This article was downloaded by:[Alerta - Chile 2005/2006 Consortium] [Alerta - Chile 2005/2006 Consortium] On: 28 June 2007

Access Details: [subscription number 778577076] Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Psychotherapy Research
Publication details, including instructions for authors and subscription information: http://www.informaworld.com/smpp/title~content=t713663589

Facial microbehavior and the emotional quality of the therapeutic relationship

Jrg Merten a a Universitt des Saarlandes.

Online Publication Date: 01 July 2005 To cite this Article: Merten, Jrg , (2005) 'Facial microbehavior and the emotional quality of the therapeutic relationship', Psychotherapy Research, 15:3, 325 - 333 To link to this article: DOI: 10.1080/10503300500091272 URL: http://dx.doi.org/10.1080/10503300500091272

PLEASE SCROLL DOWN FOR ARTICLE Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf This article maybe used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material. Taylor and Francis 2007

Downloaded By: [Alerta - Chile 2005/2006 Consortium] At: 22:46 28 June 2007

Psychotherapy Research, July 2005; 15(3): 325 /333

BRIEF COMMUNICATIONS

Facial microbehavior and the emotional quality of the therapeutic relationship

RG MERTEN JO
Universita t des Saarlandes
(Received 26 April 2002; revised 31 August 2004; accepted 06 September 2004)

Abstract The emotional quality of the therapeutic relationship has the highest and most consistent effect sizes on therapeutic outcomes found in meta-analyses on psychotherapeutic processes. In the current study, the facial-emotional (micro-) behaviors of patients and therapists are analyzed to obtain a better understanding of how relationship patterns are established in the beginning of psychotherapies and how they constitute the emotional quality of the relationship. The results provide evidence for the importance of dyadic emotional microprocesses in the psychotherapeutic process. Several significant correlations were found between the individual and more expressed dyadic emotional behavior in the first session and the therapeutic outcome reported at the end of the treatment. In particular, dyadic emotional process patterns, detected with Theme (Magnusson, 1996), correlated negatively with all outcome perspectives.

What constitutes a good or bad psychotherapeutic relationship? This far-reaching question can only be answered satisfactorily if process and interpersonal aspects of the relationship regulation are taken into consideration (Strupp, 1998). See Orlinsky, Grawe, and Parks (1994) for an extensive meta-analysis regarding studies on process and outcome. Although the operationalizations of what is meant by process differ remarkably, one in particular stands out: the well-known relationship between the quality of the therapeutic relationship and the therapeutic outcome (Luborsky et al., 2002; Rosenzweig, 1936). Commonly, a good therapeutic relationship is considered a necessary precondition for successful treatment, but the interpersonal relationship could also be conceptualized as a specific factor of change. Consequently, management of the interpersonal relationship becomes a matter of therapeutic training, and the distinction between technical and interpersonal factors is superfluous (Krause, 1997; Strupp & Binder, 1984). Furthermore, the interpersonal and spontaneous processes between patient and therapist deserve more attention than as merely a necessary precondition for the effectiveness of technical interventions and should be considered

the central factor of change themselves (Gelso & Hayes, 1998; Weiss, Sampson, & Mount Zion Psychotherapy Research Group, 1986). Orlinsky et al. (1994) report the highest and most consistent effect sizes for mutual affect and reciprocal affirmation, which characterize the quality of the therapeutic relationship. However, the procedures used to operationalize and measure aspects of the therapeutic bond are significantly heterogeneous between the summarized studies. This demonstrates the effect force of the therapeutic relationship on therapeutic outcome. Nevertheless, the measures used are so general that it is impossible to make conclusions regarding the therapeutic practice. The categories mutual affect and reciprocal affirmation used by Orlinsky et al. suggest that the studies analyze therapeutic processes and focus especially on emotional processes between patient and therapist. If one examines the original studies, these expectations are not followed through. For example, Collins et al. (1985) focus on reciprocal affirmation. What they are studying is the extent to which the staff reports affective exchange between patients in the waiting room, which is assessed using one item of the Ward

Correspondence: Jo rg Merten, Universita t des Saarlandes, Im Stadtwald, 66041, Saarbru cken. E-mail: j.merten@mx.uni-saarland.de ISSN 1050-3307 print/ISSN 1468-4381 online # 2005 Society for Psychotherapy Research DOI: 10.1080/10503300500091272

Downloaded By: [Alerta - Chile 2005/2006 Consortium] At: 22:46 28 June 2007

326

Merten disorder, the frequency of Leitaffekt correlated negatively with the experience of joy (Merten, 1997a, 1997b). In one study the same phenomenon was found for patients suffering from pain disorder (Merten & Brunnhuber, 2004). They display less joy and more negative emotions than the control group, and the Leitaffekt contempt is positively correlated to the experience of fear and shame and negatively correlated to the experience of joy. In particular, the relationship of Leitaffekt to the experience of fear and shame indicates in general the reactivation of conflicts (fear) and specifically the reactivation of conflicts of self-esteem and autarchy (shame). A comprehensive result of all these studies was that all healthy partners adapted their behavior to the facial behavior of the patients (phenomenon of adaptation; Krause, 1997; Merten, 2002). In some cases, the healthy partners behavior was even more different from that of the control groups than that of the patients (Krause, 1998). Consequently, the modeling of the social implementation of relationship patterns has to be conceptualized as a dyadic model including explicit and implicit representations, wishes, and conflicts of both participants. In the process of implementation, the representations of relationship patterns are creatively adapted and assimilated to ongoing social interactions, and attempts to change these behavior patterns are hampered by their flexibility. As the phenomenon of adaptation demonstrates, it is difficult to resist the stimulative nature of the relationship offers of the patients. The risk to complement the behavior pattern of the patient is highest if the social partners own conflictive structure fits that of the patient. However, empathic understanding of a persons goals and norms can also be used to induce anger, contempt, or other emotions that are part of conflict regulation. Dreher, Mengele, Krause, and Ka mmerer (2001) emphasize that a gradual resonance with the patients behavior pattern is not necessarily a disadvantage but can help to understand the patients social problems. They also, however, demonstrated that at a certain point in treatment the relationship patterns have to be detected and dealt with (Ba nningerHuber, 1996; Krause, 1997; Merten, 1996a, 1996b, 2001). Method Ten experienced therapists of cognitive /behavioral, psychoanalytic, and client-centered theoretical orientation treated severely disturbed patients in a brief psychotherapy setting of 15 hr during which they were videotaped by two cameras. The therapists

Atmosphere Scale (Moos, 1974). In addition, the results on process and outcome variables differ in respect to the perspective from which the ratings are done (Horvath & Luborsky, 1993; Rudolf, 1991). Orlinsky et al.s (1994) meta-analysis supplies evidence for the importance of the social-emotional side of the therapeutic relationship, but it also points out that further studies are necessary to focus on interpersonal aspects of the emotional processes in therapeutic relationships. In particular, maladaptive relationship patterns of the patient and the therapist and their relationship to one other have to be taken into account to understand the interactive processes that constitute the therapeutic relationship. The necessity is obvious in the case of patients suffering from personality disorders. The establishment of a therapeutic relationship becomes more and more difficult if the patients relationship behavior is characterized by hostility, negativism, inflexibility, and antisocial tendencies in addition to abnormal dependence (Strupp, 1998). Moreover, maladaptive relationship patterns are essential parts of the cause of all mental disorders; they perpetuate them and are auspicious starting points for psychotherapeutic changes (Ba nninger-Huber & Widmer, 1997). A series of studies by the Saarbru cken work group (Krause, 1997, 1998; Merten, 2001, 2002; Merten & Krause, 2003) deal with the question of how maladaptive relationship patterns, which are specific for certain types of mental disorders, can be detected in the spontaneous social behavior of patients and their partners. The results can be summarized by four different phenomenon: reduction, anhedonism, adaptation, and congruency. The reduction of the facial expressions of the patients is characterized by a reduction of the expression of joy and an augmentation of one negative emotion: the Leitaffekt. Although the overall reduction of facial expression is an indicator of severe mental disorders like schizophrenia (Steimer-Krause, Krause, & Wagner, 1990), the reduction of expressions of joy is a global indicator of nearly all mental disorders. The prominence of one negative emotion can be interpreted as an attempt to solve an intra- or interpersonal conflict that has been reactivated in the current social situation. Merten (1996a) describes the function of the schizophrenics Leitaffekt of contempt as an attempt to solve a conflict between wishes for closeness and the fear of being overwhelmed by the emotions of the social partner. Further evidence that Leitaffekt is tied to the reactivation of conflicts is provided by the phenomenon of congruency between facial expression and emotional feeling. In several groups of schizophrenic patients and patients suffering from psychosomatic

Downloaded By: [Alerta - Chile 2005/2006 Consortium] At: 22:46 28 June 2007

Facial microbehavior 327 diagnosed the patients as suffering from severe mental disorders, nine of whom had been treated before without success. Overall, 145 sessions involving six psychodynamic treatments, 3 cognitive / behavioral treatments, and one client-centered treatment were videotaped; of these 119 were analyzed. Outcome was measured using ratings of success, goal attainment, helpfulness, and contentment with the treatment from both patient and therapist after the 15th session. The ten therapies were ranked according to these outcome measures. The resulting ranks were consistent with changes in the scores of the Freiburger Beschwerdenliste (Fahrenberg, 1975), a list of complaints completed by the patient at the start and end of treatment. Facial behavior of both protagonists was measured using EMFACS, a technique developed by Friesen and Ekman (1984) based on the Facial Action Coding System (FACS; Ekman & Friesen, 1978). Although FACS comprehensively measures all movements in the face, EMFACS measures only movements that are potentially relevant to affect. Using a dictionary (Friesen, 1988; Wagner, 1986), the measured facial events are interpreted as expressions of the basic emotions of happiness, anger, contempt, disgust, fear, sadness, and surprise or as social smiles. The dictionary also interprets blends and masks from the raw scoring of facial events. Blends are innervations of two of the just-mentioned basic emotions displayed at the same time. Masks are patterns in which a smile is used to cover a negative expression.1 The interrater reliability varies depending on the level of analysis: whether single action units or emotional categories are considered. In the case of the emotion categories used in the current study, the mean interrater reliability was .89 (Cohens k). EMFACS distinguishes between several types of smiling, especially synchronous Duchenne smiles, which are often prematurely interpreted as indicating a good therapeutic relationship, can be part of maladaptive relationship patterns, and can serve as interactive implementations of defense mechanism (Ba nninger-Huber & Widmer, 1999; DeRoten, Gillieron, Despland, & Stigler, 2002). In addition, the negative basis emotions of anger, contempt, disgust, fear, and sadness, and mixtures of them, can be detected and used as indicators of conflicts in the regulation of the self and in the relationship between patient and therapist (Dreher et al., 2001; Krause, 1998; Merten, Ullrich, Anstadt, Krause, & Buchheim, 1997.). To compare facial expressions in psychotherapies with everyday situations, three control groups composed of healthy individuals of differing gender were gathered: male /male, male /female, female / female. The healthy individuals were not acquainted with each other and met for the first time in the laboratory to discuss political problems for 20 min. We used Magnussons (1996) algorithm to detect temporal patterns in the facial expressions of patient and therapist. To identify temporal patterns that occur more often than expected by chance, the algorithm takes into account not only the sequence in which the emotions occur but also the temporal distances between them. Such patterns are called emotional process patterns. If they include facial expression of both protagonists, they are called dyadic; if they include only facial expression of one protagonist, they are called individual. A very common pattern is composed of two Duchenne smiles of different people. This kind of smiling was first described by Duchenne de Bologne (Darwin, 2000) and consists of raising the corners of the mouth and raising the infraorbital triangle to the left and right of the nose. Negative emotions were also often elements of the detected temporal patterns. Hypotheses Hypothesis 1: Based on the results labeled as phenomenon of reduction and adaptation, the frequencies of facial expressions are reduced in the case of the patient and the therapist in comparison to healthy individuals. The most distinct differences are expected for synchronous Duchenne smiles. Hypothesis 2: The most frequent basic emotion of the patient displayed */the Leitaffekt */in the first therapy session is related to therapeutic outcome. The more distinct it is, the worse the outcome will be. Hypothesis 3: Hypothesis 2 is also true in respect to the Leitaffekt of the therapist. Based on Dreher et al. (2001), it can be assumed that the function of the therapists negative emotions is ambiguous. Negative or positive correlations of their frequency with therapeutic outcome are expected. Hypothesis 4: The proportion of positive and negative emotions of the patient and the therapist is related to the therapeutic outcome. The dyadic Leitaffekt subdivided in three categories represents this proportion. Hypothesis 5: (a) The worst outcome is expected if patient and therapist have a positive Leitaffekt. (It is more of an indicator for an everyday conversation than a therapeutic conversation.) (b) The best outcome is expected if the Leitaffekt of one compen-

Downloaded By: [Alerta - Chile 2005/2006 Consortium] At: 22:46 28 June 2007

328

Merten Hypotheses 2 and 3: The leifaffekt of patient and therapist and the therapeutic outcome Neither the valence nor the frequency of the patients Leitaffekt correlates with the therapeutic outcome. The proportion of the therapists Leitaffekt on all basic emotions negatively correlates with the therapeutic outcome rated by the therapist (Table II). The affective valence of the Leitaffekt was not taken into account; it is only considered how distinct the therapist displayed one specific basic emotion in the first therapy session. The more one specific basic emotion was displayed, the worse the therapeutic outcome was. Hypothesis 4: Therapists negative emotions and the therapeutic outcome The sum of the negative emotions anger, contempt, and disgust of the therapist correlated positively with the therapeutic outcome (see Table II). Hypothesis 5: The dyadic leitaffekt and the therapeutic outcome The functions of the negative emotions of the therapist become more apparent if one considers the proportion to the patients emotions. The dyadic Leitaffekt correlates with the therapeutic outcome, as assumed in Hypothesis 5. In particular, if the therapist compensates the patients facial expressions of joy (Duchenne smiles) with negative emotions, the therapeutic outcome was better. The proportion of therapists negative emotions to patients positive emotions correlates positively with all three perspectives of the therapeutic outcome rating. When the therapist displayed a

sates for that of the other (e.g., patient negative, therapist positive). If both have a negative Leitaffekt, the prognosis lies in between a) and b). Hypothesis 6: The extent of a successful dyadic reactivation of the patients behavior pattern in the first psychotherapeutic session is measured as the frequency of dyadic emotional process patterns and is negatively correlated to therapeutic outcome. Results Hypothesis 1: Psychotherapeutic situation and everyday conversation If one compares facial-emotional behavior in everyday conversations and psychotherapeutic situations, one yields the following results: The overall facial emotionality of patients and therapist is reduced compared with that of healthy individuals in everyday interactions (Table I, a1 /7). Joy and disgust are displayed less by patients and therapists, whereas fear and surprise are more often shown by patients in the psychotherapeutic situation. The therapists also show more surprise than participants in everyday interactions. Patients and therapists both display less contempt than healthy individuals in everyday interactions, but patients show contempt twice as frequently as therapists. In comparison to everyday conversations, synchronous Duchenne smiles are significantly reduced in psychotherapeutic interactions (M /.079, SD /.71, N /116 sessions). Healthy individuals in everyday conversations show simultaneous Duchenne smiles three times more frequently than patients and therapists (M /2.32, SD /1.91, Z / /4.67, p B/.001).

Table I. Facial expressions per 10 min in psychotherapy and everyday conversation. Basic emotions Variable Joy Disgust Contempt Sadness Anger Surprise Fear a1 /7

Psychotherapya Patients M SD Therapists M SD Control participants (n /60) M SD 5.60*** 6.31* 4.85 11.51 6.07*** 5.55 1.44*** 1.46 2.39*** 3.88 0.49*** 1.16** 1.27*** 0.51*** 17.74*** 0.79 1.17 2.31 0.78 13.59 0.57*** 0.09c 0.88 0.32 10.39*** 6.47

1.00*** 0.13*** 1.09* 1.26 0.32 1.17 Everydayb 5.29 7.10 2.19 3.39 0.98 1.63

13.88 9.50

6.43 7.10

0.28 0.70

0.09 0.37

29.14 15.32

Note. The error probabilities given in the rows Patients and Therapists refer to the comparisons with the control group. Mann-Whitney U Test. a N /119 sessions. bN /30 conversations. cNot significant. *p B/.05. **p B/.01. ***p B/.001.

Downloaded By: [Alerta - Chile 2005/2006 Consortium] At: 22:46 28 June 2007

Facial microbehavior 329

Table II. Therapists negative emotions, dyadic leitaffekt, and therapeutic outcome. Outcome perspective Facial expressions Relative frequency of therapists Leitaffekt Negative emotions therapist (anger, contempt, disgust) Negative emotions therapist/joy patient Dyadic Leitaffekt Note. Spearman rank correlations, N /10, *p B/.10. **p B/.05. ***p B/.01. Therapist /.63** .81*** .64* Patient Therapist /patient

.55*

.76** /.69**

reciprocal Leitaffekt, the therapeutic outcome was worse. Hypothesis 6: Emotional process patterns and the therapeutic outcome The application of the Magnusson algorithm to the patients and therapists facial expressions reveals many dyadic emotional process patterns that correlate with the therapeutic outcome (Table III). The frequency of dyadic emotional process patterns in the first session correlates positively with all three outcome measures. The dyadic pattern of synchronous Duchenne smiles occurs in nearly all therapy sessions, but this pattern also indicates that positive reciprocity correlates negatively with the therapeutic outcome regardless of who starts them (Table III). If the number of synchronous Duchenne smiles exceeds a certain threshold, they seem to hamper therapeutic processes more than they further them. The synchronous Duchenne smiles started by the therapist had an inverse U-shaped relation to therapeutic outcome, which did not reach the 5% significance level. Discussion These results provide evidence for the importance of dyadic emotional microprocesses in the psychotherapeutic process and show differences between the psychotherapeutic situation and everyday conversation. Dyadic emotional process patterns observed in the first session are good predictors for the therapeutic outcome at the end of the psychotherapy,

independent of the perspective when the ratings were done. The hypothesis about the differences between the psychotherapeutic situation and the everyday situation were confirmed. However, the reduction of the patients facial expressions is due not only to the effect of the situation but to the general reduction phenomenon reported early in this article. Exceptions are the emotions of anger, fear, and surprise on the part of patients and surprise on the part of therapists. The augmented emotions fear and surprise can be interpreted as indicating the information processing of new and unexpected contents and as parts of change processes in mental representations like schema accommodation. The increase of fear in the patients facial expressions signals the reactivation of specific conflicts in the current situation. The patients facial expressions of anger are signs of blocked goals, which can be understood in different ways. The anger can indicate a missing compliance with the proposals of the therapist as to how to change the patients behavior. Alternatively, the augmentation of anger could also further change processes if it is related to persons who restrict the patients wishes for autonomy. The reduction of therapists contempt expresses their professional esteem for their patients, although the patients sometimes reacted with anger to the therapists proposals and disdained their assistance (e.g., asking for simple solutions to their problems). Also in some cases, the induction of contempt in the therapist is a part of the patients maladaptive relationship pattern. The reduction of the facial expression of sadness indicates a stable relationship between patient and

Table III. Dyadic emotional process patterns, synchronous duchenne smiles, and therapeutic outcome. Outcome perspective Variable Dyadic emotional process patterns Synchronous Duchenne smiles Patient starts Therapist starts Therapist /.58* /.63* /.55* Patient /.81** Therapist /patient /.75**

Note. Spearman rank correlations, N /10. *p B/.10. **p B/.05. ***p B/ .01.

Downloaded By: [Alerta - Chile 2005/2006 Consortium] At: 22:46 28 June 2007

330

Merten patterns. Their positive function is to obtain an emotionally deeper understanding of the patients problem and to signal to the patient that the therapist grasps the problems mentally and emotionally. As mentioned, the most consistent correlations are those between dyadic variables of facial expression and the therapeutic outcome, the dyadic Leitaffekt, and the dyadic emotional process patterns. It becomes obvious that the contribution of the therapist to the quality of the therapeutic relationship is much higher than it is assumed in studies based on ratings (Krupnick et al., 1996). At least in respect to the therapeutic outcome and the facial expressions, the opposite is true; the contributions of the therapists are essential for the prediction of the outcome. The (not significant) inverse U-shaped correlation between synchronous Duchenne smiles and therapeutic outcome is mentioned because it is consistent with the results of DeRoten et al. (2002). In the groups studied by DeRoten in which the relationship was assessed as being not as good, there were either many episodes of synchronous smiling or only a few. The synchronous Duchenne smiles not only serve the function of enhancing the quality of the therapeutic bond but they can also be indicators of defense processes (Ba nninger-Huber & Widmer, 1999) if they occur too frequently. The results presented should be interpreted under the restriction that the group was very heterogeneous in respect to patient diagnoses. In addition, it was found that the dyadic gender of the therapeutic dyadic had an effect on the quality of the facial expressions (Merten, 2001). In a further study in progress, only dyads composed of female patients suffering from panic disorder with and without agoraphobia and male therapists were recruited. The correlations between dyadic Leitaffekt and the therapeutic outcome could be replicated in a sample of 20 patients by Beutel, Leithold, and Rasting (2002). To develop a technique of the management of the relationship regulation described in this study, future research must scrutinize the way patients succeed in establishing their behavior patterns and how therapists react to them. The results presented give some hints, which are discussed in more detail and demonstrated in single cases in Merten (2001). The general importance of the results presented for clinical practice is the fact that scholars as well as experienced therapists should focus more on training of the perception of the spontaneous emotional behaviors of their patients and of the therapists themselves. The therapists also send spontaneous emotional messages that they contribute essentially to the quality of the therapeutic relationship and to

therapist in comparison to the everyday conversation in which people meet for the first time and have not yet established a stable relationship. As research has demonstrated, the facial expressions of sadness serve as relationship regulating signals if one of the partners evaluates the actual perceived closeness as being not sufficient (Merten, 1997a). In cases in which the facial expressions of sadness were augmented (mostly involving patients), they were shown in the context of conflicts in the relationship and in the context of anger and fear and served to compensate the perceived weakness of the bond to the therapist. So facial expressions of anger, fear, and sadness occurred together and served as regulating signals in the context of conflicts about separation (anger) and attachment (sadness, joy), where the facial expressions of fear signal the anticipated or real loss of attachment to the therapist (Morschitzky, 1998). A single case illustrating this function of anger, fear, and sadness was described in Merten et al. (1996). The analysis of the (micromomentary) facial expressions of emotions demonstrates that the behavior patterns of patients are established already in the first session, that they bring forth changes in the behavior of therapists, and that they bias psychotherapeutic processes relevant for the whole course of the treatment, because the degree to which they occur is related to the therapeutic outcome at the end. These patterns can only be registered with the aid of the analysis of the spontaneous and mostly unconscious facial microbehavior and are not amenable to ratings about the quality of the therapeutic relationship, because neither patient nor therapist were able, in the first session, to formulate the conflicts appearing in their social behavior. Before discussing the details of the correlations between facial expressions and therapeutic outcome, one can state that only the dyadic measures consistently correlated with all three outcome measures. Hypothesis 2 could not be confirmed; the patients Leitaffekt alone was not significant to the therapeutic outcome. However, the distinctness of the therapists Leifaffekt correlated with each therapists own outcome rating. So therapists had some implicit knowledge about the processes that resulted in worse outcome situations, which is reflected in the first session in an excess of one emotion. The ambiguity of the negative emotions assumed in Hypothesis 4 could be clarified. The negative emotions of the therapists were essential parts of therapeutic understanding. However, in conjunction with the results concerning Hypothesis 5 and 6, they should be compensating elements to excesses of joy on the side of the patients, and they must not be parts of persisting maladaptive emotional process

Downloaded By: [Alerta - Chile 2005/2006 Consortium] At: 22:46 28 June 2007

Facial microbehavior 331 the effectiveness of the therapeutic processes as, for example, the enactment of maladaptive relationship patterns.
Krause, R. (1997). Allgemeine psychoanalytische Krankheitslehre. Band 1: Grundlagen [General psychoanalytic pathology: Basics] . Stuttgart: Kohlhammer. Krause, R. (1998). Allgemeine psychoanalytische Krankheitslehre. Band 2: Modelle [General psychoanalytic pathology: Models] . Stuttgart: Kohlhammer. Krupnick, J. L., Sotsky, S. M., Simmens, S., Moyer, J., Elkin, I., Watkins, J., & Pilkonis, P. (1996). The role of the psychotherapeutic alliance in psychotherapy and pharmacotherapy outcome: Findings in the NIMH Treatment of Depression Collaborative Research Program. Journal of Clinical and Consulting Psychology , 64 , 532 /539. Luborsky, L., Rosenthal, R., Diguer, L., Andrusyna, T. P., Berman, J. S., Levitt, J. T., et al. (2002). The dodo bird verdict is alive and well /mostly. Clinical Psychology: Science and Practice , 9 , 1 /11. Magnusson, M. S. (1996). Hidden real-time patterns in intra- and inter-individual behavior: Description and detection. European Journal of Psychological Assessment , 12 , 112 /123. Merten, J. (1996a). Affekte und die Regulation nonverbalen, interaktiven Verhaltens: Strukturelle Aspekte mimisch-affektiven Verhaltens und die Integration von Affekten in Regulationsmodelle [Affects and the regulation of facial-affective behavior] . Bern: Peter Lang. Merten, J. (1996b). Regulation von Na he und Distanz in dyadischen Interaktionen mit schizophrenen Patienten [The regulation of distance in dyadic interactions between healthy and schizophrenic subjects]. Gruppenpsychotherapie und Gruppendynamik , 3 , 256 /273. Merten, J. (1997a). Facial-affective behavior, mutual gaze and emotional experience in dyadic interactions. Journal of Nonverbal Behavior , 21 , 179 /201. Merten, J. (1997b). Wem gelten negative mimische Affekte? [What is the function of negative facial signs?]. In R. Posner, T. Noll, & C. Schmauser (Eds), Ko rperbewegungen und ihre Bedeutung [Body movements and their meaning] (pp. 137 /154). Berlin: Berlin Verlag Arno Spitz GmbH. Merten, J. (2001). Beziehungsregulation in Psychotherapien [Relationship regulation in psychotherapies] Maladaptive Beziehungsmuster und der therapeutische Prozess [Maladaptive patterns and the therapeutic process] . Stuttgart: Kohlhammer. Merten, J. (2002). Context-analysis of facial-affective behavior in clinical populations. In M. Katsikitis (Ed.), The human face: Measurement and meaning (pp. 131 /147). New York: Springer. Merten, J., & Brunnhuber, S. (2004). Facial expression and experience of emotions in psychodynamic interviews with patients suffering from a pain disorder. Indicators of disorders in self- and relationship-regulation in the involuntary facial expression of emotions. Psychopathology , 37 , 266 /271. Merten, J., & Krause, R. (2003). What makes good therapists fail? In P. Philippot, E. J. Coats, & R. S. Feldman (Eds.), Nonverbal behavior in clinical settings (pp. 111 /124). Oxford: Oxford University Press. Merten, J., Ullrich, B., Anstadt, T., Krause, R., & Buchheim, P. (1996). Emotional experiencing and facial expression in the psychotherapeutic-process and its relation to treatment outcome. A pilot-study. Psychotherapy Research , 6 , 198 /212. Moos, R. H. (1974). Evaluating treatment environments: A social ecological approach . New York: Wiley. Morschitzky, H. (1998). Angststo rungen: Diagnostik, Erkla rungsmodelle, Therapie und Selbsthilfe bei krankhafter Angst [Anxiety disorder, diagnostic, models, therapy and selfhelp] . New York: Springer. Orlinsky, D. E., Grawe, K., & Parks, B. K. (1994). Process and outcome in psychotherapy /noch einmal. In A. E. Bergin (Ed.), Handbook of psychotherapy and behavior change (pp. 270 /376). New York: Wiley.

Acknowledgements This project was funded by the German Research Community (Deutsche Forschungsgemeinschaft).

Note
1

Further detailed descriptions of action unit combinations and interpretations can be found on the Web site of FACSAID (Facial Action Coding System Affect Interpretation Dictionary): http://face-and-emotion.com/dataface/facsaid/description.jsp).

References
bertragung /Interaktion . Ba nninger-Huber, E. (1996). Mimik /U Die Untersuchung affektiver Prozesse in der Psychotherapie [Facial expressions /transference /interaction. Affective processes in psychotherapy]. Bern: Huber. Ba nninger-Huber, E., & Widmer, C. (1997). Affektive Beziehungsmuster. Was kann die differenzierte Betrachtung der Beziehungsmuster zum Versta ndnis psychotherapeutischer Vera nderung beitragen? [Affective relationship-patterns]. Psychotherapeut , 42 , 356 /361. Ba nninger-Huber, E., & Widmer, C. (1999). Affective relationship patterns and psychotherapeutic change. Psychotherapy Research , 9 , 74 /87. Beutel, M. E., Leithold, A., & Rasting, M. (2002, October). Dyadic affective interaction patterns in the intake interview as a predictor of outcome . Paper presented at the Clinical Emotion Research: The State of the Art conference, Saarbru cken. Collins, J. F., Ellsworth, R. B., Casey, N. A., Hyer, L., Hickey, R. H., Schoonover, R. A., et al. (1985). Treatment characteristics of psychiatric programs that correlate with patient community adjustment. Journal of Clinical Psychology , 41 , 299 /308. Darwin, C. (2000). Der Ausdruck der Gemu tsbewegungen bei dem Menschen und den Tieren [The expression of the emotions in man and animals]. Frankfurt: Eichborn Verlag. DeRoten, Y., Gillieron, E., Despland, J.-N., & Stigler, M. (2002). Functions of mutual smiling and alliance building in early therapeutic interaction. Psychotherapy Research , 12 , 193 /212. Dreher, M., Mengele, U., Krause, R., & Ka mmerer, A. (2001). Affective indicators of the psychotherapeutic process: an empirical case study. Psychotherapy Research , 11 , 99 /117. Ekman, P., & Friesen, W. V. (1978). The facial action coding system (FACS) (A technique for the measurement of facial action) . Palo Alto, CA: Consulting Psychologists Press. Fahrenberg, J. (1995). Die Freiburger Beschwerdenliste FBL [Freiburger Symptom List]. Zeitschrift fu r Klinische Psychologie , 4 , 79 /100. Friesen, W., & Ekman, P. (1984). EMFACS-7: Emotional Facial Action Coding System, Version 7 . Unpublished manuscript. Gelso, C. J., & Hayes, J. A. (1998). The psychotherapy relationship (theory, research and practice). New York: Wiley. Horowitz, L. M. (1994). Congurational analysis and the use of role-relationship models to understand transference. Psychotherapy Research , 4 , 184 /196. Horvath, A. O., & Luborsky, L. (1993). The role of the therapeutic alliance in psychotherapy. Journal of Consulting and Clinical Psychology , 61 , 561 /573.

Downloaded By: [Alerta - Chile 2005/2006 Consortium] At: 22:46 28 June 2007

332

Merten
micro-processus e motionnels dyadiques dans le processus psychothe rapeutique. Plusieurs corre lations significatives ont e te trouve es entre lindividu et un comportement e motionnel dyadique plus exprime en premie ` re se ance et le re sultat the rapeutique rapporte a ` la fin du traitement. En particulier, les patterns dyadiques e motionnels du processus de tecte s avec THEME ont corre le ne gativement avec toutes les perspectives de re sultat. Resumen Microcomportamiento facial y cualidad emocional n terape utica de la relacio La cualidad emocional de la relacio n terape utica figura como el mayor y ma s consistente efecto en los resultados terape uticos encontrados en metaana lisis de procesos psicoterape uticos. En este estudio, los comportamientos facial-emocionales (micro-) de pacientes y terapeutas fueron analizados para obtener una mejor comprensio n de co mo se establecen las pautas relacionales al principio de las psicoterapias y co mo estas constituyen la cualidad emocional de la relacio n. Los resultados dan evidencia de la importancia de los microprocesos emocionales dia dicos en el proceso psicoterape utico. Se encontraron varias correlaciones significativas entre el comportamiento individual y el dia dico ma s emocionalmente expresado en la primera sesio n, y el resultado terape utico registrado al final del tratamiento. En particular, las pautas del proceso dia dico emocional detectadas con THEME correlacionaron negativamente con todas las perspectivas de resultados. Resumo Micro Comportamentos Faciais e a Qualidade uticas es Terape Emocional das Relac o A qualidade emocional da relac utica possui a mais o terape a elevada e mais consistente magnitude do efeito sobre os resultados terape uticos encontrados em meta-ana lises de processos psicoterape uticos. No presente estudo sa o analisados os (micro) comportamentos emocionas de pacientes e terapeutas para obter a melhor compreensa o de como padro cio das psico es relacionais sa o estabelecidos no in terapias e como constituem a qualidade emocional na relac ncia para a impor o. Os resultados fornecem evide a ta ncia dos micros processos emocionais dia dicos nos processos psicoterape uticos. Foram encontradas va rias correlac es significativas entre o comportamento emocioo nal individual e dia dico expresso na primeira sessa o e o resultado terape utico relatado no final do tratamento. Em particular, os padro ades, es do processo emocional das d detectados com o THEME, correlacionam-se negativamente com todas as perspectivas de resultados terape uticos.

Rosenzweig, S. (1936). Some implicit common factors in diverse methods of psychotherapy. American Journal of Orthopsychiatry , 6 , 412 /415. Rudolf, G. (1991). Die therapeutische Arbeitsbeziehung. Untersuchungen zum Zustandekommen, Verlauf und Ergebnis analytischer Psychotherapien [The therapeutic working relationship. Beginning, process and outcome of analytic psychotherapies] . Berlin: SpringerVerlag. Steimer-Krause, E., Krause, R., & Wagner, G (1990). Interaction regulations used by schizophrenic and psychosomatic patients: Studies on facial behavior in dyadic interactions. Psychiatry: Journal for the Study of Interpersonal Processes , 53 , 209 /228. Strupp, H. (1998). The Vanderbilt I Study revisited. Psychotherapy-Research , 8 , 17 /29. Strupp, H. H., & Binder, J. L. (1984). Psychotherapy in a new key: A guide to time-limited dynamic psychotherapy . New York: Basic Books. Wagner, G. (1986). Entwicklung eines automatisierten Verfahrens Zur Affektinterpretation. [Development of an automized interpretation of emotional FACS codings]. Unpublished thesis, Universita t des Saarlands, Saarbru chen. Weiss, J., Sampson, H., & Mount Zion Psychotherapy Research Group. (1986). The psychoanalytic process: Theory, clinical observations and empirical research . New York: Guilford Press.

Zusammenfassung t der Gesichtsausdruck und die emotionale Qualita therapeutischen Beziehung Die emotionale Qualita t der therapeutischen Beziehung weist bei Meta-Analysen zum Psychotherapieprozess die ho chsten und stabilsten Effektsta rken fu r den Therapieerfolg auf. In der vorliegenden Untersuchung wurden die mimisch-emotionalen (Mikro-) Verhaltensweisen von Patient und Therapeut untersucht, um zu einem besseren Versta ndnis zu gelangen, wie Beziehungsmuster zu Beginn der Therapie entstehen, und auf welche Weise sie die emotionale Qualita t der Beziehung bestimmen. Die Ergebnisse weisen auf die Wichtigkeit von dyadischen emotionalen Mikroprozessen im psychotherapeutischen Prozess hin. Einige signifikante Korrelationen wurden zwischen individuellen und dyadischen emotionalen Verhaltensweisen in der ersten Sitzung und dem berichteten Therapieerfolg am Ende der Behandlung gefunden. Insbesondere korrelierten die mit Hilfe von THEME entdeckten dyadischen emotionalen Prozessmuster negativ mit allen Ergebnisparametern.

sume Re Le micro-comportement facial et la qualite motionnelle de la relation the rapeutique e La qualite e motionnelle de la relation the rapeutique a les effets de grandeur les plus hauts et les plus consistants sur les re sultats the rapeutiques trouve s par des me ta-analyses du processus psychothe rapeutique. Dans cette e tude, les (micro-) comportements e motionnels faciaux chez patients et the rapeutes sont analyse s pour arriver a ` une meilleure compre hension de comment des patterns relationnels sont e tablis au de but dune psychothe rapie et comment ils constituent la qualite e motionnelle dune relation. Les re sultats rendent e vidents limportance des

Sommario ` Il Microcomportamento facciale e la qualita emozionale della relazione terapeutica La qualita ` emozionale della relazione terapeutica presenta i piu ` elevati e coerenti effect size sugli outcome terapeutici che si ritrovano nelle metanalisi dei processi psicotyerapu-

Downloaded By: [Alerta - Chile 2005/2006 Consortium] At: 22:46 28 June 2007

Facial microbehavior 333

tici. Nel suddetto studio vengono analizzati i (micro-) comportamenti facciali- emozionali dei pazienti e dei terapeuti al fine di capire meglio in che modo i modelli relazionali creino allinizio delle psicoterapie e in che modo vadano a costituire la qualita ` emozionale della relazione. I risultati forniscono una prova dellimportanza dei microprocessi emozionali duali nel processo psicoterapeutico. Sono state trovate diverse correlazioni significative fra il comportamento emozionale duale individuale e maggiormente espresso nella prima sessione ed il risultato terapeutico riportato al termine della terapia. In particolare, i pattern dei processi emozionali diadici ritrovati con THEME sono correlati negativamente con tutte le prospettive delloutcome.