Le bâillement, du réflexe à la pathologie
Le bâillement : de l'éthologie à la médecine clinique
Le bâillement : phylogenèse, éthologie, nosogénie
 Le bâillement : un comportement universel
La parakinésie brachiale oscitante
Yawning: its cycle, its role
Warum gähnen wir ?
Fetal yawning assessed by 3D and 4D sonography
Le bâillement foetal
Le bâillement, du réflexe à la pathologie
Le bâillement : de l'éthologie à la médecine clinique
Le bâillement : phylogenèse, éthologie, nosogénie
 Le bâillement : un comportement universel
La parakinésie brachiale oscitante
Yawning: its cycle, its role
Warum gähnen wir ?
Fetal yawning assessed by 3D and 4D sonography
Le bâillement foetal


mise à jour du
10 mai 2009
Eur Arch Psychiatry Clin Neurosci
Empathy in schizophrenia: impaired resonance
Haker H, Rössler W
   Department of General and Social Psychiatry,
Psychiatric University Hospital of Zurich, Switzerland
Tous les articles sur la contagion du bâillement
All articles about contagious yawning


Resonance is the phenomenon of one person unconsciously mirroring the motor actions as basis of emotional expressions of another person. This shared representation serves as a basis for sharing physiological and emotional states of others and is an important component of empathy.
Contagious laughing and contagious yawning are examples of resonance. In the interpersonal contact with individuals with schizophrenia we can often experience impaired empathic resonance. The aim of this study is to determine differences in empathic resonance-in terms of contagion by yawning and laughing-in individuals with schizophrenia and healthy controls in the context of psychopathology and social functioning.
We presented video sequences of yawning, laughing or neutral faces to 43 schizophrenia outpatients and 45 sex- and age-matched healthy controls. Participants were video-taped during the stimulation and rated regarding contagion by yawning and laughing. In addition, we assessed self-rated empathic abilities (Interpersonal Reactivity Index), psychopathology (Positive and Negative Syndrome Scale in the schizophrenia group resp. Schizotypal Personality Questionnaire in the control group), social dysfunction (Social Dysfunction Index) and executive functions (Stroop, Fluency). Individuals with schizophrenia showed lower contagion rates for yawning and laughing. Self-rated empathic concern showed no group difference and did not correlate with contagion. Low rate of contagion by laughing correlated with the schizophrenia negative syndrome and with social dysfunction.
We conclude that impaired resonance is a handicap for individuals with schizophrenia in social life. Blunted observable resonance does not necessarily reflect reduced subjective empathic concern.

Empathic resonance is the phenomenon of one person unconsciously mirroring the motor actions as basis of emotional expressions of another person. This shared representation serves as a basis for the ability to share physiological and emotional states of others and makes up one component of empathy [19, 22, 42, 61]. Empathy is not a unitary function; it is more likely based on at least partially dissociable functional systems which can be divided into motor empathy, i.e., empathic resonance, a cognitive and an emotional part of empathy [4, 12, 13]. The cognitive part of empathy comprises the ability to understand and explain mental states of others (known as theory of mind, ToM), whereas the emotional part of empathy includes the own experience of the other person's actual or inferred emotional state. Resonance can be seen as a bottom-up input for the emotional and cognitive part of empathy and is mediated by shared representations in the mirror neuron system (MNS) and frontoparietal networks [14, 18, 26, 34, 47, 61]. This so-called perception-action link has also been referred to as "chameleon effect" [8, 42]: an unconscious mimicry of the postures, facial expressions, and other behaviors of one's interaction partners, such that one's behavior passively and unintentionally changes to match that of others in one's current social environment. It constitutes a basic way of getting in contact with another person.
This basic way of establishing contact can be impaired. When these subtle alterations in communication are experienced during contact with an individual suffering from schizophrenia, they subsequently lead to an intuitive diagnosis of schizophrenia. The Dutch psychiatrist Henricus Cornelis Rumke first mentioned this phenomenon in the literature in 1941. He used the term "praecox feeling" to describe "the inability to come in contact as a whole" with a person who suffers from schizophrenia [50]. We assume that this phenomenon is based on reduced resonance. This intuitive reasoning based on subliminal information is still used today by some psychiatrists in daily practice in addition to standardized diagnostic classification [21]. Impaired empathy has recently been assumed to be involved in schizophrenia [2, 19, 51, 54]. Long before, Karl Jaspers stated that a failure of empathy and understanding are common elements in diagnosing schizophrenia [23]. The "inability to come into contact as a whole" can indeed be a diagnostic tool for psychiatrists but is most notably a handicap for affected individuals in interpersonal communication in everyday life. Reduced resonance forms a barrier for interpersonal contacts and adequate social functioning [56] and has stigmatizing potential [17]. The hypothetical assumption of deficits in resonance as basis of the "praecox feeling" puts the specificity of the sign in question with respect to other diagnoses with altered social reciprocity such as schizotypal personality disorder or high functioning autism.
Contagious yawning (cY) and contagious laughing (cL) are easily observable signs of resonance as an interaction between two individuals, with one person experiencing and sharing the physiological and emotional state of the other. The implicit link between two persons in cY has been discussed in the literature as a sign reflecting the motor mimicry component of human empathy [33, 40, 41, 43, 53] and
as evidence of empathic abilities in chimpanzees [1], and dogs [24]. Platek et al. [40] showed a correlation between higher scores on the Schizotypal Personality Questionnaire (SPQ) and lower rates of cY in a sample of undergraduate students. Recently, impairment in cY in children with autism spectrum disorder was reported by Senju et al. [55]. To the best of our knowledge, no study has examined cY and cL in individuals suffering from schizophrenia so far. We hypothesize that (1) cY and cL are impaired in individuals suffering from schizophrenia compared to healthy controls. Furthermore, we hypothesize that (2) a reduced ability to resonate-measured in terms of contagion-is associated with severe psychopathology and impaired social functioning in individuals with schizophrenia.
To test our hypotheses, we assessed two observable measures of empathic resonance (cY and cL) and one selfassessment measure of empathy (Interpersonal Reactivity Index, IRI) in individuals with schizophrenia and healthy controls. In both groups, we compared the ability to resonate with their psychopathology [i.e. with the Positive and Negative Syndrome Scale (PANSS) in individuals with schizophrenia, and with the SPQ in healthy controls]. In the schizophrenia group, we additionally compared the measures of resonance and the self-reported empathic abilities with the Social Dysfunction Index (SDI).
We compared the ability to resonate (observable contagion and self-report) in individuals with schizophrenia with healthy controls in the context of psychopathology and social functioning. We hypothesized a lack of cL and cY in the schizophrenia group and an association between the lack of contagion and impairments in social functioning.
Group differences in contagion
Our first hypothesis was confirmed by the main finding in our study: the significant reduction of cL and cY in the schizophrenia group compared with healthy controls. The mean incidence rate of cY in our control group (38.3%) matches the 41.5% reported by Platek et al. [40].
The impaired contagion as a sign of empathic resonance in the schizophrenia group can be explained by different influences, in particular by psychomotor constrictions due to illness or medication. We may state, that we found no significant correlation of medication dosage (in CPZe) and contagion. Executive cognitive functions were as expected impaired in the schizophrenia group but did not correlate with contagion. The analysis of stimulus-incongruent laughing and yawning (siL, siY) allows an inference on the overall psychomotor activity of the participants. The group difference with less response in the schizophrenia group is seen in siL but not in siY. A negative influence of general illness related factors such as negative symptoms or medication (in CPZe) is also seen in siL but not in siY. Furthermore in the schizophrenia group, siL correlates significantly with the intended contagion (cL).
This means that the recorded laughing in the schizophrenia group was not specific to the phenomenon of contagion but might reflect the general level of unspecific social responsiveness, psychomotor activity, or attention. In contrast, siY-as a general disposition to yawn-correlated not with cY, and showed no influence of psychopathology or medication. We see this as an indicator of a higher specificity of the more basal yawning stimulus. The situation is inverted in the control group. Here, cL is not the continuation of the high level social responsiveness reflected by siL, but is distinct by the specific stimulation. Our interpretation is that in the control group, the (preexisting high) laughing response is highly modulated by the external stimulation. We assume that this effect is not only due to contagion but also other (possibly social-cognitive) factors which are more pronounced in the control than in the schizophrenia group. One possible explanation for the unspecific yawning response in the control group (i.e. high correlation between siY and cY) could be a gating effect: a reinforcement of the automatic process of contagion by repetitive stimulation. The initial lower level of contagiousness may hamper this gating mechanism in the schizophrenia group.
Furthermore, one might discuss if individuals with schizophrenia feel more uncomfortable in the test situation, thus reducing their ability to resonate. However, our analysis of the distractor task data showed that there was no significant difference between the two groups in their judgment how comfortable and likeable the stimulation was.
Psychopathology and social functioning
Due to the homogenous low scoring of our control group on the SPQ (compared to the original population described by Raine [45], the variance was too low to answer our question about association of reduced contagion with schizotypal symptoms in the control group. The lack of contagion in the schizophrenia group responds to the clinical impression of the illness, and is reflected in the negative correlation of cL with the PANSS negative scale. However, the overall PANSS scores of our sample were rather low, corresponding to full or partial remission, thus, restricting our conclusion to less severe psychopathological states. As stated above, cY seems to detect variances (partly) independent of obvious psychopathology and points towards an underlying phenomenon, whose embedding in the common concepts of empathy has further to be discovered. It has further to be studied, how impairments of cY are related to an experienced psychiatrist's praecox feeling.
To the best of our knowledge, this is the first empirical evidence about the contribution of empathic resonance on social functioning in schizophrenia. Other empathic abilities such as the ToM (part of the cognitive part of empathy) have already been studied and are known to be impaired in schizophrenia and to contribute to the variance of social functioning [5, 7, 35, 48]. The negative correlation of cL with the SDI gives some support to our second hypothesis. However, we see this correlation only in one stimulus condition and it is of low significance. The correlation is based on clearly significant correlations of cL and cY with only one domain: dysfunction in "Family relationships". A lack of contagion in individuals with schizophrenia seems to occur mainly in close social contacts within the family circle. Social networks of individuals with schizophrenia consist largely of unidirectional therapeutic relationships. By contrast, family members have their own interpersonal needs and desires in face of their impaired relative. This may make this domain especially vulnerable to interpersonal challenges [31, 37]. However, an interpersonal handicap due to impairments in resonance leaves room to be coped/for rehabilitation in many domains of social life.
The observable signs cY and cL are perceived in daily contact as nonverbal statements in social interaction. The absence of these signs leads us to make psychopathological interpretations about the affective state of the observed individual, and it may lead us to speculate about a deficient perception of social-emotional information. The interpretation of self-report measures in individuals with schizophrenia is limited, since deficits in ToM may affect the representation of own mental states [16]. Nevertheless, it is the most direct way to understand the subjective experience of an affected individual.
The TRI was already used for individuals with schizophrenia in a recent study by Montag et al. [36]. The results of our study replicate the group differences reported by Montag et al.: individuals with schizophrenia reported less perspective taking and more personal distress in response to difficult situations of others compared to healthy controls. No group difference was found on the Fantasy scale, in measuring the ability to fantasize about fictional characters and in empathic concern.
The reduced subjective perspective taking ability is in line with objective findings on ToM deficits reported in the literature [6, 16, 58] and evidence of a certain insight into social cognitive deficits. Unimpaired empathic concern combined with increased personal distress suggests that the perception of social-emotional content may be intact but cannot be processed adequately, suggesting a problem of the output and not predominantly of the input. This view was already expressed by Kring et al. [29, 30] and supported by facial emotional EMG findings. A core process involved in the handling of social-emotional information is the distinction between self and other. This function is known to be impaired in schizophrenia [9, 11, 20, 60]. On the one hand this can lead to a dysfunction of action attribution, contributing to positive symptoms such as verbal hallucinations or delusions of control [15, 27, 52, 62]. On the other hand, it can lead to difficulties in attributing shared representations, e.g., social-emotional information, which can produce self-oriented responses such as personal distress to difficult situations of others [32]. The positive correlation between the "personal distress" scale and the PANSS negative scale indicates the stressful component of a clinically blunted affect.
The comparison of the observed contagion with the selfreport in our sample revealed only a minor association: only one of the four subscales of the TRI, the Fantasy scale, showed a significant correlation with contagion on the behavioral level. Particularly the two scales Empathic concern and Personal distress, that one could expect to be the most influenced by motor empathy regarding contagion, showed no correlation. The absence of this correlation in both groups suggests its being a general dissociation between behavioral and experiential response, rather than an expression of illness-related lack of insight.
Clinical implications
The observation of the ability to resonate is implicitly part of each clinical examination. It may even serve as an intuitive diagnostic instrument for schizophrenia, and is possibly related to the so-called "praecox feeling" [50]. This intuitive notice is more subtle than the clinical identification of negative symptoms. To what extent our low contagion data resemble a clinician's praecox feeling has yet to be determined. Rumke, who first described this clinical phenomenon [50], already mentioned the problem of subjectivity when considering the "praecox feeling". An "objective" measurement of resonance using defined and invariant contagious stimuli (e.g. video sequences) allows controlling for (counter-) transference in personal contact-not in the Freudian sense, but in the sense of mutual resonance.
Limitations and suggestions for further research
We cannot discuss depressive numbness as reason for impaired resonance in consequence of lacking clinical rating for depression. Likewise, subjective tiredness as reason for facilitated cY was not assessed and limits the interpretation of the data.
Since this is a cross-sectional examination, we cannot predict if the so measured resonance is a state or a trait. Regarding the lack of Resonance in the (at least partial) remission state of our sample, reflected in low PANSS scores, we could speculate that this impairment is not statedependent. A retesting or longitudinal study is needed to evaluate contagion over time and in various psychopathological states.
We are aware of the problem of multiple testing. As this is the first study of this kind, we decided not to raise the significance level, as not to exclude potential interpretations for future research.
Our explanations of the differences between cY and cL are only speculative. We used these two phenomena in our study presuming a common basis. Up to now, we can only hypothesize about top down cognitive influences that make laughing more contagious and yawning a purer resonance stimulus. The examination of the two stimulation conditions together with social-/cognitive parameters could provide further information about their differences. The combination of contagion response with a broader spectrum of other empathy-related functions (e.g. ToM, selfother distinction, emotional tasks) stimuli is needed to integrate these easily observable signs in the diverse conceptions of empathy.
Our paradigm could be used to further investigate the role of the MNS in schizophrenia [2, 19, 51]. Furthermore, imaging studies could reveal differences in the neural substrates of cY and cL.
On the level of clinical use, a cY test with higher resolution in the lower range could reveal further insight into the group of most impaired individuals.