Expressive suppression (ES) is generally thought to be a maladaptive emotion regulation strategy because frequent ES use is associated with negative psychosocial functioning. However, these findings are typically based on trait-level questionnaires to capture ES use, and fail to capture functional features of ES use in the moment, such as effort and success, and emotion characteristics, such as intensity, that might motivate ES use. In order to assess potential differences in momentary ES effort and success, and emotional intensity, we recruited High (N = 99) and Low (N = 98) Suppressors (N = 197, 82.7
Dominance and prestige are two strategies people use to attain high social rank in their groups. Dominance relies on the use of coercion and force, whereas prestige relies on sharing expertise and knowledge and serving as a role model. Although we know much about the behavior of dominant and prestigious leaders, we know much less about basic mental representations that guide people’s understanding of dominance and prestige. Two experiments tested predictions regarding spatial representations that may reflect people’s understanding of the two leadership strategies. Both experiments indicated that, compared to avatars representing dominant leaders, avatars representing prestigious male leaders were placed in closer physical proximity to subordinate group members. Findings align with work in proxemics and suggest that people conceptualize prestigious (relative to dominant) male leaders as more embedded within the groups they serve. Broader discussion centers on contributions of the current work to literature on group behavior and social cognition.
This research examined whether observers can accurately detect physical exertion and how their ratings align with athletes’ self-reports and physiological indicators. Across five independent studies (conducted 2024–2025 in Germany), 276 observers evaluated exertion from short video clips of 103 athletes differing in age (children and adults), exertion stage (rested, medium, high), and activity type (sprinting, running, cycling, workouts, track and field) to test the primary hypothesis that people can detect exertion based on thin slices of behavior. Observers reliably detected exertion based on the results of ANOVA and Pearson correlation analyses: ratings increased with physical load and correlated strongly with both self-reports and heart rate data, indicating reliance on valid perceptual cues. Comparisons showed slightly stronger correlations between self-reports and heart rate than between observer ratings and heart rate. Notably, after highly strenuous tasks, observers tended to underestimate exertion relative to athletes’ self-ratings. These findings highlight the accuracy of thin-slice judgments of exertion while revealing systematic underestimation at peak effort, offering insights with practical implications for sports performance monitoring and training feedback.
Affective touch is interpersonal touch which carries an emotional valence and is important to humans and other primates for stress resilience and social bonding. It is acknowledged that infants who sleep in close proximity to a caregiver receive more touch. However, mother-infant night-time touch has not been investigated with an understanding of affective touch mechanisms and with infant night-time location treated as a continuous rather than categorical variable. In a secondary data analysis of sleep laboratory observations of 13 UK mother-infant dyads, some for multiple nights (19 total observations, averaged within participants), we used continuous coding to quantify the amount and type of touch that infants received, and how it related to the amount of time they spent in different locations during the night. At the time of the observations, infants were four months old or younger, except one infant who was observed three times, with the last observation being at 6 months of age. We found that infants who spent more time in the adult bed received significantly more touch overall, and specifically more static touch. However, the time the infants spent in the adult bed was not significantly associated with the amount of stroking touch the infants received, with very little stroking touch recorded overall. We propose that static touch received in the night-time environment during bed-sharing can be a type of affective touch, as it may communicate safety to the infant through the presence of a responsible adult caregiver.
Given the immense reproductive costs posed by pathogens, humans have evolved a behavioral immune system that identifies sickness cues in conspecifics and facilitates prophylactic behavior. Among these cues, body movements constitute a distinct channel for conveying health-related information. Previous research has shown that perceivers can distinguish sick individuals above chance by relying exclusively on variations in gait. Nonetheless, it remains unanswered whether detecting sickness via gait patterns prompts social avoidance intentions. Using animated avatars performing sick and healthy gaits, we investigated whether observers’ self-reported social avoidance intentions increase as a function of contagion potential conveyed through gait. We also assessed perceptual sensitivity to sickness-related gait features via ratings of gait regularity and perceived similarity between gait types, and tested whether these outcomes are modulated by experimentally manipulated pathogen threat salience. As predicted, participants reported stronger avoidance intentions toward avatars displaying sick gaits and judged these gaits as significantly less regular, even after controlling for age perceptions. Although pathogen salience heightened sensitivity to gait irregularity, it did not significantly affect avoidance intentions or similarity ratings. By moving beyond detection and indirect indicators of avoidance, this study provides initial empirical support for the role of dynamic bodily information in regulating self-reported social avoidance intentions, highlighting the ecological validity and functional significance of movement-based health signals in disease-avoidance psychology.
Emotional mimicry—the automatic matching of another person’s emotional display—fosters social bonding, but only when an affiliative emotion aligns with its context. We tested whether a brief Loving-Kindness Meditation (LKM) could broaden affiliative responsiveness to both congruent and incongruent emotional displays. In an online study (N = 214), participants watched targets expressing joy or disgust in scenes that were either contextually congruent (e.g., smiling at a funny clip, showing disgust at a disgusting clip) or incongruent (e.g., smiling at a disgusting clip, showing disgust at a funny clip). Facial muscle activity was assessed using OpenFace and subjective ratings of closeness and appropriateness were obtained. We found mimicry combined with high closeness ratings only for joy in a matching context, whereas incongruent and antagonistic displays evoked disapproval rather than affiliation. Critically, LKM increased mimicry to congruent joy and reduced frowning responses to incongruent disgust, without altering appropriateness judgments or perceived closeness. These findings demonstrate that even a brief LKM can selectively boost mimicry of context-appropriate affiliative emotions and temper reactive disapproval when negativity violates social norms, highlighting LKM’s potential to foster empathy in emotionally discordant situations.
A growing body of research shows medical face masks to affect face perception and recognition. Although face masks negatively impact recognition of face identities and emotion expressions, they largely positively affect trait impressions. Here, focusing on impressions of trustworthiness and responsibility, we bridge between various lines of inquiry in the extant literature. First, among American perceivers, we test whether the effect of masks varies by target faces’ race, crucially including targets from three racial groups: Asian, Black, and White. Next, we examine the extent to which effects can be explained by the presence of masks versus the absence of lower facial information. Finally, we examine the role of temporal context, testing masks’ effect on trustworthiness and responsibility judgments in early 2021 and mid 2023. Altogether, we find face masks to have a positive effect on trustworthiness and responsibility judgments. Lack of lower facial information explained some, but not all, of the effect of masking, indicating that masks provide a specific social signal that affects trait judgments beyond the effect of occlusion. Finally, we found the positive effect of masking on trait judgments to persist in 2023 but decrease in magnitude compared to 2021, indicating that the social meaning of masks has waned over time. Masking effects were moreover stronger for White than Black (2021, 2023) and Asian targets (2023) and among perceivers with more positive attitudes toward masking (especially in 2023). Altogether, we provide evidence that both bottom-up perceptual and top-down social factors influence social perceptions of masked faces.
Gestures are an important component of nonverbal communication and are often produced along with speech to aid comprehension, direct attention, and provide emphasis. This study investigates the effects of speaker and addressee visibility on gesture production in computer mediated communication. Based on previous research, we hypothesized that speakers would produce more representational gestures, which convey ideas, actions, and objects, when they could see their addressee’s face and were visible to their addressee, supporting the facial cueing and perspective taking hypotheses. These hypotheses suggest that speakers adjust their gesture use to supplement their speech by considering what their addressee can see (perspective taking) or by using cues from their addressee’s facial expressions (facial cueing). Participants viewed six cartoon clips and retold narrative descriptions of each clip to a confederate via videoconferencing, with visibility conditions varied independently for the speaker and the addressee. Speakers produced significantly more representational gestures when they were visible to the confederate than when they were not. By contrast, speakers’ representational gesture production did not differ when they could see the confederate’s face and when the confederate’s camera was off; however, speakers produced fewer representational gestures when the addressee’s torso was visible than in these conditions. The findings of this study support the perspective taking hypothesis but fail to support the facial cuing hypothesis, providing insight into how gestures are used to communicate in videoconferencing settings.
Facial expression has been proposed as a key mechanism in forming and maintaining social relationships. Humans have particularly expressive faces, but direct evidence linking individual level expressivity to group-level social bonding is limited. Here we examined whether spontaneous facial expressivity impacts social popularity during initial group formation. Participants (72 groups, N = 256) engaged in unstructured online interactions in small groups with strangers during which facial behaviour was quantified using automated analysis. Social popularity was measured by calculating social network indices based on relative liking. More facially expressive individuals were rated as warmer and more cooperative, and occupied more central positions in the group social network. In contrast, facial attractiveness did not predict social standing. These findings support the hypothesis that facial expressivity functions to promote cooperative relationships during group formation.
The current study investigated whether tactile interventions could alleviate emotional distress and reduced prosocial behavior following social exclusion. Using the Cyberball paradigm to induce social exclusion, two experiments examined the effects of distinct touch types: Study 1 implemented social affective touch (a slight pat on the back), while Study 2 employed physical affective touch (gentle stroking with a wool brush). Participants’ positive and negative emotions, along with prosocial behavior, were assessed post-intervention. Results indicated that neither form of affective touch mitigated the negative emotional consequences of social exclusion. However, both social and physical affective touch effectively attenuated the decline in prosocial behavior. Notably, the effect of social touch was moderated by attachment orientations: the buffering effect on prosocial behavior was observed only among individuals with low attachment anxiety and low attachment avoidance. In contrast, no such moderating role of attachment dimensions was found for physical touch.
Gaze behavior, being continuously accessible to interlocutors in face-to-face interactions, serves as a cue for managing turn-taking, regulating the duration of topical sequences, and supporting cognitive processing in various everyday conversational contexts. The present study seeks to enhance the understanding of the relation between two forms of interactive gaze behavior – gaze aversions and mutual gaze – and the topical development in the explanatory discourse. To do so, we analyzed 24 dyadic board game explanations in which one explainer subsequently explained a board game to three different explainees while the board game was physically absent from the shared space. The main objective of the present study was to investigate the relation of gaze aversions and mutual gaze to the topical development of explanations. For this, based on previous research (Lazarov et al., 2024; Rossano, 2012) we hypothesized that (1) gaze aversions are more likely to be associated with topic changes than topic continuations, and that (2) mutual gaze is more likely to be associated with topic continuations than topic changes. In addition, we explored how the two forms of gaze behavior are related to the interlocutor who initiates a topic change or continuation. Our proportional analysis using a Generalized linear mixed effects model revealed that gaze aversions are related to topic changes initiated by both interlocutors. In contrast, the analysis did not reveal a significant relation between mutual gaze and topic continuations, which could be explained by the feedback elicitation function of mutual gaze at the end of speakers’ utterances (Bavelas et al., 2002; Brône et al., 2017; Kendon, 1967) while monitoring the addressees’ understanding (Clark Krych, 2004) and the complexity of the analyzed fixed and random effects.
Recently developed artificial intelligence (AI) tools are now capable of generating realistic face images that are indistinguishable from photographs. However, little is known regarding how these synthetic images portray emotional expressions. Here, with only simple prompting, we generated a set of images depicting seven facial expressions using ChatGPT. These images were intermixed with posed facial expression photographs drawn from a standardised database and we collected participants’ judgements regarding which emotions were portrayed, along with their rated intensities. Our results showed that synthetic images were more accurately classified as their intended emotions, as well as more intense displays, in comparison with posed photographs. In addition, AI-generated images depicted action units associated with particular emotions to a similar extent to posed photos, although this pattern varied across emotions. Finally, both types of images were correctly classified by ChatGPT with few or no errors. Taken together, these findings provide insight into how AI tools represent, and subsequently depict, human emotions. Further, this approach may facilitate the creation of novel, individually-tailored image sets for use in future research by incorporating more complex and specific prompting.
Recognizing emotional intent from prosody (vocal emotion recognition [ER]) is found to be more challenging than the same ability with facial expressions (facial ER). However, disparities between these nonverbal modalities may be because most research compares ER accuracy for static faces and dynamic prosody. Interpreting dynamic expressions that unfold over time may be more difficult and require more advanced cognitive abilities, like working memory. We explored (a) differences in ER accuracy across tasks varying in modality (facial, vocal) and dynamism (static, dynamic), and (b) whether trait-level cognitive abilities (working memory, attentional control) differentially predicted ER accuracy across tasks. 300 adults (52.3
This special issue on the affiliative dimension of nonverbal behavior brings together cutting-edge thematically-linked empirical research examining how nonverbal cues signal, reflect, and shape affiliation and social connection. The contributions clarify the causal direction of the relationship between nonverbal cues and affiliation—whether behaviors such as laughter, mimicry, and synchrony create social closeness or emerge from it. Collectively, the papers also refine theoretical distinctions among affiliative signals by disentangling the overlapping and unique functions of backchannels, nodding, and laughter. Finally, this issue highlights the growing importance and utility of automated methods, such as the use of AI-generated avatars and motion capture technology, which allow for greater precision in isolating and measuring cues within dyadic interaction. Together, these studies advance a more nuanced and methodologically sophisticated understanding of how affiliation is expressed through our nonverbal behavior.
Pain signals immediate bodily threats. Its nonverbal expression often blends elements of other affective states, including emotions. Observers may perceive these states through the shared affective dimension of unpleasantness. However, no study has examined how the perceived unpleasantness of pain and negative emotions is prioritized when conveyed within the same facial sequence. This study examined how observers integrate facial configurations of pain and negative emotions (anger and sadness) by measuring perceived unpleasantness. It was hypothesized that facial sequences involving pain and negative emotions would be perceived as more unpleasant than those combining only emotions, regardless of their arousal. Sixty-three healthy participants (43 women, 23.52 ± 4.08 years old) watched a series of dynamic virtual agents transitioning from a short contextual facial configuration (667 ms) to a longer target configuration (1,333 ms). They then rated the unpleasantness perceived from the sequence. The results indicated that, although anger is associated with higher arousal than sadness, sequences combining anger or sadness contexts with a pain target elicited comparable levels of perceived unpleasantness. Moreover, sequences combining a pain context with anger or sadness targets were rated as more unpleasant than emotion-only sequences. These findings suggest that pain may be more salient or function as a more relevant signal than some negative emotions when embedded in facial sequences. Such evidence challenges the idea that all affective states contribute equally to the perception of another’s unpleasantness. This study highlights the importance of examining how multiple affective states are communicated over time.
Nonverbal behaviour (NVB) is a critical component of human communication, particularly in high-stakes, naturalistic settings such as sport. Despite advances in computational analysis, the extent to which objectively computable NVB features reflect subjective human judgments remains unclear. In this study, we present the inaugural investigation into the correspondence between computable NVB indicators and observer perceptions in the context of elite football. The utilisation of 3D pose data from a FIFA World Cup match enabled the extraction of volumetric, angular and kinematic features from players’ body movements during match events. Parallelly, human participants rated short video segments of these events on multiple NVB dimensions. Statistical analyses revealed significant differences in objective computable NVB indicators across play contexts (scoring, conceding, regular play), with players on the scoring team exhibiting more expansive and dynamic body language. Subjective ratings closely mirrored these differences, and repeated measures correlation analyses demonstrated strong associations between objectively computable NVB features and human judgments. These findings may provide a scientific basis for the use of 3D pose-based metrics for scalable, automated NVB analysis and may highlight their ecological validity in capturing expressive behaviours meaningful to human observers. The study has the potential to contribute to the advancement of methodological approaches for NVB research.
Gender-related acoustic features of voices may play a role in perceived aptitude for academic subjects and occupations. In the present study, parents and teachers rated the perceived competence of children they did not know with respect to a range of gender-stereotyped academic subjects and future possible occupations. Ratings were based solely on re-synthesized recordings of 8 children (4 girls, 4 boys) saying a single sentence. Voice manipulations involved changing the spacing between vocal resonances (formants), which affects perceived vocal tract length. Resynthesis involved changing format spacing only, either to the mean for each gender in the larger group from which the four speakers had been selected, or to one standard deviation above or below that mean, keeping fundamental frequency (voice pitch) unchanged. In Experiment 1, the speakers were rated for perceived competence in academic subjects by 61 parents and 38 teachers of children of a similar age to the speakers. Results showed that lowered or raised voices led to ratings of different levels of competence in traditionally masculine-stereotyped or feminine-stereotyped academic subjects, although the effects were clearer for boy speakers than for girl speakers. In Experiment 2, the speakers were rated for perceived competence in future occupations by 49 parents and 50 teachers. Results showed that perceptions of both boys’ and girls’ aptitude for gender-stereotyped occupations were even more clearly affected in stereotypical ways by the raised or lowered voices. The results have implications for possible interventions with adults to prevent unintentional stereotyping of children, and also point to the value of this method as an implicit measure of parents’ and teachers’ stereotyping of academic subjects and occupations.
Emotional mimicry, an unconscious nonverbal response, may reveal how people react to truthful versus deceptive communication. Mimicry typically increases when individuals seek connection but may be suppressed when deception reduces liking and affiliation. Emotional tone further shapes this process: positive expressions promote alignment, whereas negative ones can evoke scepticism or discomfort. This study examined how message truthfulness and emotional framing influence emotional mimicry in receivers. Using Automated Facial Expression Analysis (AFEA) in a 2 × 2 within-subjects design, 33 participants viewed 16 video messages in which four speakers delivered truthful or deceptive stories with positive or negative valence. Facial expressions were recorded and analysed to quantify emotional mimicry. Emotional mimicry was significantly lower for deceptive than for truthful messages across both positive and negative content, consistent with expectations. The decrease was strongest for positively framed deceptions, indicating that deception disrupts alignment most when positivity would otherwise encourage connection. Negative messages elicited weaker emotional mimicry overall, regardless of veracity, also in line with predictions. In contrast, presuppositions about deception were unrelated to emotional mimicry or veracity judgments, contrary to expectations. These findings suggest that emotional mimicry is not a straightforward cue to deception but an indicator of affective attunement shaped by authenticity and emotional tone. Mimicry appears driven more by emotional than cognitive processes, offering insight into unconscious receiver responses. The use of AFEA underscores both the promise and the limitations of automated tools for studying dynamic interpersonal behaviour.
Laughter is a profoundly social nonlinguistic behavior that functions primarily as a signal of affiliation and a facilitator of social bonding. Some evidence supports the anecdotal observation that laughter is contagious, but little empirical work has measured actual laughter contagion, much less the contextual variables that predict it. In Study 1, 79 participants rated the pleasantness of 93 laughter segments (31 posed laughs, 31 “stranger” laughs, and 31 “friend” laughs) while being videotaped. Two independent raters coded laughter contagion, the presence of participant laughter, in response to the 93 laughs. Friend laughter elicited more contagious laughter than stranger laughter, but friend and posed laughter did not differ. Study 2 (N = 308) attempted to replicate the results of Study 1 using some novel stimuli and additional rater judgments. Perceived authenticity and pleasantness demonstrated the predicted pattern, with friend laughter sounding the most authentic and pleasant, followed by stranger laughter, and then posed laughter. However, these laughter types did not differ by perceived contagiousness. A polynomial regression that regressed behavioral contagion ratings from Study 1 onto perceived authenticity ratings in Study 2 revealed a curvilinear pattern—behavioral contagion was highest for highly inauthentic (posed) laughs and highly authentic laughs. Acoustic analyses suggested that cues associated with greater physiological arousal predicted perceived contagiousness in Study 2. This research is among the first to demonstrate differences in behavioral laughter contagion as a function of laughter type and potentially implicates two unique theories responsible for contagious laughter—benign violation and social bonding/affiliation.
