding was observed in the single extinction group, F(1,17) ¼ 22.13, po0.001, MSE ¼ 0.11 as well as in the multiple extinction group, F(1,14) ¼ 24.36, po0.001, MSE ¼ 0.15. As expected no decrease was obtained in the control group, F(1,15) ¼ 1.49. The decrease in the single as well as in the multiple group differed significantly from the control group (single: F(1,32) ¼ 15.89, po0.001, MSE ¼ 0.13; multiple: F(1,29) ¼ 19.50, po0.001, MSE ¼ 0.14). No difference in extinction-effect was obtained between the single and the multiple groups, Fo1. As a test for renewal, responses to the spider in the old context were compared to responses in the new context. In the single extinction group, larger responses were obtained for the new context than for the old context, F(1,17) ¼ 7.07, p ¼ 0.017, MSE ¼ 0.03 indicating a renewal effect. In contrast, the presentation of a spider in the new context in the control group did not reveal an increase, Fo1, and comparing this effect with the increase in the single extinction group revealed an almost significant interaction, F(1,32) ¼ 3.92, p ¼ 0.056, MSE ¼ 0.05. As predicted, in the multiple extinction group no significant renewal effect was obtained: there was no significant difference between the responses to the spider in the old and the new context, F(1,14) ¼ 1.19, p ¼ 0.290, MSE ¼ 0.05. In addition, no significant difference with the control group was obtained, Fo1. When comparing the increase from the old to the new context in the single and the multiple extinction groups directly, a clear interaction was found between the groups, F(1,31) ¼ 5.90, p ¼ 0.021, MSE ¼ 0.04. This interaction revealed strong evidence for our hypothesis that extinction in multiple contexts attenuates renewal.
1.4 1.2 SINGLE MULT IPLE CONT ROL
SCR-amplitude
1 0.8 0.6 0.4 0.2 0 PRE
POST
NEW
Fig. 3. Mean amplitudes of the skin conductance responses to the spider before extinction (pre), at the end of extinction (post) and in a new context (new) separately for the control group, the single extinction group and the multiple extinction group. Standard deviations are presented in the error bars.
ARTICLE IN PRESS D. Vansteenwegen et al. / Behaviour Research and Therapy 45 (2007) 1169–1179
1175
Apart from testing whether a renewal effect occurred, one can also test whether the renewal was complete, or in other words whether there was a residual effect of the extinction trials when testing in a new context. To test this, we compared the responses to the spider in the new context with the responses to the spider at the beginning of the session. In the multiple extinction group, it was not surprising that a significant decrease was obtained, F(1,14) ¼ 64, po0.001, MSE ¼ 0.07 because no renewal was observed. However also in the single extinction group, a significant decrease was observed F(1,17) ¼ 8.37, p ¼ 0.008, MSE ¼ 0.15 despite the significant renewal in this group. This indicates that the renewal-effect was not complete and that there was a residual effect of extinction. As predicted, however, the impact of the extinction trials was significantly stronger in the multiple extinction group than in the single extinction group, F(1,31) ¼ 5.49, p ¼ 0.026, MSE ¼ 0.12, indicating that extinction in multiple contexts effectively facilitates generalisation of extinction to new contexts. Self-report data For 52 participants valence, fear and disgust ratings were analysed in three separate 3 (Context) 3 (Group) univariate analyses of variance. As the counterbalancing for type of contexts as well as for order of testing did not affect the crucial interactions, these variables were not included in the presented analysis. Mean valence, disgust and fear ratings for three rating moments in the single, multiple and control groups are presented in Table 1. Valence: The analysis revealed a significant main effect of Group, F(2,49) ¼ 10.29, po0.001, MSE ¼ 19.56 and of Context, F(4,98) ¼ 7.22, p ¼ 0.001, MSE ¼ 2.84. Also the interaction was significant, F(4,98) ¼ 2.51, p ¼ 0.047, MSE ¼ 2.84. Within this analysis planned comparisons were executed. The tests revealed that both extinction groups show a significant decrease in responding from pre to post (single: F(1,17) ¼ 13.67, p ¼ 0.002, MSE ¼ 2.93; multiple: F(1,16) ¼ 8.91, po0.009, MSE ¼ 2.97), whereas no decrease was obtained in the control group, Fo1. The decrease in the single as well as in the multiple group differed significantly from the control group (single: F(1,33) ¼ 8.16, p ¼ 0.007, MSE ¼ 2.66; multiple: F(1,33) ¼ 5.64, p ¼ 0.024, MSE ¼ 2.67). No difference in extinction-effect was obtained between the single and the multiple groups, Fo1. To test the renewal effect, test ratings for the spider in the old and the new contexts were compared. In the single extinction group, a significant difference was obtained between the old and the new contexts, F(1,17) ¼ 16.66, p ¼ 0.001, MSE ¼ 1.30 whereas in the control group no differences were obtained between
Table 1 Self-reported valence, fear and disgust for the spiderclip at the beginning of extinction (pre), at the end of extinction (post) and in a new context (new) separately for the control group, the single extinction group and the multiple extinction group Moment
Group
Rating
Single
Multiple
Control
Valence Pre Post New
5.89 (2.70) 3.78 (2.73) 5.33 (2.74)
5.18(3.00) 3.41(3.99) 4.00(3.46)
7.88 (2.39) 8.00 (2.35) 8.00 (2.35)
Fear Pre Post New
6.83 (1.76) 4.83 (2.60) 6.39 (2.30)
6.35 (1.97) 4.59 (2.92) 5.47 (2.83)
7.53 (2.29) 7.82 (1.74) 8.24 (1.64)
Disgust Pre Post New
8.22 (1.86) 6.11 (2.03) 7.17 (1.50)
8.12 (2.00) 6.94 (2.19) 7.41 (1.87)
8.47 (1.62) 8.94 (1.14) 8.88 (1.27)
Standard deviations are presented between brackets.
ARTICLE IN PRESS 1176
D. Vansteenwegen et al. / Behaviour Research and Therapy 45 (2007) 1169–1179
the spiders in the old and the new contexts, Fo1. Moreover, the increase in the single extinction group differed significantly from the effect that was observed in the control group, F(1,33) ¼ 6.75, p ¼ 0.013, MSE ¼ 1.52. As expected, in the multiple extinction group no significant renewal effect was obtained: there was no significant difference between the response to the spider in the old and the new contexts, Fo1 and no difference was obtained with the effect in the control group, Fo1. Testing the difference in renewal between the single and the multiple extinction groups directly, revealed no significant between-group interaction; however, F(1,33) ¼ 1, 51, p ¼ 0.23, MSE ¼ 2.71. When comparing the test responses in the new context with the responses at the beginning of the extinction trials, the return in the single extinction group seemed to be complete as no difference was obtained anymore between the response to the spider in the new context and the response to the first presentation of the spider at the beginning of the extinction, F(1,17) ¼ 1.34, p ¼ 0.260, MSE ¼ 2.07. However in the multiple extinction group no significant difference was presented either, F(1,16) ¼ 2.13, p ¼ 0.163, MSE ¼ 5.51 despite a clearly significant extinction effect in this group and a non-significant return. The lack of this effect was probably due to the fact that responding to the spider in the new context was slightly larger than in the old context, but not significantly so. When comparing the strength of this effect in the two groups no significant interaction was obtained, F(1,31)o1. Disgust: The analysis revealed a significant main effect of Group, F(2,49) ¼ 6.78, p ¼ 0.003, MSE ¼ 5.46 and of Context, F(2,98) ¼ 6.05, p ¼ 0.003, MSE ¼ 1.89. Also the interaction was significant F(4,98) ¼ 3.99, p ¼ 0.005, MSE ¼ 1.89. Planned comparisons executed within this analysis revealed that the single extinction group showed a significant decrease in disgust ratings from pre to post, F(1,17) ¼ 22.81, po0.001, MSE ¼ 1.76. In contrast to the predictions, in the multiple extinction group the decrease was not significant, F(1,16) ¼ 2.64, p ¼ 0.124, MSE ¼ 4.45. However, no significant difference in decrease was obtained between the single and the multiple groups, F(1,33) ¼ 1.24, p ¼ 0.270, MSE ¼ 3.06. In the control group no decrease between pre and post was obtained, Fo1. The decrease in the single extinction group differed significantly from the control group, F(1,33) ¼ 17.48, po0.001, MSE ¼ 1.67 but the multiple extinction group differed not significantly from the control group, F(1,33) ¼ 3.83, p ¼ 0.059, MSE ¼ 3.01. When testing for renewal in the single extinction group, a significant increase in disgust was obtained from the old to the new context, F(1,17) ¼ 13.67, p ¼ 0.002, MSE ¼ 0.73 whereas no difference between the old and the new contexts was obtained in the control group, Fo1. Moreover, this increase in the single extinction group was significantly larger than the increase in the control group, F(1,33) ¼ 7.48, po0.001, MSE ¼ 0.73 indicating a renewal effect. In the multiple extinction group, there was no significant difference between the disgust rating for the spider in the old and the new contexts, F(1,16) ¼ 1.43, p ¼ 0.249, MSE ¼ 1.32 and no difference with the ratings in the control group, F(1,32) ¼ 2.07, p ¼ 0.169, MSE ¼ 1.02. The renewal-effect was not significantly stronger in the single group than in the multiple group; however, F(1,33) ¼ 1,46, p ¼ 0.23, MSE ¼ 1.02. In the single extinction group, the renewal-effect did not seem to be complete because a significant difference was obtained between the ratings for the spider in the new context and the ratings for the first presentation of the spider at the beginning of extinction, F(1,17) ¼ 6.69, p ¼ 0.019, MSE ¼ 1.50. For the multiple extinction group, on the other hand, no difference was obtained between the ratings at the beginning of extinction and the ratings for the spider in the new context, F(1,16) ¼ 1.06, p ¼ 0.320, MSE ¼ 3.99. This is not surprising because there was also no significant extinction-effect in this group. However when comparing the two groups directly, the decrease in disgust due to the extinction-trials when tested in the new context did not differ significantly, F(1,33)o1. Fear: The analysis revealed a significant main effect of Group, F(2,49) ¼ 6.49, p ¼ 0.003, MSE ¼ 12.37 and of Context, F(2,98) ¼ 12.68, po0.003, MSE ¼ 1.56. Also the interaction was significant, F(4,98) ¼ 4.77, p ¼ 0.001, MSE ¼ 1.56. Planned comparisons revealed that the single extinction group as well as the multiple extinction group showed a significant decrease in responding from pre to post (single: F(1,17) ¼ 21.86, po0.001, MSE ¼ 1.66; multiple F(1,16) ¼ 8.06, p ¼ 0.012, MSE ¼ 3.28) and no difference in decrease was obtained between the single and the multiple groups, Fo1. In the control group no decrease between pre and post was obtained, Fo1. The decrease in the single group as well as in the multiple group differed significantly from the control group, single: F(1,33) ¼ 17.35, po0.001, MSE ¼ 1.33; multiple: F(1,32) ¼ 8.44, p ¼ 0.007, MSE ¼ 2.13.
ARTICLE IN PRESS D. Vansteenwegen et al. / Behaviour Research and Therapy 45 (2007) 1169–1179
1177
When testing the spider in the single extinction group a significant difference was found between the old and the new contexts, F(1,17) ¼ 22.22, po0.001, MSE ¼ 0.95, whereas no differences were obtained in the control group, Fo1. Moreover, the increase in the single extinction group was larger than the increase observed in the control group, F(1,33) ¼ 7.77, po0.01, MSE ¼ 0.74, indicating a renewal effect. In the multiple extinction group, there was in contrast with the predictions, a significant difference between the response to the spider in the old and the new contexts, F(1,16) ¼ 8.22, p ¼ 0.011, MSE ¼ 0.81; however this increase was not significantly larger than the increase in the control group, F(1,32) ¼ 1.44, p ¼ 0.24, MSE ¼ 0.65. When comparing the single and the multiple extinction groups directly, no significant interaction was obtained, F(1,33) ¼ 2.25, p ¼ 0.143, MSE ¼ 0.88. When looking at the impact of the extinction trials on extinction of fear in a new context, no significant difference was obtained in the single extinction group indicating that the renewal in fear ratings was complete, F(1,17) ¼ 1.24, p ¼ 0.279, MSE ¼ 1.42. Also for the multiple extinction group no significant difference was obtained between the first rating at the beginning of extinction and the rating to the spider in a new context, F(1,17) ¼ 1.89, p ¼ 0.187, MSE ¼ 3.49. The between-group interaction was also not significant, F(1,33)o1.
Discussion This study was designed to evaluate the effect of extinction in multiple contexts on renewal of fear for spiders. First of all, our procedure was effective in reducing fear of spiders in spider-anxious students when comparing beginning and end of the session. In the extinction groups electrodermal responding, fear and valence ratings decreased over extinction trials whereas no such decrease was observed in the control group that did not receive the repeated confrontation with a spider in the videoclips. Only for the disgust ratings in the multiple exposure group no significant decrease was obtained. Hence, we can conclude that the repeated presentation of the spider videoclips effectively produced a within-session reduction. In contrast with the clear decreases in electrodermal responding, the rather small decreases (usually 1–2 points on an 11-point scale) in the ratings might be surprising. However, one cannot expect that spider-anxious students are released from their fear of spiders after a single session of simply watching video clips of a spider accompanied with only minimal instructions. In this study, we wanted to create an analogue extinction procedure that clearly had an impact on naturally occurring fear for spiders and that allowed us to test renewal and methods to attenuate renewal. It was not our intention to provide the students with a real treatment for their fear. Hence, our first attempt was successful. Second, the context-manipulation in this procedure seemed to be effective as well and produced the predicted renewal-effect. In the single extinction group, fear as indexed by electrodermal responding, fear ratings, valence ratings and disgust ratings returned when a videotape was presented showing the spider in a new context compared with the spider in the old context. When the control group received a similar contextchange, no increase in fear was obtained. Hence, we were able to demonstrate that changing a context after extinction can indeed cause return of fear in a group of spider-anxious students. This observation is in line with experimental studies on renewal in animals (e.g., Bouton & King, 1983) and humans (e.g., Vansteenwegen et al., 2005) as well as with some clinical studies investigating renewal of fear among spider phobics (e.g., Rodriguez et al., 1999). The present data contribute to the existing literature on renewal of naturally occurring fear in humans with regard to the following points: (1) an extinction control group was included, (2) a strictly controlled extinction procedure was used instead of an exposure treatment and (3) clear evidence was obtained not only in the verbal reports but also in the psychophysiological index, electrodermal responding. Finally and most importantly, the present experiment provided evidence for the fact that renewal can be attenuated by conducting extinction in multiple contexts. Spider-anxious students that received videotapes with the spider presented in different locations of a house showed less return of fear when confronted with the spider in a new context than students that saw similar tapes but with the spider presented in one and the same context. Hence, these data corroborate the findings in the animal conditioning literature (Chelonis et al., 1999; Gunther et al., 1998) and provide new evidence for the effectiveness of this method in order to reduce renewal of naturally occurring fears in humans.
ARTICLE IN PRESS 1178
D. Vansteenwegen et al. / Behaviour Research and Therapy 45 (2007) 1169–1179
Based on failures to replicate the multiple context extinction effect in their lab, Bouton, Moody, Garcia-Gutierrez, and Zilski (2006) warn for presenting this method as the ‘magic bullet’ to prevent relapse. They argue that several factors that are possibly influencing the effect are not yet known. One of these factors is whether or not acquisition took place in different contexts. Gunther et al. (1998) showed in their second experiment that extinction in three different contexts does not prevent renewal when acquisition took also place in three different contexts. Based on this finding one can question the clinical relevance of the multiple extinction method as the acquisition of fears treated in therapy is often unknown. However, in the present study, we worked with a naturally occurring fear and nevertheless demonstrated the impact of extinction in multiple contexts on the return of fear in a new context. This finding underlines the possible usefulness of this method for clinical practice. However, we agree with Bouton et al. (2006) that more research is needed to further unravel the preconditions for this phenomenon and the mechanisms that can explain it. Also, other methods to prevent return of fear after successful treatment, such as for example the use of retrieval cues (Collins & Brandon, 2002) should be further explored. Some limitations of the present study should be highlighted. First, the data of the verbal indices were not as convincing as the data of the electrodermal responses. Although as expected, renewal was only obtained in the single extinction group and not in the multiple extinction group for the valence and the disgust ratings, for the fear ratings in the multiple extinction group a return was nevertheless obtained. Additionally, the between-group interactions that compared the ratings for the single and the multiple extinction groups were not significant. The return in the fear ratings can easily be explained by the fact that only for this rating, the control group also showed a small increase when comparing the spider in the old and the new contexts. In fact when taking this baseline-increase into consideration, no additional increase for the multiple extinction group was left over. The lack of between-group interactions in the ratings is more difficult to explain and might possibly be due to the moderate impact of the extinction presentations on the verbal indices in general. As the data of the electrodermal responses do show all necessary interactions and as this index is considered to be less controllable and less vulnerable to demand effects, the observation that the data pattern of the ratings is highly similar to the data pattern of the electrodermal responses can be considered satisfying. Second, in comparing the single and the multiple extinction groups, we did not control for equal exposure to all the contexts in the two groups. We did however control for novelty in two other ways. As skin conductance responses, that are often regarded as orienting responses, are extremely sensitive to novelty and changes of all types, we controlled for the different reactions to the context itself. On every trial the context was first presented for 15 s without the spider, and then the spider occurred. These 15 s were used as a baseline that was subtracted from the largest response to the spider of that specific trial. Hence it is expected that most reactions to the novelty of the context as such were habituated before the spider occurred. Secondly, we always compared the extinction groups with the control group for which the new context was as new as for the single extinction group. However, future studies can include an even better control and equate the number of exposures to the different contexts for all groups. Third, there is the analogue nature of this study. All participants were to some extent afraid of spiders but their fear was not disturbing other aspects of life and they were not seeking treatment. Additionally, there was only a minimal extinction intervention restricted to watching videotapes. Future studies could use real spider phobics and include other aspects of the actual treatment for spider phobics, in order to guarantee that this method will useful in a real clinical context. Based on the data in this study, we can conclude that extinction in multiple contexts can prevent the return of naturally occurring fear in humans caused by the confrontation with the fear-eliciting object in a new context. This finding is of clinical importance as it suggests that conducting treatment in multiple contexts can improve the generalisability of exposure-based therapies. Acknowledgements The authors would like to thank An Schrauwen, An Dylst and Tine Lauwers for their assistance with running the experiments.
ARTICLE IN PRESS D. Vansteenwegen et al. / Behaviour Research and Therapy 45 (2007) 1169–1179
1179
References Arcediano, F., Ortega, N., & Matute, H. (1996). A behavioural preparation for the study of human pavlovian conditioning. Quarterly Journal of Experimental Psychology: Comparative and Physiological Psychology, 49, 270–283. Bouton, M. E., Moody, E. W., Garcia-Gutierrez, A., & Zilski, J. (2006). Extinction in multiple contexts does not necessarily make extinction less vulnerable to relapse. Behaviour Research and Therapy, 44(7), 983–994. Bouton, M. E., & King, D. A. (1983). Contextual control of the extinction of conditioned fear: Tests for the associative value of the context. Journal of Experimental Psychology: Animal Behavior Processes, 9, 248–265. Bouton, M. E., & Ricker, S. T. (1994). Renewal of extinguished responding in a 2nd context. Animal Learning and Behavior, 22, 317–324. Brooks, D. C., & Bouton, M. E. (1994). A retrieval cue for extinction attenuates response recovery (renewal) caused by a return to the conditioning context. Journal of Experimental Psychology: Animal Behaviour Processes, 20, 366–379. Chelonis, J. J., Calton, J. L., Hart, J. A., & Schachtman, T. R. (1999). Attenuation of the renewal effect by extinction in multiple contexts. Learning and Motivation, 30, 1–14. Collins, B. N., & Brandon (2002). Effects of extinction context and retrieval cues on alcohol cue reactivity among nonalcoholic drinkers. Journal of Consulting and Clinical Psychology, 70(2), 390–397. Eelen, P., Hermans, D., & Baeyens, F. (2001). Learning perspectives on anxiety disorders. In E. J. L. Griez, C. Faravelli, D. Nutt, & J. Zohar (Eds.), Anxiety disorders: An introduction to clinical management and research (pp. 249–264). London: Wiley. Garcı´ a-Gutie´rrez, A., & Rosas, J. M. (2003a). Empirical and theoretical implications of additivity between reinstatement and renewal after interference in causal learning. Behavioral Processes, 63, 21–31. Garcı´ a-Gutie´rrez, A., & Rosas, J. M. (2003b). Context change as the mechanism of reinstatement in causal learning. Journal of Experimental Psychology: Animal Behavior Processes, 29(4), 292–310. Gunther, L. M., Denniston, J. C., & Miller, R. R. (1998). Conducting exposure treatment in multiple contexts can prevent relapse. Behaviour Research and Therapy, 36, 75–91. Havermans, R. C., Keuker, J., Lataster, T., & Jansen, A. (2005). Contextual control of extinguished conditioned performance in humans. Learning and Motivation, 36(1), 1–19. Klorman, R., Weerts, T. C., Hastings, J. E., Melamed, B. G., & lang, P. J. (1974). Psychometric descriptions of some specific fear questionnaires. Behavior Therapy, 5, 401–409. Lykken, D., & Venables, P. H. (1971). Direct measurement of skin conduction: A proposal for standardization. Psychophysiology, 8, 656–672. Merckelbach, H., De Jong, P. J., Arntz, A., & Schouten, E. (1993). The role of evaluative conditioning and disgust sensitivity in the etiology and treatment of spider phobia? Advances in Behaviour Research and Therapy, 15(4), 243–255. Mineka, S., Mystkowski, J. L., Hladek, D., & Rodriguez, B. I. (1999). The effects of changing contexts on return of fear following exposure treatment for spider fear. Journal of Consulting and Clinical Psychology, 67, 599–604. Muris, P., & Merckelbach, H. (1996). A comparison of two spider fear questionnaires. Journal of Behaviour Therapy and Experimental Psychiatry, 27(3), 241–244. Myers, K. M., & Davis, M. (2002). Behavioral and neural analysis of extinction. Neuron, 36, 567–584. Mystkowski, J. L., Craske, M. G., & Echiverri, A. M. (2002). Treatment context and return of fear in spider phobia. Behaviour Therapy, 33, 399–416. Mystkowski, J. L., Mineka, S., Vernon, L. L., & Zinbarg, R. E. (2003). Changes in caffeine states enhance return of fear in spider phobia. Journal of Consulting and Clinical Psychology, 71(2), 243–250. Neumann, D. L. (2006). The effects of physical context changes and multiple extinction contexts on two forms of renewal in a conditioned suppression task with humans. Learning and Motivation, 37(2), 149–175. Rachman, S. (1989). The return of fear: Review and prospect. Clinical Psychology Review, 9, 147–168. Rescorla, R. A. (2001). Experimental extinction. In R. R. Mowrer, & S. B. Klein (Eds.), Handbook of contemporary learning theories (pp. 119–154). Mahwah, NJ: Lawrence Erlbaum Associates. Rescorla, R. A. (2005). Spontaneous recovery of excitation but not inhibition. Journal of Experimental Psychology: Animal Behavior Processes, 31(3), 277–288. Rescorla, R. A., & Heth, C. D. (1975). Reinstatement of fear to an extinguished conditioned stimulus. Journal of Experimental Psychology: Animal Behavior Processes, 104(1), 88–96. Rodriguez, B. I., Craske, M. G., Mineka, S., & Hladek, D. (1999). Context-specificity of relapse: Effects of therapist and environmental context on return of fear. Behaviour Research and Therapy, 37, 845–862. Rowe, M. K., & Craske, M. G. (1998). Effects of varied-stimulus exposure training on fear reduction and return of fear. Behaviour Research and Therapy, 36, 719–734. Szymanski, J., & O’Donohoe, W. (1995). Fear of spiders questionnaire. Journal of Behaviour Therapy and Experimental Psychiatry, 26(1), 31–34. Vansteenwegen, D., Hermans, D., Vervliet, B., Francken, G., Beckers, T., et al. (2005). Return of fear in a human differential conditioning paradigm caused by a return to the original acquisition context. Behaviour Research and Therapy, 43(3), 323–336. Vansteenwegen, D., Vervliet, B., Hermans, D., Beckers, T., Baeyens, F., & Eelen, P. (2006). Stronger renewal in human fear conditioning when tested with an acquisition retrieval cue than with an extinction retrieval cue. Behaviour Research and Therapy, 53, 131–142. Vervliet, B., Vansteenwegen, D., & Eelen, P. (2004). Generalization of extinguished skin conductance responding in human fear conditioning. Learning and Memory, 11(5), 555–558. Vila, N. J., & Rosas, J. M. (2001). Reinstatement of acquisition performance by the presentation of the outcome after extinction in causality judgements. Behavioral Processes, 56, 147–154.