- Email: [email protected]
Effects of social anxiety and facial expression on habituation of the electrodermal orienting response
Biological Psychology 33 (1992) 21 l-223 0 1992 Elsevier Science Publishers B.V. All rights reserved
211 0301-0511/92/$05.00
Effects of social anxiety and facial expression on habituation of the electrodermal orienting response Belinda
M. Clark
‘, David A.T. Siddle
’ and Nigel W. Bond
’
’ Macquarie University, Macquarie, NSW 2109, Australia 2 Department of Psychology, University of Queensland, St. Lucia, Queensland 4072, Australia
The present research examined electrodermal orienting to happy and angry faces as a function of social anxiety and threat of shock. A preliminary study using 569 undergraduate participants developed an adequate set of normative data of social anxiety for the Willoughby questionnaire (WQ) for use in subject selection. Electrodermal activity was measured in both high and low socially anxious subjects (N = 85) during exposure to 10 presentations of an angry face intermixed with 10 presentations of a happy face. Threat of shock (no-shock, shock work-up only, and shock work-up plus threat) was also manipulated. Skin conductance responses (SCRs) which occurred within l-4 s of stimulus onset and trials-to-habituation constituted the data of primary interest. Although trials-to-habituation did not differ between angry and happy facial expressions, SCRs were larger to the angry face than to the happy face in both high and low socially anxious subjects. No differences in SCR magnitude were found as a function of threat of shock. The implications of these results for 6hman’s functional-evolutionary model of social phobia are discussed, and alternative explanations in terms of prepotency and prior learning are examined.
1. Introduction ohman’s (1986) functional-evolutionary approach to human phobias analyses phobias in terms of the behavioural systems within which the fear has evolved (ijhman, Dimberg, & Ost, 1985). According to Ohman (1986), social phobias originate in a social submissive system of ancient evolutionary origin. In order to maximize the benefits which may be gained from social living, conspecifics must maintain a relatively stable social order where agonistic encounters are minimal (6hman et al., 1985). This is usually achieved by means of a dominance/submissiveness hierarchy (Ohman, 1986) in which dominance is established through agonistic encounters. Such encounters are usually restricted to a symbolic level, and are characterized by intricate symbolic displays. Correspondence to: Dr. David Siddle, Lucia, Queensland 4072, Australia.
Department
of Psychology,
University
of Queensland,
St.
212
B.M. Clark / Orienfing and social anxiety
In primates, facial expression plays a definitive role in the nonverbal signals used in agonistic encounters (Ghman, 1986). For example, the facial threat display of dominant male primates, including man, is unmistakable as an angry, silent glare, with lowered corners to the mouth over bared teeth, with wide open staring eyes and frowning, lowered brows (Ekman, 1973). Following Mayr (1974), ehman and his colleagues (ohman, 1986; bhman et al., 1985) have argued that the symbolic displays evidenced in conspecific agonistic encounters are most likely to be characterized by relatively innate behaviours because of the high selective premium for these signals to be unmistakable. Further, the behavioural context within which symbolic displays are expressed has remained stable throughout evolutionary history. For example, aversive emotions such as anger signal the threat of impending attack. Happiness, on the other hand, always appears in a context antagonistic to anger, signalling play or filial attachment. Thus, from a functionalevolutionary perspective, the signalling function of an emotion could be expected to place biological constraints on its associability (hhman & Dimberg, 1984). It follows, therefore, that as a result of natural selection, anger and fear will be biologically prepared to signal aversive events, and happiness will be biologically unprepared. In a series of studies of Pavlovian conditioning, Dimberg and bhman (1983; ohman & Dimberg, 1978) have reported that, if a picture of an angry facial expression is used as the conditioned stimulus (CS) and shock as the unconditioned stimulus (US), the conditioned response (CR) is more resistant to extinction and less amenable to instructional manipulation than is a response conditioned to a picture of a neutral or a happy facial expression. In addition, Orr and Lanzetta (1980) demonstrated superior differential acquisition when pictures of fear expressions were compared with happy expressions. Not only were skin conductance responses (SCRs) larger to CS + than to CS - when a fearful face was the CS + , but more rapid acquisition of the differential response was also demonstrated. Thus, Orr and Lanzetta’s (1980) results are consistent with the theory that angry or fearful expressions are selectively associable with a shock US. The basic assumption underlying 6hman’s (1986) functional-evolutionary approach is that angry faces are selectively associable with an aversive US, and not that SCRs are more likely or more persistent with angry faces than with happy faces. However, some studies have found that other fear-relevant stimuli such as snakes and spiders elicit more vigorous orienting and require more trials to reach a criterion of habituation than do fear-irrelevant stimuli such as flowers and mushrooms (6hman, Eriksson, Fredriksson, Hugdahl, & Olofsson, 1974). If angry faces can be shown to elicit larger and more slowly habituating SCRs than happy faces prior to acquisition trials, then the results of conditioning studies might be explained in terms of the intrinsic stimulus properties of angry faces. In view of the importance of electrodermal orient-
B.M. Clark /Orieniing
and social anxiety
213
ing and habituation to an understanding of the mechanisms underlying selective associations, one of the aims of the present study was to examine orienting responses (ORs) to pictures of angry and happy facial expressions of emotion in an habituation paradigm. Because autonomic responding to facial threat displays is said to reflect the conditioning history of the subject (hhman & Dimberg, 1984), a second aim was to investigate the effect of prior conditioning on orienting to angry and happy facial expressions. The role of learning is emphasized explicitly in accounts of the functional-evolutionary model of social phobias. As bhman to point out that we are talking about et al. (1985) state, “it is important readiness to learn rather than about biologically given responses” (p. 142). Thus, an individual who has repeatedly experienced humiliation and defeat in social encounters with dominant individuals will, according to the model, preattentively “tag” facial threat displays as significant, leading to the affective reaction of fear. On the other hand, individuals who repeatedly experience victory and dominance in the same situation will not tag threat displays as significant, even though they share a biological readiness to associate fear with the same social stimuli. Thus, a fear response will not occur in these individuals following the perception of facial threat displays. According to ehman (1986), a conditioning history of repeated failures in dominance conflicts results in a conditioned fear response as indicated by marked social anxiety in social encounters and increased electrodermal responding to social threat cues. Thus, subjects were preselected in terms of self-reported level of social anxiety. It was hypothesized that the high socially anxious subjects, with their assumed history of social defeat, will display larger autonomic responses to pictures of social threat cues than will low socially anxious subjects who are assumed to have a conditioning history of success in agonistic encounters. In addition to these variables, we also manipulated threat of shock in an attempt to examine the possible nonassociative effects of threat on differential responding to fear-relevant social stimuli. Although threat of shock has been found to potentiate responding to certain fear-relevant stimuli (e.g. Kartsounis & Pickersgill, 1981; ohman et al., 1974), the effects of selective sensitization on social threat cues has not been examined. Subjects with high and low social anxiety scores were allocated to one of three groups. In the first group (no-shock) subjects were exposed to repeated presentations of two facial stimuli. The second group (threat-of-shock) underwent a shock work-up procedure and was informed that further shock would occur during the habituation phase. The third group (shock work-up) underwent shock work-up, but were not threatened with further shock. This group constituted a control for nonspecific arousal. Because most conditioning studies with facial stimuli have used differential conditioning in which subjects receive both CS + and CS - presentations, facial expression was
214
B. M. Clark /Orienting
and social anxie@
manipulated within subjects. Thus, the design was a 2 X 2 X 3 (Social Anxiety X Stimulus x Group) factorial design, with repeated measures on type of stimulus. A trials factor was included as a repeated measure in the statistical analyses.
2. Method 2.1. Subject selection In a preliminary study, the Willoughby questionnaire (Willoughby, 1932) was administered to 569 undergraduate volunteers (406 women and 163 men). Mean age was 22.8 years and the range 17-65 years. Of these, 237 (48 males and 189 females) were tested after an interval of 22 weeks. Both split-half (r = 0.87) and test-retest reliability (r = 0.80) were satisfactory. Factor analysis revealed three factors which were labelled “hypersensitivity to interpersonal relationships” (30% of variance), “negative affect” (7.6% of variance) and “hypersensitivity” (5.4% of variance). These and other analyses indicated that the Willoughby questionnaire is an internally consistent and reliable instrument which primarily measures hypersensitivity to interpersonal relationships. The apparent relevance of the extracted factors to the symptoms of social phobia, together with the construct validity reported by Turner, Meles, and DiTomassa (19831, support the utility of the Willoughby questionnaire as an instrument for identifying subjects high and low in interpersonal (social) anxiety. Subjects in the main study were 85 undergraduates (age range 17-32 years) who were contacted by telephone and asked to volunteer. The sample consisted of 45 low socially anxious subjects (19 males and 26 females) who obtained a midpoint percentile ranking less than 14.02% on the Willoughby questionnaire (M score = 13.961, and 40 high socially anxious subjects (7 males and 33 females) who obtained a midpoint percentile ranking greater than 87.53% (M score = 58.97). Subjects provided informed consent and participated as partial fulfilment of a course requirement. They were allocated randomly to the three experimental conditions with the restriction that within each social anxiety level all groups contained the same proportion (low, 2 : 3; high, 2 : 11) of men and women. The experiment was double-blind in that neither the subjects nor the experimenter knew the subjects’ social anxiety score. 2.2. Apparatus Skin conductance was recorded directly by using a constant voltage of 0.5 V applied across domed Ag-AgC1 electrodes in conjunction with a 0.05 M NaCl electrolyte. Conductance was recorded from masked areas on the distal
B.M. Clark /Orienting
and social anxiety
215
phalanges of the index and second fingers of the subject’s left hand using a Grass 7PIF preamplifier and a chart speed of 2.5 mm s-‘. Recording sensitivity was 0.05 microsiemens &S) mm-’ of pen deflection. Respiration was recorded using a chest pneumograph bellows (Phipps and Bird, model RPBA) in conjunction with a Grass 7PIF preamplifier. A locally constructed constant voltage stimulator (O-90 V) produced pulsed shocks. Pulse duration was 2 ms and the repetition rate was 50 Hz. Shock was delivered via a 36 mm diameter concentric electrode in which contact between the skin and metal elements was made through normal saline-soaked sponge pads (Tursky, Watson, & O’Connell, 1965). The shock electrode was placed on the dorsal surface of the right forearm, 17 cm above the ulna styroid. The electrode site was prepared with EC2 Electrode Cream (Grass Instruments) prior to electrode attachment to ensure stability of the electrode-skin interface. The visual stimuli were two black-and-white slides, one showing a male with an angry facial expression and the other showing a different male with a happy facial expression. These two slides were selected from Ekman and Friesen’s (1976) Pictures of facial affect on the basis of their inter-rater reliability score of r = 1.0 (Ekman & Friesen, 1976). The slides were projected through a small window in the wall of the experimental chamber by a Pradovit projector fitted with a tachistoscopic shutter (Gerbrands, Model G1166). A 60 cm X 95 cm image was produced on a white screen situated 200 cm in front of the subject at eye level. 2.3. Procedure Each subject in the threat-of-shock and shock work-up groups was fitted with the shock electrode and given a shock work-up procedure in the laboratory control room. The subject was asked to report when the shock level was “uncomfortable, but not painful”. Shocks of 0.5 s duration were then delivered, commencing at 0 V and incrementing by 0.5 V, with the subject reporting at each shock presentation. The shock electrode was then removed from subjects in the shock work-up group, who were informed that they would not receive any further shocks. Subjects were then seated in a semi-reclining padded chair in the sound-attenuated experimental room. Ambient illumination was 2 lux, the temperature between 19°C and 25°C and the relative humidity within the range 48-56%. The stimulus equipment and recording apparatus were housed in an adjoining room. Following attachment of the skin conductance electrodes and chest pneumograph bellows, the subjects were informed that the first part of the experiment involved a 5 min rest period during which they were to relax with their eyes open.
216
B.M. Clark /Orienting and social anxiety
Following the prestimulation period all subjects were informed that some pictures of faces would appear from time to time on the screen. They were instructed to remain as still as possible, and to watch the screen at all times. Subjects in the threat-of-shock group were informed that some of the pictures might be followed by a brief electric shock at the intensity level set earlier. During the habituation phase there were 10 presentations each of an angry face and a happy face. The order of the first two stimulus presentations was counterbalanced. Half the subjects received the angry face followed by the happy face, whereas for the other half, the order was reversed. The remaining stimuli were presented randomly, with the restriction that not more than two consecutive slides were the same. All stimuli were of 2 s duration, and were presented at randomly ordered intervals of 20, 2.5, and 30 s (offset to onset). 2.4. Scoring An SCR greater than 0.05 PUS which occurred within 1-4 s of stimulus onset was scored as a response to that stimulus. In order to normalize the SCR data, they were subjected to a square root transformation prior to analysis. Any SCR greater than 0.05 PUSwhich occurred in the absence of a stimulus, or outside the l-4 s window following stimulus onset, was considered a nonspecific response (NSR). Skin conductance level (SCL) was scored at 1 min intervals during the prestimulation phase, and a mean SCL was obtained for that period. During the habituation phase, tonic SCL was scored immediately prior to stimulus onset. A measure of trials-to-habituation was obtained separately for angry and happy faces. In each case, it was measured as the number of angry and happy stimulus presentations prior to 3 successive SCRs less than 0.05 pS.
3. Results A rejection region of p < 0.05 was used for analysis of main effects and interactions, and adjusted Bonferroni alpha levels were used for multiple comparisons. All main effects and interactions which involved repeated measures were tested using degrees of freedom that were adjusted in terms of the degree of asymmetry in the covariance matrices (Greenhouse & Geisser, 1959). Mean shock levels chosen by the high socially anxious (M = 37.62 V) and the low socially anxious (M = 43.88 V) subjects in the shock work-up and threat-of-shock groups were analyzed using a 2 x 2 (Group x Anxiety) ANOVA. Neither of the main effects nor the interaction was significant [Group : F(1,53) < 1; Anxiety: F(1,53) = 2.24; Group x Anxiety: F(1,53) < 11.
B.M. Clark /Orienting
and social anxiety
217
0.6
0.5 -
-
Angry
-o-
Happy
0.4 -
0.3 -
0.2 -
0.1 -
I
I
I
I
I
1
2
3
4
5
BLOCKS OF TRIALS
Fig. 1. Mean SCR magnitude to the angry and happy face as a function of emotion and blocks of 2 trials during the habituation phase.
3.1. Prestimulation phase The prestimulation phase occurred following the shock work-up procedure and prior to the habituation series. Mean SCL was examined using a 3 X 2 (Group X Anxiety) ANOVA. There were no significant differences as a function of either experimental group or anxiety level; the Group X Anxiety interaction was not significant (all Fs < 1). The NSR data were subjected to a 3 x 2 (Group X Anxiety) ANOVA, which revealed that the frequency of NSRs was influenced by experimental group [ F(2,79) = 3.03, MSe = 164.8, p = 0.051. Mean NSR frequencies were 11.96, 8.9, and 17.19 for the no-shock, shock work-up, and threat-of-shock groups respectively. NSR frequency did not differ as a function of anxiety level (F < 1) and the Group X Anxiety interaction was not significant (F < 1). 3.2. Habituation phase Mean SCR magnitude to the angry and to the happy face was arranged into 5 blocks of 2 trials (see Fig. 1). A 3 x 2 x 2 x 5 (Group x Anxiety x Emotion X Block) ANOVA revealed reliable effects for Emotion [F(1,79) =
218
B.M. Clark /Orienting and social amiety
4.30, MSe = 0.021, Block [F(3,271) = 131.04, MSe = 0.03,], and for the Emotion X Block interaction [F(3,261) = 5.06, MSe = 0.011. Further analysis indicated that responding was significantly greater to the angry face than to the happy face during block 1 [t(316) = 4.22, p < 0.011, but not during blocks 2-5. Mean SCR magnitude did not vary as a function of Anxiety [F(1,79) = 2.261, Group [F(2,79) = 1.091, or any of the interactions in which they participated. The trials-to-habituation scores were assessed using a 3 x 2 X 2 (Group x Anxiety X Emotion) ANOVA which revealed no significant difference in rate of habituation between the angry and the happy face (F < 1). Although the threat-of-shock group (M = 6.12) was slower to habituate to both faces when compared with the no-shock and shock work-up groups (M = 4.86 and M = 4.48 respectively), the differences were not statistically significant [F(2,79) = 2.51, MSe = 16.65, p = 0.093. Similarly, the high socially anxious group (M = 4.60) appeared to habituate more quickly than did the low socially anxious group (M = 5.70), although this difference was also nonsignificant [F(1,79) = 3.06, MSe = 16.65, p = 0.081. No interactions approached significance. Mean SCL was arranged into 5 blocks of 4 trials. A 3 X 2 x 5 (Group x Anxiety X Block) ANOVA revealed a significant decline in SCL across blocks [F(2,169) = 21.94, MSe = 0.061. No other main effects or interactions approached significance. A 3 X 2 (Group X Anxiety) ANOVA of NSR frequency revealed a similar pattern of results to those found in the prestimulation phase. The mean frequency of NSRs (13.997, 10.79, and 21.498 for the no-shock, shock work-up and threat-of-shock groups respectively) differed significantly between groups [ F(2,79) = 3.9, MSe = 220.91. The threat-ofshock group displayed a significantly higher mean NSR frequency than the shock work-up group [t(79) = 2.72, p < 0.011. NSR frequency did not differ as a function of Anxiety [F(1,79) = 1.561 or the Group x Anxiety interaction [F(2/79) < 11.
4. Discussion The results can be summarized as follows. First, SCR magnitude was significantly larger to the angry face than to the happy face on the first block of habituation trials. Second, SCRs to the facial expressions were not influenced by either level of social anxiety or the experimental manipulation of shock threat. Similarly, neither SCL nor the trials-to-habituation measure was found to vary as a function of either social anxiety or shock threat. Third, differences between shock threat groups in terms of frequency of NSRs were evident in both the prestimulation and the habituation phases. The threat-of
B. M. Clark / Orienting and social anxiety
219
shock group displayed more NSRs than did the shock work-up group. Again, frequency of NSRs did not differ as a function of social anxiety. The major finding was that angry faces elicited larger SCRs than happy faces. This finding held across level of social anxiety and across threat conditions. Thus, the effects of an angry facial expression appear to be independent of conditioning history and to occur in the absence of an aversive threat. Moreover, the enhanced responding to the angry face occurred in the physiological response system argued to be the most sensitive to conspecific anxiety. These findings are clearly inconsistent with Ghman’s predictions, and question the role of associative mechanisms in enhanced electrodermal responding to social threat cues. If the role of associative mechanisms in enhanced responding to social threat cues is to be questioned, other explanations must be considered. One possibility involves the notion of stimulus salience. According to general process theories of associative learning (e.g. Mackintosh, 1975; Rescorla & Wagner, 1972), the variable which is of prime importance in determining the rate at which learning occurs is stimulus salience. In acquisition trials, a stimulus that acquires associative strength more rapidly than another stimulus, despite equal reinforcement, is said to be more salient. Salience has been defined operationally in terms of OR amplitude and the rate at which habituation occurs (ohman, Fredrikson, Hugdahl, & Rimmo, 1976). A highly salient stimulus evokes a large OR and requires more trials to reach a criterion of habituation. Thus, it may be argued that the enhanced SCR amplitude to the angry face observed here reflects the greater salience of the angry face, leading even the low-fear subjects to respond strongly in the early trials. However, there is no obvious reason why the two faces presented in this study differed with regard to stimulus salience, as both were human faces of the same complexity and brightness. Additionally, stimulus salience does not provide a convincing explanation for the results of conditioning studies in the preparedness literature. The most robust empirical finding is resistance to extinction of CRs conditioned to fear-relevant stimuli. This finding is in direct contrast to the salience hypothesis which predicts faster extinction to highly salient stimuli (Rescorla & Wagner, 1972). A second explanation can be fashioned in terms of the frequency with which angry and happy faces are encountered outside the laboratory. It could be argued, for example, that exposure to happy faces is more prevalent than exposure to angry faces. If this is the case, we might expect more habituation to have occurred to the former than to the latter. One implication of this argument is that other less common facial expressions (e.g. disgust) will also elicit more vigorous orienting than will happy faces. Clearly, this issue needs to be pursued further. A third explanation can be offered in terms of prepotency. The prepotency hypothesis states that humans are biologically predisposed to attend
220
B.M. Clark /Orienting and social anxiety
selectively to particular types of stimuli and to respond to such stimuli with high levels of arousal (Marks, 1981). Although prepotency is conceptually related to the concept of preparedness (Seligman, 1971) in that both emphasize biologically determined behavioural responses, they differ in terms of their theoretical nuances. For example, prepotent stimuli such as direct gaze promote nonspecific arousal which is not necessarily connected with a particular response. In contrast, preparedness theory predicts that direct gaze is biologically prepared to elicit the specific emotional response of fear following associations with an aversive event (McNally, 1987). Another important difference between the two theories concerns the role of associative learning in the enhanced responding seen to fear-relevant stimuli. Whereas preparedness explains enhanced responding in terms of selective associability, the prepotency hypothesis actually questions the role of associative learning mechanisms. The enhanced electrodermal responding to the angry face found in both the high and low socially anxious groups in this study is consistent with the prepotency hypothesis in that it suggests a performance effect that is not dependent on associative learning. Indeed, the need to consider nonassociative performance effects has been highlighted by the results of a conditioning study reported by Dimberg (1987). Although Dimberg reported superior resistance to extinction to the angry face, he obtained no reliable differential conditioning for either the angry or the happy group during acquisition. Thus, Dimberg’s results are clear evidence of a performance effect, displayed in the absence of conditioned acquisition. The assumption underlying the preselection of subjects in terms of social anxiety was that high- and low-fear subjects may differ with regard to their conditioning histories in social encounters. According to Ghman (19861, a conditioning history of repeated failures in dominance conflicts results in a conditioned fear response, as indicated by marked social anxiety in social encounters, and increased electrodermal responding to social threat cues. Thus, it was hypothesized that increased electrodermal responding to the angry face will be evidenced only in the high socially anxious group. The present finding that both high- and low-fear groups responded more to the angry than to the happy face is clearly inconsistent with the experimental hypothesis. Although it could be argued that the predicted pattern of results might have been obtained with other measures (e.g. cardiac activity), ehman (1986; Ghman et al., 198.5) has specifically predicted SCR differences to social stimuli in high and low socially anxious individuals. Based on the functional-evolutionary perspective, Ohman proposed that interspecific and conspecific fears will be characterized by different emotional response patterns. Contact with interspecific phobic stimuli can be expected to elicit autonomic activation appropriate to a fight or flight response. This activation includes cardiac acceleration and cephalic vasoconstriction. In contrast, the
B.M. Clark /Orienting
221
and social anxiety
responses to conspecific fear stimuli are not tightly organized around vigorous escape or avoidance responses, and ohman et al. (1985) have suggested that the detection of threatening social stimuli by socially anxious individuals will result in “signs of anxious vigilance for potential attacks” (p. 163). This, they argue, is evidenced in enhanced SCRs to threat cues. The present results failed to replicate previous data concerning the effects of threat of shock on SCL and trials-to-habituation (e.g. Bohlin, 1976; Gatchel & Gaas, 1976). They also failed to replicate previous findings that threat of shock in an habituation paradigm potentiates the ORs to fear-relevant stimuli (e.g. hhman et al., 1974). There are, however, important differences in experimental procedure between the present study and other studies which have used shock threat. It is possible that anticipation of an “uncomfortable, but not painful” level of shock intensity was not a sufficiently aversive threat. Of the earlier experiments which found reliable effects for shock threat, two did not include an initial shock work-up procedure (Gatchel & Gaas, 1976; Hodges & Spielberger, 19661, and several other studies threatened the subjects with receiving a more aversive shock than they had selected during the shock work-up procedure (e.g. Bridger & Mandel, 1964; Hugdahl & Ghman, 1977). It may be that these procedures produced a greater increase in arousal than was achieved in the present study. In conclusion, the present study demonstrated that elevated electrodermal responding to social threat cues can occur even in the absence of associative learning. The data appear to question the conclusion that conditioning differences between angry and happy faces result from selective associations. Rather, the results encourage a closer examination of prepotency and prior learning hypotheses. However, the within-subject manipulation of facial expression may have produced sensitization or adaptation level effects, and additional work using a between-groups design is clearly required.
Acknowledgements This research was supported by a grant from the Australian Research Council to the second and third authors. We wish to thank Len Glue for his assistance with computer programming and Jeanette Packer and Evalynn Mazurski for their advice and assistance.
References Bohlin, G. (1976). Delayed habituation of the electrodermal orienting increased level of arousal. Psychophysiology, 13, 345-351.
response
as a function
of
222
B.M. Clark /Orienting and social anxiety
Bridger, W.H., & Mandel, I.J. (1964). A comparison of GSR fear responses produced by threat and electric shock. Journal of Psychiatric Research, 2, 31-40. Dimberg, U. (1987). Facial reactions, autonomic activity, and experienced emotion: A three component model of emotional conditioning. Biological Psychology, 24, 105-122. Dimberg, U., & Ghman, A. (1983). The effects of directional facial cues on electrodermal conditioning to facial stimuli. Psychophysiology, 20, 160-167. Ekman, P. (1973). Cross-cultural studies of facial expression. In P. Ekman (Ed.), Darwin and facial expression (pp. 169-222). New York: Academic Press. Ekman, P., & Friesen, W.V. (1976). Pictures of facial affect. Palo Alto: Consulting Psychologists Press. Gatchel, R.J., & Gaas, E. (1976). Effects of arousal level on short-and long-term habituation of the orienting response. Physiological Psychology, 4, 66-68. Greenhouse, SW., & Geisser, S. (1959). On methods in the analysis of profile data. Psychometrika, 24, 95-112. Hodges, W.F., & Spielberger, CD. (1966). The effects of threat of shock on heart rate for subjects who differ in manifest anxiety and fear of shock. Psychophysiology, 2, 287-294. Hugdahl, K., & Ghman, A. (1977). Effects of instruction on acquisition and extinction of electrodermal responses to fear-relevant stimuli. Journal of Experimental Psychology: Human Learning and Memory, 3, 608-618. Kartsounis, L.D., & Pickersgill, M.J. (1981). Orienting responses to stimuli others fear. British Journal of Clinical Psychology, 20, 261-273. Mackintosh, N.J. (1975). A theory of attention: Variations in the associability of stimuli with reinforcement. Psychological Review, 82, 276-298. Marks, I.M. (1981). Cure and care of neuroses: Theory and practice of behavioral psychotherapy. Toronto: Wiley. Mayr, E. (1974). Behavior programs and evolutionary strategies. American Scientist, 62, 650-659. McNally, R.J. (1987). Preparedness and phobias: A review. Psychological Bulletin, ZOI, 283-303. Ghman, A. (1986). Face the beast and fear the face: Animal and social fears as prototypes for evolutionary analyses of emotion. Psychophysiology, 23, 123-145. Ghman, A., & Dimberg, U. (1978). Facial expressions as conditioned stimuli for electrodermal responses: A case of “Preparedness”? Journal of Personality and Social Psychology, 36, 1251-1258. Ghman, A., & Dimberg, U. (1984). An evolutionary perspective on human social behavior. In W.M. Waid (Ed.), Sociophysiology (pp. 47-86). New York: Spruzer-Verley. Ghman, A., Dimberg, U., & Ost, L.-G. (1985). Animal and social phobias: Biological constraints on learned fear responses. In S. Reiss, & R.R. Bootzin (Eds.), Theoretical issues in behatlior therapy. Orlando: Academic Press. Ghman, A., Eriksson, A., Fredriksson, M., Hugdahl, K., & Olofsson, C. (1974). Habituation of the electrodermal orienting reaction to potentially phobic and supposedly neutral stimuli in normal human subjects. Biological Psychology, 2, 85-93. Ghman, A., Fredrikson, M., Hugdahl, K., & Rimmo, P.-A. (1976). The premise of equipotentiality in human classical conditioning: Conditioned electrodermal responses to potentially phobic stimuli. Journal of Experimental Psychology: General, 105, 3133337. Orr, S.P., & Lanretta, J.T. (1980). Facial expressions of emotion as conditioned stimuli for human autonomic responses. Journal of Personality and Social Psychology, 38, 2788282. Rescorla, R.A., & Wagner, A.R. (1972). A theory of Pavlovian conditioning: Variations in the effectiveness of reinforcement and nonreinforcement. In A.H. Black, & W.F. Prokasy (Eds.), Classical conditioning II: Current theory and research (pp. 64-99). New York: Appleton-Century-Crofts. Seligman, M.E.P. (1971). Phobias and preparedness. Behacior Therapy, 2, 307-320.
B.M. Clark /Orienting
and social anxiety
223
Turner, R.M., Meles, D., & DiTomasso, R. (1983). Assessment of social anxiety: A controlled comparison among social phobics, obsessive-compulsives, agoraphobics, sexual disorders and simple phobics. Behauiour Research and Therapy, 21, 181-183. Tursky, B., Watson, P.D., & O’Connell, D.N. (1965). A concentric shock electrode for pain stimulation. Psychophysiology, I, 296-298. Willoughby, R.R. (1932). Some properties of the Thurstone Personality Schedule and a suggested revision. Journal of Social Psychology, 3, 401-424.
211 0301-0511/92/$05.00
Effects of social anxiety and facial expression on habituation of the electrodermal orienting response Belinda
M. Clark
‘, David A.T. Siddle
’ and Nigel W. Bond
’
’ Macquarie University, Macquarie, NSW 2109, Australia 2 Department of Psychology, University of Queensland, St. Lucia, Queensland 4072, Australia
The present research examined electrodermal orienting to happy and angry faces as a function of social anxiety and threat of shock. A preliminary study using 569 undergraduate participants developed an adequate set of normative data of social anxiety for the Willoughby questionnaire (WQ) for use in subject selection. Electrodermal activity was measured in both high and low socially anxious subjects (N = 85) during exposure to 10 presentations of an angry face intermixed with 10 presentations of a happy face. Threat of shock (no-shock, shock work-up only, and shock work-up plus threat) was also manipulated. Skin conductance responses (SCRs) which occurred within l-4 s of stimulus onset and trials-to-habituation constituted the data of primary interest. Although trials-to-habituation did not differ between angry and happy facial expressions, SCRs were larger to the angry face than to the happy face in both high and low socially anxious subjects. No differences in SCR magnitude were found as a function of threat of shock. The implications of these results for 6hman’s functional-evolutionary model of social phobia are discussed, and alternative explanations in terms of prepotency and prior learning are examined.
1. Introduction ohman’s (1986) functional-evolutionary approach to human phobias analyses phobias in terms of the behavioural systems within which the fear has evolved (ijhman, Dimberg, & Ost, 1985). According to Ohman (1986), social phobias originate in a social submissive system of ancient evolutionary origin. In order to maximize the benefits which may be gained from social living, conspecifics must maintain a relatively stable social order where agonistic encounters are minimal (6hman et al., 1985). This is usually achieved by means of a dominance/submissiveness hierarchy (Ohman, 1986) in which dominance is established through agonistic encounters. Such encounters are usually restricted to a symbolic level, and are characterized by intricate symbolic displays. Correspondence to: Dr. David Siddle, Lucia, Queensland 4072, Australia.
Department
of Psychology,
University
of Queensland,
St.
212
B.M. Clark / Orienfing and social anxiety
In primates, facial expression plays a definitive role in the nonverbal signals used in agonistic encounters (Ghman, 1986). For example, the facial threat display of dominant male primates, including man, is unmistakable as an angry, silent glare, with lowered corners to the mouth over bared teeth, with wide open staring eyes and frowning, lowered brows (Ekman, 1973). Following Mayr (1974), ehman and his colleagues (ohman, 1986; bhman et al., 1985) have argued that the symbolic displays evidenced in conspecific agonistic encounters are most likely to be characterized by relatively innate behaviours because of the high selective premium for these signals to be unmistakable. Further, the behavioural context within which symbolic displays are expressed has remained stable throughout evolutionary history. For example, aversive emotions such as anger signal the threat of impending attack. Happiness, on the other hand, always appears in a context antagonistic to anger, signalling play or filial attachment. Thus, from a functionalevolutionary perspective, the signalling function of an emotion could be expected to place biological constraints on its associability (hhman & Dimberg, 1984). It follows, therefore, that as a result of natural selection, anger and fear will be biologically prepared to signal aversive events, and happiness will be biologically unprepared. In a series of studies of Pavlovian conditioning, Dimberg and bhman (1983; ohman & Dimberg, 1978) have reported that, if a picture of an angry facial expression is used as the conditioned stimulus (CS) and shock as the unconditioned stimulus (US), the conditioned response (CR) is more resistant to extinction and less amenable to instructional manipulation than is a response conditioned to a picture of a neutral or a happy facial expression. In addition, Orr and Lanzetta (1980) demonstrated superior differential acquisition when pictures of fear expressions were compared with happy expressions. Not only were skin conductance responses (SCRs) larger to CS + than to CS - when a fearful face was the CS + , but more rapid acquisition of the differential response was also demonstrated. Thus, Orr and Lanzetta’s (1980) results are consistent with the theory that angry or fearful expressions are selectively associable with a shock US. The basic assumption underlying 6hman’s (1986) functional-evolutionary approach is that angry faces are selectively associable with an aversive US, and not that SCRs are more likely or more persistent with angry faces than with happy faces. However, some studies have found that other fear-relevant stimuli such as snakes and spiders elicit more vigorous orienting and require more trials to reach a criterion of habituation than do fear-irrelevant stimuli such as flowers and mushrooms (6hman, Eriksson, Fredriksson, Hugdahl, & Olofsson, 1974). If angry faces can be shown to elicit larger and more slowly habituating SCRs than happy faces prior to acquisition trials, then the results of conditioning studies might be explained in terms of the intrinsic stimulus properties of angry faces. In view of the importance of electrodermal orient-
B.M. Clark /Orieniing
and social anxiety
213
ing and habituation to an understanding of the mechanisms underlying selective associations, one of the aims of the present study was to examine orienting responses (ORs) to pictures of angry and happy facial expressions of emotion in an habituation paradigm. Because autonomic responding to facial threat displays is said to reflect the conditioning history of the subject (hhman & Dimberg, 1984), a second aim was to investigate the effect of prior conditioning on orienting to angry and happy facial expressions. The role of learning is emphasized explicitly in accounts of the functional-evolutionary model of social phobias. As bhman to point out that we are talking about et al. (1985) state, “it is important readiness to learn rather than about biologically given responses” (p. 142). Thus, an individual who has repeatedly experienced humiliation and defeat in social encounters with dominant individuals will, according to the model, preattentively “tag” facial threat displays as significant, leading to the affective reaction of fear. On the other hand, individuals who repeatedly experience victory and dominance in the same situation will not tag threat displays as significant, even though they share a biological readiness to associate fear with the same social stimuli. Thus, a fear response will not occur in these individuals following the perception of facial threat displays. According to ehman (1986), a conditioning history of repeated failures in dominance conflicts results in a conditioned fear response as indicated by marked social anxiety in social encounters and increased electrodermal responding to social threat cues. Thus, subjects were preselected in terms of self-reported level of social anxiety. It was hypothesized that the high socially anxious subjects, with their assumed history of social defeat, will display larger autonomic responses to pictures of social threat cues than will low socially anxious subjects who are assumed to have a conditioning history of success in agonistic encounters. In addition to these variables, we also manipulated threat of shock in an attempt to examine the possible nonassociative effects of threat on differential responding to fear-relevant social stimuli. Although threat of shock has been found to potentiate responding to certain fear-relevant stimuli (e.g. Kartsounis & Pickersgill, 1981; ohman et al., 1974), the effects of selective sensitization on social threat cues has not been examined. Subjects with high and low social anxiety scores were allocated to one of three groups. In the first group (no-shock) subjects were exposed to repeated presentations of two facial stimuli. The second group (threat-of-shock) underwent a shock work-up procedure and was informed that further shock would occur during the habituation phase. The third group (shock work-up) underwent shock work-up, but were not threatened with further shock. This group constituted a control for nonspecific arousal. Because most conditioning studies with facial stimuli have used differential conditioning in which subjects receive both CS + and CS - presentations, facial expression was
214
B. M. Clark /Orienting
and social anxie@
manipulated within subjects. Thus, the design was a 2 X 2 X 3 (Social Anxiety X Stimulus x Group) factorial design, with repeated measures on type of stimulus. A trials factor was included as a repeated measure in the statistical analyses.
2. Method 2.1. Subject selection In a preliminary study, the Willoughby questionnaire (Willoughby, 1932) was administered to 569 undergraduate volunteers (406 women and 163 men). Mean age was 22.8 years and the range 17-65 years. Of these, 237 (48 males and 189 females) were tested after an interval of 22 weeks. Both split-half (r = 0.87) and test-retest reliability (r = 0.80) were satisfactory. Factor analysis revealed three factors which were labelled “hypersensitivity to interpersonal relationships” (30% of variance), “negative affect” (7.6% of variance) and “hypersensitivity” (5.4% of variance). These and other analyses indicated that the Willoughby questionnaire is an internally consistent and reliable instrument which primarily measures hypersensitivity to interpersonal relationships. The apparent relevance of the extracted factors to the symptoms of social phobia, together with the construct validity reported by Turner, Meles, and DiTomassa (19831, support the utility of the Willoughby questionnaire as an instrument for identifying subjects high and low in interpersonal (social) anxiety. Subjects in the main study were 85 undergraduates (age range 17-32 years) who were contacted by telephone and asked to volunteer. The sample consisted of 45 low socially anxious subjects (19 males and 26 females) who obtained a midpoint percentile ranking less than 14.02% on the Willoughby questionnaire (M score = 13.961, and 40 high socially anxious subjects (7 males and 33 females) who obtained a midpoint percentile ranking greater than 87.53% (M score = 58.97). Subjects provided informed consent and participated as partial fulfilment of a course requirement. They were allocated randomly to the three experimental conditions with the restriction that within each social anxiety level all groups contained the same proportion (low, 2 : 3; high, 2 : 11) of men and women. The experiment was double-blind in that neither the subjects nor the experimenter knew the subjects’ social anxiety score. 2.2. Apparatus Skin conductance was recorded directly by using a constant voltage of 0.5 V applied across domed Ag-AgC1 electrodes in conjunction with a 0.05 M NaCl electrolyte. Conductance was recorded from masked areas on the distal
B.M. Clark /Orienting
and social anxiety
215
phalanges of the index and second fingers of the subject’s left hand using a Grass 7PIF preamplifier and a chart speed of 2.5 mm s-‘. Recording sensitivity was 0.05 microsiemens &S) mm-’ of pen deflection. Respiration was recorded using a chest pneumograph bellows (Phipps and Bird, model RPBA) in conjunction with a Grass 7PIF preamplifier. A locally constructed constant voltage stimulator (O-90 V) produced pulsed shocks. Pulse duration was 2 ms and the repetition rate was 50 Hz. Shock was delivered via a 36 mm diameter concentric electrode in which contact between the skin and metal elements was made through normal saline-soaked sponge pads (Tursky, Watson, & O’Connell, 1965). The shock electrode was placed on the dorsal surface of the right forearm, 17 cm above the ulna styroid. The electrode site was prepared with EC2 Electrode Cream (Grass Instruments) prior to electrode attachment to ensure stability of the electrode-skin interface. The visual stimuli were two black-and-white slides, one showing a male with an angry facial expression and the other showing a different male with a happy facial expression. These two slides were selected from Ekman and Friesen’s (1976) Pictures of facial affect on the basis of their inter-rater reliability score of r = 1.0 (Ekman & Friesen, 1976). The slides were projected through a small window in the wall of the experimental chamber by a Pradovit projector fitted with a tachistoscopic shutter (Gerbrands, Model G1166). A 60 cm X 95 cm image was produced on a white screen situated 200 cm in front of the subject at eye level. 2.3. Procedure Each subject in the threat-of-shock and shock work-up groups was fitted with the shock electrode and given a shock work-up procedure in the laboratory control room. The subject was asked to report when the shock level was “uncomfortable, but not painful”. Shocks of 0.5 s duration were then delivered, commencing at 0 V and incrementing by 0.5 V, with the subject reporting at each shock presentation. The shock electrode was then removed from subjects in the shock work-up group, who were informed that they would not receive any further shocks. Subjects were then seated in a semi-reclining padded chair in the sound-attenuated experimental room. Ambient illumination was 2 lux, the temperature between 19°C and 25°C and the relative humidity within the range 48-56%. The stimulus equipment and recording apparatus were housed in an adjoining room. Following attachment of the skin conductance electrodes and chest pneumograph bellows, the subjects were informed that the first part of the experiment involved a 5 min rest period during which they were to relax with their eyes open.
216
B.M. Clark /Orienting and social anxiety
Following the prestimulation period all subjects were informed that some pictures of faces would appear from time to time on the screen. They were instructed to remain as still as possible, and to watch the screen at all times. Subjects in the threat-of-shock group were informed that some of the pictures might be followed by a brief electric shock at the intensity level set earlier. During the habituation phase there were 10 presentations each of an angry face and a happy face. The order of the first two stimulus presentations was counterbalanced. Half the subjects received the angry face followed by the happy face, whereas for the other half, the order was reversed. The remaining stimuli were presented randomly, with the restriction that not more than two consecutive slides were the same. All stimuli were of 2 s duration, and were presented at randomly ordered intervals of 20, 2.5, and 30 s (offset to onset). 2.4. Scoring An SCR greater than 0.05 PUS which occurred within 1-4 s of stimulus onset was scored as a response to that stimulus. In order to normalize the SCR data, they were subjected to a square root transformation prior to analysis. Any SCR greater than 0.05 PUSwhich occurred in the absence of a stimulus, or outside the l-4 s window following stimulus onset, was considered a nonspecific response (NSR). Skin conductance level (SCL) was scored at 1 min intervals during the prestimulation phase, and a mean SCL was obtained for that period. During the habituation phase, tonic SCL was scored immediately prior to stimulus onset. A measure of trials-to-habituation was obtained separately for angry and happy faces. In each case, it was measured as the number of angry and happy stimulus presentations prior to 3 successive SCRs less than 0.05 pS.
3. Results A rejection region of p < 0.05 was used for analysis of main effects and interactions, and adjusted Bonferroni alpha levels were used for multiple comparisons. All main effects and interactions which involved repeated measures were tested using degrees of freedom that were adjusted in terms of the degree of asymmetry in the covariance matrices (Greenhouse & Geisser, 1959). Mean shock levels chosen by the high socially anxious (M = 37.62 V) and the low socially anxious (M = 43.88 V) subjects in the shock work-up and threat-of-shock groups were analyzed using a 2 x 2 (Group x Anxiety) ANOVA. Neither of the main effects nor the interaction was significant [Group : F(1,53) < 1; Anxiety: F(1,53) = 2.24; Group x Anxiety: F(1,53) < 11.
B.M. Clark /Orienting
and social anxiety
217
0.6
0.5 -
-
Angry
-o-
Happy
0.4 -
0.3 -
0.2 -
0.1 -
I
I
I
I
I
1
2
3
4
5
BLOCKS OF TRIALS
Fig. 1. Mean SCR magnitude to the angry and happy face as a function of emotion and blocks of 2 trials during the habituation phase.
3.1. Prestimulation phase The prestimulation phase occurred following the shock work-up procedure and prior to the habituation series. Mean SCL was examined using a 3 X 2 (Group X Anxiety) ANOVA. There were no significant differences as a function of either experimental group or anxiety level; the Group X Anxiety interaction was not significant (all Fs < 1). The NSR data were subjected to a 3 x 2 (Group X Anxiety) ANOVA, which revealed that the frequency of NSRs was influenced by experimental group [ F(2,79) = 3.03, MSe = 164.8, p = 0.051. Mean NSR frequencies were 11.96, 8.9, and 17.19 for the no-shock, shock work-up, and threat-of-shock groups respectively. NSR frequency did not differ as a function of anxiety level (F < 1) and the Group X Anxiety interaction was not significant (F < 1). 3.2. Habituation phase Mean SCR magnitude to the angry and to the happy face was arranged into 5 blocks of 2 trials (see Fig. 1). A 3 x 2 x 2 x 5 (Group x Anxiety x Emotion X Block) ANOVA revealed reliable effects for Emotion [F(1,79) =
218
B.M. Clark /Orienting and social amiety
4.30, MSe = 0.021, Block [F(3,271) = 131.04, MSe = 0.03,], and for the Emotion X Block interaction [F(3,261) = 5.06, MSe = 0.011. Further analysis indicated that responding was significantly greater to the angry face than to the happy face during block 1 [t(316) = 4.22, p < 0.011, but not during blocks 2-5. Mean SCR magnitude did not vary as a function of Anxiety [F(1,79) = 2.261, Group [F(2,79) = 1.091, or any of the interactions in which they participated. The trials-to-habituation scores were assessed using a 3 x 2 X 2 (Group x Anxiety X Emotion) ANOVA which revealed no significant difference in rate of habituation between the angry and the happy face (F < 1). Although the threat-of-shock group (M = 6.12) was slower to habituate to both faces when compared with the no-shock and shock work-up groups (M = 4.86 and M = 4.48 respectively), the differences were not statistically significant [F(2,79) = 2.51, MSe = 16.65, p = 0.093. Similarly, the high socially anxious group (M = 4.60) appeared to habituate more quickly than did the low socially anxious group (M = 5.70), although this difference was also nonsignificant [F(1,79) = 3.06, MSe = 16.65, p = 0.081. No interactions approached significance. Mean SCL was arranged into 5 blocks of 4 trials. A 3 X 2 x 5 (Group x Anxiety X Block) ANOVA revealed a significant decline in SCL across blocks [F(2,169) = 21.94, MSe = 0.061. No other main effects or interactions approached significance. A 3 X 2 (Group X Anxiety) ANOVA of NSR frequency revealed a similar pattern of results to those found in the prestimulation phase. The mean frequency of NSRs (13.997, 10.79, and 21.498 for the no-shock, shock work-up and threat-of-shock groups respectively) differed significantly between groups [ F(2,79) = 3.9, MSe = 220.91. The threat-ofshock group displayed a significantly higher mean NSR frequency than the shock work-up group [t(79) = 2.72, p < 0.011. NSR frequency did not differ as a function of Anxiety [F(1,79) = 1.561 or the Group x Anxiety interaction [F(2/79) < 11.
4. Discussion The results can be summarized as follows. First, SCR magnitude was significantly larger to the angry face than to the happy face on the first block of habituation trials. Second, SCRs to the facial expressions were not influenced by either level of social anxiety or the experimental manipulation of shock threat. Similarly, neither SCL nor the trials-to-habituation measure was found to vary as a function of either social anxiety or shock threat. Third, differences between shock threat groups in terms of frequency of NSRs were evident in both the prestimulation and the habituation phases. The threat-of
B. M. Clark / Orienting and social anxiety
219
shock group displayed more NSRs than did the shock work-up group. Again, frequency of NSRs did not differ as a function of social anxiety. The major finding was that angry faces elicited larger SCRs than happy faces. This finding held across level of social anxiety and across threat conditions. Thus, the effects of an angry facial expression appear to be independent of conditioning history and to occur in the absence of an aversive threat. Moreover, the enhanced responding to the angry face occurred in the physiological response system argued to be the most sensitive to conspecific anxiety. These findings are clearly inconsistent with Ghman’s predictions, and question the role of associative mechanisms in enhanced electrodermal responding to social threat cues. If the role of associative mechanisms in enhanced responding to social threat cues is to be questioned, other explanations must be considered. One possibility involves the notion of stimulus salience. According to general process theories of associative learning (e.g. Mackintosh, 1975; Rescorla & Wagner, 1972), the variable which is of prime importance in determining the rate at which learning occurs is stimulus salience. In acquisition trials, a stimulus that acquires associative strength more rapidly than another stimulus, despite equal reinforcement, is said to be more salient. Salience has been defined operationally in terms of OR amplitude and the rate at which habituation occurs (ohman, Fredrikson, Hugdahl, & Rimmo, 1976). A highly salient stimulus evokes a large OR and requires more trials to reach a criterion of habituation. Thus, it may be argued that the enhanced SCR amplitude to the angry face observed here reflects the greater salience of the angry face, leading even the low-fear subjects to respond strongly in the early trials. However, there is no obvious reason why the two faces presented in this study differed with regard to stimulus salience, as both were human faces of the same complexity and brightness. Additionally, stimulus salience does not provide a convincing explanation for the results of conditioning studies in the preparedness literature. The most robust empirical finding is resistance to extinction of CRs conditioned to fear-relevant stimuli. This finding is in direct contrast to the salience hypothesis which predicts faster extinction to highly salient stimuli (Rescorla & Wagner, 1972). A second explanation can be fashioned in terms of the frequency with which angry and happy faces are encountered outside the laboratory. It could be argued, for example, that exposure to happy faces is more prevalent than exposure to angry faces. If this is the case, we might expect more habituation to have occurred to the former than to the latter. One implication of this argument is that other less common facial expressions (e.g. disgust) will also elicit more vigorous orienting than will happy faces. Clearly, this issue needs to be pursued further. A third explanation can be offered in terms of prepotency. The prepotency hypothesis states that humans are biologically predisposed to attend
220
B.M. Clark /Orienting and social anxiety
selectively to particular types of stimuli and to respond to such stimuli with high levels of arousal (Marks, 1981). Although prepotency is conceptually related to the concept of preparedness (Seligman, 1971) in that both emphasize biologically determined behavioural responses, they differ in terms of their theoretical nuances. For example, prepotent stimuli such as direct gaze promote nonspecific arousal which is not necessarily connected with a particular response. In contrast, preparedness theory predicts that direct gaze is biologically prepared to elicit the specific emotional response of fear following associations with an aversive event (McNally, 1987). Another important difference between the two theories concerns the role of associative learning in the enhanced responding seen to fear-relevant stimuli. Whereas preparedness explains enhanced responding in terms of selective associability, the prepotency hypothesis actually questions the role of associative learning mechanisms. The enhanced electrodermal responding to the angry face found in both the high and low socially anxious groups in this study is consistent with the prepotency hypothesis in that it suggests a performance effect that is not dependent on associative learning. Indeed, the need to consider nonassociative performance effects has been highlighted by the results of a conditioning study reported by Dimberg (1987). Although Dimberg reported superior resistance to extinction to the angry face, he obtained no reliable differential conditioning for either the angry or the happy group during acquisition. Thus, Dimberg’s results are clear evidence of a performance effect, displayed in the absence of conditioned acquisition. The assumption underlying the preselection of subjects in terms of social anxiety was that high- and low-fear subjects may differ with regard to their conditioning histories in social encounters. According to Ghman (19861, a conditioning history of repeated failures in dominance conflicts results in a conditioned fear response, as indicated by marked social anxiety in social encounters, and increased electrodermal responding to social threat cues. Thus, it was hypothesized that increased electrodermal responding to the angry face will be evidenced only in the high socially anxious group. The present finding that both high- and low-fear groups responded more to the angry than to the happy face is clearly inconsistent with the experimental hypothesis. Although it could be argued that the predicted pattern of results might have been obtained with other measures (e.g. cardiac activity), ehman (1986; Ghman et al., 198.5) has specifically predicted SCR differences to social stimuli in high and low socially anxious individuals. Based on the functional-evolutionary perspective, Ohman proposed that interspecific and conspecific fears will be characterized by different emotional response patterns. Contact with interspecific phobic stimuli can be expected to elicit autonomic activation appropriate to a fight or flight response. This activation includes cardiac acceleration and cephalic vasoconstriction. In contrast, the
B.M. Clark /Orienting
221
and social anxiety
responses to conspecific fear stimuli are not tightly organized around vigorous escape or avoidance responses, and ohman et al. (1985) have suggested that the detection of threatening social stimuli by socially anxious individuals will result in “signs of anxious vigilance for potential attacks” (p. 163). This, they argue, is evidenced in enhanced SCRs to threat cues. The present results failed to replicate previous data concerning the effects of threat of shock on SCL and trials-to-habituation (e.g. Bohlin, 1976; Gatchel & Gaas, 1976). They also failed to replicate previous findings that threat of shock in an habituation paradigm potentiates the ORs to fear-relevant stimuli (e.g. hhman et al., 1974). There are, however, important differences in experimental procedure between the present study and other studies which have used shock threat. It is possible that anticipation of an “uncomfortable, but not painful” level of shock intensity was not a sufficiently aversive threat. Of the earlier experiments which found reliable effects for shock threat, two did not include an initial shock work-up procedure (Gatchel & Gaas, 1976; Hodges & Spielberger, 19661, and several other studies threatened the subjects with receiving a more aversive shock than they had selected during the shock work-up procedure (e.g. Bridger & Mandel, 1964; Hugdahl & Ghman, 1977). It may be that these procedures produced a greater increase in arousal than was achieved in the present study. In conclusion, the present study demonstrated that elevated electrodermal responding to social threat cues can occur even in the absence of associative learning. The data appear to question the conclusion that conditioning differences between angry and happy faces result from selective associations. Rather, the results encourage a closer examination of prepotency and prior learning hypotheses. However, the within-subject manipulation of facial expression may have produced sensitization or adaptation level effects, and additional work using a between-groups design is clearly required.
Acknowledgements This research was supported by a grant from the Australian Research Council to the second and third authors. We wish to thank Len Glue for his assistance with computer programming and Jeanette Packer and Evalynn Mazurski for their advice and assistance.
References Bohlin, G. (1976). Delayed habituation of the electrodermal orienting increased level of arousal. Psychophysiology, 13, 345-351.
response
as a function
of
222
B.M. Clark /Orienting and social anxiety
Bridger, W.H., & Mandel, I.J. (1964). A comparison of GSR fear responses produced by threat and electric shock. Journal of Psychiatric Research, 2, 31-40. Dimberg, U. (1987). Facial reactions, autonomic activity, and experienced emotion: A three component model of emotional conditioning. Biological Psychology, 24, 105-122. Dimberg, U., & Ghman, A. (1983). The effects of directional facial cues on electrodermal conditioning to facial stimuli. Psychophysiology, 20, 160-167. Ekman, P. (1973). Cross-cultural studies of facial expression. In P. Ekman (Ed.), Darwin and facial expression (pp. 169-222). New York: Academic Press. Ekman, P., & Friesen, W.V. (1976). Pictures of facial affect. Palo Alto: Consulting Psychologists Press. Gatchel, R.J., & Gaas, E. (1976). Effects of arousal level on short-and long-term habituation of the orienting response. Physiological Psychology, 4, 66-68. Greenhouse, SW., & Geisser, S. (1959). On methods in the analysis of profile data. Psychometrika, 24, 95-112. Hodges, W.F., & Spielberger, CD. (1966). The effects of threat of shock on heart rate for subjects who differ in manifest anxiety and fear of shock. Psychophysiology, 2, 287-294. Hugdahl, K., & Ghman, A. (1977). Effects of instruction on acquisition and extinction of electrodermal responses to fear-relevant stimuli. Journal of Experimental Psychology: Human Learning and Memory, 3, 608-618. Kartsounis, L.D., & Pickersgill, M.J. (1981). Orienting responses to stimuli others fear. British Journal of Clinical Psychology, 20, 261-273. Mackintosh, N.J. (1975). A theory of attention: Variations in the associability of stimuli with reinforcement. Psychological Review, 82, 276-298. Marks, I.M. (1981). Cure and care of neuroses: Theory and practice of behavioral psychotherapy. Toronto: Wiley. Mayr, E. (1974). Behavior programs and evolutionary strategies. American Scientist, 62, 650-659. McNally, R.J. (1987). Preparedness and phobias: A review. Psychological Bulletin, ZOI, 283-303. Ghman, A. (1986). Face the beast and fear the face: Animal and social fears as prototypes for evolutionary analyses of emotion. Psychophysiology, 23, 123-145. Ghman, A., & Dimberg, U. (1978). Facial expressions as conditioned stimuli for electrodermal responses: A case of “Preparedness”? Journal of Personality and Social Psychology, 36, 1251-1258. Ghman, A., & Dimberg, U. (1984). An evolutionary perspective on human social behavior. In W.M. Waid (Ed.), Sociophysiology (pp. 47-86). New York: Spruzer-Verley. Ghman, A., Dimberg, U., & Ost, L.-G. (1985). Animal and social phobias: Biological constraints on learned fear responses. In S. Reiss, & R.R. Bootzin (Eds.), Theoretical issues in behatlior therapy. Orlando: Academic Press. Ghman, A., Eriksson, A., Fredriksson, M., Hugdahl, K., & Olofsson, C. (1974). Habituation of the electrodermal orienting reaction to potentially phobic and supposedly neutral stimuli in normal human subjects. Biological Psychology, 2, 85-93. Ghman, A., Fredrikson, M., Hugdahl, K., & Rimmo, P.-A. (1976). The premise of equipotentiality in human classical conditioning: Conditioned electrodermal responses to potentially phobic stimuli. Journal of Experimental Psychology: General, 105, 3133337. Orr, S.P., & Lanretta, J.T. (1980). Facial expressions of emotion as conditioned stimuli for human autonomic responses. Journal of Personality and Social Psychology, 38, 2788282. Rescorla, R.A., & Wagner, A.R. (1972). A theory of Pavlovian conditioning: Variations in the effectiveness of reinforcement and nonreinforcement. In A.H. Black, & W.F. Prokasy (Eds.), Classical conditioning II: Current theory and research (pp. 64-99). New York: Appleton-Century-Crofts. Seligman, M.E.P. (1971). Phobias and preparedness. Behacior Therapy, 2, 307-320.
B.M. Clark /Orienting
and social anxiety
223
Turner, R.M., Meles, D., & DiTomasso, R. (1983). Assessment of social anxiety: A controlled comparison among social phobics, obsessive-compulsives, agoraphobics, sexual disorders and simple phobics. Behauiour Research and Therapy, 21, 181-183. Tursky, B., Watson, P.D., & O’Connell, D.N. (1965). A concentric shock electrode for pain stimulation. Psychophysiology, I, 296-298. Willoughby, R.R. (1932). Some properties of the Thurstone Personality Schedule and a suggested revision. Journal of Social Psychology, 3, 401-424.