Different types of laughter and their function for emotion regulation in dyadic interactions

For our research, we are mainly interested in investigating the psychotherapeutic process. We are particularly interested in which role certain verbal and nonverbal interactive behaviors play in facilitating productive therapeutic work. Specifically, we explore the relation between processes of affective regulation on a micro-level of verbal and nonverbal behavior and psychotherapeutic outcome. To achieve this goal, it is important to have enough knowledge about the different nonverbal phenomena which are relevant for affective regulation in interpersonal interactions. Therefore, we examine the phenomenology and function of different nonverbal phenomena such as smiling, laughter or crying not only in psychotherapy but also in everyday interactions (e.g., conversations of couples, mother-daughter dyads or male and female friends). In the field of social psychology, there already exist a number of studies dealing with laughter. These studies, however, have examined the phenomenology and functions of laughter mostly within an individual and neglect the one aspect that is particularly important for understanding laughter in psychotherapeutic interaction: the relationship. In part I, we will give a short introduction of our research approach. We will briefly describe our model of affective regulation, which differentiates between intrapsychic and interactive processes of emotion regulation. The role of facial behavior is emphasized. Smiling and laughing are described from a phenomenological and functional perspective. Afterwards, we will introduce our micro-analytical approach of behavioral observation as well as the Facial Action Coding System (FACS) as a way of objectively capturing facial behavior. So-called Prototypical Affective Microsequences (PAMS) including smiling and laughter are presented as an example of the interaction between intrapsychic and interactive forms of affect regulation. In part II, we summarize two of our empirical studies on the subject of laughter. The first of these studies deals with the phenomenology and function of laughter episodes in everyday interactions. We analyzed laughter episodes occurring in conversations between mothers and fathers and their adolescent daughters and sons about guilt feelings in their relationship. This study has already been published in extracts (Bänninger-Huber & Gruber, 2010). The second empirical study focuses on laughter in psychotherapeutic interaction. The data come from an extensive research project in which we examined videotaped psychoanalytic longterm psychotherapies (e.g., Bänninger-Huber & Monsberger, 2016). Finally, we will discuss the relevance of our results for the understanding of laughter as an important phenomenon of affective regulation in interpersonal relationships.


Intrapsychic and interactive process of emotion regulation are linked
According to newer conceptualizations in emotion psychology, emotions are no longer understood exclusively as subjective processes taking place in an individual (e.g., Frijda, 1986).Emotions mostly occur in interpersonal relationships and are elicited and regulated between two or more persons.We therefore differentiate between intrapsychic processes and interactive forms of emotion regulation and focus on the specific ways in which they are inter-connected (Bänninger-Huber, 2006; The project "Emotional processes of emotion regulation in psychoanalytic psychotherapies" was approved by the Board for Ethical Questions in Science of the University of Innsbruck, October 10, 2014. Bänninger -Huber & Widmer, 1996, 1999;Moser & von Zeppelin, 1991).In dyadic interactions, we distinguish between the domain of self-regulation and the domain of relationshipregulation.Both domains are linked in the sense that success in self-regulation is a prerequisite for high emotional involvement in a relationship with another person (Bänninger- Huber & Huber, 2020).Furthermore, a successful interaction can have a positive influence of a person's self-regulation.For instance, my guilt feelings may be reduced if my partner forgives me for having offended him and will thus lead to a more relaxed interaction between the two of us.This allows us to reestablish and intensify our affective relatedness.
Despite the importance of interpersonal relationships for the elicitation and regulation of emotions, most empirical studies in emotion research focus on intra-individual processes without including the aspect of relationship regulation; for example, emphasis is placed solely on the relation between a person's subjective experience of an emotion, his or her expression and, more recently, his or her physiological and neurological components.In our research, we adhere to the view that emotional processes can only be understood when they are conceptualized as interpersonal phenomena and are studied accordingly.

The importance of facial expressions for affective regulation
The examination of facial behavior is interesting for psychotherapy research, because facial expressions reflect the individual's intrapsychic processes, while simultaneously conveying communicative meaning to the partner.Thus, facial behavior may be viewed as an "interface" between intrapsychic and interactive processes of affective regulation fulfilling several functions simultaneously.Smiling, for instance, can be the expression of the inner state "happiness".At the same time, smiling is an important signal in terms of relationship regulation and may convey to the interacting partner: "I like you, please keep talking!"Though facial behavior mostly occurs unconsciously and automatically, it may influence the interactive partner's affective state.

Phenomenology of laughter
A bout of laughter involves different phenomena such as respiration, vocalization, facial behavior and body movements.Laughing starts with an intensified exhalation.The laughter vocalization consists of laugh cycles, i.e., repetitive laugh-pulses modulated in an individual specific way.These vocalizations are mostly accompanied by innervations of the zygomaticus major muscle, which pulls up the lip corners.During laughter, the lips are usually parted and space can be seen between the teeth.Specific body movements such as shaking the head, trunk movements or hitting the knees with the hands are also observable, especially in conjunction with intense forms of laughter (Hochgruber, 2001; for a more comprehensive phenomenological description see Ruch & Ekman, 2001;Vettin & Todt, 2004).

The differentiation between smiling and laughing is important
Smiling and laughing are insofar related as laughing as well as smiling are accompanied by a movement of the zygomaticus-major-muscle or "pulling up the lips".In the case of laughter, this muscle innervation is accompanied by a specific vocalization.Not surprisingly, laughing and smiling often get mixed up in everyday life.Laughing is frequently viewed as the same facial phenomenon as smiling only with a more intense expression (e.g., Darwin, 1872).There is meanwhile enough empirical evidence for the differentiation between smiling and laughing in terms of function and effects.If we are being greeted by somebody with loud laughter instead of a friendly smile, for example, we perceive this behavior as inappropriate and we react confused.

Functions of smiling and laughing
Smiling is highly contagious and may be used to establish affective resonance.This means that, during simultaneous or mutual smiling, both partners may share a positive affective state (Bänninger-Huber et al., 2004).This experience may again enhance the affective relatedness between the interacting partners.Smiling may also have a behavior controlling effect.Accordingly, we can produce a deliberate smile as an excuse for our own inappropriate behavior and, in this way, appease our partner.
Contrary to popular belief, laughing is not only an expression of exhilaration but has other functions in the domains of self-and relationship regulation.In the case of an individual, laughing often serves as a means of regulating negative emotions.Thus, laughing can reduce a person's anxiety, stress, illness or pain (Freud, 1905;Gregory, 1924;Zweyer et al., 2004).Furthermore, laughing is considered a "safety valve" that allows persons to reduce states of tension in a socially acceptable form, e.g., in highly intimate situations (proximity, eye contact) (Chapman & Foot, 1976;Frijda, 1986).But laughing is also closely linked to aggression.Human etiologists regard laughing as a ritualized threat signal developed from teeth display in animals (Eibl-Eibesfeldt, 1978, 1984;Lorenz, 1983;Van Hoof, 1972).Thus, other persons' laughter can be seen as a cue of social threat and rejection (Platt et al., 2012;Ruch & Proyer, 2008;Papousek et al., 2014).Lorenz (1983) stresses that laughter may replace an attack because it enhances the connection between the laughing persons and excludes the person not laughing.

Smiling and laughter as important elements of affective regulation
Smiling is especially useful to regulate negative emotions, because it communicates basic relatedness in spite of the expressed negative emotions.We therefore could call a smile a "thread of resonance".When a relationship is threatened by a verbal confrontation, smiling may help to reaffirm the emotional bond between the interacting partners (Bänninger-Huber, 2020).
Laughing plays an important role in self-and relationship regulation.Laughing can interrupt one's own negative affective states as well as the partner's (Bänninger-Huber, 1992;Bänninger-Huber et al., 1990).The brief interruption of the dialogue permits internal affective reorganization.Laughter is not only important for self-regulation but also for maintaining the relationship between two persons.Laughing is contagious and may elicit positive affective states in the interacting partners.According to our model, such experiences of affective resonance strengthen the emotional bond between two individuals and increase feelings of wellbeing in a relationship.Laughter results in a new affective state in the partners (Bänninger-Huber, 1992).

Micro-analytic approach
In order to describe the phenomena listed above, we developed an interdisciplinary approach combining knowledge and methods from traditionally separate fields such as clinical psychology, emotion psychology, psychotherapy research, interaction research, developmental psychology and psychoanalysis.Methodologically, we pursued an approach of systematic behavior observation.Dyadic interactions were video recorded and facial interactive behavior was analyzed using the Facial Action Coding System (FACS) by Ekman et al., 2002).FACS is anatomically based and provides a "map" of a person's facial muscular behavior.This allows for the detailed and objective description of all visually distinguishable facial movements.Observational elements are the action units (AUs).The AUs are given numbers.AU 10, for example, describes the pulling up of the upper lip.AU 12, however, pulls the lip corners up and is commonly referred to as a smile.FACS distinguishes 44 action units (AUs).The intensity of an expression may be scored on a 5-point intensity scale.For the basic emotions (anger, contempt, disgust, fear, happiness, sadness, and surprise), there exist so-called tables of emotion prediction that relate specific AU-combinations to these emotions.Using FACS, 12 different types of smiles have also been identified with each differing in its appearance and function (e.g., Bänninger-Huber & Rauber-Kaiser, 1989;Ekman, 1985;Ekman & Friesen, 1982).A genuine smile, for example, is a spontaneous and involuntary expression of positive emotions and is defined by the innervation of zygomaticus major in combination with an orbicularis oculi innervation which raises the cheeks and tightens the eye lids (AU-combination 6 + 7 + 12).So-called "masking smiles" function to hide, control, or dampen negative emotions.In this case, indicators of negative emotions become visible (e.g., an innervation of the levator labii superioris muscles which wrinkle the nose or raise the upper lips (AUs 9 + 10) in addition to the smiling (AU 12).The use of a differentiated system for recording facial expressions is important, because facial behavior usually occurs very rapidly.The resolution of FACS is very high.Precise coding is possible up to 1/50th of a second.

Analyzing processes of affective regulation in dyadic interactions
As mentioned above, we not only analyzed psychotherapeutic interactions but also everyday interactions such as couples or mother-daughter and father-son dyads.There are two reasons for this procedure.The comparison between everyday interactions and therapies should help to identify the specific characteristics of productive psychotherapeutic interventions.Furthermore, we wanted to learn more about the postulated functions of affective regulation processes in social interactions, especially in view of the lack of studies on this topic.For this purpose, we developed a standardized observation setting.Subjects were given instructions to talk about situations in which they experienced a specific negative emotion (such as anger, guilt feelings or jealousy) in their relationship/towards their partner.The interactions were videotaped.In order to record aspects of subjective experiences different questionnaires were used (e.g., the Adult Attachment Projective (AAP), George et al., 1999).Instructions were based on the following theoretical assumption.The remembering and reporting of episodes that originally elicited negative emotions are reactivated in the here and now of the interaction (Bänninger- Huber & Widmer, 1999).This reactivation leads to a perturbation in the affective self-and relationship regulation of an individual which may be regulated interactively.This process of affective regulation can be observed, among others, in the form of so-called Prototypical Affective Microsequences (PAMs).

Key patterns of affective regulation: Prototypical Affective Microsequences (PAMs)
PAMs fulfil the function to balance out perturbations in affective regulation involving the interacting partner.PAMs are short sequences of affective regulation that are usually beyond the conscious awareness of the participants.PAMs are expressed mainly nonverbally and are characterized by smiling and laughing.As described above, smiling and laughing are contagious phenomena, which may create positive affective resonance between the persons involved.This experience strengthens the emotional bond between the two individuals, allowing them to re-establish a secure relationship.In psychotherapy, PAMs also function to enhance the affective relationship between interacting partners, which is important in establishing and maintaining a secure working alliance.

Operationalization of PAMs
The prototypical process of a PAM is as follows: On the verbal level we may observe indicators of a perturbation in patients' affective regulation such as repeating and correcting words, using filling sounds, or pausing in the middle of the sentence.On the nonverbal level, we may observe facial indicators of negative emotions (e.g., nose wrinkler (AU 9), upper lip raiser (AU 10) or lip corner depressor (AU 15) or adaptors such as lip pressing (AUs 23,24) or touching the face.The initiator of the PAM makes a joke or an ironical remark which increases the affective relatedness between the partners.He or she gazes towards the interacting partner in order to check the impact of his or her statement, a process that we could conceptualize as "social referencing" (Campos & Stenberg, 1981;Klinnert et al., 1983).He or she then begins to smile, which we interpret as a kind of relationship offer.

Different types of PAMs
Depending on the reaction of the interacting partner, socalled successful, unsuccessful, plus-minus and participation PAMs can be distinguished (e.g., Bänninger-Huber, 1996;Bänninger-Huber et al., 1990, 2016;Huber et al., 2016).In successful PAMs, the interacting partner reciprocates the initiator´s smile or laughter.Both partners are in a positive affective state.The negative affect is regulated and the relationship is secure.The subsequent course of interaction may be experienced positively due to the joint experience of mutual responsivity (Bänninger-Huber & Steiner, 1992).
In unsuccessful PAMs, however, the interacting partner does not reciprocate the smiling or laughing.Dealing with the perturbation is delegated to the respective self-regulation domains.The individual preoccupation with self-regulation results in a reduction of emotional involvement.Thus, the negative affect remains and the relationship is experienced as insecure.(The terms "successful " and "unsuccessful" do not evaluate the productivity of a regulation process in a certain context but refer to whether the initiator of a PAM succeeds in eliciting a smile or laughter in his or her interacting partner or not.) In participation PAMs, the interacting partner reciprocates the smile but only in a weak form.Because the smiling is only weak, the negative affect is not fully regulated and the initiator of the PAM still has to deal with his or her perturbation in affective self-regulation.The initiator's smiling has the function of a "thread of resonance" giving the partner a sense of security.
In plus-minus PAMs, the interacting partner reciprocates the partner's smiling but shows at the same time facial indicators of negative emotions as well.Here too, the smiling response has the function to maintain the partners' affective relatedness.The partner, however, simultaneously experiences negative emotions.In these situations, the negative affect is not fully regulated, but the relationship is nevertheless secure.

PAMs are frequent phenomena in interactions
During the last twenty years, we conducted many studies in which we could validate the functions of PAMs (e.g., Bänninger-Huber, 2016, 2021).In everyday interactions, we observed PAMs in the context of several negative emotions (anger, shame, jealousy, guilt feelings) and in different types of dyads (mothers with their two-year-old children, couples, female and male friends, mothers with their adolescent daughters, fathers with their adolescent sons, mothers and their anorexic daughters).Overall, successful PAMs occurred most frequently and thus contributed to the experience of wellbeing in a relationship.In couples, we found a relation between the frequency of successful PAMs and relationship satisfaction.
The postulated functions also apply to psychotherapeutic interactions (Bänninger-Huber, 1992;Bänninger-Huber & Widmer, 1999, 2020;Bänninger-Huber et al., 2016;Huber et al., 2016, Bänninger-Huber, 2021).Successful PAMs provide a reliable working alliance in giving the patient a basic sense of security.This enables the patient to explore his or her problematic experiences and to understand and accept the therapist's interventions.Unsuccessful PAMs, on the other hand, are essential for maintaining a certain level of conflictive tension as a prerequisite for working on the patient's conflicts (Bänninger-Huber & Widmer, 1999, 2000, 2020).

Part II: Identifying different types of laughter: Two empirical studies
As laughter is a core element of PAMs, we wanted to learn more about the phenomenology and function of this relevant phenomenon.For this purpose, we conducted two studies, in which we tried to differentiate between various types of laughter.The first study analyzed father-son and mother-daughter dyads talking about guilt feelings towards each other (Bänninger-Huber & Gruber, 2010); the second examined psychotherapeutic interactions (Salvenauer, 2018;Salvenauer & Bänninger-Huber, 2018).
Quite a number of attempts to differentiate various types of laughter already exist.Ruch and Ekman (2001), for example, differentiated between spontaneous and voluntary laughter.Spontaneous laughter is an impulsive, uncontrolled laughing reaction.The voluntary type, on the other hand, is an artificial expression and the laughing persons do not experience any pleasure.Volitional and spontaneous expressions are perceived differently (e.g., Bryant et al., 2018).Other authors tried to describe different types of laughter by relating them to specific emotional states (Drack et al., 2009).A differentiation between vocalized and un-vocalized laughter also exists (Bachorowski & Owren, 2001).Gupta et al. (2018) made a distinction based on the attributed affective state of patients in a therapeutic context and differentiated, for example, between the categories "cheerfulness, "politeness" or "reflectiveness" laughter.The description of these categories also includes nonverbal behaviors such as gazing, hand movements, tone of voice or body positions.An objective and systematic observation based on FACS-Codes, however, is missing.

Laughter in guilt feeling situations
For this study, ten mother-daughter and ten father-son dyads were analyzed (Bänninger-Huber & Gruber, 2010).The participants were asked to discuss situations in which they had experienced guilt feelings towards each other.The mean age of the daughters and sons was about 20 years.The discussions lasted about 14 min on average.For this investigation, we were interested in examining the following basic questions: How frequent are laughter episodes?Who initiates the laughter?What reactions of the interacting partner to the laughter can be observed?What function does the laughter in terms of self-and relationship regulation serve?

Method
We analyzed the first ten laughter episodes occurring after the instruction (to talk about guilt feelings in the relationship) in six dyads.Laughter was defined as innervations of the zygomaticus major muscle (AU 12) accompanied by a specific laughter vocalization.Onset, apex and offset were coded for each laughter episode.Additionally, the initiator of the laughter as well as his or her gazing behavior (towards or away from the partner) was noted.The partners' responses to the laughter were captured by the categories "laughter", "no laughter", and "smiling".Gazing behavior (towards and away from the partner) was also coded.The interactive context was captured by noticing the verbal utterance preceding the laughter.In order to detect possible functions of laughing, we developed a category system based on earlier empirical work (e.g., Bänninger-Huber, 1996;Bänninger-Huber et al., 2004).The following categories were distinguished: Exhilaration, reduction of tension, interruption, enhancing affective relatedness, dampening laughter and laughing at a person.The category "exhilaration" occurs in the context of an amusing behavior or a humorous remark and is an expression of a positive emotional state."Reduction of tension" includes observable laughter in situations in which the person is tense and nervous.This type of laughter helps to reduce the nervousness elicited by an emotionally difficult situation.Laughing as "interruption" allows a reorganization of self-regulation by interrupting the dialog for a short moment.Laughter from the type "enhancing affective relatedness" occurs in situations in which relationship regulation is disturbed.This laughter is a kind of relationship offer enhancing the affective relatedness between the persons."Dampening laughter" modifies the verbalization, especially when it has an aggressive component and has the function to avoid harming the other."Laughing at a person" is an aggressive comment on the verbalization or the behavior of the interacting partner.As nonverbal expressions always have different meanings at the same time (e.g., Bänninger-Huber, 1996), the categories tried to capture the most dominant function with respect to either self-or relationship regulation.

Results
Frequency and initiators of laughter episodes 60 laughter episodes occurred, which lasted 2 to 3 s on average.In 39 episodes, the initiator was gazing towards the interacting partner, whereas only 21 laughter episodes were initiated without gazing towards the partner.The frequent gazing towards the interacting partner supports our hypothesis that laughing often serves relationship regulation.
The time needed to capture ten laughter episodes varied from dyad to dyad.We may explain these differences in the frequency of laughter by the existence of different levels of conflictive tension during the conversations.The differences in initiating the laughter episodes were attributed to the individuals' "regulation pressure", which has to do with the quality of guilt feelings discussed.In dyad 1, for example, the conflictive tension was high.The daughter was rather confrontational and started the discussion with the verbalization "you have to have a lot of bad conscience, mum".During the whole conversation, the mother stayed in a defensive position.Several nonverbal indicators (such as the frequent expression of negative affects or adaptors) suggested that her self-regulation was highly disturbed and guilt feelings were reactivated.In the course of further interaction, the mother tried to get the daughter's absolution from her guilt feelings in the sense of "you are a good mother; you need not feel guilty in any way, because there is nothing you have done wrong".The daughter, however, did not react in the desired way and did not relieve her mother from her guilt feelings.This dynamic was reflected in the observation that the mother initiated the laughter 8 times, whereas the daughter only responded to it once.

Simultaneous laughter
Simultaneous laughter occurred in all dyads and were far more frequent in the mother-daughter-dyads than in the father-son dyads (12 vs. 6).In father-son dyads, however, laughter was often responded to with smiling, a weaker form of emotional reaction than laughing (14 smiling reactions vs. 6 laughter reactions).In mother daughter-dyads, we found only 3 smiling reactions.These findings fit with observations in other studies that have shown that in male dyads conflictive tension was generally kept on a lower level.

Not reciprocated laughter
The phenomenon of not reciprocated laughter could also be observed in all dyads.In mother-daughter the daughters did not respond to their mothers' laughter in 10 instances, but in only five cases did mothers not reciprocate their daughters' laughter.In the male dyads, not reciprocated laughter occurred less frequently than in the female dyads (10 vs. 15).It is remarkable that fathers tend not to reciprocate their sons' laughter (10 vs. 0).These findings may again reflect the observation that daughters are more confrontational than sons, thus often refusing their mothers' relationship offers.The sons, however, try to avoid conflictive tension by relieving their fathers from their guilt feelings very rapidly.The fathers' reduced responsiveness may be due to their preoccupation with the self-regulation of negative emotions.

Frequency of the functional categories
The most frequent categories found in our study were "reduction of tension" (21), followed by "enhancing affective relationship" (17) and interruption (9).These three categories covered almost 80% of all laughter episodes."Dampening laughter" was observed 5 times and "laughing at someone" and the "exhilaration-type of laughter" occurred only twice.Thus, the types of laughter observed in this study mainly served the function of regulating negative emotions (47).Laughter as an expression of exhilaration is rather rare (2).

Conclusion from the guilt study
From these results, we may conclude that laughing in social interactions is not primarily an expression of positive emotions but plays an important role in the intrapsychic and interactive regulation of negative affects.The 'guilt study' has given evidence that we may distinguish between different types of laughter according to their function in everyday interactions.

Analyzing types of laughter in psychoanalytic psychotherapies
In this study, we wanted to learn more about the function and phenomenology of laughter in psychoanalytic psychotherapies.We started from the functional categorization of laughter developed in the guilt study described above.We assumed that the categories "reduction of tension" and "enhancing affective relatedness" would be particularly relevant in psychotherapy, because, in these clinical contexts, difficult topics are usually discussed and are accompanied by negative emotions that need to be regulated.We assumed that laughter plays a special role in this process because it is contagious and enhances the affective relatedness between two individuals.This experience of affective closeness leads to a feeling of security in patients.This in turn enables them to continue working on their conflicts.Based on these two main categories, we tried to explore other functional types of laughter that differ in terms of their appearances, e.g., facial expression, vocalization, and gaze.

Data
The data were taken from the research project "Processes of emotion regulation in psychoanalytic psychotherapies", which examined psychoanalytic psychotherapies in a naturalistic setting (Bänninger-Huber & Monsberger, 2016).The therapies took place at our psychotherapy unit and each session was videotaped.The project comprises a total of seven completed treatments ranging between 64 and 330 sessions.After each session, patient and therapist completed the Helping Alliance Questionnaire (HAQ) (Bassler et al., 1995), which assesses relationship and outcome satisfaction of patient and therapist.Additionally, comprehensive clinical diagnostics at the beginning and at the end of the treatments were conducted (Bänninger-Huber, 2016).

The therapeutic dyad analyzed
The patient under scrutiny was a male student in his midtwenties.During therapy, he complained about relationship problems and learning difficulties.He also suffered from other depressive symptoms such as lack of motivation or feelings of worthlessness.The therapist was an experienced psychoanalyst in his fifties.The psychotherapy lasted a total of 330 sessions.From this data material, we selected seven therapy sessions coming from the early, middle, and late phases of therapy.The seven sessions lasted 53 min on average.

Operationalization of laughter and categories
Laughter was operationalized as an innervation of the Zygomatic major muscle (AU 12) together with an open mouth (AUs 25, 26 or 27) and a forceful exhale.As a first step, we identified all laughter episodes, including who initiated the laughter.We watched the video sequences repeatedly, identified typical verbal and nonverbal behaviors, and formulated a preliminary categorization, which was revised several times.As a final step, we formulated the following criteria (see also Table 1): Verbal context, gaze direction, indicators of tension regulation (adaptors), intensity of expression and potential probability of affect induction.The verbal context captures the verbal content shortly before a laughter episode.A distinction is made, for example, between therapist-related and therapyrelated content.The category "gaze behavior" captures the direction in which the initiator looks during the laughter episode.The categories include "looking towards the therapist", "looking away from the therapist" and "social referencing".The following indicators were used to evaluate the internal state of tension, namely hand touches on the face and head, blushing, lip presses (AUs 23 and 24), and lip licking (AU37), incoherent speech, restless sitting behavior and frequent changes of gaze.Intensity of expression was evaluated on a scale of 1-3.Facial intensity was determined by the FACS intensity coding of AU 12 (a-e).
Vocal intensity was subjectively assessed on the basis of the duration of the laughter as well as the strength of the exhalation.The category "affect induction" was intended to determine whether the patient created an atmosphere in which the therapist was seduced to laugh.This was mainly measured by the therapist's reaction categories "laughing", "smiling" or "no reaction" (Salvenauer, 2018;Salvenauer & Bänninger-Huber, 2018).

Frequency of laughter episodes
During the seven sessions, we could identify a total of 167 episodes of patient's laughter.The patient initiated 160 episodes and the therapist only 7.This corresponds to the study by Marci et al. (2004), who have also found that the majority of laughing episodes in psychotherapy are initiated by the patient.In 128 laughter episodes initiated by the patient, the therapist showed no reaction.By contrast, a smile response occurred in 24 episodes and laughter occurred in only 5 episodes.These results support our findings when analyzing PAM-types.Therapists mostly do not reciprocate the patients' smiling and laughing but stay abstinent.With this, they are able to keep up conflictive tension which allows the patients to further explore their conflicts (see balance hypothesis above).This is in contrast to our analyses of everyday interactions.In our guilt feeling study, for example, interacting partners did

Identifying types of laughter
In our data, we could distinguish five types of laughter (see Table 1): "Self-regulative laughter" is characterized by an increase of adaptors and missing humorous elements.Moreover, the patient turned his gaze away.With this, he created a personal space, which helped him to regulate his negative emotions.This type occurred most frequently after the discussion of conflictive topics concerning either the patient himself or the therapeutic relationship.In these cases, the process of emotion regulation predominantly took place on an individual level without involving the therapist.Example: The patient is wondering about his motives for choosing his field of study.He denies that a desire for recognition by others could be a motive.In the next sentence, however, he admits that this could be the case after all and that he may be deluding himself and then he laughs.The type "affective distancing" occurred after a silence when the intimacy between the patient and the therapist reached an aversive level from the patient's perspective.Here also, the signs of tension increased and the patient gazed away in order to reduce affective relatedness and then forcefully exhaled.Example: The therapist emphasizes the importance of the patient's cooperation and the patient agrees.Afterwards, there is a 30-s period of silence, during which the patient seems increasingly agitated, especially after moments of eye contact.He begins to laugh very forcefully and turns away from the therapist.
In "relatedness laughter", the therapeutic relationship was enhanced.The intensity of vocalization and facial expression was higher and eye contact lasted longer.Adaptors were quite rare.They mostly occurred when the therapist's reaction did not correspond with the patient's expectations.The verbal context involved the therapy or the therapeutic relationship, e. g, countertransference feelings.The patient tried with his behavior to please the therapist.Example: After a 20-s silence, which seems to be very unpleasant for the patient, he resumes the conversation by saying to the therapist: "Now we have collected a few important things" and laughs.With this remark he addresses a common process, which intensifies the affective relatedness.
In the context of the type "social orienting laughter", the patient typically talked about difficult topics, which he assumed that the therapist would judge negatively.He gazed towards the therapist in order to check the impact of his statement on the therapist.By looking at the therapist, the probability for a reciprocated laughter induction increased.The mutual laughter enhanced the affective relatedness and helped the patient to overcome his uncertainty.In contrast to "self-regulative laughter", in which affective regulation took place within the patient, here the therapist got involved.Example: The therapist confronts the performance-oriented patient with the fact that he intends to complete his studies in the summer semester.The patient jokingly asks "in which?" and laughs.After no reaction from the therapist, signs of tension appear.
"Exhilaration laughter" is characterized by its spontaneous appearance and often occurred due to surprising or funny remarks.Eye behavior was usually characterized by eye contact.Example: The therapist topicalizes the cold weather at the beginning of the lesson, whereupon the patient points to his winter jacket and says: "That's why I wear a "Wolfskin"" and laughs.
The most frequent category in our study was "self-regulative laughter" (73), followed by "exhilaration" (39), "social orienting" (31), and "relatedness" laughter (15).Two episodes could not be categorized.The distribution of the categories seems to be strongly related to the therapeutic situation.Both, "self-regulative" and "socially oriented" laughter, occur in the context of discussing conflictual topics.This is, of course, often the case in psychotherapies.The frequency of exhilaration laughter, however, is remarkable.Humor seems to have an important function in psychotherapy.This assumption is supported by the fact that the few laughter episodes (n = 7) initiated by the therapist could be assigned to the "exhilaration" type.Exhilaration laughter not only helps to regulate negative emotions.Since it is usually contagious, the affective bond between the laughers is increased.This in turn gives the patient the necessary security to deal with difficult issues.
From a phenomenological perspective, it is also interesting, that in all laughter episodes, the orbicularis oculi muscle (AU 6) was innervated in addition to AU 12. Thus, in the patient studied, laughter is always accompanied by a felt smile.In a first, not systematic analysis of the facial behavior (apart from the AUS 6, 25, 26 and 27 and the signs of tension) we did not find significant differences in terms of FACS-Coding between the types of laughter.It seems that the variation of facial expressions is very limited in the moment of laughter.Variations are more likely to occur before and after laughter.Another interesting observation concerns the duration of laughter in the therapeutic dyad examined: In 90% of the cases, a laughter episode lasted less than 1 s.This means that almost no laughing cycles occur, but laughter consists only of a single exhalation.This is very different from everyday interactions.In these, several vocalizations are often emitted in succession and the laughter episodes last correspondingly longer.This is especially true for exhilaration laughter and supports the assumption that laughter in psychotherapies predominantly fulfils a regulative function.

Discussion
Laughter is a complex phenomenon.Both the appearance and the corresponding functions vary with the type of social interaction.How laughter is shaped depends on various factors.A primary influence concerns the personality of the involved individuals, especially their ability and the way they regulate their emotions (Ruch, 1997).Relevant is also the role and power distribution between the interacting partners.In everyday interactions, the nature of the relationship varies strongly.While in couple interactions the distribution of roles is, in principle, symmetrical, this is not necessarily the case in mother-daughter or father-son dyads.Furthermore, the topic of conversation is also of great importance.When discussing difficult topics, regulatory pressure is higher and probably leads to more frequent and to different types of laughter.The quality of the relationship also influences the frequency and type of laughter.Shared laughter, which increases the affective bond between people, occurs more often in "good" relationships.This is demonstrated by our PAM studies (e.g., Huber & Bänninger-Huber, 2018).
How laughter manifests itself vocally and nonverbally is also influenced by the so-called display rules.These rules, which are culturally shaped and gender specific, state what inner state of mind you may show in which situation, form and intensity (e.g., Ekman, 1985).Therefore, cultural and gender comparative studies would be an important supplement to this existing research.
In psychotherapies, the frequency and types of laughter are influenced even more by the setting and the persons involved.Roles tend to be fixed in psychotherapy.Patients talk about difficult topics that are of emotional concern.The therapist usually refrains from making personal statements.The personality of patients and their ability to regulate emotions strongly determine the nature of the interaction.The quality of the therapeutic relationship is essential, because a good therapeutic relationship forms a necessary basis for a successful therapeutic process.Especially in psychoanalysis, the relationship is an important impact factor.It is the therapist's task to establish a form of relationship that enables patients to deal with their problems.Smiling and laughter play an important role in this process.In order to find out which nonverbal behaviors contribute to a productive therapeutic process, it is therefore necessary to describe these phenomena in a differentiated way.
Methodologically, we may conclude from these findings that it is important to study the phenomenon of laughter not only in an individual but in social interactions that consider the ongoing relationship situation.It also seems important not only to differentiate between the phenomena "laughing" and "smiling" but also to describe the different functional types of smiling and laughter according to their appearances, e.g., facial expression, vocalization, gaze and head movements.The differentiation of these types of laughter is still a work in progress.In a next step, we will systematically analyze the reactions of the therapist and include more psychotherapeutic dyads from our study.This might contribute to a more comprehensive understanding of the very multifaceted phenomenon "laughter".
Funding Open access funding provided by University of Innsbruck and Medical University of Innsbruck.The research was conducted at the Institute of Psychology at the University of Innsbruck without any external funding.

Table 1
Classification criteria and types of laughter; * AU6 = Cheek Raiser/Orbicularis Oculi, AU12 = Lip Corner Puller/Zygomaticus Major, can be scored on a 5-point scale ranging from a (low) to e (high)