The Components of Interpersonal Synchrony in the Typical Population and in Autism: A Conceptual Analysis. — TEI-tagged view

Claire Bowsher-MurraypersName, Sarah GersonpersName, Elisabeth von dem HagenpersName, Catherine R. G. JonespersName (2022date). “The Components of Interpersonal Synchrony in the Typical Population and in Autism: A Conceptual Analysis..” Frontiers in Psychologytitle·j.

Object of study — the component processes of non-verbal interpersonal synchronyterm·dom and their moderators

Research question — where interpersonal synchronyterm·dom may diverge in autismterm·key, given its components in the typical population

Contribution — a framework of how component mechanisms contribute and interact to bring about interpersonal synchronyterm·dom

Full text — tagged

The Components of Interpersonal Synchronyterm·dom in the Typical Population and in Autismterm·key: A Conceptual Analysisterm·mth.

Abstract Interpersonal synchronyterm·dom – the tendency for social partners to temporally co-ordinate their behaviour when interacting – isterm·dom a ubiquitous feature of social interactions. Synchronous interactions play a key role in development, and promote social bonding and a range of pro-social behavioural outcomes across the lifespan. The process of achieving and maintaining interpersonal synchronyterm·dom isterm·dom highly complex, with inputs required from across perceptual, temporal, motor, and socio-cognitive domains. In this conceptual analysisterm·mth, we synthesise evidence from across these domains to establish the key components underpinning successful non-verbal interpersonal synchronyterm·dom, how such processes interact, and factors that may moderate their operation. We also consider emerging evidence that interpersonal synchronyterm·dom isterm·dom reduced in autistic populations. We use our account of the components contributing to interpersonal synchronyterm·dom in the typical population to identify potential points of divergence in interpersonal synchronyterm·dom in autismterm·key. The relationship between interpersonal synchronyterm·dom and broader aspects of social communication in autismterm·key are also considered, together with implications for future research.

Keywords: interpersonal synchronyterm·dom, behavioural co-ordination, social interaction, synchronisation, social motor behaviour, autismterm·key Introduction The tendency for social partners to temporally co-ordinate their behaviour, known as interpersonal synchronyterm·dom (ISterm·dom), isterm·dom a common feature of social interactions ( Bernieri et al., 1988 ; Delaherche et al., 2012 ). It isterm·dom sometimes the product of conscious effort, such as when we shake hands, high five, or dance together. It may also arise spontaneously: social partners might fall into step ( Zivotofsky and Hausdorff, 2007 ), align their postural positions ( Shockley et al., 2003 ; Gaziv et al., 2017 ), or entrain their body movements ( Hadar et al., 1984 ) or facial expressions ( Louwerse et al., 2012 ). ISterm·dom may display rhythmical properties (e.g., walking in step; nodding), but equally may be less structured in nature (e.g., sporadic gestures or postural adjustment). Although ISterm·dom may arise via a broad range of behavioural processes including gesture, gaze, facial expression, speech, and vocalisation, the current review isterm·dom focussed on the synchrony of non-verbal behaviours. Temporal co-ordination of social behaviour emerges shortly after birth ( Condon and Sander, 1974 ; Dominguez et al., 2016 ) and becomes more temporally accurate, more complex, and less reliant on adult facilitation during infancy ( Hilbrink et al., 2015 ; Meyer and Hunnius, 2020 ). Mother-child ISterm·dom isterm·dom believed to positively influence self-regulation and empathy, and to promote later cognitive, social and emotional development and secure attachment relationships ( Harrist and Waugh, 2002 ; Feldman, 2007 ; Evans and Porter, 2009 ). Throughout the lifespan, ISterm·dom serves as a social signifier and promotes various social outcomes, including increased affiliation ( Hove and Risen, 2009 ; Tunçgenç et al., 2015 ), rapport ( Vacharkulksemsuk and Fredrickson, 2011 ), bonding ( Tarr et al., 2015 ; Tunçgenç and Cohen, 2016 ), helping ( Tunçgenç and Cohen, 2018 ), and co-operation ( Rabinowitch and Meltzoff, 2017 ). Such effects are present both when ISterm·dom isterm·dom spontaneous and when it isterm·dom intentional, although there isterm·dom mixed evidence as to whether they are enhanced when partners share an intention to co-ordinate ( Reddish et al., 2013 ; Howard et al., 2021 ). The full range of social outcomes arising from ISterm·dom has been documented in recent reviews ( Rennung and Goritz, 2016 ; Vicaria and Dickens, 2016 ; Mogan et al., 2017 ; Cross et al., 2019 ; Hoehl et al., 2021 ).

Difficulties with social communication and social interaction, including with non-verbal communicative behaviour and building and maintaining relationships, are hallmarks of autismterm·key ( American Psychiatric Association, 2013 ). As such, there has been particular interest in how autistic people engage in and experience ISterm·dom ( McNaughton and Redcay, 2020 ). Evidence indicates that ISterm·dom isterm·dom less accurate and/or less frequent in interactions involving autistic people, in both spontaneous ( Marsh et al., 2013 ; Fitzpatrick et al., 2016 ; Kaur et al., 2018 ; Georgescu et al., 2020 ; Zampella et al., 2020 ) and intentional ( Fitzpatrick et al., 2016 ) contexts. Several studies have found an association between lower levels of ISterm·dom and higher levels of autistic traits ( Brezis et al., 2017 ; Cheng et al., 2017 ; Fitzpatrick et al., 2017b ; Zampella et al., 2020 ; Granner-Shuman et al., 2021 ; although cf. Kaur et al., 2018 ). There isterm·dom also some evidence that the social significance of ISterm·dom may be attenuated for many autistic people ( Koehne et al., 2016 ).

ISterm·dom emerges as a function of multiple mechanisms operating in concert with each other ( Konvalinka et al., 2010 ; Delaherche et al., 2012 ; Mills et al., 2019 ; McNaughton and Redcay, 2020 ). However, research commonly focuses on the role of individual mechanisms, such as attention (e.g., Temprado and Laurent, 2004 ; Richardson et al., 2007 ), perceptual processing (e.g., Noel et al., 2018 ), motor behaviour (e.g., Hart et al., 2014 ; Monier and Droit-Volet, 2019 ) and social factors (e.g., Kirschner and Tomasello, 2009 ; Lumsden et al., 2012 ; Honisch et al., 2021 ). To better understand ISterm·dom there isterm·dom a need to synthesise findings from across the perceptual, sensorimotor, social and cognitive domains. Understanding how the component processes underlying ISterm·dom operate together isterm·dom also necessary for understanding why ISterm·dom manifests differently in autismterm·key. In addition to core differences in social functioning, differences between autistic and typical populations have been observed across domains relevant to ISterm·dom, including attention ( Frazier et al., 2017 ; Hedger et al., 2020 ), temporal perception (e.g., Allman and Falter, 2015 ); perceptual processing (e.g., Feldman et al., 2018 ; Meilleur et al., 2020 ), and motor behaviour (e.g., Fournier et al., 2010 ). Therefore, characterising ISterm·dom in autismterm·key necessarily requires a holistic understanding of how differences in functioning across multiple underlying processes operate together.

The first part of this conceptual analysisterm·mth begins with a synthesis of the key component mechanisms that contribute to ISterm·dom ( Figure 1 ), including an exploration of how such processes interact, and the factors that may moderate their operation. Where useful, we additionally draw on the related concepts of imitationterm·dom and joint actionterm·dom. Imitationterm·dom, like ISterm·dom, involves behavioural matching in form, albeit not in time ( Hove and Risen, 2009 ; Catmur and Heyes, 2013 ). Joint actionterm·dom involves the conscious co-ordination of complimentary behaviour to achieve a shared goal ( Sebanz et al., 2006 ). The parallels between these two phenomena and ISterm·dom mean that their underlying processes can shed light on the role of equivalent processes in ISterm·dom. The second part of this conceptual analysisterm·mth describes how each of the identified component mechanisms operates in autismterm·key, and considers the extent to which differences in the functioning of these mechanisms may explain differences in ISterm·dom. In considering each mechanism not only individually but also as part of a wider system, we aim to build an understanding of how relevant mechanisms collectively underpin reduced ISterm·dom in autismterm·key.

Schematic illustration of the components of non-verbal interpersonal synchronyterm·dom, as discussed in this review.

Part 1: Interpersonal Synchronyterm·dom in the Typical Population Social Orientingterm·dom In order for an individual to synchronise their movements with an external stimulus, they must have perceptual access to the stimulus ( Richardson et al., 2007 ; Oullier et al., 2008 ; Miyata et al., 2017 ; Oh Kruzic et al., 2020 ). Perception of the stimulus, in turn, requires that sufficient attentional resources are allocated to it ( Aubin et al., 2021 ). In the case of ISterm·dom, the relevant stimulus isterm·dom one’s interaction partner. During social interaction, the perception of one’s partner and their movements isterm·dom facilitated by an inherent tendency to orient toward social stimuli, that isterm·dom, the inherent preferential allocation of visual attention to social rather than non-social cues ( Fletcher-Watson et al., 2008 ; Gluckman and Johnson, 2013 ; Leppanen, 2016 ; Rösler et al., 2017 ). Experimental manipulation of the extent to which partners are oriented toward each other has demonstrated its importance in facilitating ISterm·dom. Partners spontaneously synchronised their movements at above chance levels when looking directly at each other, but not when they had only peripheral visual access to each other ( Richardson et al., 2007 ).

It isterm·dom also important to consider what aspects of a social partner’s behaviour are particularly informative. The direction of a partner’s gaze and the kinematic cues provided by their limb movements convey information about their future movements ( Sartori et al., 2011 ; Ansuini et al., 2014 ; Khoramshahi et al., 2016 ), particularly where the end of point of an action isterm·dom otherwise unclear ( Fulceri et al., 2018 ). Such cues are likely to facilitate both spontaneous and intentional ISterm·dom. Mutual gazeterm·dom between partners may also support the emergence of ISterm·dom. Although there isterm·dom no direct evidence that mutual gazeterm·dom plays a role in ISterm·dom, mutual gazeterm·dom isterm·dom considered to facilitate social engagement, and isterm·dom known to promote spontaneous partner imitationterm·dom ( Wang and Hamilton, 2012 ; Prinsen et al., 2017 ). The shared element of behavioural matching in ISterm·dom and in imitationterm·dom suggests that mutual gazeterm·dom may have a similar role in promoting ISterm·dom.

Flexibility in attending to social cues isterm·dom another key consideration for effective ISterm·dom. Naturalistic interactions almost always contain multiple social cues, occurring either simultaneously or in quick succession. To capitalise on their predictive value, it seems likely that attention must be flexibly directed and redirected toward relevant cues. A further consideration isterm·dom that social cues often consist of limb or whole-body movements, which are inherently dynamic. For example, in order for a handshake to be successfully synchronised, each partner must match their eye movements to the movement of the other’s hand, such that the motion of the other partner isterm·dom efficiently tracked via smooth pursuit ( Johnson et al., 2016 ). In summary, social orientingterm·dom isterm·dom a core process necessary for ISterm·dom to take place. Dynamically attending to a partner’s behavioural cues, particularly kinetic and eye gaze cues, isterm·dom likely to be the specific means by which attending to a social partner supports ISterm·dom.

Multisensory Processing Interacting with a social partner isterm·dom a multisensory experience, involving the integration of social information delivered via multiple modalities. Sensory inputs from two or more modalities that occur with sufficient temporal proximity are perceptually combined via multisensory integrationterm·dom ( Alais et al., 2010 ; Murray et al., 2016 ). There isterm·dom evidence that efficient multisensory integrationterm·dom supports ISterm·dom. For example, synchronisation of both fine and gross motor activity was achieved with significantly greater accuracy when multisensory stimuli were available, as compared with when stimuli were unisensory only ( Elliott et al., 2010 ; Su, 2014 ; Wright et al., 2014 ; Khan et al., 2020 ). Further, increased multisensory acuity was significantly associated with increased spontaneous ISterm·dom ( Noel et al., 2018 ). Multisensory integrationterm·dom of social information isterm·dom likely to support ISterm·dom in a number of ways. First, it likely enhances the perceptual cohesion and salience of the behaviour ( Bahrick and Todd, 2012 ) with which synchrony isterm·dom to be achieved. Second, the combination of information from multiple streams enables the individual to make enhanced statistical predictions about the stimulus ( Elliott et al., 2010 ). Third, multisensory integrationterm·dom may promote perception action couplingterm·dom, whereby audio-visual cues are likely to engage the observer’s own sensorimotor system more than single modality cues, which then simulates the observed movement as a means of predicting its ongoing trajectory ( Su, 2014 ). Fourth, in addition to supporting the processing of incoming social stimuli, multisensory integrationterm·dom isterm·dom likely to support the effective planning, monitoring and adjustment of one’s own actions, which requires the effective integration of visual and proprioceptive information about one’s own motor behaviour ( Foster, 2019 ). Overall, therefore, efficient multisensory processing isterm·dom a core process that likely supports effective ISterm·dom via multiple pathways.

Action Predictionterm·key A defining feature of ISterm·dom isterm·dom that partners move together simultaneously. In order to achieve this, a partner’s movements must be anticipated, both spatially and temporally ( Sebanz and Knoblich, 2009 ; Meyer and Hunnius, 2020 ; Gvirts Probolovski and Dahan, 2021 ), rather than merely reacted to. The tendency to make spontaneous, online predictions about the actions of others isterm·dom present throughout the lifespan ( Flanagan and Johansson, 2003 ; Reddy et al., 2013 ; Sebanz and Knoblich, 2021 ), and isterm·dom likely influenced by low level kinematic information ( Ansuini et al., 2014 ) and gaze cues ( Khoramshahi et al., 2016 ), as well as higher-level cognitive processes, such as assessing a partner’s likely intentions in light of other contextual information ( Falck-Ytter, 2012 ). A number of studies suggest a role for action predictionterm·key in intentional ISterm·dom. For example, participants who displayed more accurate temporal prediction in a solo task showed significantly greater accuracy and stability of synchronisation during a joint finger tapping task ( Pecenka and Keller, 2011 ). Similarly, toddlers who made more temporally accurate predictions in an observation task moved with greater temporal stability during a joint actionterm·dom task with a partner ( Meyer et al., 2015 ). Thus, the evidence suggests that being able to form accurate predictions about a partner’s movements isterm·dom a core process facilitating intentional ISterm·dom. There isterm·dom no direct evidence as to how action predictionterm·key abilities inform spontaneous ISterm·dom. However, some form of prediction must necessarily occur in order for behaviour to become temporally aligned ( Gvirts Probolovski and Dahan, 2021 ). Further research isterm·dom required to establish exactly how predictions are made and integrated with other component processes in the context of spontaneous ISterm·dom.

Motor Behaviour In addition to anticipating the movement of a partner, it isterm·dom necessary to plan and execute one’s own complimentary movement sequence, both temporally and spatially. The ability to synchronise simple motor movements with an external stimulus isterm·dom limited in young children, partly by a relatively limited capacity to adapt motor behaviour to the tempo of an external stimulus, but develops during childhood ( Drake et al., 2000 ; McAuley et al., 2006 ; Monier and Droit-Volet, 2018 ) as a function of developing motor skills ( Monier and Droit-Volet, 2019 ). Immature motor skills are therefore believed to limit young children’s levels of ISterm·dom ( Trainor and Cirelli, 2015 ), although there isterm·dom limited direct evidence as to the role of motor skills in ISterm·dom in typical populations. In support of a role for motor planningterm·key in intentional ISterm·dom, reduced performance on a motor planningterm·key task was significantly associated with reduced intentional ISterm·dom in a hand movement task ( Granner-Shuman et al., 2021 ). However, the contribution to ISterm·dom of other aspects of motor abilities are yet unknown.

Just as motor skills are likely important to the achievement of ISterm·dom, so too isterm·dom the form of movement people tend to produce. People tend to display an “individual motor signatureterm·dom,” which isterm·dom a distinct and stable pattern of movement that isterm·dom personal to the individual in terms of direction, range and velocity of movement ( Richardson and Johnston, 2005 ; Hart et al., 2014 ; Słowiński et al., 2016 ). Some individual motor signaturesterm·dom convey more predictive information than others ( Koul et al., 2016 ), which isterm·dom likely to make them easier to synchronise with. Further, evidence suggests that partners with similar individual motor signaturesterm·dom are better at predicting the timing of each other’s movements ( Colling et al., 2014 ) and achieve a higher degree of co-ordination when moving together ( Słowiński et al., 2016 ), relative to partners whose motor signatures are relatively dissimilar. As well as displaying individualised patterns of movement, people also tend to exhibit a preferred pace of movement, or spontaneous motor tempoterm·dom ( Delevoye-Turrell et al., 2014 ). Just as the (dis)similarity of partners’ motor signatures influences the degree of co-ordination they achieve when interacting, it seems likely that those with relatively similar spontaneous motor temposterm·dom would achieve higher levels of ISterm·dom than those with relatively dissimilar motor tempos. In sum, interacting partners’ motor abilities, their natural movement patterns as well as, potentially, their relative pace of movement, all contribute to ISterm·dom.

Monitoring and Adaptation The component processes of ISterm·dom have so far been considered independently. However, during dynamic real-world interactions involving both intentional and spontaneous ISterm·dom, these processes are believed to be embedded together in a continuous feedback loop ( Shamay-Tsoory et al., 2019 ; Gvirts Probolovski and Dahan, 2021 ). Specifically, it isterm·dom proposed that predictions about a partner’s movement and one’s own plan to align with it are integrated into a forward modelterm·dom of the shared movement between partners. As motor commands are executed, “error monitoringterm·dom” occurs, whereby both one’s own and one’s partner’s actual motor output isterm·dom compared to the forward modelterm·dom. “Error detection” occurs when either partner’s actual movement does not match the generated prediction. Error detection precipitates “error correction,” where the predictive model and movement plan are updated ( Shamay-Tsoory et al., 2019 ).

The component processes of ISterm·dom, described above, may contribute to the effectiveness of the feedback loop in a number of ways. For example, effective error monitoringterm·dom isterm·dom likely underpinned by continued social orientingterm·dom and dynamic attendance to behavioural cues. Further, error monitoringterm·dom, in the context of ISterm·dom, consists of detecting asynchrony between the actions of partners. Thus, perceptual sensitivity to the temporal alignment of events likely contributes to the achievement of ISterm·dom. Error correction isterm·dom likely to draw on action predictionterm·key and motor abilities. The effectiveness of the updating process also depends on how quickly it occurs ( Vishne et al., 2021 ). The faster the internal model and movement plan can be updated and implemented, the more closely aligned partners’ behaviour will be over time. Efficient updating isterm·dom likely to be critical in real-world social interactions, in which the form and speed of partners’ movements change over time, placing persistent demands on adaptive mechanisms.

The process of continuous mutual adaptation during ISterm·dom, described by the above account, isterm·dom supported by behavioural evidence. For example, dyads required to synchronise their finger tapping adjusted the time between their taps in opposite directions to one another, on a tap-by-tap basis ( Konvalinka et al., 2010 ), suggesting that each partner continuously accounted for the pace of the other and modified the pace of their own tapping accordingly. During more complex interactions, there isterm·dom evidence that partners spontaneously adapt both the spatial ( Sartori et al., 2009 ; Candidi et al., 2015 ) and temporal ( Vesper et al., 2011 ) qualities of their movements, so as to make them more predictable to their partner. The result of partners’ mutually adaptive behaviour isterm·dom that they coalesce into a third movement pattern distinct from either of their individual motor signaturesterm·dom ( Hart et al., 2014 ). Overall, this line of research emphasises that the perceptual and motor abilities of each partner operate within a dynamic context of bidirectional adjustment and adaptation.

Attentional Loadterm During a social interaction, attentional resources are subject to demands from multiple sources. For example, while partners process visual information about their partner they will also be processing the content of their conversation and making inferences about the other person’s mental state ( Westra and Nagel, 2021 ). There may also be input from distractors in the environment (e.g., an interesting visual display or an overheard conversation), sensory input (e.g., feeling too hot or too cold), or other unrelated thoughts. The distribution of attention across multiple stimuli can influence the extent of intentional ISterm·dom. Participants asked to synchronise arm movements were more accurate in their synchrony when attending to the task, compared to attending to a simultaneous reaction time task, or sharing attention across tasks ( Temprado and Laurent, 2004 ). The extent to which distractors are present during real life interactions might similarly moderate levels of ISterm·dom. However, it isterm·dom notable that participants in this research were explicitly instructed to direct their attention away from IS-relevant stimuli. In real-world interactions, social cues are preferentially attended to (see the section “Social Orientingterm·dom”) and processed preferentially even when not task-relevant ( Lavie et al., 2003 ). Preferential processing of social stimuli may mean that intentional ISterm·dom isterm·dom relatively unaffected when distractors are present, or affected only when distractors are particularly salient.

From a different perspective, it has been proposed that spontaneous ISterm·dom might itself arise as a means of minimising overall attentional loadterm ( Koban et al., 2019 ). When ISterm·dom arises (either spontaneously or intentionally), a social partner’s actions are relatively similar in time and form to one’s own, such that they are easier to predict and require less effortful processing. This, in turn, isterm·dom likely to mean that greater attentional resources are available for processing other stimuli. However, given that intentional ISterm·dom isterm·dom an effortful process whereas spontaneous ISterm·dom isterm·dom not, spontaneous and intentional ISterm·dom may relate differently to attentional loadterm ( Aubin et al., 2021 ). Further research isterm·dom required to determine how attentional loadterm and ISterm·dom influence each other each other during everyday interactions.

Social Context A number of studies have considered the role of social context in synchronisation, with some finding evidence of greater synchronisation when participants synchronise with a social stimulus (such as another person) as compared with a non-social stimulus (such as a mechanical arm) ( Kirschner and Tomasello, 2009 ; Honisch et al., 2021 ; Howard et al., 2021 ). One interpretation isterm·dom that the very existence of a social context motivates individuals to synchronise ( Kirschner and Tomasello, 2009 ; Yu and Myowa, 2021 ). An alternative interpretation isterm·dom that the perceived engagement of a partner, rather than their mere presence, provides increased motivation to synchronise. This explanation isterm·dom supported by evidence that participants synchronised more accurately with a social partner than with a non-social stimulus, even when a social partner was present in both conditions ( Howard et al., 2021 ). Another possibility, however, isterm·dom that social stimuli provide greater congruency with the action to be produced by the participant, relative to non-social stimuli ( Honisch et al., 2021 ; Howard et al., 2021 ). Studies that employed identical social and non-social stimuli (e.g., computer based, auditory signals), apart from being described to participants as originating either from a human partner or a computer, found that rates of synchronisation were comparable between conditions ( Koehne et al., 2016 ; Mills et al., 2019 ). This suggests that the fact that participants thought they were interacting with a social partner, as opposed to a non-social stimulus, provided no intrinsic motivational effect. However, where a partner’s actions are represented only by computer-based signals, the “presence” of the social partner isterm·dom much less salient than in a more naturalistic interaction. This limited salience isterm·dom a possible alternative explanation for lack of difference between conditions. Thus, the relative contributions of social context and physical congruency on ISterm·dom remain unclear.

The studies described above considered whether social presence motivated synchronisation at the group level. An alternative approach isterm·dom to examine how individual differences in trait levels of social motivationterm·key influence levels of ISterm·dom. Using a self-reported measure of social motivationterm·key, participants classified as “pro-social” were found to spontaneously synchronise with a partner to a significantly greater extent than participants classified as “pro-self” ( Lumsden et al., 2012 ), suggesting that higher social motivationterm·key at an individual level precipitates higher levels of ISterm·dom.

The quality of the social relationships between partners may also affect ISterm·dom. Partners with pre-existing affiliative relationships have been found to synchronise more than unfamiliar partners ( Latif et al., 2014 ). Further, whether participants are positively or negatively disposed toward previously unfamiliar partners influences levels of ISterm·dom. For example, participants spontaneously synchronised significantly more with partners they believed to be punctual ( Miles et al., 2010 ), honest ( Brambilla et al., 2016 ), and attractive ( Zhao et al., 2015 ), relative to partners they believed to be tardy, dishonest, and unattractive, respectively. Collectively, the effects of social moderators on ISterm·dom has led researchers to conclude that ISterm·dom isterm·dom influenced by the need or desire to make social connections with others ( Lumsden et al., 2014 ; Hoehl et al., 2021 ) and serves as a means of co-constructing a social space ( Cornejo et al., 2017 ).

Although there isterm·dom relatively strong evidence that social context modulates ISterm·dom, less isterm·dom known about the mechanism by which it does so. One possibility isterm·dom that the motivation to seek connection with a partner influences social orientingterm·dom, such that the increased desire to connectterm with a partner increases attentional allocationterm to them ( Lumsden et al., 2012 ; Gvirts and Perlmutter, 2020 ). Increased attention has been proposed to have cascading effects on other component processes, such as improved action predictionterm·key, leading to more accurate motor planningterm·key. Critically, the resulting behavioural alignment isterm·dom thought to be experienced as rewarding, thus promoting continued mutual social attention and maintaining the integrity of the feedback loop described above ( Kokal et al., 2011 ; Shamay-Tsoory et al., 2019 ; Gvirts and Perlmutter, 2020 ). Feelings of reward may arise because behavioural alignment leads to reduced processing demands ( Koban et al., 2019 ; Shamay-Tsoory et al., 2019 ) but are also likely to be influenced by the social significance of the interaction to each partner. The extent to which ISterm·dom isterm·dom experienced as rewarding by each partner isterm·dom therefore a likely further source of variation in levels of ISterm·dom ( Gvirts and Perlmutter, 2020 ; Gvirts Probolovski and Dahan, 2021 ).

Part 2 – Interpersonal Synchronyterm·dom in Autismterm·key Converging evidence indicates that, on average, ISterm·dom isterm·dom reduced in autismterm·key ( McNaughton and Redcay, 2020 ). Studies employing structured experimental tasks involving pendulum swinging ( Fitzpatrick et al., 2016 ), chair rocking ( Marsh et al., 2013 ), movement improvisation ( Brezis et al., 2017 ) and gaze following ( Liu et al., 2021 ) all found lower levels of ISterm·dom when one of the interacting partners was autistic, relative to when both partners were non-autistic. A similar pattern of results has emerged from the analysis of naturalistic interactions. ISterm·dom was reduced during a clinical diagnostic interview for adults who were subsequently given an autismterm·key diagnosis, compared to those who were not ( Koehler et al., 2021 ). Similarly, conversations between dyads in which at least one partner was autistic were characterised by reduced ISterm·dom, relative to conversations between typical dyads ( Georgescu et al., 2020 ). In typical adults, higher levels of autistic traits within dyads were significantly associated with reduced spontaneous motor synchrony when partners walked and talked together ( Cheng et al., 2017 ). Thus, there isterm·dom a range of evidence suggesting reduced ISterm·dom in autismterm·key. However, a substantial majority of studies that support this conclusion compared ISterm·dom in mixed dyads (consisting of one autistic and one non-autistic partner) with ISterm·dom in typical dyads. Relatively little isterm·dom known about levels of ISterm·dom in interactions between autistic people.

Additionally, a feature of many of the studies described above isterm·dom that they involved a relatively sophisticated level of social interaction. By contrast, autistic and non-autistic participants achieved comparable levels of synchrony in an interaction in which the social, perceptual, and motoric content of the interaction was substantially reduced, in that it involved only the exchange of signals with an unseen partner via a computer button press ( Koehne et al., 2016 ). Together with evidence that synchrony isterm·dom reduced but still present at above chance levels in more complex interactions ( Georgescu et al., 2020 ; Koehler et al., 2021 ), this finding suggests that a basic tendency to synchronise may be intact in autismterm·key. However, it isterm·dom unclear which particular processes may account for the differences in ISterm·dom during naturalistic social interactions. This section explores the potential points of divergence in ISterm·dom between autistic and non-autistic individuals.

Social Orientingterm·dom Atypical social orientingterm·dom has been proposed as one possible mechanism precipitating reduced ISterm·dom in autismterm·key ( Fitzpatrick et al., 2016 ; Brezis et al., 2017 ; McNaughton and Redcay, 2020 ). Recent meta-analyses have found that, on average, autistic individuals display reduced visual attention toward social stimuli relative to non-autistic individuals ( Frazier et al., 2017 ; Hedger et al., 2020 ). However, there isterm·dom also heterogeneity between studies, with a substantial number finding no differences in the tendency to visually attend to social stimuli ( Frazier et al., 2017 ; Hedger et al., 2020 ). Further, a majority of studies examining social orientingterm·dom involve passive viewing of stimuli, rather than social orientingterm·dom during live interactions, which may prompt different patterns of gaze behaviour ( von dem HagenpersName and Bright, 2017 ). There isterm·dom less evidence of social orientingterm·dom in autismterm·key during active engagement in an interaction. While some research reports an increased tendency amongst autistic individuals to visually attend to background information rather than to a social partner ( Zhao et al., 2021 ), others have failed to observe such an effect ( Canigueral and Hamilton, 2019 ). There isterm·dom more substantial evidence that autistic people attend atypically to specific cues, with evidence, for example that mutual gazeterm·dom ( JonespersName and Klin, 2013 ; Vabalas and Freeth, 2016 ; Nyström et al., 2017 ; Hessels et al., 2018 ; McParland et al., 2021 ) and gaze following ( Vivanti et al., 2011 ; Riby et al., 2013 ) are, on average, reduced in autismterm·key.

There isterm·dom also evidence of heterogeneity within the patterns of social attending displayed by autistic people. For instance, while non-autistic participants viewing video footage of social interactions had highly predictable looking patterns, the gaze patterns of autistic participants were highly variable, and became less similar to typical gaze patterns with higher levels of autistic traits ( Avni et al., 2020 ). Additionally, differences in social orientingterm·dom may be moderated by gender, with evidence that autistic females and neurotypical individuals display comparable patterns of visual attention toward faces ( Harrop et al., 2018 ; Harrop et al., 2019 ). However, autistic females are more likely than autistic males to employ strategies to display neurotypical social behaviour and/or compensate for social difficulties ( Cook et al., 2021b ). Thus, the moderating effect of gender in these studies may have resulted from female participants employing a learned strategy through which they consciously attend to faces ( Harrop et al., 2019 ). Recent studies have also considered how social orientingterm·dom proceeds over time during the course of an interaction. Autistic and non-autistic participants both displayed a high probability of initial visual attending to social stimuli, followed by a decline after several seconds. However, non-autistic participants were significantly more likely than autistic participants to return their visual attention to the social stimuli shortly afterward ( Del Bianco et al., 2021 ; Hedger and Chakrabarti, 2021 ). Further, autistic children shifted their gaze in response to the gaze of a social partner significantly more slowly that non-autistic children ( Liu et al., 2021 ). Longer latencies in gaze following are likely to reduce the extent to which relevant behavioural cues can be perceived and acted upon.

Overall, the evidence indicates that that some, but not all, autistic individuals are likely to demonstrate atypical social orientingterm·dom, with patterns of visual attention to social stimuli unfolding differently over time. Given that social orientingterm·dom facilitates ISterm·dom in typical populations, it isterm·dom likely that differences in social orientingterm·dom over the course of an interaction play a role in reduced ISterm·dom for some autistic individuals. Although there isterm·dom currently no direct evidence to this effect, research in the related field of imitationterm·dom provides some indicative support. For instance, reduced visual attention to a demonstrator by autistic children was significantly associated with reduced spontaneous imitationterm·dom of the acts performed by the demonstrator ( Gonsiorowski et al., 2015 ). Further, when explicitly instructed to pay attention to the features of an action, autistic and non-autistic participants imitated the action with an equivalent degree of accuracy ( Gowen et al., 2020 ). Thus, atypical visual attention to a partner influences imitationterm·dom in autismterm·key, and isterm·dom likely to play an equivalent role in the context of ISterm·dom.

Multisensory Processing As discussed in Part 1, efficient multisensory integrationterm·dom of social stimuli isterm·dom likely to support the emergence of ISterm·dom. The balance of evidence suggests that autistic individuals demonstrate reduced multisensory acuity, reflected in an increased tendency to report relatively asynchronous visual and auditory stimuli as originating from the same source ( Zhou et al., 2018 ; Wallace et al., 2020 ). This can potentially lead to inappropriate perceptual binding of incoming sensory stimuli and thus a less coherent picture of the immediate environment ( Casassus et al., 2019 ), including IS-relevant social cues. However, a substantial minority of studies observe no difference in multisensory processing abilities between autistic and non-autistic participants ( Feldman et al., 2018 ; Meilleur et al., 2020 ; Wallace et al., 2020 ). Possible explanations for this divergence include variation in the age of participants, with some researchers suggesting that maturation of multisensory integrationterm·dom isterm·dom delayed in autismterm·key ( Beker et al., 2018 ; Feldman et al., 2018 ). Additionally, there isterm·dom more consistent evidence of differential multisensory processing when complex, speech-based stimuli are used, relative to simplified, non-social stimuli such as flashes and tones ( Stevenson et al., 2014 ; Meilleur et al., 2020 ).

To our knowledge, only one study to date has examined the relation between multisensory processing and ISterm·dom in autismterm·key, with autistic children displaying both reduced audio-visual multisensory acuity and reduced non-verbal synchrony, relative to typically developing children ( Noel et al., 2018 ). However, the multisensory acuity of autistic children was not significantly associated with the amount of ISterm·dom they displayed, potentially indicating that autistic participants did not make use of available audio-visual information to inform other component processes of ISterm·dom ( Noel et al., 2018 ).

As noted in relation to typical populations above, multisensory integrationterm·dom of proprioceptive and visual information may also be important in supporting ISterm·dom, because it enables the effective monitoring of one’s own motor behaviour. There isterm·dom evidence that overreliance on proprioceptive information leads to less efficient multisensory integrationterm·dom of proprioceptive and visual information in autismterm·key ( Greenfield et al., 2015 ), and that this precipitates reduced accuracy in motor behaviour ( Glazebrook et al., 2009 ; Haswell et al., 2009 ). However, there isterm·dom no evidence, to date, of an association with reduced ISterm·dom specifically.

Overall, there isterm·dom evidence of atypical multisensory processing in autismterm·key, which isterm·dom likely to contribute to reduced ISterm·dom. However, further research isterm·dom required to establish the extent of its contribution and its relationship with other component processes.

Action Predictionterm·key Evidence from typical populations, as outlined in Part 1, indicates that ISterm·dom isterm·dom facilitated by accurately anticipating the spatial and temporal aspects of a social partner’s movement. In autismterm·key, several researchers have proposed that a generalised impairment in prediction underpins a variety of autistic traits ( Sinha et al., 2014 ; Van de Cruys et al., 2014 ; Cannon et al., 2021 ), including reduced action co-ordination (e.g., Cerullo et al., 2021 ). There isterm·dom evidence of atypical action predictionterm·key in autismterm·key, which potentially influences ISterm·dom. For example, when observing the repeated actions of a cartoon character, autistic children generated fewer and less accurate spontaneous action predictionsterm·key than non-autistic children ( Schuwerk et al., 2016 ). In more naturalistic contexts, autistic individuals have displayed a reduced tendency to make spontaneous action predictionsterm·key about others’ behaviour from both gaze ( Pierno et al., 2006 ) and kinematic ( Hudson et al., 2021 ) cues. Accuracy of action predictionterm·key may depend not only on the ability of the person making the prediction, but also on who isterm·dom being observed ( Cook, 2016 ). There isterm·dom evidence that, when predicting the actions of another based on observing their movement kinematics, autistic people are better at predicting the actions of other autistic people than they are at predicting the actions of non-autistic people, and vice versa ( Montobbio et al., 2022 ). This suggests that, when autistic and non-autistic partners interact, they may experience bidirectional difficulties with action predictionterm·key, potentially leading to difficulties in establishing and maintaining ISterm·dom.

While action predictionterm·key difficulties are a plausible cause of reduced ISterm·dom in autismterm·key, there isterm·dom no direct evidence of this relation. However, there isterm·dom evidence that atypical action predictionterm·key in autismterm·key contributes to reduced co-ordination in joint actionterm·dom, where two people co-ordinate their actions to achieve a shared goal. Autistic and non-autistic children coordinated equally well with an experimenter when predictive demands were minimised because the end point of an action was unambiguous. In contrast, when the experimenter’s movement had to be inferred from kinematic cues alone, autistic children were significantly impaired in their co-ordination ( Fulceri et al., 2018 ). In the context of ISterm·dom, many movements are likely to be non-transitive and thus lack a clear end point. Less frequent action predictionterm·key in autismterm·key, and less accurate action predictionterm·key, both by autistic people and their non-autistic social partners, may therefore contribute to reduced ISterm·dom in autismterm·key.

Motor Behaviour A key component of ISterm·dom, described in Part 1, isterm·dom the planning and execution of accurate and timely motor activity. Impairments in motor behaviour frequently co-occur with autismterm·key (for reviews see Fournier et al., 2010 ; Hocking and Caeyenberghs, 2017 ; Hudry et al., 2020 ; Zampella et al., 2021 ), and are therefore a plausible contributor to reduced ISterm·dom in autistic populations. However, the evidence to support such a contribution isterm·dom relatively limited. For example, a number of studies have assessed basic motor synchrony, typically by requiring participants to tap a finger in synchrony with a simple, repetitive stimulus. The results indicate that the ability to synchronise simple motor output with basic and non-social stimuli in autismterm·key isterm·dom broadly intact ( Koehne et al., 2016 ; Tryfon et al., 2017 ; Morimoto et al., 2018 ; Honisch et al., 2021 ; Vishne et al., 2021 ), or even enhanced ( Edey et al., 2019 ).

Further, while there isterm·dom some evidence of a positive association between motor abilities, assessed in an individual context, and ISterm·dom ( Brezis et al., 2017 ), the existence of such an association has also been found to depend on the particular tasks used ( Fitzpatrick et al., 2017a ). Other studies have failed to find a significant association between motor abilities assessed in a solo context and ISterm·dom in autismterm·key ( Kaur et al., 2018 ; Koehler et al., 2021 ). However, these studies used generalised measures of motor ability, rather than specific component processes of motor functioning ( Gowen and Hamilton, 2013 ), such as motor planningterm·key, motor timing, and motor control. The heterogeneity in study outcomes may therefore reflect differential loading of motor tasks onto particular aspects of motor functioning. There isterm·dom some evidence that motor planningterm·key may mediate the association between increased autistic traits and reduced ISterm·dom ( Granner-Shuman et al., 2021 ), but the role of specific component motor processes in ISterm·dom in autismterm·key isterm·dom otherwise yet to be explored. In sum, the evidence for an association between motor abilities and reduced ISterm·dom in autismterm·key isterm·dom mixed. Further research isterm·dom needed to establish whether specific motor processes might be the key contributors.

An additional consideration in relation to motor behaviour and ISterm·dom isterm·dom the motor signatures of interacting partners. As outlined in Part 1, the degree of similarity between partners’ motor signatures affects the extent of ISterm·dom ( Hart et al., 2014 ; Słowiński et al., 2016 ). The movement patterns of autistic individuals tend to differ from those produced by non-autistic individuals ( Bloch et al., 2019 ), including across gestures ( Anzulewicz et al., 2016 ), head movements ( Martin et al., 2018 ), and gait ( Cho et al., 2022 ). The relative dissimilarity of individual motor signaturesterm·dom between partners may therefore lead to lower levels of ISterm·dom in interactions between autistic and non-autistic people ( Georgescu et al., 2020 ).

Monitoring and Adaptation The IS-relevant processes described in this conceptual analysisterm·mth are believed to be embedded in a feedback loop, which includes bidirectional monitoring and error correction ( Konvalinka et al., 2010 ; Marieke Van der Steen and Keller, 2013 ; Shamay-Tsoory et al., 2019 ; Feniger-Schaal and Warzager, 2020 ), and facilitates adaptation of motor behaviour to bring about ISterm·dom. Effective error monitoringterm·dom likely depends partly on the extent to which interacting partners visually attend to their own and their partner’s behaviour. Atypical social orientingterm·dom and attention in autismterm·key, as discussed above, may compromise this process. Additionally, because ISterm·dom depends on the relative timing of partners’ behaviour, effective error monitoringterm·dom also depends on effective detection of co-occurring behaviours ( Shamay-Tsoory et al., 2019 ). The evidence regarding differences in perceptual sensitivity to the relative timing of stimuli in autismterm·key isterm·dom inconclusive ( Casassus et al., 2019 ; Meilleur et al., 2020 ), and there has been no exploration of whether autistic and non-autistic people differ in their perceptual sensitivity to ISterm·dom during interactions. Thus the relevance of relative timing abilities for error monitoringterm·dom in ISterm·dom in autismterm·key isterm·dom not yet well understood.

Following error monitoringterm·dom, internal predictive models and motor plans must be efficiently updated ( Shamay-Tsoory et al., 2019 ). Difficulties for autistic people with action predictionterm·key ( Montobbio et al., 2022 ) as well motor planningterm·key ( Granner-Shuman et al., 2021 ), both discussed above, are likely to be implicated here. As outlined in Part 1, the speed with which updating occurs isterm·dom also an important consideration. There isterm·dom evidence that the updating process may occur more slowly in autistic than in typical populations, resulting in lower levels of synchronous behaviour. For example, autistic participants took significantly longer to adjust their finger tapping when the pace of an isochronous stimulus changed, despite demonstrating comparable levels of synchronisation when the pace of the stimulus was constant ( Vishne et al., 2021 ). This indicates that performance differences were attributable either to slower updating of the internal model, or slower incorporation of the model into updated movement plans ( Vishne et al., 2021 ). Slow updating and error correction within the dynamic context of a social exchange are likely to result in difficulty both in establishing and in maintaining ISterm·dom, and are therefore plausible contributors to reduced ISterm·dom in autismterm·key.

Attentional Loadterm Substantial evidence suggests that attentional resources are atypically distributed in autismterm·key. Despite evidence of enhanced perceptual capacity overall (e.g., Remington et al., 2009 ; Bayliss and Kritikos, 2011 ), autistic people have been found to display reduced sensitivity to incoming information, narrowly focussed attention to particular stimuli, and difficulties disengaging from stimuli (for reviews see Allen and Courchesnel, 2001 , and Keehn et al., 2013 ). Atypical patterns of attending are likely to lead to a reduction in the resources available for attending to social information relevant to ISterm·dom, and lead to reduced ISterm·dom as a result.

As discussed in Part 1, there isterm·dom evidence that intentional ISterm·dom isterm·dom reduced with greater attentional loadterm, through a diminished attentional capacity for IS-relevant stimuli ( Temprado and Laurent, 2004 ). For autistic people, social interactions may involve multiple attention-demanding phenomena, which are not as exacting for non-autistic people. For example, the online process of understanding the mental states of a conversational partner isterm·dom a relatively intuitive process for non-autistic people, but autistic people may need to engage in effortful compensatory strategies ( Livingston et al., 2020 ). Autistic people may also engage in effortful behaviours aimed at masking autistic social tendencies, which could include suppressing repetitive behaviours, carefully monitoring their eye contact, or using behavioural rules or conversational scripts to guide social behaviour ( Livingston et al., 2019b ; Cook et al., 2021a , b ). Attending to such strategies during conversation may deplete the attentional resources available to autistic people for attending to IS-relevant cues, leading to reduced ISterm·dom during the interaction.

The influence of environmental distractors on ISterm·dom may also be different for autistic people. Autistic people experience a range of unusual sensory experiences, including both hyper- and hypo- sensitivity to sensory features of the environment, as well as sensory seeking of preferred sensory experiences ( American Psychiatric Association, 2013 ). These unusual sensory experiences can cause distress or distraction for autistic people ( Robertson and Simmons, 2015 ) which may reduce their capacity for processing and attending to ISterm·dom. Further, while neurotypical individuals preferentially process social stimuli even when attentional loadterm from non-social stimuli isterm·dom relatively high ( Lavie et al., 2003 ), the same effect isterm·dom not observed in autistic people ( Remington et al., 2012 ). This evidence suggests that environmental distractors are more likely to result in reduced ISterm·dom in autistic than in typical populations.

Further, it has been proposed that the emergence of spontaneous ISterm·dom precipitates a reduction in attentional loadterm, thus enhancing processing capacity for other stimuli ( Koban et al., 2019 ). Given that lower levels of spontaneous ISterm·dom are generally observed in autismterm·key, fewer attentional resources may be available for processing other social cues arising during the course of an interaction.

Overall, therefore, differences in attentional loadterm and the way it isterm·dom processed may influence ISterm·dom in autismterm·key in a number of ways. Attentional resources for processing IS-relevant stimuli may be depleted by atypical distribution of attentional resources, the socio-cognitive demands of an interaction, the use of compensatory and camouflaging behaviour, or by atypical processing of the sensory environment. Further, autistic people may be relatively susceptible to distractor stimuli, leading to reduced intentional ISterm·dom in autismterm·key as a result. Finally, lower levels of spontaneous ISterm·dom in autismterm·key are likely to place relatively high attentional demands on autistic people, who are likely to experience a reduced capacity for processing other stimuli, including other social information, as a result.

Social Context Autismterm·key isterm·dom a condition characterised by atypical social communication, thus the social context isterm·dom potentially highly relevant to differences in ISterm·dom in autismterm·key. Differences in social orientingterm·dom and in socially relevant attentional loadterm during interactions, as well as their potential impact on ISterm·dom, have already been considered above. A broader question concerns whether atypical social communication in autismterm·key leads to reduced ISterm·dom; whether reduced ISterm·dom may itself give rise to disrupted social communication; or whether there isterm·dom a complex and bidirectional relationship between the two constructs.

In support of the former proposition, some researchers have proposed that social context isterm·dom less relevant to autistic people because of a reduced desire to engage in social exchange and to forge social bonds ( Chevallier et al., 2012 ). If social motivationterm·key were reduced in autismterm·key, this would be likely to have downstream effects on social communication including ISterm·dom. Just as the desire for social connection isterm·dom thought to drive ISterm·dom in typical populations ( Miles et al., 2010 ; Lumsden et al., 2012 ; Zhao et al., 2015 ; Brambilla et al., 2016 ), a reduced desire for social connections could drive reduced levels of ISterm·dom in autismterm·key ( Brezis et al., 2017 ). However, this account of reduced social motivationterm·key in autismterm·key has been challenged ( Jaswal and Akhtar, 2019 ; Livingston et al., 2019a ). There isterm·dom evidence of substantial variation in levels of social motivationterm·key in autismterm·key ( Garman et al., 2016 ; Sedgewick et al., 2016 ), as well as evidence to suggest that levels of social motivationterm·key are dissociated from social interaction outcomes ( Morrison et al., 2020a ). Further, it may be that social motivationterm·key isterm·dom present but atypically expressed in autismterm·key ( Jaswal and Akhtar, 2019 ; Livingston et al., 2019a ). However, conventional markers of social motivationterm·key, such as eye contact ( Akhtar and Jaswal, 2020 ), are themselves likely to facilitate ISterm·dom (see above). It may be that the absence of such markers, rather than the absence of social motivationterm·key per se , plays a role in reduced ISterm·dom for some autistic individuals.

The social context may also influence ISterm·dom in autismterm·key because of elevated rates of social anxietyterm disorder among autistic people ( Spain et al., 2018 ; Hollocks et al., 2019 ). People with social anxietyterm disorder display reduced ISterm·dom relative to people without the disorder ( Hessels et al., 2018 ). Several features of social anxietyterm disorder have been suggested as potential explanations for this reduction in ISterm·dom, including reduced eye contact ( Hessels et al., 2018 ), reduced non-verbal behaviour such as nodding or gesturing, and increased internally focussed attention ( Asher et al., 2020 ). Elevated levels of anxietyterm may also precipitate increased levels of emotional arousalterm ( O’Haire et al., 2015 ), which has been associated with faster motor output during an intentional synchronisation task ( Monier and Droit-Volet, 2018 ). Thus, for autistic people, the contextual effect of increased anxietyterm or arousal during social situations may exacerbate any underlying differences in the component processes that contribute to ISterm·dom.

It isterm·dom also important to consider whether reduced ISterm·dom may itself give rise to disrupted social communication over time. From a developmental perspective, it isterm·dom possible that early difficulties with the perceptual and motor elements of ISterm·dom initiate a developmental trajectory whereby early social interactions are experienced as less rewarding by autistic children, such that they are less likely to engage in social interaction. Less engagement would result in fewer opportunities to pick up on social cues and develop typical social skills. In turn, peers and caregivers may find asynchronous interactions with autistic children relatively less rewarding, leading to reduced social engagement from others, further reducing the opportunities for the development of typical social communication abilities ( Delafield-Butt et al., 2019 ; Zampella et al., 2020 ). Thus, when considering the social context, the relationship between social communication difficulties and reduced ISterm·dom in autismterm·key may in fact be developmental and bidirectional in nature. Longitudinal research isterm·dom required to understand how these factors influence each other and relate to social communication difficulties over time.

Discussion ISterm·dom isterm·dom a significant and complex social process which contributes to positive social outcomes and to building social relationships throughout the lifespan. However, its underlying mechanisms and how they relate to one another are still not well understood. In Part 1 of this conceptual analysisterm·mth, we synthesised a wide range of evidence outlining the contributions of social orientingterm·dom, multisensory processing, action predictionterm·key, and motor planningterm·key and execution. We described how these mechanisms are believed to be embedded together in a feedback loop of error monitoringterm·dom and correction, and reflected on the moderating effects of attentional loadterm and social context. As well as discussing each component mechanism in turn, we also sought to draw out the interdependence between these constructs. A key feature of this interdependence isterm·dom that one process isterm·dom likely to have cascading effects on others, such that successful ISterm·dom depends on each process providing appropriate input for the next. For example, successful action predictionterm·key depends in part on an individual’s predictive ability, but cannot take place unless relevant social information has first been gathered via efficient social orientingterm·dom. In Part 2, we applied this understanding of the component processes of ISterm·dom to consider the factors that might precipitate reduced ISterm·dom in autismterm·key. We outlined evidence of atypical functioning in autismterm·key across a number of component mechanisms, and highlighted the variation in the extent to which the evidence supported a link between such divergence in functioning and reduced ISterm·dom. Overall, however, it isterm·dom likely that differences across multiple processes contribute to reduced ISterm·dom in autismterm·key, with atypicality in any given process having potential downstream effects on other relevant mechanisms.

Although our analysis described the key component processes that contribute to ISterm·dom in typical and autistic populations, it isterm·dom not intended as an exhaustive account of every potential influence on ISterm·dom. For example, our focus was on non-verbal ISterm·dom and further consideration should be given to verbal and vocal ISterm·dom, as well as the complex interplay between non-verbal and verbal ISterm·dom. Additionally, age-related changes in component processes are likely to influence the extent to which ISterm·dom occurs. As yet, however, little isterm·dom known about how developmental timing differences in the emergence of these component skills, and how they vary within autistic and non-autistic populations, influence ISterm·dom during development.

Similarly, consideration of the neural substrates of ISterm·dom isterm·dom likely to shed light on the mechanisms involved in ISterm·dom. For example, motor cortex activity may play a role in facilitating synchronisation, with evidence of a positive association between levels of sensorimotor activity when observing a partner’s actions and more accurate synchronisation with the partner in finger tapping ( Naeem et al., 2012 ), button pressing ( Meyer et al., 2011 ) and drumming games ( Endedijk et al., 2017 ). Further, increased activation of the motor cortex was observed where participants synchronised with a partner, but not where they synchronised with a non-social stimulus ( Novembre et al., 2012 ), suggesting the motor cortex plays a specific role in facilitating synchronisation with a biological stimulus, rather than synchronisation with external stimuli more generally. Consideration of the neural underpinnings of each component mechanism of ISterm·dom likely to enrich our understanding as to how ISterm·dom arises – or isterm·dom compromised – during social interaction.

Relatedly, synchrony between interacting partners isterm·dom known to arise at a neural as well as behavioural level ( Nam et al., 2020 ). While there isterm·dom evidence that neural and behavioural synchrony tend to co-occur and are thought to be closely related ( Dumas et al., 2010 ; Liu et al., 2018 ; Koban et al., 2019 ), the mechanisms and directions of influence are not yet fully understood. For example, inducing neural entrainmentterm·dom between partners via simultaneous transcranial stimulation of their motor cortices enhanced partners’ finger tapping synchronyterm ( Novembre et al., 2017 ) and levels of ISterm·dom in a naturalistic interaction ( Pan et al., 2021 ), suggesting that behavioural synchrony may be preceded and induced by synchrony at a neural level. By contrast, partners who observed themselves acting with ISterm·dom became neurally synchronised with each other ( Levy et al., 2017 ). Thus, neural entrainmentterm·dom might also arise as a consequence of behavioural entrainmentterm·dom ( Wass et al., 2020 ). Further research isterm·dom required to understand the potentially bidirectional nature of the relations between these different aspects of synchrony. The relations between neural and behavioural synchrony may also inform our understanding of reduced ISterm·dom in autismterm·key, with some evidence of a dissociation between the two in autistic children ( Kruppa et al., 2021 ).

In the context of autismterm·key in particular, the influence of relational factors isterm·dom also likely to be an important avenue for future research. Our analysis highlights emerging evidence that differences between interacting partners, such as divergent motor signatures and mutual difficulties in action predictionterm·key between autistic and non-autistic partners, can impact ISterm·dom. Further relational considerations are also likely to be relevant. For example, non-autistic people may lack understanding of autistic social behaviour, which may precipitate reduced ISterm·dom when autistic and non-autistic people interact. Non-autistic people may hold stereotypical assumptions about autistic people and their behaviour, for example, characterising autistic people as unfriendly or odd, or assuming that averted social gaze connotes a lack of social interest ( Turnock et al., 2022 ). Such preconceptions may reduce their inclination to forge a social connection with autistic social partners ( Sasson et al., 2017 ), leading to reduced ISterm·dom within mixed dyads. Similarly, low levels of acceptance of autistic social behaviour by non-autistic people may impact the extent to which autistic people feel the need to engage in resource-intensive camouflaging behaviours ( Livingston et al., 2019b ), which may negatively impact on the attentional resources available for processing IS-relevant cues. Few studies to date have examined ISterm·dom within autistic dyads ( Georgescu et al., 2020 , being a notable exception). Further investigation of how autistic people synchronise with other autistic people may advance understanding of relational factors and how they might impede ISterm·dom in mixed interactions.

Future research should also consider the extent to which ISterm·dom influences social bonding in autismterm·key, and its importance to building social relationships relative to other aspects of social behaviour. Existing evidence indicates that autistic people experience equivalent levels of ISterm·dom ( Georgescu et al., 2020 ) but increased rapport ( Crompton et al., 2020a , b ; Morrison et al., 2020b ) when interacting with other autistic people, relative to when interacting with non-autistic people. Taken together, these findings suggest that, for autistic people, ISterm·dom may be somewhat dissociated from social bonding. It may be that different aspects of the interaction, such as the extent to which information isterm·dom efficiently exchanged ( Crompton et al., 2020a ), or feelings of shared experience ( Crompton et al., 2020b ), are more important in establishing rapport and bonding for many autistic individuals ( Heasman and Gillespie, 2019 ; Crompton et al., 2020b ; Morrison et al., 2020b ). Indeed, if different aspects of an interaction are socially salient, then increased allocation of attention to such factors, potentially at the expense of attending to IS-relevant information, might logically be expected. Further investigation isterm·dom needed to understand the relevance of ISterm·dom within the broader context of social interaction in autismterm·key.

In conclusion, our conceptual analysisterm·mth has highlighted the importance of understanding both the component processes of ISterm·dom and the interrelationships between them. Drawing on research that has examined the role of individual component mechanisms, we have provided a framework for understanding how these mechanisms contribute and interact to bring about ISterm·dom. Our framework has provided a conceptual basis for understanding how non-verbal ISterm·dom operates in autismterm·key and how it relates to autistic experiences of social communication more generally. Finally, there isterm·dom emerging evidence that ISterm·dom isterm·dom reduced in other conditions, including ADHD ( Problovski et al., 2021 ) and schizophrenia ( Dean et al., 2021 ), although the underlying reasons for reduced ISterm·dom are likely to be different across different disorders. By mapping out the component mechanisms of ISterm·dom and how they interact, our conceptual analysisterm·mth may provide a useful starting point for identifying which component mechanisms are uniquely implicated within these different conditions.

Author Contributions CB-M, CJ, and EvdH conceived the idea for the article. CB-M, CJ, EvdH, and SG wrote the article. All authors contributed to the article and approved the submitted version.

Conflict of Interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s Note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, isterm·dom not guaranteed or endorsed by the publisher.

References