Investigating mother–child inter‐brain synchrony in a naturalistic paradigm: A functional near infrared spectroscopy (fNIRS) hyperscanning study

Successful social interactions between mothers and children are hypothesised to play a significant role in a child's social, cognitive and language development. Earlier research has confirmed, through structured experimental paradigms, that these interactions could be underpinned by coordinated neural activity. Nevertheless, the extent of neural synchrony during real‐life, ecologically valid interactions between mothers and their children remains largely unexplored.

Overall, we demonstrate the feasibility of measuring inter-brain synchrony between mothers and children in a naturalistic environment.These findings can inform future study designs to assess inter-brain synchrony between parents and pre-lingual children and/or children with communication needs.
K E Y W O R D S communication behaviours, cooperation, development, mother-child pair, neural synchrony, social interaction 1 | INTRODUCTION A child's social, emotional and cognitive development has been shown to be directly influenced by the quality of parent-child interactions in early childhood.Indeed, behavioural studies have shown a strong correlation between parent-child interactions and children's emotional regulation, language and social competence outcomes (Cartmill et al., 2013;Cooke et al., 2019;Hollenstein et al., 2017;Justice et al., 2019;Osterling et al., 2002;Romeo et al., 2018).For these reasons, parent-child interactions have been the target of numerous treatment plans for children, particularly toddlers (Jeong et al., 2021).A global systematic review and metaanalysis showed that interventions that targeted parentchild interactions improved the child's cognitive, language and motor development (Jeong et al., 2021).
Successful mother-child interactions are thought to rely on behavioural, physiological and neural synchrony (Davis et al., 2017;Feldman, 2007b;Harrist & Waugh, 2002;Leclère et al., 2014).In this context, synchrony describes the phenomenon where interacting partners mutually adapt their behaviour in real-time in response to one another (Delaherche et al., 2012).Behavioural synchrony refers to the coordination of verbal and non-verbal communication (e.g., eye gaze, posture, etc.) whereas physiological synchrony encompasses the coordination of biological rhythms such as heart rate and breathing patterns (Feldman, 2007a(Feldman, , 2007b;;Hoehl et al., 2020;Miles et al., 2009;Valdesolo & DeSteno, 2011).Neural synchrony is defined as the temporal alignment of concurrent brain activity between interacting partners (Dumas et al., 2010).Neural synchrony between pairs can be measured using hyperscanning, i.e., the simultaneous recording of brain activities of two or more individuals (Dumas et al., 2010).
A wealth of evidence exists already demonstrating the relationship between behavioural and physiological synchrony and parent-child interactions (Carollo et al., 2021).Currently, the research landscape is witnessing a continuous expansion in the number of studies investigating neural synchrony within parent-child interactions.This collective body of research has, as a primary finding, established the existence of neural synchrony in parent-child pairs and underscored its importance in the context of child development.Neural synchrony, in this context, has the potential to promote parent-child interactions, foster behavioural coordination and boost performance in collaborative tasks (Atzil & Gendron, 2017;Leong et al., 2017;Miller et al., 2019).Therefore, it is postulated that there is a bidirectional relationship in which interactions lead to increased neural synchrony, and increased neural synchrony, in turn, may support these interactions.Specifically, neural synchrony between parents and children appears to be supported by nonverbal behavioural reciprocity cues such as joint attention and mutual eye gaze, as well as the participants' current mental states and personality traits (Azhari et al., 2019;Azhari et al., 2020;Hasson et al., 2012;Nguyen et al., 2020;Reindl et al., 2018).Conversely, neural synchrony could potentially enhance mother-child interactions by enabling individuals to make more accurate predictions about their own and others' behaviour, thereby facilitating behavioural self-regulation (Dai et al., 2018).Furthermore, both behavioural and neural synchronies are believed to contribute to children's capacity to regulate their emotional states, thus allowing them to participate in reciprocal interactions (Atzil & Gendron, 2017).
To date, the majority of this research has been conducted using tasks that do not accurately reflect naturalistic social interactions between parents and children.Typically employed paradigms are either passive (book reading to a child; for instance, see (Piazza et al., 2020)) or goal-directed (solve a puzzle, play a computer game, etc.; e.g.(Liao et al., 2015)).Very few studies have examined neural synchrony between parents and children during ecologically valid interactions that represent real-life scenarios.Namely, Piazza et al. (2020) used a paradigm that included free play, book reading and rhythm signing.They showed that there was an alignment in neural synchrony between infants and unrelated adults which appeared to be supported by mutual eye gaze and changes in the adult's infant-directed speech patterns (Piazza et al., 2020).However, the authors did not discuss the specific structure of their task.Subsequently, it was unclear how long each element lasted for and what the duration of the free play was (Piazza et al., 2020).In another study with a naturalistic paradigm, Nguyen, Schleihauf, Kayhan, et al. (2021) demonstrated that neural synchrony increased as a function of the number of conversational turns taken during conversations between toddlers and their mothers.Investigations of neural synchrony during adversity and child irritability have utilised unstructured free play as a recovery condition (Hoyniak et al., 2021;Quiñones-Camacho et al., 2020).In both studies, a custom-made metric of behavioural interactions was created, measuring interactions by aggregating turn-taking, mutual attention and mutual eye gaze instances.Nevertheless, in neither study utilising this method to quantify behavioural synchrony during freeplay interactions did the authors identify a correlation with neural synchrony patterns.Recently, Norton et al. developed a "social EEG" paradigm where they were able to successfully measure neural synchrony between parent-child pairs using a naturalistic set-up where participants solved age-appropriate puzzles, read books and watched movies whilst joint engagement to common objects or activities was coded as a metric of a behavioural measure of the parent-child interaction (Norton et al., 2022).
These studies confirm the presence of neural synchrony during more natural interactions.However, the question remains whether this neural synchrony can manifest between parents and children during unstructured play, thereby contributing to our comprehension of the elements that underpin parent-child neural synchrony in real-life settings.Furthermore, such interpersonal exchanges hold substantial ecological relevance, especially in the context of investigating neural synchrony among preverbal and clinical populations with communication needs.These groups, marked by intrinsic communication difficulties, may have restricted attention spans and may exhibit avoidance behaviours, such as avoiding mutual eye contact (e.g.Osterling et al., 2002).Hence, we emphasise the importance of investigating patterns of neural synchrony within parent-child dyads characterised by typical development, allowing for maximal freedom in their interactions.In the future, it will be essential to contrast this research direction with analogous studies involving clinical populations.
The primary objective of this study is to assess the feasibility of measuring neural synchrony patterns between mothers and their children by employing a naturalistic free-play paradigm.Within this framework, participants were encouraged to engage in play similar to their interactions at home, using toys that did not impose any performance expectations, thus affording them the freedom to interact to the extent they desired.In addition, we introduced an independent condition where participants played with the same toy but separately, simulating a scenario in which the parent and child both attended to the same stimulus without direct interaction.This design aimed to verify that any observed neural synchrony was not merely a consequence of shared motor and attentional demands associated with the stimuli.Our hypothesis was that, in the naturalistic setting, neural synchrony between parent-child pairs would be higher during interactive play compared to the independent condition.
Neural synchrony was measured bilaterally over frontal and temporoparietal brain regions in line with previous hyperscanning studies involving verbal communication and problem-solving (Nguyen et al., 2020;Nguyen, Schleihauf, Kayhan, et al., 2021).Neural synchrony was measured using functional nearinfrared synchrony (fNIRS).FNIRS is a safe, non-invasive neuroimaging method employed for functional mapping of the human cortex that offers tangible advantages compared to other neuroimaging techniques particularly when conducting hyperscanning studies.Namely, fNIRS is resilient against motion artefacts and easily transportable, enabling experiments to be conducted in a range of different settings (Harrison & Hartley, 2019;Pinti et al., 2020;Quaresima & Ferrari, 2019).Especially in the context of hyperscanning studies, it can facilitate the assessment of neural activity simultaneously in two or more individuals during face-to-face experiments where participants have the flexibility to move relatively freely.This capability allows us to design paradigms that can replicate real-life, ecologically valid interactions.
As a secondary, exploratory objective in this study, we investigated what might be driving the hypothesised increased neural synchrony (or lack thereof) in the interactive condition we examined turn-taking patterns between the pairs and personality traits of both the mother and the child.Previous hyperscanning studies have shown that neural synchrony is positively correlated with turn-taking (Ahn et al., 2018;Dollar & Stifter, 2012;Nguyen, Schleihauf, Kayhan, et al., 2021;Pérez et al., 2019), we were interested in examining this behaviour, given that participants were free to talk with each other but, unlike previous studies, they were not required to do so.We hypothesised that higher turn-taking would be positively associated with patterns of neural synchrony.Additionally, both behavioural and neuroimaging studies have reported that personality traits influence the quality of behavioural and neural synchrony in mother-child dyads (Azhari et al., 2019;Reindl et al., 2018).Thus, we also studied the association between child temperament and neural synchrony, as well as maternal emotion regulation.We hypothesised that adaptive child temperament dimensions (surgency, effortful control) would be positively associated with patterns of neural synchrony during mother-child interactions.In turn, maladaptive dimensions (negative affect) would elicit the opposite outcome.Similarly, we postulated that maternal adaptive strategies of emotion regulation (cognitive reappraisal) would be correlated with higher levels of neural synchrony between mothers and children whereas maladaptive strategies (expressive suppression) would have the reverse effect.

| Participants
Twelve pairs of mothers (mean age = 35.42years, SD = 5.21) and children between the ages of 3 and 4;11 years old (7 female, mean age = 3.97 years, SD = 0.62) took part in this study.All participants were native English speakers with normal or corrected to normal vision and no known hearing, language or cognitive problems; children with a history of cognitive or motor impairment, as reported by the parents, were excluded.Participants were identified from the National Institute of Health Research Nottingham Biomedical Research Centre Hearing Sciences participant database, and via online advertisements on parent Facebook groups in the Nottinghamshire area.This investigation was approved by the University of Nottingham Faculty of Medicine and Health Sciences Research Ethics Committee.

| Behavioural assessments
To assess the mothers' and children's personality traits, mothers were asked to fill in two questionnaires prior to the hyperscanning session.Child temperament was quantified using the Very Short form of the Early Childhood Behaviour Questionnaire (VS-ECBQ) (Putnam, Ellis, & Rothbart, 2001;Putnam & Rothbart, 2006).The VS-ECBQ is comprised of 36 questions that relate to 3 child temperament dimensions: Surgency, Effortful Control and Negative Effect.Surgency refers to levels of activity, impulsivity, shyness and positive anticipation.Effortful Control encompasses levels of attention and inhibition control.Negative affectivity describes levels of discomfort, anger/frustration, sadness and fear.Each VS-ECBQ item is rated on a Likert scale with a score of 1 indicating "never" and a score of 7 indicating "always".A "Not Applicable" (NA) option was available if the parent had not encountered a particular situation.Items included statements such as: "When playing indoors, how often did your child like rough and rowdy games?" Maternal emotion regulation was assessed using the Emotion Regulation Questionnaire (ERQ) (Gross & John, 2003).The ERQ is a 10-question survey with 6 questions comprising the Cognitive Reappraisal subscale and 4 the Expressive Suppression subscale.Questions were rated on a 7-point Likert scale where a score of "1" indicated "strongly disagree" and a score of "7" "strongly agree".Items included statements like: "When I am feeling negative emotions, I make sure not to express them."

| Paradigm
The mother-child dyads attended one research appointment where the mothers provided written informed consent and the children expressed verbal assent.Pairs were comfortably seated at a 90 angle at a table suitable for toddlers.They were given age-appropriate toys that included a potato head with a variety of accessories or building blocks to play with.These toys were chosen as they did not pose any performance demands and would allow the pairs to play freely.The researchers explained the task to the pair and allowed them to familiarise themselves with the toys and the fNIRS cap.The cap was always placed on the mother first and then the child to minimise the amount of time the child would have to wear the cap.After both participants were wearing the cap, they were given time to adjust while the researchers assessed that optode placement was correct and that FNIRS signals were captured correctly.
Each session was recorded by two cameras placed to capture both participants.The experiment lasted for approximately 20 minutes and was comprised of two conditions (interactive and independent).Each condition lasted for 5 minutes and was repeated twice.Each condition was separated by approximately 2 minutes of rest during which the dyad was offered the opportunity to take a longer break and the mother was moved to an adjacent table if the following condition was independent play.The order of the conditions was pseudorandomised across pairs.
During the interactive condition, the dyads were instructed to play together for 5 minutes "as they would at home".The researchers observed the interactions from outside the room and mothers were allowed to remove their face coverings to allow for more organic interactions.
During the independent condition, the mother and child were separated by an opaque screen and were instructed to play with their respective toys silently for 5 minutes.One researcher remained in the room to ensure the child's safety (Figure 1).

| Turn-taking analysis
Turn-taking was manually quantified based on the recordings of the interactive condition by EP and NP.One conversational turn was defined as a continuous pair of utterances between mother and child spoken in any order within 5 s between speakers (e.g., Mother: "Do you like the blue arms?" Child: "Yes, blue arms!") (Bishop et al., 1998;Gilkerson et al., 2018;Quiñones-Camacho et al., 2021;Romeo et al., 2018).Laughs and acknowledgement sounds were counted as part of a turn if they were produced in response to something the other participant said.A total of 12.5% of the video recordings were also coded independently by a third researcher with inter-rater reliability at 83%.

| FNIRS data acquisition
Two continuous wave fNIRS systems (Hitachi ETG-4000, Japan; sampling rate at 10 Hz) were used to measure brain activities in each dyad (one for each participant).Both systems were connected to an independent computer running a custom-made MATLAB script, which synchronised them and simultaneously initiated the experiment for both systems.The ETG-4000 used two wavelengths of light (695 and 830 nm) to allow for the estimates of changes in both oxy-(HbO) and deoxyhaemoglobin (HbR).The mother's cap was composed of 48 optodes arranged in three arrays: one 3 Â 5 array over the PFC and two 3 Â 3 arrays bilaterally over the TPJ (see Figure 1).The child's cap was comprised of 16 optodes arranged in four 2 Â 2 arrays over the same areas.Fewer optodes were used for the children to make the cap lighter and more tolerable for the children, whilst retaining coverage over the ROIs.The distance between optodes was 3 cm in both the mother's and the child's cap.Before placing the probes, the child's head circumference was measured using a tape measure to account for variability in head shape and size.Placement of the optodes was standardised using the international 10-20 System (left TPJ around position CP5, right TPJ around position CP6, left PFC around position AF3 and right PFC around position AF4) (Jasper, 1958).package (Huppert et al., 2009) incorporating customised scripts (Anderson et al., 2017;Lawrence et al., 2018;Lawrence et al., 2021;Mushtaq et al., 2019;Wiggins & Hartley, 2015).The raw fNIRS intensity signals were first converted into changes in optical density followed by correction of motion artefacts using wavelet filtering (via the HOMER2 hmrMotionCorrectionWavelet function that removed outlying wavelet coefficients outside the 0.725 inter-quantile range) (Molavi & Dumont, 2012).The signals were then bandpass filtered between 0.01 and 0.5 Hz (via a zero-phase 3rd-order Butterworth filter) to attenuate low-frequency drifts and cardiac oscillations.Using the modified Beer-Lambert Law, optical density was converted to estimated changes in the concentration of HbO and HbR (Huppert et al., 2009).The haemodynamic modality separation (HMS) algorithm was used to extract cortical activation (Yamada et al., 2012).This was done by trying to isolate the functional component of the haemodynamic signals from systemic physiological interference (Yamada et al., 2012) assuming that changes in HbO and HbR are negatively correlated in the functional responses but positively correlated in the motion and physiological noises (Yamada et al., 2012).Using this algorithm has demonstrated greater reliability of fNIRS signal quality (Wiggins et al., 2016).Moreover, to further improve the signal quality, channels with potentially poor signals were detected using the scalp coupling index (Pollonini et al., 2014).Signals were bandpass filtered at 0.5-2.5 Hz and channels with poor signal quality were excluded for subsequent analyses.

| FNIRS data analysis
Furthermore, due to the naturalistic free-play paradigm in the current study, we considered that systemic physiological confounds may not be eliminated completely after the preprocessing.We thus further divided the preprocessed signals into frequency subbands that correspond to specific types of possible physiological confounds and then focused on the band which led to significant neural synchrony (i.e., synchrony that was significantly greater for the interactive than the independent play).This was conducted so that we could detect which frequency range may be least contaminated by these confounds that led to optimal measures for neural synchrony.To avoid arbitrary choices of frequency bands, we chose three sub-bands based on the previous literature: 0.01-0.05Hz, 0.05-0.2Hz and 0.2-0.5 Hz which reflect ranges for the autonomic, myogenic and respiratory activities, respectively (Fernandez Rojas et al., 2017;Rossi et al., 2006).Zero-phase 3rd-order Butterworth filters were additionally applied to obtain fNIRS signals at each band (N.B., to avoid duplicate cut-off slopes applied on the lower/upper bound of the alreadypre-processed signal, lowpass and high pass filters at 0.05 and 0.2 Hz was used to obtain the sub-bands at 0.01-0.05and 0.2-0.5 Hz, respectively).We also included the 0.01-0.5 Hz range (i.e., without further filtering the preprocessed signals).Here, we focused on the range at 0.05-0.2Hz, which was the only frequency band that showed significant neural synchrony.This reflected that myogenic confounds could be most effectively attenuated at this frequency range.
Neural synchrony between the mother-child dyads was measured using PTE (Lobier et al., 2014;Ursino et al., 2020).PTE calculates the degree of certainty in one signal given the past values of another signal using phase entropies.It thus quantifies the extent to which two different neural activities causally influence one another (Lobier et al., 2014).Compared to other methods typically used in fNIRS hyper scanning studies such as wavelet coherence transformation, PTE allows for measuring the directionality of neural synchrony that considers potential time lags between brain activities in each participant and quantifies the degree of interbrain synchrony in different directionalities (i.e., how the brain of one participant follows the other and vice versa) (Cao et al., 2018;Marriott Haresign et al., 2022;Wang & Chen, 2020).The haemodynamic response functions (HRF) could be different between children and adults (Minagawa-Kawai et al., 2011), PTE has a specific advantage for which it does not require the assumption of the same HRF for both participants as required by other methods like wavelet coherence transformation and Granger Causality.We applied the open-accessed MATLAB codes that calculate the PTE (Fraschini & Hillebrand, 2017) using the following formulas (Lobier et al., 2014): where PTE x !y (formula (1)) calculates the synchrony between two signals x(t) and y(t) with information flowing from x to y (i.e., y follows x).θ x (t) and θ y (t) refer to the instantaneous phase (via Hilbert transform) at time t of the two signals, respectively.δ (formula (2)) denotes the maximum time lag between the two signals.The time lag δ can be predetermined or it can be modulated by the duration of the frequency cycle.In this instance, we followed the open-accessed MATLAB codes by Fraschini and Hillebrand (2017) and defined the time lag as the time duration of approximately one cycle of the selected frequency (0.01 Hz-0.5 Hz) (Fraschini & Hillebrand, 2017).p refers to the probability of instantaneous phases and H is the phase entropy where the summation was implemented over the period between t and t' (i.e., from 0 to δ) (formula (3)).Greater PTE x !y reflects smaller entropy (i.e., greater certainty) of θ y given in the past values of θ x , hence informing greater information flows from x to y.Here, δ was set at 4 seconds according to previous reports that showed that neural synchrony between adults and children peaked when the time lag between the two signals (i.e.signal from adult and signal from child) was approx.4 s (Zhao et al., 2021;Piazza et al., 2020).The pre-processed signals were averaged across channels within each ROI in the time domain before PTE was applied to measure the neural synchrony between the dyads.As there were four ROIs (bilateral TPJ and bilateral prefrontal cortices) for both the child and parent, this resulted in 16 PTE values for each directionality (information flow from child to parent or from parent to child) for each dyad.

| Statistical analysis
Statistical analysis was performed on IBM SPSS Statistics (version 27.0).Pearson's correlations were used to explore the relationships between the behavioural measures and turn-taking and linear regressions were used to explore how behavioural measures and turn-taking influenced the neural synchrony between the pairs.Repeated measures ANOVAs were performed to explore the effects of condition, direction and ROI on neural synchrony.PTE was the dependent variable, and condition (interactive vs. independent play), directionality (child to parent vs. parent to child) and ROI (left TPJ, right TPJ, left PFC and right PFC) were the independent variables.The threshold for statistical significance was set at p < 0.05.
To counteract the problem of multiple comparisons the Bonferroni corrections were applied to the correlation analyses and the post hoc multiple comparisons.
The effects of the condition were also significant after controlling for child's gender (F[1,10] = 14.12, p = .024).
We also performed an exploratory post hoc analysis for the interaction between condition and direction (even though it was not statistically significant).Descriptive statistics showed that in both the interactive and independent conditions child-directed neural synchrony (Interactive mean dif = 1.16,SD = .020,independent mean dif = 1.08,SD = .023)was lower compared to the mother-directed neural synchrony (Interactive mean dif = 1.17,SD = .016,independent mean dif = 1.147,SD = .022).Paired t-tests showed that child-directed and mother-directed neural synchrony was not statistically significantly different in either condition (independent: mean dif = À.07, p = .052;Interactive: mean dif = À.012, p = .557).However, when comparing neural synchrony in each condition for each direction, childdirected neural synchrony was statistically significantly higher in the interactive compared to the independent condition (mean dif = .08,p < .001).There was no statistically significant difference between conditions for the mother-directed neural synchrony (mean dif = .025,p = .337)(Figure 2).However as stated above, since there was not a statistically significant main effect of direction and ROI or a statistically significant interaction between direction, condition and ROI, further analysis was conducted using the average neural synchrony between the two directions across all ROIs.Subsequent analysis to determine whether there was an effect of region in the neural synchrony between the mother's and the child's brain areas found no evidence in either condition (child hemisphere: (F[1,11] = 2.44, p = .147),child region (F[1,11] = 4.37, p = .61),mother hemisphere (F[1,11] = 3.24, p = .099),mother region (F [1,11] = 2.29, p = .159).

| Neural synchrony in relation to turn-taking
We assessed whether the observed increased neural synchrony in the interactive condition could be attributed to the conversational patterns of the pairs.A linear regression showed that turn-taking was not a predictor of neural synchrony (F[1,10] < .001,p = .986)(Table 1).

| DISCUSSION
This study examined the neural synchrony between pairs of mothers and their young children under naturalistic conditions.In a departure from prior studies that have F I G U R E 2 Boxplots of mean neural synchrony across all ROIs in the interactive and independent condition.Bars represent neural synchrony for the two directions (child to mother & mother to child) and the average of the two.Neural synchrony in the interactive condition was statistically significantly higher compared to the independent condition (mean = .053,p = .003).Child-directed neural synchrony was significantly higher in the interactive compared to the independent condition (mean = .08,p < .001).**p < 0.01, ***p < 0.001 (Bonferroni corrected).
predominantly focused on the presence of neural synchrony in parent-child pairs during task-oriented problem-solving activities, we employed a free-play paradigm, allowing participants the freedom to interact without a specific objective.As hypothesised, we showed higher levels of neural synchrony between the dyads in the interactive condition compared to the independent condition.This finding validates the use of free-play paradigms in investigations of neural synchrony between parents and their children.We also explored the influence of turn-taking and maternal and child personality traits on neural synchrony, however, our analysis did not yield any robust associations between the relationships explored, except for a negative correlation between neural synchrony and child surgency.
Neural synchrony was significantly higher in the interactive compared to the independent condition over bilateral prefrontal areas and the bilateral temporoparietal junction.This is in line with previous research that has reported synchrony in prefrontal areas during parent-child cooperative performance (Miller et al., 2019;Nguyen et al., 2020;Reindl et al., 2018).These regions are also involved in attention and executive functioning, indicating that these processes support neural synchrony, especially during cooperative interactions (Azhari  et al., 2019).The TPJ is also recruited during interpersonal interactions as it is associated with language processing, self-reference and processing of one's own and others' mental states (Monticelli et al., 2021).We examined the potential interactions among our ROIs and their influence on the overall neural synchrony of the pairs.Nevertheless, our findings did not support any discernible effect of ROIs, as well as any disparity stemming from hemispheric (left versus right) or region (PFC vs TPJ) variations.This lack of discernible effects suggests that these ROIs may be equivalently involved in facilitating neural synchrony during pair interactions.
In line with prior research, our findings similarly indicated that child gender had no discernible impact on neural synchrony (Azhari et al., 2019;Nguyen et al., 2020;Reindl et al., 2018).Nonetheless, earlier investigations into neural synchrony among adults have indicated variations in synchrony between same-sex pairs when contrasted with mixed-gender pairs (Cheng et al., 2015).Additionally, Miller and colleagues reported that in their sample mother-son dyads exhibited lower neural synchrony than mother-daughter pairs in the control task but no differences in the cooperation task.This divergence could potentially signify variations in the approaches taken by mother-son pairs when engaging in tasks (Miller et al., 2019).Nevertheless, it remains intriguing to delve deeper into the potential influence of child gender on parent-child interactions and the accompanying neural synchrony.Expanding the study with larger sample sizes and considering diverse developmental stages would be valuable in this regard.It is also important to note that our study exclusively focused on pairs consisting of children and their biological mothers.Initial findings suggest that comparable neural synchrony exists during interactions between biological fathers and their children.(Azhari et al., 2021;Nguyen, Schleihauf, Kungl, et al., 2021).Consequently, it presents a compelling avenue for scientific and ecological exploration to extend this research and investigate neural synchrony in father-child pairs, as well as caregiver-child pairs, where the adult figure may be someone other than the child's biological parent, such as an adoptive parent, grandparent or another primary caregiver.
Neuroimaging data from the pairs were analysed using PTE, which allowed us to investigate the directionality of the neural synchrony.Even though the effect of direction did not reach the assigned levels for statistical significance, it is worth mentioning that child-directed neural synchrony in the interactive condition was statistically significant compared with the independent condition.This might be an indication that the neural synchrony between the dyads could be primarily driven by the child.Similarly, Quiñones-Camacho et al. (2020) suggested that the neural synchrony in their set-up might be driven by child-related and not maternal characteristics.Notably, a similar difference was not found between mother-directed neural synchrony in the interactive versus the independent condition.This suggests that mothers within our sample remained "tuned in" with their children even when they engaged in independent play.This is substantiated by the fact that the lack of difference in mother-directed neural synchrony between conditions is attributed to the elevated levels of mother-directed neural synchrony observed during independent play.However, larger sample sizes and further work would be required to properly explore these observations.
In this study, participants were instructed to play as they would at home, interactively or separately; they were not explicitly told to collaborate, communicate or aim for a specific goal.Nonetheless, despite the pairs being permitted to approach the task as they pleased, and despite varying degrees of communication between each dyad, higher neural synchrony was measured in the interactive condition compared to the independent one.Even though it is reported that levels of neural synchrony are affected by the intensity and type of the interaction (Gvirts & Perlmutter, 2020), our findings are also corroborated by investigations in both adult dyads and parent-child dyads with similar paradigms, demonstrating significant levels of neural synchrony when dyads interact in low demand contexts (Nguyen, Schleihauf, Kayhan, et al., 2021;Piazza et al., 2020;Quiñones-Camacho et al., 2020).Quiñones-Camacho et al. (2020) showed similar levels of neural synchrony when parents and toddlers completed a goal-oriented task and an unstructured free-play task.Above chance-level neural synchrony was also observed during face-to-face motherchild and adult pair conversations (Jiang et al., 2012;Nguyen, Schleihauf, Kayhan, et al., 2021).A possible explanation of the presence of high neural synchrony in the absence of a complex, goal-oriented task is the setting of and the interacting partners in, these low-demand interactions (Gvirts & Perlmutter, 2020).Neural synchrony appears to be enhanced when partners are facing one another (Jiang et al., 2012;Jiang et al., 2017) and are sharing gazes and smiles (Nguyen et al., 2020;Piazza et al., 2020).Additionally, neural synchrony is fostered when interacting with "familiar" and "significant" partners and when exhibiting joint attention to mutually important stimuli (Djalovski et al., 2021;Kinreich et al., 2017;Pan et al., 2017).Nonetheless, neural synchrony among individuals does not necessarily demand the presence of all these behaviours.For instance, it has been shown that neural synchrony can arise during verbal communication and joint computer games even if participants are unable to see each other (Ahn et al., 2018;Pérez et al., 2019).Conversely, it has been suggested that neural synchrony can exist only due to mutual eye gaze (Noah et al., 2020).Hence, the mechanisms supporting neural synchrony appear intricate and multifaceted.In our study, the occurrence of neural synchronisation can likely be attributed to the unique dynamics of biological mother-child interactions, encompassing both verbal and non-verbal cues.These factors may have served as a compensatory mechanism, bridging the absence of goal-oriented tasks and sustained faceto-face engagement.This finding holds profound significance, especially when contemplating its implications for investigating neural synchrony within clinical populations with speech, communication and behavioural disorders.Children with communication challenges often face obstacles in undertaking complex tasks demanding verbal communication and sustained attention, owing to their linguistic limitations and behavioural difficulties that hinder their ability to complete such activities (Hage et al., 2021;Hollo & Chow, 2015;Wintgens, 2013).Additionally, behavioural studies have highlighted time and time again that low-quality parent-child interactions can have a negative impact on childhood development (Baumwell et al., 1997;Bornstein & Tamis-LeMonda, 1989;Justice et al., 2019;Scheiber et al., 2022;Simpkins et al., 2006;Takeuchi et al., 2015).As a result, parent-implemented therapies, that target parent-child interactions and train parents on how to optimally interact with their children, have been shown to be highly effective (Allen & Marshall, 2011;Curtin et al., 2021;Koly et al., 2021;Lyons et al., 2022;McConachie & Diggle, 2007;Rieth et al., 2018;Thomas et al., 2017;van Noorden et al., 2022).Utilising research paradigms tailored for examining neural synchrony in these specific populations can yield valuable insights into the neural foundations of communication disorders and the mechanisms underlying the effectiveness of parentimplemented therapies.Such insights, in turn, have the potential to empower clinicians in assessing the impact of their parent-implemented therapies more comprehensively and in tailoring treatment goals to enhance outcomes for each individual child.
A secondary aim of this study was to explore the relationship between turn-taking and neural synchrony.Previous studies have found a positive correlation between turns and neural synchrony in adult dyads (Ahn et al., 2018;Hasson et al., 2012;Pérez et al., 2019;Stephens et al., 2010;Wilson & Wilson, 2005).Nyugen, Schleihauf, Kayhan, et al. ( 2021) also presented similar findings during conversations between mothers and children.Nonetheless, the current study did not reveal such a relationship.One potential explanation could be the approach used for coding turn-taking.In this study, due to the participants' relatively young age, only pertinent and alternating turns were incorporated into the analysis, as these more definitively indicated active engagement between the participants, particularly the children.Another possible reason for the absence of a connection between neural synchrony and turn-taking might be attributed to the nature of the task.The dyads were instructed to engage in unstructured play, where direct communication between the pair was not a prerequisite for successful task completion, thus resulting in fewer turns taken.Therefore, the specific activities assigned to the dyads could significantly influence the extent of the correlation between neural synchrony and turn-taking.It would be valuable to investigate this finding further with a larger sample size.
A further exploratory aim of this work was to examine whether the personality traits of both the mother and the child influenced neural synchrony.Previous studies have reported that emotional regulation as well as child temperament can affect neural and behavioural synchrony between parent-child dyads (Azhari et al., 2019;Hoyniak et al., 2021;Quiñones-Camacho et al., 2021;Santamaria et al., 2020).In this study, we observed that child surgency was negatively correlated to neural synchrony.Surgency characterises individuals who are cheerful, responsive and impulsive (Oldehinkel et al., 2004;Putnam, Ellis, Rothbart, et al., 2001).Surgency in children can support relationships with their peers and help children to be socially competent (Putnam & Stifter, 2005;Rimm-Kaufman & Kagan, 2005).However, behavioural studies also suggest that surgency can be maladaptive, leading to higher levels of externalising behavioural problems such as aggression (Berdan et al., 2008;Gunnar et al., 2003;Stifter et al., 2008).Additionally, children with high levels of surgency might employ distraction/self-soothing behaviours, which in some contexts might be advantageous (for example when an outcome is delayed) but problematic when faced with goal-oriented situations (Dollar & Stifter, 2012).Since interpersonal synchrony requires mutual attentiveness, it is possible that high levels of activity and spontaneity can make a child more distracted by their environment and hinder their ability for neural synchrony with their parent.It's crucial to emphasise that additional research is necessary to validate this finding, especially considering our small sample size.
We did not find any evidence that suggested that maternal emotional regulation and the other child temperament dimensions (i.e., effortful control and negative affect) were related to neural synchrony.In contrast, Reindl et al. reported a positive relationship between parent and child emotional regulation to their neural synchrony during collaboration, and Azhari et al. (2019) found that maternal stress levels were negatively correlated with mother-child neural synchrony when the pairs watched a movie (Azhari et al., 2019;Reindl et al., 2018).In our study, we did not measure maternal stress so we cannot speculate on how it might be associated with our neural synchrony measurements.We might not have been able to detect an association between maternal emotional regulation and neural synchrony due to our small sample size.Additionally, Reindl and colleagues deployed a paradigm with higher-demands i.e., cooperative computer game which might have resulted in a higher degree of neural synchrony compared to our lower-demands interactive free play paradigm.That might have also accounted for the lack of association between maternal factors and neural synchrony in our study.
Our study had a few limitations.To begin with, participant recruitment was performed during the summer of 2021 amidst the ongoing COVID-19 pandemic, which led to our relatively small sample size.Even so, despite the relatively small sample size of 12 dyads, the study accomplished its aim of exploring the feasibility of measuring the relationship between neural synchrony and turn-taking using a free-play paradigm.Additionally, it is well documented in behavioural investigations of synchrony and parent-child interactions that the nature of the parent-child relationship evolves as the child grows up (Davis et al., 2018;Farran & Kasari, 1990), thus longitudinal studies would be valuable in understanding how neural synchrony changes over a child's development.Thirdly, it is well-documented that mutual eye gaze promotes interpersonal and inter-brain synchronisation (Dravida et al., 2020); Hirsch et al., 2017;Kelley et al., 2020;Noah et al., 2020).However, our video recording setup did not allow us to perform analysis on non-verbal behavioural cues such as mutual eye gaze that could have offered additional information on the factors influencing neural synchrony.For that reason, in subsequent studies, we will employ eye-tracking glasses that continue to allow for naturalistic experimental setups but also capture eye-gaze data (Saravanan et al., In preparation).Lastly, fNIRS has poor depth penetration that does not allow imaging of subcortical regions (Fukui et al., 2003;Harrison & Hartley, 2019).Additionally, our fNIRS machines are wired, so they require participants to stay in one place.Truer naturalistic play may be recorded via the use of wireless fNIRS setups that do not have such strict restrictions on movement around a room.

| CONCLUSION
This study explored neural synchrony within motherchild dyads during a low-demand free play paradigm.We observed a significant increase in neural synchrony when the pairs engaged in joint play as opposed to playing individually.These results offer valuable evidence supporting the existence of neural synchrony in naturalistic interactions.Notably, a negative correlation emerged between neural synchrony and child surgency, underscoring the significance of a child's engagement with their parent in shaping the neural features of parent-child interactions, nonetheless, to confirm this discovery, a larger sample size is necessary.
Overall, given the pivotal role of mother-child interactions in childhood development, these findings pave the way for future investigations into the neural foundations of parent-implemented therapies and neural synchrony among pre-lingual children and those with communication needs.

AUTHOR CONTRIBUTIONS
FNIRS data were pre-processed via MATLAB 2019b (The MathWorks, 2019, Natick, MA) using the HOMER2 F I G U R E 1 (a) Experimental set-up during the interactive (top panel) and independent conditions (bottom panel).(b) Illustration of the probes overlaying the ROIs (yellow circles) for the mother (top panel) and the child (bottom panel) bilaterally over the prefrontal cortex and the temporoparietal junction (only the left hemisphere depicted here).The red circles represent the emitter optodes, the blue represents the detector optodes and the black lines represent the channels.The figures are for illustrative purposes only, not to scale.
Regression analysis summary for child surgency effortful control and negative affect predicting neural synchrony.
T A B L E 2Note: R 2 adjusted <.001, CI = confidence interval for B.