Differences between Mimicking and Non-Mimicking laughter in Child-Caregiver Conversation: A Distributional and Acoustic Analysis

Abstract

Despite general agreement that laughter is crucial in social in- teractions and cognitive development, there is surprisingly lit- tle work looking at its use through childhood. Here we inves- tigate laughter in middle childhood, using a corpus of online calls between child and parent and between the (same parent) and another adult. We focus on laughter mimicry, i.e., laugh- ter shortly following laughter from the partner, and we com- pare mimicking and non-mimicking laughter in terms of dis- tribution and acoustic properties using spectrotemporal modu- lation measures. Our results show, despite similar frequencies in laughter production, different laughter mimicry patterns be- tween Parent-Child and Parent-Adult interactions. Overall, in comparison with previous work in infants and toddlers, our results show laughter mimicry is more balanced between par- ents and school-age children. At the acoustic level, we observe differences between mimicking and non-mimicking laughter in children, but not in adults. Moreover, we observe signifi- cant differences in laughter acoustics in parents depending on whether they interact with children or adults, which highlights a strong interlocutor effect on laughter mimicry.

How do you laugh in an fMRI scanner? Laughter distribution, mimicry and acoustic analysis

“…Although slightly more deviance was explained by the spectral modulation GAMM (93%), no significant differences between laughter types were observed. These findings are consistent with those reported in [27], where no significant differences between mimicking and non-mimicking laughter produced by adults were observed. It is possible that mimicking laughter is more strongly associated with modulations of amplitude envelopes rather than changes in spectral features, such as formants and fundamental frequency, which have been correlated with arousal in humans [45,46,43] and in non-human animals [47,48].…”

“…As studies have reported temporal modulations in the 1-2 Hz range are associated with prosody [22,44], one interpretation of this finding suggests that participants might prosodically mark their responsive (mimicking) laughter, crucial to showing affiliation and bonding [2], to compensate for the experimental setting, i.e., one partner in an fMRI machine, communication via computer-mediation, etc. A similar trend was observed in [27], which reported children communicating with their parent via similar audiovisual means, i.e., interaction through a screen, produced mimicking laughter with increased temporal modulations in the 0-4 Hz range. Although the authors reported no differences in temporal modulations between laughter types for adults, mimicking laughter produced by parents interacting with their child were increased in comparison with other adults.…”

“…Following the annotation scheme proposed in [7], the level of arousal displayed by the laughter's form (low/medium/high), temporal sequence in relation to speech (laughter/speech-laughter), and others' laughs (non-/mimicking laughter) were annotated. We define mimicking laughter as laughter that shortly follows a laughter from the partner, either overlapping or beginning within one second from the end of the initiating laughter (see [30,27] for calculation methods). The term non-mimicking laughter refers to any laughter not meeting this criteria.…”

“…[25,26] showed that modulations of the amplitude envelope and F0 could be used to distinguish speech, laughter, and speech-laughter. [27] relied on such measures to examine whether mimicking and non-mimicking laughter are characterised by specific modulation rates. Since previous work suggests mimicking laughter might involve different pathways in comparison to non-mimicking laughter [28], a spectrotemporal modulation analysis was selected to analyse the acoustic differences between the two.…”

Automatic Detection of Gaze and Smile in Children’s Video Calls

“…Finally, in addition to insights we gained from the above modeling experiments, one broader goal of this work is to provide the basis of a collaborative toolkit for the automatic detection of children's communicative signals in video chat, thus facilitating and speeding up research on children's online multimodal interactions. While the current work focused on smiles and gaze (the code of which will be made available to the community to use/optimize), in future work we will tackle other important signals like head nods/shakes [28], interactive alignment [26], prosody [10], Backchannel [5], laughter [24], dialog acts [27], and contingency [1].…”