Content uploaded by Sam Wass
Author content
All content in this area was uploaded by Sam Wass on Feb 28, 2020
Content may be subject to copyright.
Review
Interpersonal Neural Entrainment during Early
Social Interaction
Sam V. Wass,
1,
*M. Whitehorn,
1
I. Marriott Haresign,
1
E. Phillips,
1
and V. Leong
2,3
Currently, we understand much about how children’s brains attend to and learn
from information presented while they are alone, viewing a screen –but less
about how interpersonal social influences are substantiated in the brain. Here,
we consider research that examines how social behaviors affect not one, but
both partners in a dyad. We review studies that measured interpersonal neural
entrainment during early social interaction, considering two ways of measuring
entrainment: concurrent entrainment (e.g., ‘when A is high, B is high’–also
known as synchrony) and sequential entrainment (‘changes in A forward-
predict changes in B’). We discuss possible causes of interpersonal neural
entrainment, and consider whether it is merely an epiphenomenon, or whether
it plays an independent, mechanistic role in early attention and learning.
Tracking Dynamic, Social Influences on Early Attention and Learning
During the first years of our lives, in particular, our waking hours are spent almost entirely in the
company of others. Yet currently, and paradoxically, most of our knowledge of how the developing
brain functions during social interaction comes from studies that examine individual humans in iso-
lation [1–4]. From early life, however, we know that social factors influence how we allocate our at-
tention and learn. For example, 9-month-old infants learn new speech sounds better through live
interaction with an adult than through watching an equivalent video of someone speaking [5].
When a 16-month-old infant initiates an exchange by pointing to an object, their memory retention
for functions subsequently demonstrated on that object is increased [6]. And when a parent pays
attention to a particular object while interacting with their 12-month-old infant, this immediately in-
creases the infant’s own duration of attention to that object [7]. Yet, we presently understand little
about how these transient, interpersonal influences are substantiated in the brain.
Recent research, building on advances in adult [3]andanimal[8]socialneuroscience,hasbegunto
explore these dynamic, social influences by doing two things differently: first, rather than recording
from one individual brain in isolation, they record from two interacting brains concurrently (sometimes
known as hyperscanning, see Glossary)[9]; second, rather than examining unidirectional
influences using predesigned, screen-presented experimental stimuli, they examine naturally occur-
ring moments of reciprocal influence during free-flowing interpersonal naturalistic interactions. This
research is starting to uncover a range of important new discoveries about which brain regions
are active during social interaction, that illustrate the importance of studying social interaction in eco-
logically valid contexts. For example, mentalizing and reward networks show markedly different pat-
terns of activity during live interaction, compared to when passively viewing equivalent social stimuli
on a screen [3,4].
Here, we examine a different question: what this research has taught us about the neural tempo-
ral dynamics of early social interactions. In the first part of this review, we consider two important
and well-researched aspects of social interaction: ostensive cues (signals that are sent during
social interaction to indicate communicative intent), and contingent responsiveness (behaviors
Highlights
Social factors exert transient influences
on the brains of both partners during an
interaction.
Interpersonal neural entrainment during
early learning interactions has been doc-
umented at multiple temporal scales,
replicating findings with adults and
animals.
Neural entrainment can be a conse-
quence of behavioral coordination, but it
can also arise in the absence of behav-
ioral coordination: shared understanding
involves temporally co-occurring pat-
terns of brain activity.
Neural entrainment may influence learn-
ing in multiple ways; for example, by
allowing the sender to ensure information
delivered arrives at an optimal time for
encoding by the receiver.
1
Department of Psychology, University
of East London, London, UK
2
Department of Experimental Psychol-
ogy, University of Cambridge, Cam-
bridge, UK
3
School of Social Sciences, Nanyang
Technological University, Singapore
*Correspondence:
s.v.wass@uel.ac.uk (S.V. Wass).
Trends in Cognitive Sciences, Month 2020, Vol. xx, No. xx https://doi.org/10.1016/j.tics.2020.01.006 1
© 2020 Elsevier Ltd. All rights reserved.
Trends in Cognitive Sciences
TICS 2012 No. of Pages 14
that indicate communicative sensitivity within an interaction). We conclude that the unidirectional
neural correlates of both –that is, how one partner transiently influences the other, with both
partners considered independently –are well understood; but that our understanding of how
ostensive cues and contingent responsiveness alter the interpersonal neural dynamics of the
interaction –that is, how the partners inter-relate to one another –is currently limited.
In the second part of the review, we consider research that has directly examined interpersonal
neural dynamics by measuring interpersonal neural entrainment during social interaction. We
describe key methodological challenges in measuring entrainment and outline the evidence for
the different types of entrainment that emerge during early social interaction. Building on the ev-
idence for unidirectional influences described in the first section, we also consider the mecha-
nisms through which bidirectional interpersonal neural entrainment could be achieved and
maintained. In the final section, we discuss whether interpersonal neural entrainment is merely
an epiphenomenon; or whether it may play a mechanistic role during early attention and learning.
Ostensive Signals
Social interactions are complex, fast-moving, and multilayered: they require the brain to process
rapidly changing information from multiple visual and auditory sources in a time-sensitive manner.
During social interaction, we use signals known as ostensive cues to indicate communicative in-
tent; these tend to be concentrated on moments where the sender wants to convey particularly
important information to the receiver [10]. Historically, the majority of previous research has exam-
ined how adults use ostensive cues towards children, consistent with pedagogical approaches
that primarily emphasize a flow of information from an adult sender to a child receiver. However,
more recent research has recognized that even young infants also use ostensive cues [11,12];
and that, rather than acting as purely passive recipients of information sent by an adult, they
also play an active role as senders of information (such as interrogative cues), which influence
learning exchanges [13–17]. Thus, in addition to studying the direction of influence from adult
sender to child receiver, recent approaches are acknowledging that children can also act as
senders of social information, and adults as receivers [13].
Ostensive cues lead to a range of changes in behavior during the time period immediately follow-
ing the cue [10]. Although ostensive signals are unidirectional by definition (insofar as they are sig-
nals sent from Partner A to Partner B), extensive research suggests that they also affect the
relationship between Partner A and Partner B. Specifically, research has shown that ostensive
cues lead to increased behavioral entrainment (Box 1) in the time period following the ostensive
cues. For example, ostensive cues such as direct eye gaze lead to increases in behavioral mimicry
[18] and the mirroring of facial affect [19]. Similarly, in language, increased vocal mirroring is
observed following the use of child-directed speech contours [20]. Direct eye gaze [21], child-
directed speech [22], and pointing [23] all lead to increases in gaze following, which is another
form of sequential behavioral entrainment (Box 2).
Considerable research has investigated the transient unidirectional effects of ostensive cues
(i.e., how the receiver, considered independently, is affected by the sending of a social signal).
This research has suggested that both child and adult brains are highly sensitive to ostensive
cues [24]: for example, infants show larger neural evoked-response or event-related potential
(ERP) responses (specifically, a larger amplitude N170 component) to images of faces showing
direct compared to averted gaze even shortly after birth [25]. During live adult–infant play, cortical
activity (measured from the level of oxyhemoglobin in the medial prefrontal cortex) increases in
7-month-old infants during direct gaze compared with averted gaze [26,27]. Similarly, child-
directed speech evokes greater neural responses (a larger amplitude N250 ERP component)
Glossary
Allostasis: process by which internal
equilibrium is maintained.
Contingent responsiveness:
behaviors that indicate communicative
sensitivity within an interaction.
Cross-correlation: measure of the
similarity between two time-series as a
function of the displacement of one
relative to the other. Cross-correlations
examine whether changes in one time-
series tend to anticipate, or follow on
from, changes in another.
Dyadic attention: two-way sharing of
attention –either between a person and
an object, or a person and another
person.
Electroencephalography: method for
recording naturally occurring electrical
brain activity.
Entrainment: in this paperwe consider
two forms of entrainment. The first is
concurrent entrainment (a zero-lag
relationship between two time series,
e.g., ‘when A is high, B is high’or ‘when
A is high, B is low’), which is commonly
known as synchrony. The second is
sequential entrainment (‘changes in A
forward-predict changes in B’). See Box
1for further details of how these terms
are quantified.
Functional near-infrared
spectroscopy: method for recording
blood oxygenation levels near the scalp.
Granger causality: method for
quantifying sequential entrainment by
analyzing how one time-series
influences another; similar to cross-
correlations, but incorporating
information about the self-similarity of
each time series.
Hyperscanning: neuroimaging studies
that record brain activity in two
individuals at the same time.
Ostensive cues: signals from a
communicator to generate an
interpretation of communicativeintention
in an addressee.
Partial directed coherence:
technique to examine cross-spectral
Granger-causal relationships in
multivariate time series.
Phase entrainment: concurrent
entrainment in the phase of ongoing
oscillatory activity.
Phase resetting: abrupt shift
(e.g., advancement or delay) in the
phase of ongoing oscillatory activity,
usually in response to perturbation by a
sensory cue.
Trends in Cognitive Sciences
2Trends in Cognitive Sciences, Month 2020, Vol. xx, No. xx
in 6–12-month-old infants compared with adult-directed speech [28], and during live
interactions, fluctuations in the child-directedness of speech correlate with fluctuations in
prefrontal cortex activity in 9–15-month-old infants [29]. In addition to neural activity directly
in response to the gaze cue, neural responsiveness is also increased for objects presented
immediately following an ostensive cue. For example, when an adult gazes first to a
9-month-old infant’s face before looking to a screen-presented object, the infant’sevoked
neural responses to the subsequently presented object are greater (measured as a larger
amplitude Nc ERP component) [30,31].
Currently, however, little research has investigated the bidirectional neural effects of ostensive
cues (i.e., how ostensive cues alter the relationship between the two partners’neural activity in
the time period following the cue). In order to assess how ostensive cues alter the interpersonal
neural dynamics of the interaction, it is necessary to examine change from the perspective of
not one, but both members of the dyad.
Phase-locking value: technique for
estimating concurrent entrainment
between thephase series of two signals.
Synchrony: in this paper, we treat the
term synchrony as synonymous with
concurrent entrainment. See Box 1 for
further details.
Triadic attention: three-way sharing of
attention –generally between two
people and an object.
Wavelet transform coherence:
technique that can be used to measure
both concurrent and sequential
synchrony of two signals in the time-
frequency plane.
Tre nd s
Tre nd s
in
in
Cognitive
Cognitive
Sciences
Sciences
Figure 1. Behavioral Paradigms Used to Measure Interpersonal Neural Entrainment. (A) examples of the experimental setups described in the main text, used
by: left [43,126]; middle [60]; right [52]. (B) Examples of raw data collected using the paradigms from [43,126]; even from this raw data sample, some significant patterns
noted that overall data can be seen –such as, parents pay more attention to infants during joint play than vice versa [114], and infants are more inattentive during solo play
than joint play [126]. (C) Example of a 30-s segment of dat a illustrating the further range of different types of events that can be identified in naturalistic interactions.
Abbreviation: EEG, electroencephalography.
Trends in Cognitive Sciences
Trends in Cognitive Sciences, Month 2020, Vol. xx, No. xx 3
Contingent Responsiveness
Another strand of research has investigated the role of contingent responding during a social
interaction: how the receiver indicates communicative sensitivity within an interaction by consis-
tently responding to behaviors from the sender. Considerable evidence suggests that both
children and adults are highly sensitive to whether or not their partner is responding contingently
to their social signals [14,24,32,33]. For example, one study observed 6.5- and 9.5-month-old
infants’reactions to adults who either responded to the infants’gaze cues by following their
gaze towards an object (congruent looking) or looked in the opposite direction (incongruent).
Older, but not younger, infants showed a visual preference for the congruent actor and showed
greater neural reactivity to the stimulus cued by the congruent (compared to the incongruent)
adult [34].
Behavioral research also suggests that interactions featuring greater behavioral contingency in
both members of the dyad are also more effective as teaching exchanges. For example, in a
task in which adults presented word labels either contingently in response to infant vocalizations
or noncontingently, only infants who received labels contingently in response to their own
attention learned the association [6,13,14,35,36]. However, not all forms of contingency are equally
effective: during videoed interaction only particular types of contingent responding (mirroring, and
marking with a smile) were predictive of the growth of these behaviors over time [18].
Considerable research has investigated the unidirectional correlates of contingent responding
(i.e., how the receiver of a social signal is influenced by the sender). This research has suggested
that observing someone else perform an action involves neural activity in the observer becoming
more like neural activity in the person performing the action [37,38]. This pattern is similar to the
actor–observer correspondences documented while watching and performing actions [39], and
while watching someone else experience pain, anger, and reward [40,41]. There is also some ev-
idence that, in more contingently responsive social partners, these actor–observer correspon-
dences are stronger [38,42].
Recent research also suggests that similar principles might also apply in different contexts, such
as when considering how adults’brain activity tracks infants’attention patterns during naturalistic
play. The study recorded dual-electroencephalography (EEG) from parents and 12-month-old
infants during free-flowing play (Figure 1A,B). By tracking the continuous fluctuations of brain
Tre nd s
Tre nd s
in
in
Cognitive
Cognitive
Sciences
Sciences
Figure 2. Neural Activity in the Parent Entrains to Attentional Fluctuations in the Child. (A) Schematic illustrating the analysis; (B) cross-correlation between child
EEG and child gaze, showing an association between child neural activity and child attention; (C) cross-correlation between adult EEG and adult gaze, also showing an
association between adult neural activity and adult attention; (D) cross-correlation between adult EEG and child gaze, showing an association between the adult’s
neural activity and the child’s attention. Subsequent analyses showed that the association between the adult’s neural activity and the child’s attention was independent
of the adult’s own attention. Whereas the child EEG–child gaze and adult EEG–adult gaze relationships were predictive (i.e., strongest associations were found
between neural activity at a given moment and attention c. 750 ms after that moment), the adult EEG–child gaze associations were reactive (i.e., strongest associations
were found between the child’sattentionatagivenmomentandtheadult’sneuralactivityc.750msafterthatmoment). Reproduced, with permission, from [43].
Abbreviation: EEG, electroencephalography.
Trends in Cognitive Sciences
4Trends in Cognitive Sciences, Month 2020, Vol. xx, No. xx
activity in the adult and attention patterns in the infant, their results suggested that adults’neural
activity entrained to fluctuations in the child’s attention, independent of the adult’s own attention
[43] (see Figures 1 and 2). When the adult shows greater neural entrainment to the child’s atten-
tion, the child is more attentive [43].
Currently, however, and although contingent responsiveness is inherently a bidirectional
behavioral phenomenon (because it describes the relationship between the two partners’
behaviors), little research has investigated how contingent responding alters the bidirec-
tional neural dynamics of the interaction. Previous research has mainly considered actor–
observer similarities by recording separately from actors and observers; in order to inves-
tigate interpersonal neural dynamics, it is necessary to record from both partners
concurrently.
Box 1. Types of Interpersonal Entrainment and Measurement Caveats
The concept of ent rainment, and the related concept of synchrony, are imme nsely rich topics that have been studied
across all domains of knowledge [82–85]. Within cognitive neuroscience, extensive previous research has examined
how different units within a brain, (from individual neurons to brain regions), entrain to one another (intrapersonal
entrainment) [86–88]. Previous research has also examined entrainment between individual brains and temporal structures
in the environment (stimulus–brain entrainment) [89].
In this article, we focus on two forms of interpersonal entrainment [19,90]:
•Concurrent entrainment –a zero-lag relationship, for example, ‘when X is high, Y is high’–measured using techniques
including Phase-locking value [91], wavelet coherence [92], and others. Concurrent synchronyis undirected (A→Bis
indistinguishable from B→A). Concurrent entrainment is treated as synonymous with the term synchrony.
•Sequential entrainment –‘changes in X forward-predict changes in Y’–measured using techniqu es derived from
Granger causality [93], including generalized partial directed coherence [94]. Sequential entrainment is directional
(A→B≠B→A).
Entrainment can also be estimated based on different aspects of the signal:
•Amplitude. Some fNIRS studies [29] and most fMRI studies [95] measure cofluctuations in the amplitude of the signal –
which, depending on the method, mea sures blood oxygenation/deoxygenation (fNIRS), the BO LD signal (fMRI), or
voltage (for EEG).
•Phase. Many EEG studies measure the alignment of oscillatory phase between two signals. [59,61].
•Combinations. Many fNIRS studies measure wavelet coherence, which includ es both amplitude and phase [51,52].
Other studies measure, for example, phase–amplitude coupling [96].
In the case of child–adult interactions, one complication is that ad ults and children have different dominant frequency
bands of naturally occurring brain activity [97]. Although the majority of studies hitherto have not addressed this, techniques
for measuring cross-frequency coupling are available [55].
Researchers measuring interpersonal neural entrainment face a number of methodological caveats [90]:
The first is that common intrinsic properties of the neural signal itself can create a false impression of entrainment [90]. For
example, two adults, each with a dominant alpha rhythm of 10 Hz, might show consistent phase relationships between
their alpha rhythms even inthe absence of any communication [90]. Similar considerationsapply when considering variability
in amplitude and power of the signal, where co-occurring oscillatory activity may be attributable to other sources (such as
autonomic activity [98]).
A second problem is that neural activity is also influenced by common environmental perturbations. For example, neural ac-
tivity synchronizes to temporal structures in speech [89], and differentiating neural interpersonal entrainment from synchrony
attributable to shared external perturbation can be highly challenging [99]. This is particularly true for naturalistic paradigms
where factors such as the acoustic environment cannot be completely controlled for.
Common techniques for addressing these problems often use bootstrapping approaches in which corresponding epochs
from each dyad are either temporally translocated (‘shuffled’) or phase scrambled and the entrainment analysis is performed
repeatedly. In this way, it is possible to estimate how the observed entrainment would differ from the entrainment observed
by chance [29]. Of note, however, this approach does not control for environmental influences in naturalistic studies where
the environment differs between dyads. Becauseof this inherent problem, most hyperscanning studies do not measure ab-
solute levels of synchrony, but rather examine relative changes in coupling between different experimental conditions (such
as the presence or absence of mutual gaze) –while keeping other factors constant (such as the acoustic environment).
Trends in Cognitive Sciences
Trends in Cognitive Sciences, Month 2020, Vol. xx, No. xx 5
Measuring Interpersonal Neural Entrainment in Development
So far, we have considered two well-researched topics within early social interaction: ostensive
cues and contingent responsiveness. We have concluded that the unidirectional neuralcorrelates
of both of these –that is, influences of the sender on the receiver, with the two partners consid-
ered independently –are well understood. However, we have also concluded that our under-
standing of how ostensive cues and contingent responsiveness alter the interpersonal neural
dynamics of the interaction –that is, how the partners’neural activity inter-relates –is currently
limited.
In this section, we consider another strand of research that has directly investigated interpersonal
neural dynamics, by measuring interpersonal neural entrainment. In Box 1 we present a formal
definition of entrainment, distinguishing two ways of measuring entrainment: concurrent entrain-
ment (e.g., ‘when A is high, B is high’–also known as synchrony), and sequential entrainment
(‘changes in A are followed by changes in B’). We discuss several crucial methodological caveats
in measuring entrainment. In Boxes 2 and 3 we summarize recent research into interpersonal en-
trainment at the other levels –behavior (Box 2) and physiology (Box 3). As we emphasize in these
boxes, previous research suggests that behavioral and physiological entrainment is not all or
none. Rather, during social interaction, the parent–child dyad oscillates between states of high
Box 2. Behavioral Entrainment
Research into concurrent and sequential behavioral entrainment in parent–child dyads has a long history [100,101], and
includes investigations using both qualitative [102] and quantitative [103] methods. Entrainment has been investigated
at multiple levels of behavior, including:
Vocalizations
Patterns of sequential entrainment during vocal exchanges between adults and infants have been identified at multiple
scales [104,105]. Weaker adult–infant coordination has been associated, for example, with increased attachment problems
and poorer cognitive outcomes [105,106].
Facial Affect
Concurrent and sequential entrainment in facial affect has been identified during tabletop play [19] that changes over time
and differs between fathers and mothers [107]. Stronger child→parent and parent→child influences associate with, for
example, later child self-control [108] and symbolic competence [109]. Not all findings associate increased entrainment
with positive outcomes [110](Box 3).
Eye Gaze
Two principal types of eye gaze entrainment are of interest. (i) Concurrent partner gaze, referred to as mutual gaze. Mutual
gaze during infancy positively correlates with later attention control [111] and is reduced in some atypical dyads [112,113].
(ii) Sequential entrainment in object gaze, referred to as gaze following –that is, that one partner’s look towards an object
forward-predicts the other partner’slook.Althoughgazefollowinghasbeenextensivelyinvestigatedusingsimplified
screen-based paradigms, research suggests that in real-world naturalistic settings (see Figure 1 in main text) infants
actually follow parents’gaze only rarely [114].
Both types of gaze entrainment, concurren t mutual gaze and sequential gaze following, are often combined as joint, or
triadic attention –the three-way sharing of attention between a partner and an object, which involves both mutual gaze
and gaze following [115].
Touch
Current research may overemphasizethe role of gaze during shared parent–child attention and learning: gaze is a predominant
feature of western parent–child interaction, but less so in other cultures [116]. Similar tothe di stinctionbetw eenmutual ga zeand
joint attention, research has examined both touching one another during parent–child interactions [117], and combined touch
to an object [114].
Inducing Behavioral Entrainment
Some research has experimentally induced behavioral entrainment –suggesting, for example, that this can be effective at
promoting shared understanding in adult–child dyads [118,119].
Trends in Cognitive Sciences
6Trends in Cognitive Sciences, Month 2020, Vol. xx, No. xx
and low synchrony [44]. Different types of entrainment are observed at different spatiotemporal
scales, and interactions can show excessive as well as insufficient entrainment.
Research with adults [3,45] and animals [8,46] has also already shown that interpersonal entrain-
ment develops during social interaction, independent of features such as shared entrainment to
common environmental influences (Box 1). Previous research with adults has also suggested that
interpersonal neural entrainment influences learning [47,48].
Interpersonal Entrainment at the Second-to-Minute Scale
Of the range of methods available to researchers for studying in vivo neural activity in infants
and children, the two most commonly used techniques are functional near-infrared
spectroscopy (fNIRS) and EEG. fNIRS, which measures changes in blood oxygenation in
the cortex [49], has a high spatial resolution but a low temporal resolution: the hemodynamic
response lags neural activity by ~2 s and takes ~5 s to reach its peak value [50], meaning
that this technique is best equipped to examine the cofluctuation of brain activity over
timescales of seconds or minutes.
Building on a rapidly emerging body of research in adults [45], a numberof studies have used fNIRS
to examine how brain activity cofluctuates between children and adults during social interaction
(see Figure 1). For example, one study [29] used fNIRS to examine entrainment between 9 and
15-month-old infants and an unfamiliar adult, and to examine how this differed between social in-
teraction and when conducting separate activities in the same room. Concurrent and sequential
entrainment was measured by calculating the cross-correlation in deoxyhemoglobin levels. Rel-
ative to bootstrapping analyses (Box 2), significant concurrent entrainment was observed only dur-
ing interaction in 11 of 57 channel pairs (mainly in frontal areas). Of note, however, the
bootstrapping analyses would not have controlled for shared entrainment to the audiovisual envi-
ronment (Box 1), which was more similar during the interaction condition. Peak associations were
observed with infant brain activity forward-predicting adult brain activity by ~3 s.
Box 3. Physiological Entrainment
Whereas some research into parent–child interactions has examined how, for example, ind ividual heart beats become
coordinated over time [120], most research has studied how autonomic arousal covaries across time windows (both
concurrent, and sequentially –see Box 1). Some research has administered how patterns of change covary within dyads
by administering experimental stressors in the laboratory [121,122]; other research has examined how autonomic arousal
levels cofluctuate in naturalistic, home settings [123,124].
Parents Use Diverse Tactics to Maintain Allostasis
One central aim of social interaction is thought to be to help individuals (particularly young individuals) to maintain a stable
state –a process known as allostasis [75]. How parents respond to changes in their child in order to maintain allostasis is
thought to vary contingent on context. Short-term increa ses in concurrent parent–child physiological entra inment were
observed, for example, nega tive affect vocalizations from the child [123]. When the initial arousal level of the parents is
low, parents increase their own arousal in response to increases in child arousal –matching their own arousal state to their
child’s; but when the initial arousal level of the parent is high, parents respond to increases in the child’s arousal in the op-
posite way –by decreasing their own arousal [123]. This suggests that adu lts use diverse tactics to maintain allostasis
within the dyad [72,75]–dynamically connecting, or disconnecting, their own level of arousal from their child’s[76].
‘There When You Need Me’versus ‘Always On’
Arousal levels in typical dyads do not routinely cofluctuate in naturalistic settings [123]. Instead, typical parents selectively
respond to peak changes in their child’s arousal, but not otherwise. Parents with depression under-respond to peak mo-
ments of child arousal [110](see[125] for comparable neur oimaging results). By contrast, parents with higher anxiety
showed no difference in responsivity to peak child arousal moments but were more responsive to small-scale fluctuations
in their child, and showed higher parent–child entrainment overall [124]. These observations echo similar behavioral findings
[110] and question whether optimal outcomes always associate with greater parent–child entrainment.
Trends in Cognitive Sciences
Trends in Cognitive Sciences, Month 2020, Vol. xx, No. xx 7
Another study [51] used wavelet transform coherence to examine concurrent entrainment in
the 0.02–0.1-Hz range between 5-year-old children and their parents while solving a Tangram
puzzle either together or individually, separated by a screen. They recorded from right and left
frontal and temporoparietal areas and observed entrainment thatwas strongest in the right frontal
and temporoparietal areas. Stronger neural entrainment correlated with increased behavioral
synchrony, and with better problem-solving success while working together. Of note, however,
visual sensory input would also have been more similar during the cooperation condition (see
Box 1 and further discussions below). A third study controlled for this by positioning 5–9-year-old
children and adults facing forwards, in silence, conducting a computer task that involved either co-
operative or competitive behavior (Figure 1A). Using wavelet transform coherence, they observed
greater concurrent entrainment in 0.08–0.5-Hz power fluctuations across left prefrontal and
frontopolar optode pairings between children and their parents only during the cooperative condi-
tion, and not with an unfamiliar adult –even though the conditions were otherwise tightly matched:
auditory and visual information would have been identical between conditions, mutual gaze was not
permitted, and movements were not more synchronous during the cooperation condition [52].
These replicate other findings using the same paradigm [53,54].
Interpersonal Synchrony at the Subsecond Scale
EEG measures electrical brain activity at the subsecond scale but has a low spatial resolution, making
strong conclusions about the neural generators of entrainment hard to draw [55]. Studies using EEG
generally decompose neural activity into frequency bands, the most commonly studied of which in
developmental research are theta (3–6/4–7Hzinyoungchildren/adults)andalpha(6–9/8–12 Hz in
children/adults). Activity in these bands has been associated with attention and learning [56,57].
Building on a large body of dual EEG research with adults [58], one recent study examined the
relationship between social learning and concurrent neural entrainment in adult–12-month-old in-
fant dyads [59]. Parents modeled positive or negative emotions towards objects, and infants’
subsequent choices were examined. Some infants tended to choose the positively modeled ob-
ject, and others the negatively modeled; but for both groups, parent–infant neural entrainment
(phase synchrony in 6–9 Hz, corresponding to the infant alpha band) during teaching predicted
the likelihood of social learning on a given trial. Interpersonal entrainment was most predictive
of learning across central and parietal electrodes [59]. Finally, trial-to-trial increases in interper-
sonal neural entrainment were associated with greater maternal use of ostensive signals such
as eye contact and speech pitch modulation.
Another recent study recorded concurrent and sequential entrainment in neural activity in
8-month-old infants and adults while an adult recited nursery rhymes while alternating between di-
rect gaze and indirect gaze with the infant [60](Figure 1). A control condition, direct-oblique, was
also presented in which adults’face angle was the same as for indirect gaze, but their eyes were
looking at the infant. Recording at the vertex only, separate bidirectional Granger-causal influ-
ences (child→adult and adult→child) were identified during live interaction that were stronger dur-
ing direct and direct-oblique compared to indirect gaze in both theta and alpha bands. The auditory
environment did not differ between conditions, and speech–brain entrainment also did not differ
between conditions, suggesting that entrainment was independent of the shared environment.
Infants who vocalized for longer also had a stronger neural influence on the adult [60].
A third study measured how both concurrent and sequential entrainment differed between parental
positive and negative affect [61]. Graph theory analyses suggested that parents’and 12-month-old
infants’interpersonal neural networks were more closely connected during maternal positive affect,
and that mother to infant directional influences were stronger during positive affect.
Trends in Cognitive Sciences
8Trends in Cognitive Sciences, Month 2020, Vol. xx, No. xx
These studies have shown that fine-grained (subsecond) neural entrainment develops during so-
cial interaction. Next, we consider how this entrainment develops.
How is Entrainment Achieved and Maintained?
Building on work already conducted with adults andanimals,thestudiesreviewedabovehavesug-
gested that aspects of interpersonal neural entrainment develop during early adult–child interaction.
As yet, however, our understanding of how entrainment is achieved and maintained is limited.
Here, building on the discussion of unidirectional influences in the first section of the review, we spec-
ulate about two complementary, but distinct, possible causes of neural entrainment. First, we discuss
how the evoked responses that both children and adults show in response to ostensive cues such as
gaze onsets could involve concurrently phase resetting,leadingtophase entrainment.Inthesec-
ond section, we discuss how actor–observer correspondences could lead to concurrent patterns of
neural activity, potentially causing neural entrainment even in the absence of behavioral entrainment.
Neural Entrainment as a Consequence of Behavioral Cues
Social interactions involve the development of behavioral entrainment, both concurrent and
sequential –including movements, gaze patterns, vocalizations, and facial expressions (Box 2).
Evidence reviewed above also suggests that ostensive cues cause immediate, transient in-
creases in behavioral entrainment. Since brain–behavior correspondences are equivalent across
different individuals, this behavioral entrainment is also likely to cause neural entrainment. Indeed,
some of the studies we reviewed have noted significant correlations between the degree of be-
havioral synchrony observed in dyads, and the neural synchrony observed [51].
The EEG studies reviewed above have, consistent with adult [58] and animal studies [8,46], also
documented phase entrainment during social interaction at much finer timescale (up to 9Hz)
than the second-to-second scale over which behavioral entrainment has been observed. This
more fine-grained entrainment may also, however, have behavioral causes. As described above,
behavioral ostensive cues are known to cause strong neural evoked responses, even in newborn
infants. One possibility is that ostensive cues might operate as ‘edges’in a similar way to the acous-
tic edges (i.e., sharp increases in signal intensity) in the speech amplitude envelope that are known
to drive theta- and delta-rate phase entrainment to cause speech–brain synchrony [62]. Phase re-
setting could take place in both partners to ostensive cues such as gaze onsets and vocalizations,
and this could be one driver that allows phase entrainment to be achieved and maintained
(Figure 3). One prediction that would test this hypothesis would be to assess whether interpersonal
neural entrainment, on both a second-to-second and a subsecond scale, shows transient in-
creases in the time window immediately following ostensive cues (see Outstanding Questions).
Neural Entrainment Arising from Higher-Order Cognitive Processes?
Some of the studies reviewed above [52,54] have, however, also documented neural entrainment
that cannot be explained solely by behavioral entrainment –consistent with recent animal
research that observed neural activity in socially interacting mice under conditions in which
behavioral synchrony and shared entrainment to external sensory input were tightly controlled
for [8]. Using in vivo recordings from populations of neurons in the dorsomedial prefrontal cortex,
the results of this animal research suggested that concurrent entrainment (synchrony) was driven
by behavior-encoding neurons that show overlapping activity when an action is performed by
themselves, and when the same action is performed by a social partner [8] (see section above,
on contingent responsiveness and actor–observer correspondences).
Adult studies have further built on this, by suggestinginadditionthatneural entrainment may reflect
higher-order cognitive processes such as comprehension, engagement, and shared understanding
Trends in Cognitive Sciences
Trends in Cognitive Sciences, Month 2020, Vol. xx, No. xx 9
[63,64]. In one study, for example, concurrent interpartic ipant entrainment in neural activity was
recorded while adult participants listened to a real-life auditory narrative compared to a tempo-
rally scrambled version; interparticipant entrainment was increased in default mode network
areas (including medial prefrontal cortex) when participants had a shared understanding of a
story [65,66]. These results suggest that entrainment is not just a consequence of concurrent
brain–behavior correspondences and sensory cue-based phase-resetting, as described
above; rather, that neural entrainment may also be a consequence of temporally concurrent
patterns of activity-driven shared understanding in addition to shared entrainment to sensory
cues [37,67,68]. As yet, however, no research has investigated this from the perspective of
early learning.
As we describe further in the Outstanding Questions, further work also remains to uncover
whether, and if so how, these separate causes contribute to neural entrainment during early
learning. For example, one area for future investigation is the degree to which interpersonal en-
trainment may potentially affect later stages of information processing more than earlier stages
[38,69]. Research with 5-month-old infants has shown that different ostensive cues (eye gaze
and vocalizations) show differing patterns of activation in low-level processing regions, but over-
lapping patterns of activation in frontal areas [70]. It is possible that similar patterns would be ob-
served for interpersonal entrainment –but no research has yet investigated this.
Is Synchrony Just an Epiphenomenon?
In this final section, we consider whether interpersonal neural dynamics, quantified by measuring
entrainment, are best seen simply as epiphenomena –as secondary consequences of common
entrainment to behavioral cues and of actor–observer correspondences. Alternatively, we hy-
pothesize two possible mechanistic routes through which interpersonal neural entrainment
might play a causal role during early learning.
First, there are inherent differences in brain function between infants and adults: developing brains
are intensely stochastic [71,72], with altered intrabrain connectivity [73] and entrainment to exter-
nal stimuli [74]. One key function of social interaction is allostasis (helping to maintain a stable
state) [75], achieved via bidirectional, dynamical mutual adaptation within the dyad [19,76]. For
Tre nd s
Tre nd s
in
in
Cognitive
Cognitive
Sciences
Sciences
Figure 3. Schematic Illustrating a Mechanistic Role for Interpersonal Entrainment during Early Learning. In a mutual responsive interaction, there is a mutual
timely exchange of phase-resetting cues between partners. Social ostensive signals may act as synchronizing cuesthattriggertransientincreasesin interpersonal
entrainment through phase-resetting, leading subsequently produced maternal speech to arrive at a high receptivity phase for optimal encoding by the infant.
Trends in Cognitive Sciences
10 Trends in Cognitive Sciences, Month 2020, Vol. xx, No. xx
children, then, entrained states might involve more mature patterns of functional activity; transient
phases of child–adult entrainment could thus serve as a transitional stage towards mature
function.
Second, human perception is known to rely on oscillatory processes which shape conscious
experience [77]. Research has suggested that the phase of neural activity at the time of stimu-
lus presentation may relate systematically to the excitability of neural populations and the mag-
nitude of event-related responses [78,79] (although see [80]); accordingly, perceptual stimuli
that are delivered during a high excitability oscillatory phase may be more likely to be detected
and encoded than stimuli that arrive at an inhibitory oscillatory phase [79,81]. During an inter-
action, we may use social cues to nudge our partner into a transient state of entrainment –
such that, for example, parent-initiated mutual gaze might trigger a short-term increase in par-
ent–child phase synchrony. The effect of this would be to ensure that, for the duration of the
existence of a high synchrony state, high excitability oscillatory phases co-occur, thus ensuring
that information (e.g., word labels) is presented at optimal phases for encoding by the receiver
(Figure 3). In the Outstanding Questions section, we outline some predictions to test these
hypotheses.
Concluding Remarks
In this review, we have evaluated the evidence for how social behaviors affect not one, but
both partners in a dyad. We have concentrated on two important and well-researched as-
pects of early social interaction: ostensive cues (signals that are sent during social interaction
to indicate communicative intent) and contingent responsiveness (behaviors that indicate
communicative sensitivity within an interaction). We have concluded that the unidirectional ef-
fects of each –that is, how one partner transiently influences the other, with both partners con-
sidered independently –are well understood. However, we have concluded that our
understanding of how ostensive cues and contingent responsiveness alter the interpersonal
neural dynamics of the interaction –that is, how the partners inter-relate to one another –is
currently limited.
We have also reviewed a smaller corpus of more recent research that has taken a different
approach, by recording from two individuals concurrently during social interaction and mea-
suring interpersonal neural entrainment. We concluded that, consistent with animal and
adult research, this evidence suggests that interpersonal neural entrainment does develop
during social interaction. Building on the discussion of unidirectional influences in the first
section of the review, we have discussed how concurrent and sequential neural entrainment
may arise as a result of two causes: (i) as a consequence of shared entrainment to behav-
ioral cues such as ostensive cues, and (ii) as a consequence of actor–observer correspon-
dences and shared understanding. And we have hypothesized two possible mechanistic
routes through which interpersonal neural entrainment may play a causal role during early
learning.
Our understanding of how early social interaction affects the bidirectional neural dynamics of
the two partners (i.e., how the two patterns of brain activity relate to one another) is still at an
early stage. Many important and fundamental questions remain (see Outstanding Questions).
Perhaps the two most important aspects of the results hitherto are, first, that social influences
affect early learning exchanges at a variety of different temporal scales, including both
subsecond as well as second-to-minute temporal scales; and, second, that these interactions
involve bi-directional sensitivity, with both partners - child, as well as adult - influencing one
another.
Outstanding Questions
Intrabrain entrainment has been
shown to play an important role in
attention and learning; how does
interbrain entrainment relate to
intrabrain entrainment?
Are later stages of stimulus processing
relatively more influenced by
interpersonal influences than earlier
stages?
Do concurrent and sequentialentrainment
reflect the same or distinct phenomena,
in terms of underlying causes and
consequences?
Does interpers onal entrainme nt show
transient increases in the time window
immediately following an ostensive
cue? And are these increases driven
by temporally co-occurring phase re-
setting in response to the ostensive
cue?
Does greater adult–child neural en-
trainment at the time of a learning
event associate with more effective
learning? If so, are differences medi-
ated by an increased likelihood of
learning items (e.g., word labels)
being presented during high excitatory
oscillatory phases?
How does interpersonal entrainment
change over development? Other
aspects of development show a
transition from coregulation (within the
dyad) to self-regulation (within the
individual) over time. Is the same true
for early learning? If so, is interpersonal
neural entrainment more important
during early learning than later on?
Does atypical development manifest
unusual patterns of neural responsivity
and entrainment? Certain clinical
populations show excessive behavioral
and physiological entrainment (Boxes
2 and 3); is more interp ersonal neural
entrainment always better?
Trends in Cognitive Sciences
Trends in Cognitive Sciences, Month 2020, Vol. xx, No. xx 11
Acknowledgments
This research was funded by a Project Grant from the Leverhulme Trust, number RPG-2018-281. We wish to thank Chiara
Bulgarelli, Trinh Nguyen, Carina de Klerk, Vanessa Reindl, and Louise Goupil for reading and commenting on earlier versions
of this manuscript.
References
1. Schilbach, L. et al. (2013) Toward a second-person neurosci-
ence 1. Behav. Brain Sci. 36, 393–414
2. Wheatley, T. et al. (2019) Beyond the isolated brain: the
promise and challenge of interacting minds. Neuron 103,
186–188
3. Redcay, E. and Schilbach, L. (2019) Using second-person
neuroscience to eluc idate the mechanisms of soc ial interac-
tion. Nat. Rev. Neurosci. 20, 495–505
4. Redcay, E. and Warnell, K.R. (2018) A social-interactiveneuro-
science approac h to understanding the dev eloping brain. In
Advances in Child Development and Behavior,pp.1–44,
Elsevier
5. Kuhl, P.K. et al. (2003) Foreign-language experience in in-
fancy: effects of short-term exposure and social interaction
on phonetic learning. Proc. Natl. Acad. Sci. U. S. A. 100,
9096–9101
6. Begus, K. et al. (2014) Infant s learn what they want t o learn:
Responding to infant pointing leads to superior learning.
PLoS One 9, e108817
7. Yu, C. and Smith, L.B. (2016) The social ori gins of sustained
attention in one-year-old human infants. Curr. Biol. 26,
1235–1240
8. Kingsbury, L. et al. (2019) Correlated neural activity and
encoding of behavior across brains of socially interacting ani-
mals. Cell 178, 429–446
9. Dumas, G. et al. (2010) Inter-brain synchronization during so-
cial interaction. PLoS One 5, e12166
10. Csibra, G. and Gergely, G. (2009 ) Natural pedagogy. Trends
Cogn. Sci. 13, 148–153
11. Albert, R.R. et al. (2018) The social func tions of babbling:
acoustic and contextual characteristics that facilitate maternal
responsiveness. Dev. Sci. 21, e12641
12. Lucca, K. and Wilbourn, M.P. (2018) Communicating to learn:
Infants’pointing gestures result in optimal learning. Child Dev.
89, 941–960
13. Begus, K. and Southgate, V. (2018) Curious learners: how
infants’motivation to learn shapes and is shaped by infants’in-
teractions with the social world. In Active Learning from Infancy
to Childhood, pp. 13–37, Springer
14. Murray, L. and Trevarthen, C. (1986) The infant’srolein
mother–infant communications. J. Child Lang. 13, 15–29
15. Stahl, A.E. and Feigenson, L. (2015) Observing the unexpected
enhances infants’learning and exploration. Science 348,
91–94
16. Goupil, L. et al. (2016) Infants ask for help when they know they
don’t know. Proc. Natl. Acad. Sci. 113, 3492–3496
17. Goupil, L. and Kouider, S. (2016) Behavioral and neural indices
of metacognitive sensitivity in preverbal infants. Curr. Biol. 26,
3038–3045
18. Murray, L. et al. (2016) The functional architecture of mother-
infant communication, and the development of infant social ex-
pressiveness in the first two months. Sci. Rep. 6, 39019
19. Feldman, R. (2007) Parent–infant synchrony and the construc-
tion of shared timing; physiological precursors, developmental
outcomes, and risk conditions. J. Child Psychol. Psychiatry
48, 329–354
20. Gratier, M. and Devouche, E. (2011) Imitation and repetition of
prosodic contour in vocal interaction at 3 months. Dev.
Psychol. 47, 67
21. Arco, C.M. et al. (1979) The effect of increased maternal visual
regard on neonatal visual behavior. J. Clin. Child Adolesc.
Psychol. 8, 117–120
22. Senju, A. and Csibr a, G. (2008) Gaze f ollowing in huma n in-
fants depends on communicative signals. Curr. Biol. 18,
668–671
23. Butterworth, G. and Itakura, S. (2000) How the eyes, head and
hand serve definite reference. Br. J. Dev. Psychol. 18, 25–50
24. Young, K.S. et al. (2017) The neural basis of responsive care-
giving behaviour: investigating temporal dynamics within the
parental brain. Behav. Brain Res. 325, 105–116
25. Farroni, T. et al. (2002) Eye contact detection in humans from
birth. Proc. Natl. Acad. Sci. 99, 9602–9605
26. Urakawa, S. et al. (2015) Selective medial prefrontal cortex re-
sponses during live mutual gaze interactions in human infants:
an fNIRS study. Brain Topogr. 28, 691–701
27. Grossmann, T. et al. (2008) Early cortical specialization for
face-to-face comm unication in human infa nts. Proc. R. Soc.
Lond. B Biol. Sci. 275, 2803–2811
28. Zhang, Y. et al. (2011) Neural coding of formant-exaggerated
speech in the infant brain. Dev. Sci. 14, 566–581
29. Piazza, E.A. et al. (2020) Infant and adult brains are coupled to
the dynamics of natural communication. Psychol. Sci. 31, 6–17
30. Striano, T. et al. (2006) Neural mechanisms of joint attention in
infancy. Eur. J. Neurosci. 23, 2819–2823
31. Hoehl, S. et al. (2014) Eye contact during live social interaction
modulates infants’oscillatory br ain activity. Soc. Neurosci. 9,
300–308
32. Grossmann, T. et al. (2013) Brain responses reveal young
infants’sensitivity to when a social partner follows their gaze.
Dev. Cogn. Neurosci. 6, 155–161
33. Bourvis, N. et al. (2018) Pre-linguistic infants employ complex
communicative loops to engage mothers in social exchanges
and repair interaction ruptures. R. Soc. Open Sci. 5, 170274
34. Rayson, H. et al. (2019) Building blocks of joint attention: early
sensitivity to having one’sowngazefollowed.Dev. Cogn.
Neurosci. 100631
35. Goldstein, M.H. et al. (2010) Learning while babbling:
prelinguistic object-directed vocalizations indicate a readiness
to learn. Infancy 15, 362–391
36. Lavelli, M. and Foge l, A. (2013) Interdyad diff erences in early
mother–infant face-to-face communication: Real-time dynam-
ics and developmental pathways. Dev. Psychol. 49, 2257
37. Gallotti, M. et al. (2017) Alignment in social interactions.
Conscious. Cogn. 48, 253–261
38. Hasson, U. and Frith, C.D. (2016) Mirroring and beyond:
coupled dynamics as a gener alized framework for mode lling
social interactions. Philos. Trans. R. Soc. B Biol. Sci. 371,
20150366
39. Meltzoff, A.N. and Marshall, P.J. (2018) Human infant imitation
as a social survival circuit. Curr. Opin. Behav. Sci. 24, 130–136
40. Bernhardt, B.C. and Singer, T. (2012) The neural basis of
empathy. Annu. Rev. Neurosci. 35, 1–23
41. Singer, T. et al. (2004) Empathy for pain involves the affective
but not sensory components of pain. Science303, 1157–1162
42. Leerkes, E.M. et al. (2016) Pathways by which mothers’phys-
iological arousal and regulation while caregiving predict sensi-
tivity to infant distress. J. Fam. Psychol. 30, 769
43. Wass, S.V. et al. (2018) Parental neural responsivity to infants’
visual attention: how mature brains influence immature brains
during social interaction. PLoS Biol. 16, e2006328
44. Gianino, A. et al. (1988) The mutual regulation model: the
infant’s self and interactiv e regulation and coping and defen-
sive capacities. In Stress and Coping across Development
Field, pp. 47–68, Lawrence Erlbaum Associates
45. Gvirts, H.Z. and Perlmutter, R. (2019) What guides us
to neurally and behaviorally align with anyone specific?
ANeurobiological Model Based on fNIRS hyperscanning
studies. Neuroscientist. Published online July 11, 2019.
https://doi.org/10.1177/1073858419861912
46. Zhang, W. and Yartsev, M.M. (2019) Correlated neural activity
across the brains of socially interacting bats. Cell 178, 413–428
47. Davidesco, I . et al. (2019) Brain-t o-brain synch rony predicts
long-term memory retention more accurately than individual
brain measures. bioRxiv 644047
Trends in Cognitive Sciences
12 Trends in Cognitive Sciences, Month 2020, Vol. xx, No. xx
48. Pan, Y. et al. (2018) Int erpersonal sync hronization o f inferior
frontal cortices tracks social interactive learning of a song.
Neuroimage 183, 280–290
49. Reindl, V. et al. (2019) Conducting hyperscanning experiments
with function al near-infrar ed spectroscopy . J. Vis. Exp. 143,
e58807
50. Hong, K.-S. et al. (2015) Classification of prefrontal and motor
cortex signals for three-class fNIRS–BCI. Neurosci. Lett. 587,
87–92
51. Nguyen, T., Schleihauf, H., Kayhan, E., Matthes, D., Vrtička, P.
& Hoehl, S., Th e effects of inte raction qual ity on neural syn-
chrony durin g mother-chil d problem solv ing. Cereb. Co rtex,
124, 235–249
52. Reindl, V. et al. (2018) Brain-to-brain synchrony in parent-child
dyads and the relationship with emotion regulation revealed by
fNIRS-based hyperscanning. Neuroimage 178, 493–502
53. Pan, Y. et al. (2017) Cooper ation in lovers: an f NIRS -based
hyperscanning study. Hum. Brain Mapp. 38, 831–841
54. Miller, J.G. et al. (2019) Inter-brain sy nchrony in mother-child
dyads during cooperation: an fNIRS hyperscanning study.
Neuropsychologia 124, 117–124
55. Noreika, V. et al.14challengesandtheirsolutionsfor
conducting social neuroscience and longitudinal EEG research
with infants. Infant Behav. Dev., 58, 101393
56. Jones, E.J. et al. (2015) Developmental changes in infant brain
activity during naturalistic social experiences. Dev. Psychobiol.
57, 842–853
57. Orekhova, E.V. et al. (20 06) EEG theta rhythm in infant s and
preschool children. Clin. Neurophysiol. 117, 1047–1062
58. Liu, D. et al. (2018) Interactive brain activity: review and prog-
ress on EEG-based hyperscanning in social interactions.
Front. Psychol. 9, 1862
59. Leong, V. et al. (2019) Mother-infant interpersonal neural con-
nectivity predicts infants’soci al learning. PsyArXiv. Published
online July 25, 2019. http://dx.doi.org/10.31234/osf.io/gueaq
60. Leong, V. et al. (2017) Speaker gaze increases information
coupling between infant and adult brains. Proc. Natl. Acad.
Sci. 114, 13290–13295
61. Covarrubias, L.S. et al. (2020) Emotion al valence modu lates
the topology of the parent-infant inter-brain network.
NeuroImage 207, 116341
62. Doelling, K.B. et al. (2014) Acoustic landmarks drive delta–
theta oscillations to enable speech comprehension by facilitat-
ing perceptual parsing. Neuroimage 85, 761–768
63. Fishburn, F.A. et al. (2018 ) Putting our heads together: in ter-
personal neural synchronization as a biological mechanism
for shared intentionality. Soc. Cogn. Affect. Neurosci. 13,
841–849
64. Stolk, A. et al. (2014) Cerebral coherence between communi-
cators marks the emergence of meaning. Proc. Natl. Acad.
Sci. 111, 18183–18188
65. Simony, E. et al. (2016) Dynamic reconfiguration of the default
mode network during narrative comprehension. Nat. Commun.
7, 12141
66. Liu, Y. et al. (2017) Measuring speaker–listener neural coupling
with functional near infrared spectroscopy. Sci. Rep. 7, 43293
67. Nummenmaa, L. et al. (2012) Emotions promote social interac-
tion by sync hronizing b rain activi ty across in dividuals . Proc.
Natl. Acad. Sci. 109, 9599–9604
68. Hu, Y. et al. (2018) Inter-brain synchrony and cooperation con-
text in interactive decision making. Biol. Psychol. 133, 54–62
69. Konvalinka, I. and Roepstorff, A. (2012) The two-brain ap-
proach: how can mutually interacting brains teach us
something about social interaction? Front. Hum. Neurosci.
6, 215
70. Parise, E. and Csibra, G. (2013) N eural responses to multi-
modal ostensive signals in 5-month-old infants. PLoS One 8,
e72360
71. Robertson, S.S. (2014) Empty-headed dynamical model of in-
fant visual foraging. Dev. Psychobiol. 56, 1129–1133
72. Beebe, B. et al. (2016) A systems view of mother–infant face-
to-face communication. Dev. Psychol. 52, 556
73. Grayson, D.S. and Fair, D.A. (2017) Development of large-
scale functional networks from birth t o adulthood: a guide to
the neuroimaging literature. Neuroimage 160, 15–31
74. Vakorin, V.A. et al. (2011) Variability of brain signals processed
locally transforms into higher connectivity with brain develop-
ment. J. Neurosci. 31, 6405–6413
75. Atzil, S. et al. (2018) Growing a social brain. Nat. Hum. Behav.
2, 624–636
76. Fogel, A. (2017) Two principles of communication: co-regulation
and framing. In New Perspectives in Early Commun icative
Development, pp. 9–22, Routledge
77. Buzsaki, G. (2006) Rhythms of the Brain,OxfordUniversity
Press
78. Lakatos, P. et al. (2008) Entrainment of neuronal oscillations as
a mechanism of attentional selection. Science 320, 110–113
79. Busch, N.A. et al. (2009) Th e phase of ongoing EEG oscilla -
tions predicts visual perception. J. Neurosci. 29, 7869–7876
80. Ruzzoli, M. et al. (2019) The relevance of alpha phase in human
perception. Cortex 120, 249–268
81. Mathewson, K.E. et al. (2009) To see or not to see: prestimulus α
phase predicts visual aware ness. J. Neurosci. 29, 2725–2732
82. Hagen, E.H. and Bryant, G.A. (2003) Music and dance as a co-
alition signaling system. Hum. Nat. 14, 21–51
83. Johnson, S. (2009) Emergence: the Connected Lives of Ants,
Brains, Cities, and Software, Simon and Schuster
84. Di Paolo, E.A. (2001) Rhythmic and non-rhythmic attractors in
asynchronous random Boolean networks. BioSystems 59,
185–195
85. Kennedy, J. (2006) Swarm intelligence. In Handbook of Nature-
Inspired and Innovative Computing, pp. 187–219, Springer
86. Engel, A.K. and Sing er, W. (2001) Temporal bi nding and the
neural correlates of sensory awareness. Trends Cogn. Sci. 5,
16–25
87. Calderone, D.J. et al. (2014) Entrainment of neural oscillations
as a modifiable substrate of attention. Trends Co gn. Sci. 18,
300–309
88. Regev, M. et al. (2018) Prop agation of informat ion along the
cortical hierarchy as a function of attention w hile reading and
listening to stories. Cereb. Cortex 29, 4107-4034
89. Giraud, A.-L. and Poeppel, D. (2012) Cortical oscillations and
speech processing: emerging computational principles and
operations. Nat. Neurosci. 15, 511–517
90. Burgess, A.P. (2013) On the interpretation of synchronization in
EEG hypersc anning studi es: a cautiona ry note. Front. Hum.
Neurosci. 7, 881
91. Lachaux, J.P. et al. (1999) Measuring phase synchrony in brain
signals. Hum. Brain Mapp. 8, 194–208
92. Grinsted, A. et al. (2004) Application of the cross wavelet
transform and wavelet coherence to geophysical time series.
Nonlinear Process. Geophys. 11, 561–566
93. Granger, C.W. (1969) Investigating causal relations by econo-
metric models and cross-spectral methods. Econometri ca
424–438
94. Baccalá, L.A. and Same shima, K. (2001) Partial dire cted co-
herence: a new concept in neural structure determination.
Biol. Cybern. 84, 463–474
95. Babiloni, F. and Astolfi,L.(2014)Socialneuroscienceand
hyperscanning techniques: past, present and future. Neurosci.
Biobehav. Rev. 44, 76–93
96. Cohen,M.X. (2014) Analyzing Neural Time Series Data: Theory
and Practice, MIT Press
97. Dustman, R.E. et al. (1999) Life-span changes in EEG spectral
amplitude, amplitude variability and mean frequency. Clin.
Neurophysiol. 110, 1399–1409
98. Shmueli, K. et al. (2007) Low-frequency fluctuations in the car-
diac rate as a source of varianc e in the resting-state fMRI
BOLD signal. Neuroimage 38, 306–320
99. Fairhurst, M. and Dumas, G. (2019) Reciprocity and alignment:
quantifying coupling in dynamicinteractions.PsyArXiv. Published
online 25 March 2019. http://doi.org/10.31234/osf.io/nmg4x
100. Kaye, K. and Fogel, A. (1980) The temporal structure of face-
to-face communication between mothers and infants. Dev.
Psychol. 16, 454
101. Tronick, E. (1982) Social interchange in Infancy: Affect,
Cognition, and Communication, University Park Press
102. Leclère, C. et al. (2014) Why synchrony matters during
mother-c hild inter actions: a sy stematic r eview. PLoS One 9,
e113571
Trends in Cognitive Sciences
Trends in Cognitive Sciences, Month 2020, Vol. xx, No. xx 13
103. Cohn, J.F. and Tronick, E.Z. (1988) Mot her-infant face-to-
face interaction: influence is bidirectional and unrelated to
periodic cycles in either partner’sbehavior.Dev. Psychol.
24, 386–392
104. Trevarthen, C. (1979) Communication and cooperation in early
infancy: a description of primary intersubjectivity.Before
speech: The beginning of interpersonal communication 1
pp. 530–571
105. Jaffe, J. et al. (2001) Rhythms of dialogu e in infancy:
coordinated timing in developme nt. Monogr. Soc. Res. Child
Dev. i-149
106. Beebe, B. et al. (2010) The origins of 12-month attachment: a
microanalysis of 4-month mother–infant interaction. Attach
Hum. Dev. 12, 3–141
107. Feldman, R. (2003) Infant-mother and infant-father synchrony:
the coregulation of positive arousal. Infant Ment. Health J. 24,
1–23
108. Feldman, R. et al. (1999) Mother–infant affect synchrony as an
antecedent of the emergence of self-control. Dev. Psychol. 35,
223
109. Feldman, R. and Greenbaum, C.W. (1997) Affect regulation
and synchrony in mother–infant play as precursors to the de-
velopment of symbolic competence. Infant Ment. Health J.
18, 4–23
110. Granat, A. et al. (2017) Maternal depression and anxiety, social
synchrony, and infant regulation of negative and positive
emotions. Emotion 17, 11
111. Niedźwiecka, A. et al. (2017) Mutual gaze during early mother–
infant interactions promotes attention control development.
Child Dev. 89, 2230–2244
112. Harel, H. et al. (2011) Gaze behaviors of preterm and full-term
infants in nonsocial and social contexts of increasing dynamics:
visual recognition, attention regulation, and gaze synchrony.
Infancy 16, 69–90
113. Field, T. et al. (1990) Behavior-state matching and synchrony in
mother-infant interactions of nondepressed versus depressed
dyads. Dev. Psychol. 26, 7
114. Yu, C. and Smith, L.B. (2013) Joint attention without gaze
following: human infants and their parents coordinate visual at-
tention to objects through eye-hand coordination.PLoS One 8,
e79659
115. Siposova, B. and Carpenter, M. (2019) A new look at joint at-
tention and common knowledge. Cognition 189, 260–274
116. Akhtar, N. and Gernsbacher, M.A. (2008) On privileging the role
of gaze in infant social cognition. Child Dev. Perspect. 2, 59–65
117. Gliga, T. et al. (2018) Social touch: a new vista for developmen-
tal cognitive neuroscience? Dev. Cogn. Neurosci. 35, 1–4
118. Cirelli, L.K. (2018) How interpersonal synchrony facilitates early
prosocial behavior. Curr. Opin. Psychol. 20, 35–39
119. Levy, J. et al. (2017) Perc eption of social syn chrony induces
mother–child gamma coupling in the social brain. Soc. Cogn.
Affect. Neurosci. 12, 1036–1046
120. Feldman, R. et al. (2011) Mo ther and infant coo rdinate heart
rhythms through episodes of interaction synchrony. Infant
Behav. Dev. 34, 569–577
121. Ham, J. and Tronick, E. (2009) Relational psychophysiology: les-
sons from mother–infant physiology research on dyadically ex-
panded statesof consciousness. Psychother. Res. 19, 619–632
122. Waters, S.F. et al. (2 014) Stress contagion: physiolo gical co-
variation between mothers and infants. Psychol. Sci. 25,
934–942
123. Wass, S.V. et al. (2019) Parent s mimic and influence their in-
fant’sautonomicstatethroughdynamicaffectivestate
matching. Curr. Biol. 29, 2415–2422
124. Smith, C.G. et al. (2019) Anxi ous parents show higher
physiological synchrony with their infants. PsyArXiv. Published
online 20 June 2019. http://doi.org/10.31234/osf.io/
m2rsghttp://dx.doi.org/
125. Azhari, A. et al. (2019) Parenting Stress undermines mother-child
brain-to-brain synchrony: a hyperscanning study. Sci. Rep. 9, 1–9
126. Wass, S.V. et al. (2018) I nfants’visual s ustained att ention is
higher during joint play than solo play: is this due to increased
endogenous attention control or exogenous stimulus capture?
Dev. Sci. 21, e12667
Trends in Cognitive Sciences
14 Trends in Cognitive Sciences, Month 2020, Vol. xx, No. xx
