Interperformer coordination in piano-singing duo performances: phrase structure and empathy impact

Abstract

Musicians’ body motion plays a fundamental role in ensemble playing, by supporting sound production, communication, and expressivity. This research investigates how Western classical musicians’ head motion during ensemble performances relates to a piece’s phrase structure and musicians’ empathic perspective taking (EPT) profile. Twenty-four advanced piano and singing students took part in the study, and their EPT score was pre-assessed using the Interpersonal Reactivity Index. High and low EPT duos were formed, and musicians were paired with a co-performer from the same and the other EPT group. Musicians rehearsed Fauré’s Automne and Schumann’s Die Kartenlegerin, and performed the pieces one time before and three times after rehearsal. Motion capture data of the musicians’ front head, audio, and MIDI recordings of the performances were collected and analysed. Similarity in musicians’ head motion and tendency to lead/lag their co-performer were computed by extracting, respectively, power and phase difference of the cross-wavelet transforms of the velocity curves of each paired marker. Results demonstrate that the power of interperformer coordination corresponds to the piece’s phrase levels and that singer’s EPT can impact the leader-follower relationships between musicians, depending on piece and take number. In the Fauré piece, the higher the singer’s EPT score, the higher the tendency for the singer to lead and pianist to follow in take 3, and the lower the tendency for the singer to lead and pianist to follow in take 2. These results contribute to a further understanding of the mechanisms underpinning social interactions, by revealing the complexity of the association between empathy and body motion in ensembles in promoting and diffusing leadership between musicians.

Full text

Vol.:(0123456789)
1 3
Psychological Research
https://doi.org/10.1007/s00426-023-01818-8
ORIGINAL ARTICLE
Interperformer coordination inpiano‑singing duo performances:
phrase structure andempathy impact
SaraD’Amario1 · HaraldSchmidbauer2· AngiRoesch3· WernerGoebl1· AnnaMariaNiemand1· LauraBishop4,5
Received: 29 March 2022 / Accepted: 13 March 2023
© The Author(s) 2023
Abstract
Musicians’ body motion plays a fundamental role in ensemble playing, by supporting sound production, communication,
and expressivity. This research investigates how Western classical musicians’ head motion during ensemble performances
relates to a piece’s phrase structure and musicians’ empathic perspective taking (EPT) profile. Twenty-four advanced piano
and singing students took part in the study, and their EPT score was pre-assessed using the Interpersonal Reactivity Index.
High and low EPT duos were formed, and musicians were paired with a co-performer from the same and the other EPT
group. Musicians rehearsed Fauré’s Automne and Schumann’s Die Kartenlegerin, and performed the pieces one time before
and three times after rehearsal. Motion capture data of the musicians’ front head, audio, and MIDI recordings of the perfor-
mances were collected and analysed. Similarity in musicians’ head motion and tendency to lead/lag their co-performer were
computed by extracting, respectively, power and phase difference of the cross-wavelet transforms of the velocity curves of
each paired marker. Results demonstrate that the power of interperformer coordination corresponds to the piece’s phrase
levels and that singer’s EPT can impact the leader-follower relationships between musicians, depending on piece and take
number. In the Fauré piece, the higher the singer’s EPT score, the higher the tendency for the singer to lead and pianist to
follow in take 3, and the lower the tendency for the singer to lead and pianist to follow in take 2. These results contribute to
a further understanding of the mechanisms underpinning social interactions, by revealing the complexity of the association
between empathy and body motion in ensembles in promoting and diffusing leadership between musicians.
Introduction
Music ensemble playing is a peculiar, complex and naturalis-
tic form of nonverbal joint action that is of scientific interest
to researchers in music psychology, performance science
and neuroscience (Keller etal., 2014). In order to achieve
a cohesive and expressive performance, trained ensemble
musicians in the Western classical tradition coordinate and
adjust their tempo, sound and bodily gestures to align with
that of their co-performer(s), agree on a shared understand-
ing of the composer’s intentions, predict co-performer(s)
individual intentions, and often monitor audience responses.
Musicians’ body motion during ensemble performance is
continuous and multifunctional. Some aspects of motion are
directly involved in sound production; other aspects, often
referred to as “ancillary motion”, support sound-producing
motion, facilitate interpersonal communication and inter-
actions, help with achieving expressive goals, and provide
visual expressive cues to audience/co-performers (Jensenius
etal., 2010). Ancillary motion relates to the performer’s
understanding of the piece’s structural significance, to the
coordination of musical phrases (Thompson & Luck, 2011),
and to expressive intentions (Dahl & Friberg, 2007; Thomp-
son & Luck, 2011). In sum, body motion can help co-per-
formers to achieve their intended musical interpretation and
intentions and ease coordination.
In this study, we analysed interperformer coordination,
as manifested in musicians’ continuous head motion during
duo performances, in relation to the empathic perspective
* Sara D’Amario
damario@mdw.ac.at
1 Department ofMusic Acoustics-Wiener Klangstil,
mdw–University ofMusic andPerforming Arts Vienna,
Anton-von-Webern-Platz 1, 1030Vienna, Austria
2 Shanxi University ofFinance andEconomics, Taiyuan,
Shanxi, China
3 FOM University ofApplied Sciences, Munich, Germany
4 RITMO Centre forInterdisciplinary Studies inRhythm, Time
andMotion, University ofOslo, Oslo, Norway
5 Department ofMusicology, University ofOslo, Oslo,
Norway
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
taking ability of the musicians and the structure of the music
performed.
In the following sections, we discuss the current under-
standing of the role of body motion in ensemble perfor-
mance and the relationships between interperformer coor-
dination and empathy, review methodologies for the analysis
of musicians’ bodily coordination, and then pose hypotheses
for the current study.
Body motion inensemble performance
Musicians’ body motion plays a fundamental role in ensem-
ble playing. Some studies have shown that head movements
can reflect the emotionally expressive intentions that musi-
cians aim to convey. For example, the information flow
between members of a professional trio, as measured in
their anterior-posterior head sway, was found to be higher
when performing with emotional expression than without
(i.e., a mechanical performance) (Chang etal., 2019). Pia-
nists’ head movement velocities have also been found to dif-
fer depending on the pianists’ expressive intent, and to be
higher in expressive serene conditions than in sad, allegro,
and overexpressive conditions (Castellano etal., 2008).
Some emotional intentions that musicians aim to con-
vey during solo performances can also be communicated to
observers through body motion. In an experiment involv-
ing participants’ ratings of silent videos of marimba clips
in which a musician was instructed to express different
intentions, Dahl and Friberg (2007) found that the inten-
tions of happiness, sadness and anger of marimba perfor-
mances were successfully communicated by the musician
to participants. Anger was mostly represented through
jerky movements, happiness was represented through large
movements, and sadness was represented through slow and
smooth movements.
Musicians’ body motion can also reveal leader-follower
relationships between musicians during ensemble playing.
Designated leaders in piano duos tend to raise their fingers
higher than designated followers (Goebl & Palmer, 2009),
anda study of coupling in the body sway of string quar-
tet players shows that assigned leaders influence the others
more and are less influenced by others than are followers
(Chang etal., 2017). Studies on leadership in string quartets
in ecologically more valid situations (i.e., without research-
ers assigning the role of leader) highlight the complexity
and differentiated patterns of dependencies rather than the
more traditional role allocation that attributes leadership to
the first violin (Glowinski etal., 2012; Timmers etal., 2014).
These results demonstrate that musicians’ body movements
reflect musical roles of leader or follower, and suggest that
the leader-follower relationship can impact the way musi-
cians adapt their behaviours to that of the co-performers
during ensemble performance. These findings also imply
that leader-follower roles are flexible and may be exchanged
back and forth during a piece.
Furthermore, body motion in music ensembles can facili-
tate interpersonal communication, coordination and interac-
tion. Certain acceleration patterns in leaders’ head gestures,
comprising the period of deceleration following acceleration
peaks, were found to communicate beat position in piano
duos synchronizing piece entrances (Bishop & Goebl, 2017).
Also found were increases in interperformer coordination,
quantity of head movements and explicit cueing gestures in
piano and clarinet duos during irregularly timed passages
compared to other parts of a piece, demonstrating a tendency
to interact visually during periods of temporal instability
(Bishop etal., 2019). Pianists’ head movements in piano
duos were found to be more synchronized when auditory
feedback was reduced (i.e., both pianists could hear only
themselves though playing together), compared to when des-
ignated leaders could hear only themselves, whilst followers
had full feedback, and also compared to when both pianists
had full auditory feedback. These results demonstrate that
musicians can adapt their body motion as a way to maintain
successful interpersonal coordination if auditory information
is incomplete (Goebl & Palmer, 2009).
Taken together, these studies demonstrate that musi-
cians’ body motion can relate to higher-order piece struc-
ture, reflect the dynamics of nonverbal information flow
and facilitate communication between co-performers during
ensemble performances.
Relationships betweeninterperformer interactions
andempathy
Joint action activities can enhance the capacity to under-
stand what another person is experiencing, which is gen-
erally referred to as “empathy”, a term coined in 1909 by
Edward Titchener as translation from “Einfühlung” (“feeling
into”) used in the German aesthetic theory of the nineteenth
century (Titchener, 1909). Empirical investigations in joint
action activities have recently revealed the bidirectional rela-
tionship between interpersonal coordination and empathy,
which represents a fundamental social-psychological com-
ponent of ensemble playing.
A body of research has analysed whether and how inter-
personal interactions in joint music making enhance empa-
thy related skills in children and adults. Results demonstrate
that long-term participation in group-based, interactive
musical activities increases emotional empathy scores in
school-aged children (Rabinowitch etal., 2012), and sym-
pathy and pro-social skills in children having poor pro-social
skills before the musical intervention took place (Schellen-
berg etal., 2015). Interestingly, it has been demonstrated that
the ensemble experience of college music students sampled
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
in the United States and in South Korea relates to the stu-
dent’s empathy skills (Cho, 2019; Cho & Han, 2021).
In addition to studies analysing the impact of ensemble
playing on empathy, research has also demonstrated the
impact of empathy on joint action. Empathy impacts the
three core cognitive-motor skills (i.e, anticipation, adapta-
tion, and attention) underpinning interpersonal coordina-
tion in expressive ensemble performance (Keller, 2014).
In a multidimensional approach to empathy, the Empathic
Perspective Taking (EPT) trait,1 a component of cognitive
empathy referring to the individual tendency to adopt the
psychological point of view of other(s) (Davis, 1980, 1983),
has been the focus of a number of investigations in ensemble
performance.
Cognitive neuroscience studies, analysing joint-action
in “simulated” piano duos (i.e., pianists who believe that
they are playing along with a second pianist, though per-
forming with a pre-recorded performance), show that higher
empathic perspective-taking scores are correlated with
higher microtiming adaptation (Novembre etal., 2014).
Similar studies in this field have shown that more empathic
musicians rely on motor simulation to a higher degree, since
EPT was found to be positively correlated with neurophysi-
ological measures (i.e., corticospinal excitability recorded
by means of electromyography) indicating pianists’ ability
to represent their co-performer’s actions in their own motor
system (Novembre etal., 2012). A recent study further
investigated the role of EPT during a joint music-making
task, by demonstrating that this promotes interpersonal syn-
chronization accuracy measured at low-order note-to-note
synchronization, and that designated followers with high
EPT scores show greater predictive skills than the low EPT
followers, as they lagged behind leaders to a smaller degree
(Novembre etal., 2019).
In summary, these results demonstrate that empathy
improves synchronization skills. However, these results
have not been corroborated in the context of ecologically
valid ensemble performances. Empathy may also facilitate
coordination at a deep expressive level, leading to effects
on unintentional coordination of musicians’ head motion
in ensembles. Empathy might also impact leader-follower
relationships by promoting followers’ greater abilities in
anticipating leaders’ behaviour, due to the enhancement of
predictive skills and simulation mechanisms. The goal of
our study was to shed some more light on this aspect by
seeking evidence of the impact of EPT on interperformer
coordination of piano singing-duos, operationalised in terms
of musicians’ head motion.
Interperformer coordination measures
Measures of interpersonal coordination are meant to char-
acterise the synchronicity of two person-related time series
of sensor readings. The investigation can be performed in
the time- and/or frequency-domain (Issartel etal., 2015).
Time-lagged cross-correlations methods may allow for an
adequate measurement of the synchrony in applications
to event timing data. However, cross-correlation methods
are prone to producing spurious results when applied to
sensor readings of musicians’ body movements (Dean &
Dunsmuir, 2016): in response to the flow of the musical
score, the sensor data stream is smoothly changing, hence
auto-correlated as well as non-stationary, i.e.with statisti-
cal properties changing over time. Frequency-domain analy-
ses of time-series data are of particular interest to studies
of ensemble performance, in which ensemble musicians’
expressive movements often reflect the hierarchical struc-
ture of the music, defined by subdivisions which are found
within beats, within bars, within phrases, within sections,
within pieces (Demos & Chaffin, 2017; Demos etal., 2017).
A well-known mathematical method for spectrum analy-
sis is the Fourier transform, which computes the power of
individual sinusoidal components. This specific method is
very efficient with interactions relying on stable frequencies
across time, as it assumes the stationarity of the processes in
time (Issartel etal., 2006); however, it can present a practical
limitation in the case of movement interactions in ensemble
performances, where a dominant rhythm cannot always be
set and does not readily translate into fixed-frequency body
oscillations, as faster and slower bodily oscillations can fre-
quently occur.
An alternative method that circumvents the rigid assump-
tions of Fourier analysis is wavelet analysis. It allows for
variable frequencies across time and can thus capture also
intermittent oscillations in the time-frequency domain (Tor-
rence & Compo, 1998) as well as nested rhythmic structures
(Schmidt etal., 2012; Washburn etal., 2014). When applied
to a single time series, the wavelet transform will provide
information about the time-frequency structure of the series;
that is, which frequency is important at what time.
The joint wavelet transformation of two time series yields
the so-called cross-wavelet transform (CWT), which pro-
vides information about which frequency is an important
constituent in both series. Thus, applied to time series of
two musicians’ sensor readings, CWT analysis gives insight
into the strength, or power, of this joint frequency in bod-
ily oscillations. In addition, the oscillations at a joint fre-
quency can be in phase (e.g., both musicians moving forward
and backward in sync), or out of phase (e.g., one musician
1 Davis (1980, 1983) defined four different constructs of empathy,
including Fantasy (the tendency to imagine oneself into fictional situ-
ations), Empathic Concern (the tendency to experience feelings of
sympathy for others), and Personal Distress (the tendency to feel dis-
comfort in response to distress in others) in addition to Perspective
Taking defined above.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
moving forward while the other is moving backward). At any
point in time, CWT analysis allows to measure the degree
of synchronization of joint-frequency oscillations by means
of the phase difference, or phase shift. Therefore, the con-
cept of phase difference will not only indicate whether the
two oscillations are in phase or out of phase, but also which
oscillation is leading (i.e., reaches its peak or trough first,
within a cycle). CWT analysis thus permits the identifica-
tion of patterns of coordination between two musicians and
also provides an indication of the tendency to lead or follow.
The CWT has been found useful in a variety of disciplines,
including geophysics (Grinsted etal., 2004), electroenceph-
alographic studies (De Carli etal., 2004), and also social
psychology measuring interpersonal interactions in dance
settings (Washburn etal., 2014) and between co-actors dur-
ing joke-telling tasks (Schmidt etal., 2014).
A few studies have already demonstrated the potential of
using CWT in the context of music ensemble performances.
Walton etal. (2015) showed different patterns in the lateral
forearms movements of piano players emerging as a func-
tion of the musical context, when piano duos improvised
and played in synchrony with an ostinato backing track.
Specifically, coordination between pianists was stronger
when playing in synchrony rather than improvising, and syn-
chronization power was strongest at the four second period
length, corresponding to the melodic phrase of four ascend-
ing chords repeating every four seconds. This aspect was
further investigated by Eerola etal. (2018) using wavelet and
cross-wavelet transform analysis with computer vision tools;
the researchers demonstrated a range of periodic behaviours
in each performer with frequency peaks that differ for non-
pulsed and pulsed jazz duo performances, being at higher
frequencies (0.75 and 0.40 Hz, faster movements) for the
former, and lower frequencies (0.50 and 0.33 Hz, slower
movements) for the latter. The strength of the interperformer
interactions in the non-pulsed music, as measured by CWT
power, was found to predict audience perception of commu-
nicative interactions between co-performers. In a later study,
Jakubowski etal. (2020) found that synchrony judgments of
jazz duo performances were related to the regularity of the
musical pulse, and synchrony ratings increased when musi-
cians’ periodic movements were at similar frequency bands.
CWT has been also used in the context of Indian clas-
sical instrumental music, and results show that interper-
sonal coordination of movements was greater at metrical
boundaries and mostly related to cadential than other met-
rical instances (Clayton etal., 2019). Furthermore, Dotov
etal. (2021) found that interpersonal synchrony between
audience members, as manifested in their head movements,
was tighter with music higher in groove and when audi-
ence members could see each other rather than in absence
of visual contact, suggesting that social context and musical
features impact how the music is embodied. Taken together,
these studies demonstrate the successful application of CWT
analysis for a better understanding of interpersonal inter-
actions in ensemble playing, by offering new insights into
the correspondence between interpersonal coordination
and musical features such as structural hierarchy and music
styles.
Based on a plurifrequential approach to motion analysis,
we investigated interpersonal movement coordination by
means of cross-wavelet analysis in piano-singing duos. More
specifically, we operationalised the strength of coordination
as the power of common periodic oscillations in musicians’
head motion, and the tendency to lead and follow a co-
performer in terms of the phase difference between these
periodic oscillations. We expected that power and phase dif-
ference at nested time scales, from micro structure (i.e., half
bar) to macro structure (i.e., four bars and form sections),
would reveal the hierarchical nature of head motion in line
with the hierarchical structure of the music and the dynami-
cal aspects of leadership.
The current study
The current study aims to investigate if and how musicians’
body motion coordination relates to the phrase structure of
the piece and musicians’ empathy profile. This study focused
on Lied duos, i.e. piano-singing duos in which the pianist
accompanies the singer performing the melodic line. Lied
duos are small ensembles without a conductor, and are of
particular interest as they act as self-managed teams, with
all ensemble members contributing to the group (D’Amario
etal., 2018a; Timmers etal., 2014; Volpe etal., 2016). The
duos in the current study rehearsed and performed two songs
from the Romantic period. The inclusion of these two pieces
allowed us to investigate the evolution of interperformer
coordination in a context where the relationships between
musicians are dictated by the music to some extent. The
following research questions (RQs) and hypotheses were
formulated:
RQ1 How does head motion coordination relate to musi-
cal structure? We hypothesise that peaks in the similar-
ity between head trajectories correspond to the phrase
level of the piece performed, based on previous studies
highlighting changes in body motion as a function of the
piece structure (Eerola etal., 2018; Jakubowski etal.,
2020; Walton etal., 2015).
RQ2 Does musicians’ head motion reflect leader-fol-
lower roles induced by the structure of the music per-
formed? We hypothesise that singers’ head motion tends
to lead the head motion of pianists. This hypothesis is
motivated by research showing that leading and follow-
ing relationships emerge in the body motion of string
quartet musicians when leadership is assigned and
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
manipulated experimentally (Chang etal., 2017). Our
hypothesis tests whether similar patterns emerge when
leadership is not assigned, but implied by the structure
of the music (in Lied duos, singers mostly perform the
melodic line and pianists play the accompaniment role;
Frăţilă, 2021). We also expect to see that these leader-
follower dynamics are stronger in joint sections than in
prelude/interlude moments where only the pianist per-
forms.
RQ3 How do the strength and leadership dynamics
of musicians’ head motion coordination change with
rehearsal? We expect to see that the strength of their
similarity becomes greater as more practice is acquired
after rehearsing, building on results of previous studies
of a string quartet (Wood etal., 2022). Furthermore, in
line with results of studies investigating body motion in
music duos in the context of rehearsal sessions (Bishop
etal., 2019), we also expect that the overall quantity of
motion, a measure of the energy of the performance, is
greater after rehearsal (when the duo has established
familiarity with the music and each other) than before.
The musicians’ tendency to lead/follow is hypothesised
to be stronger after rehearsal than before rehearsal, as
leadership might emerge with time spent interacting and
rehearsing.
RQ4 Does EPT impact interperformer coordination? We
expect to see that the strength of interperformer coordi-
nation, as measured in musicians’ head motion, is higher
in more empathic musicians, as it has been shown that
empathy promotes note-to-note synchronization skills
in individuals without music training (Novembre etal.,
2019). We expect to show that higher EPT scores of the
singers, who traditionally act as leaders in piano-singing
duos, are hypothesised to be related to a stronger ten-
dency to lead.
Method
Ethics statement
The experiment was conducted in accordance with the Dec-
laration of Helsinki, and ethical approval for the study, with
reference EK Nr: 05/2020, was obtained from the Ethics
Committee at mdw—University of Music and Performing
Arts Vienna (Austria).
Participants
A total of 24 participants (age
M=24.9
years old, SD =
2.3 years old; 14 females, 10 males) took part in the study.
Twelve of them were singing students and 12 piano stu-
dents attending undergraduate and postgraduate courses at
the departments of Piano Performance, Chamber Music or
Vocal Studies and Music Theatre at mdw. They reported
having on average 13.6 years of formal training (SD = 5.6
years) and practicing on average 3.3h/day (SD =1.2h). Par-
ticipants received a compensation of 200 Euros.
Empathy pre‑assessment
Before taking part in the experiment, the empathic perspec-
tive taking (EPT) trait of the participants was pre-assessed.
EPT was measured using the 7-item “Perspective Taking”
sub-scale of the Interpersonal Reactivity Index (IRI) ques-
tionnaire (Davis, 1980, 1983), which is the most widely used
measure of empathy (Grevenstein, 2020). The EPT sub scale
was also provided in a German version (Paulus, 2009) to
those participants preferring the German translation to the
original English questionnaire. Scores for the German ver-
sion were then scaled to match the scores of the English
version.
The mean EPT score for females was 20.6 (SD = 4.1),
and that of males was 19.2 (SD = 3.5). Based on the pia-
nists’ and singers’ median EPT (Mdn = 20, Range = 13–28
for pianists and Mdn = 22, Range = 15–27 for singers),
musicians were then split in low (i.e., below the median)
and high (i.e., above the median) EPT pianist/singer groups.
Then, piano-singing duos were formed by pairing musicians:
• with a similar EPT score within the same EPT group (i.e.,
absolute EPT difference was not larger than two), and
• with a different EPT score from different EPT groups
(i.e., absolute EPT difference
M=6.3
, SD = 2.1),
resulting in the following four groups of 6 duos each:
• LL group (EPT
M=17.1
, SD = 2.3): pianists and singers
from the low EPT group
• HH group (EPT
M=23.9
, SD = 1.9): pianists and sing-
ers from the high EPT group
• LH group (EPT
M=20.4
, SD = 4.3): pianists from the
low EPT group and singers from the high EPT group
(absolute EPT difference
M=7.5
, SD = 1.6)
• HL group (EPT
M=20.6
, SD = 3.8): pianists from the
high EPT group and singers from the low EPT group
(absolute EPT difference
M=6.1
, SD = 2.3)
Materials
Two contrasting pieces for voice and piano from the Western
classical musical tradition were used in the study: Automne
Op. 18 N. 3 composed by Gabriel Fauré (presented in the
original key of b minor and in the transposed key of c sharp
minor) and Die Kartenlegerin Op. 31 N. 2 composed by Rob-
ert Schumann (presented in the original key of E flat major
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
and in the transposed key of F major). Full scores of the
Fauré and Schumann piece in the original key are reported
in Supplementary material Figs.1 and 2, respectively.
Both pieces are from masters of the Lied duo repertoire
of the XIX century, featuring a distinct role distribution
between musicians, with the pianist being the accompa-
niment of the singer performing the melodic line (Frăţilă,
2021). Nevertheless, the pieces differ substantially in
melodic contour and harmonic structure, with the Fauré
piece being slower and with a more stable tempo. The pieces
also differ in their formal structure, being a clear A, B and A′
structure for the Fauré piece and more elaborated (A, A′, B,
A″, C, A′′′) for the Schumann piece (form analysis reported
in Table1a and b, respectively). Whilst the piano in the
Fauré piece always accompanies the singer, the piano in the
Schumann piece sometimes plays the melody alongside the
singer (e.g., bar 17, Supplementary material Fig.2), thus
contributing to the performance of the leading theme. The
Schumann piece features several solo sections (as shown in
Table1a) and tempo changes (e.g., rit. at bars 17–19, 25,
36–38, etc.), which were challenging for duos to coordinate.
The Fauré piece also presents solo moments, but fewer in
number, as shown in Table1b. The Fauré piece comprises
39 bars and the performance was on average 169s long; the
Schumann piece consists of 156 bars and its performance
was on average 201s long. The selection of pieces for the
study was deliberated with the aid of a professional pianist
and a professional singer. They were chosen because they
pose diverse challenges for the performers’ coordination,
while still being feasible to master during a short rehearsal
session. Singers were allowed to choose the key for each
piece that better fits their singing range.
Design
This study used a 2 (rehearsal sessions)
×
2 (pieces)
×
4
(takes, take 0 pre-rehearsal and takes 1, 2 and 3 post-
rehearsal)
×
4 (EPT group)
×
6 (duos) design, comprising a
total of 24 duos and 8 repeated performances recorded per
duo. Take 0, being the first time duos performing together,
and take 3, the last performance after rehearsals, were
expected to show most contrasts in interperformer coor-
dination patterns; take 2 was also analysed to investigate
the consistency of any results found in take 3. Take 1 was
expected to be more similar to takes 2 and 3 than take 0, so
to reduce the number of statistical tests, we excluded Take
1 from our analyses.
The order of rehearsing and performing the two pieces
was counterbalanced across sessions and EPT groups. Meas-
ures of interperformer coordination (i.e., power and phase
difference of the CWT) were the response variables, whilst
empathy measures (i.e., EPT group and musicians’ own EPT
score) were the explanatory variables. Duo number was the
random variable.
Apparatus
The experiment took place in a seminar room of the Depart-
ment of Music Acoustics at mdw—University of Music and
Performing Arts Vienna (Austria). Each pianist performed
sitting at the piano bench and each singer stood in front of
the piano facing the experimenters (and imagined audience),
as if they were performing on stage.
Three sets of data were collected: (1) audio from the
piano, from a head-mounted close proximity microphone
(Sennnheiser ew 100 G2) and two stereo condenser micro-
phones (Neumann KM A P48) providing right and left out-
puts; (2) MIDI from a Yamaha Clavinova; and, (3) body
motion data from an optical motion capture system.
The head-mounted microphone was placed on the cheek
of the singer approximately 2.5cm from the lips. The stereo
microphones pointed towards the pianist and the singer. Four
audio outputs (1 head-mounted microphone, 1 mono audio
from the piano, 2 stereo microphone picking room sound)
fed into a multi-channel audio interface (Focusrite Scarlett
18i8), which was connected to a PC. The 4 outputs, along
with MIDI from the piano, were then recorded using a digi-
tal audio workstation (Ableton Live) as separate tracks at a
sampling frequency of 44.1 kHz and 32-bit depth.
Body motion was recorded at 240 Hz using a 12-cam-
era (Prime 13) OptiTrack motion capture system. Pianists’
body motion data consisted of trajectories from 14 reflective
markers placed on the head and upper body, as follows: 3
markers on the head, 2 on the back, 1 per shoulder, arm,
wrist, and hand, and 1 on the chest. Singers’ body motion
data included trajectories of four extra markers placed
around the hip, for a total of 18 markers.
Audio recordings were synchronized with OptiTrack
recordings using an audiovisual signal produced by a film
clapboard, marked with reflective markers and placed in
view of the motion capture cameras, near to the two stereo
microphones collecting audio from the room. The clapboard
was struck at beginning of each recording, and all record-
ings were synchronized retrospectively to this point. Video
recordings of rehearsal sessions, eye-gaze/pupillometry data
from participants, and breathing and cardiac activity during
the ensemble performances were also collected but are not
part of the current report.
Procedure
Participants were invited to take part in two sessions within
the same week. Each session was about 2h long. At the
beginning of the first session, participants received spo-
ken and written explanations of the research project and
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
the tasks, then they gave written consent to take part in the
study. In each session, they rehearsed each piece for about
10min with their assigned duo partner, and performed
them one time before and three times after rehearsal. Post-
rehearsal performances were blocked by piece; that is, the
duos played three times through one piece, then three times
through the other piece. They had about 1min rest between
performances within blocks and about 2min rest between
blocks.
Musicians were instructed to use the rehearsal time to
prepare a camera-ready interpretation of each piece, and
to give their post-rehearsal performances as though on a
concert stage. These instructions were meant to encourage
naturalistic behaviour from the musicians and high-quality,
musically expressive performances.
Musicians received a digital copy of the scores one week
prior to the experiment to practice and learn the pieces by
themselves, but were not allowed to practice with their
Table 1 Form analysis of the
pieces used in the study: (a) the
Schumann piece and (b) the
Fauré piece
Section Bar n Musician(s)
A 1-19
solo 1-2 pianist
together 3-19 both
A’ 20-39
solo 20-21 pianist
together 22-39 both
B 40-77
solo 40 pianist
together 41-72 both
solo 73-77 pianist
A” 78-96
solo 78-79 pianist
together 80-96 both
C 97-129
together 97-112 both
recitativo 113-121 both
solo 122-129 pianist
A”’ 130-156
solo 130-131 pianist
together 132-151 both
solo 152-156 pianist
(a
)Schumann piece
Section Barn Musician(s)
A 1-14
solo 1-3 pianist
together 4-6.
5b
oth
solo 6.5-7.75 pianist
together 7.75-14 both
B 15-25.5
solo 15-16.7
5p
ianist
together 16.75-25.
5b
oth
A’ 25.5-39
together 25.5-36 both
solo 37-39 pianist
(b)Fauré piece
Labels and bar n of the main sections are presented in italics
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
co-performers outside of the experimental sessions. Co-
performers within the same duo did not know each other
before the experimental session took place, and met for the
first time in the laboratory. This was done so that we could
record changes in the development of interperformer interac-
tions that emerged as a result of their exchanges in the labo-
ratory. Musicians played from scores placed on the music
stand; some of them used the paper copies, others the digital
versions displayed on their own personal tablet. They were
not aware of the purpose of the study and an audience was
never present.
Analysis
Interperformer coordination in piano-singing duo perfor-
mances was measured using cross wavelet transform (CWT)
analysis and investigated in relation to the musical structure
and the empathic profile of the musicians, by implementing
a three-step analytical approach as follows:
1. Musical tempo computation:
(a) Alignment of MIDI and score.
(b) Computation of average inter-beat-intervals (IBI)
from pianists’ MIDI data.
(c) Calculation of mean duration of multiple hierar-
chical phrase levels (i.e., half bar, one bar, two
bars, three bars, and four bars) for each take/piece/
duo.
2. Power and phase difference of head motion coordination
across a broad frequency range:
(a) Extraction of the front head marker position from
the MoCap data.
(b) Computation of 3d velocity trajectories of the
front head markers of each musician.
(c) Extraction of the CWT power and phase differ-
ence on the chosen broad-band.
3. Analysis of musicians’ coordination strength and lead-
ing/lagging behaviours across phrase levels:
(a) Computations of CWT power and phase difference
in selected narrow-bands corresponding to the dif-
ferent hierarchical phrase levels.
(b) Analysis of CWT power and phase difference as a
function of phrase levels, and identification of the
most dominant phrase level, measured as grand
average of CWT power.
(c) Analysis of the impact of piece, take and musical
section on CWT power and phase difference at the
most dominant phrase level.
(d) Analysis of the impact of the empathic profile of
the musicians on mean CWT power and phase
difference, in relation to the whole take, to dif-
ferent music sections, and to solo versus together
moments.
A flowchart of this approach is shown in Fig.1 and the
details are provided step-by-step in the sections that follow.
Step 1: Musical tempo computation
MIDI data from the Clavinova were aligned to digital ver-
sions of the scores using the performance-score matching
algorithm developed by Nakamura etal. (2017) (Fig.1, step
1a). Output was parsed and alignments were corrected using
the Python package “partitura” (Grachten etal., 2019) and
Parangonada interface.2 Alignments were then imported into
Fig. 1 Flow chart of the stepwise analytical framework used for the
analysis of interperformer coordination by means of cross wavelet
transform (CWT) power and phase difference in motion of the front
head markers placed on both musicians. In addition to these steps, the
analysis also included computation of the overall musicians’ quantity
of motion, in order to measure the overall energy of the performance
as related to a total of 14 markers applied to each musician
2 https:// silda ter. github. io/ paran gonada/.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
R, and tempo curves were obtained for each performance by
computing the average onset time per beat and the series of
intervals (in seconds) between each beat onset. Wherever a
beat had no performed onsets (either because the score con-
tained a rest or because the performer missed a note), a lin-
ear interpolation was run to estimate the missing beat times.
An averaged inter-beat interval (IBI) was then computed
per performance as the measure of musical tempo (step 1b).
The meter is 12/8 in the Fauré piece (compound quad-
ruple meter) and 2/8 in the Schumann piece (simple duple
meter). Therefore, for the purpose of analysis, beats were
measured at the 1/8 note level for both pieces (acknowl-
edging that performers may not have felt the beat at this
fine-grained level; see below). The mean IBI for Schumann
across all performances was 0.49s (SD = 0.04s, range =
0.4–0.6s) and the mean IBI for Fauré was 0.38s (SD =
0.03s, range = 0.32–0.5s). Finally, mean duration of dif-
ferent hierarchical phrase levels (half bar, one bar, two bars,
three bars, and four bars) were computed for each piece and
performance based on the mean IBI of that performance
(step 1c). These values are reported in Fig.2.
Step 2: Power andphase difference ofhead motion
coordination acrossabroad frequency range
The analysis of interperformer coordination focused on head
motion, which is closely tied to visual expressivity (Bishop
etal., 2019; Glowinski etal., 2013; Goebl & Palmer, 2009;
Keller & Appel, 2010).
Front head marker position data were smoothed and
velocity was derived using a Savitzky–Golay filter (poly-
nomial order 3, window size 25), through the “prospectr”
package (Stevens & Ramirez-Lopez, 2021) in R (R Core
Team, 2013)—(step 2a). Then, the Euclidean norm of 3D
data was calculated (step 2b). The analysis was based on
velocity, which can be conceptualized in terms of “quantity
of motion” (i.e., displacement per unit of time) or motion
energy or intensity. Velocity is commonly used as a measure
of motion in studies of musical movement (Bishop etal.,
Fig. 2 Display of the phrase lev-
els of the Fauré and Schumann
piece, in relation to the corre-
sponding mean duration, width
of the CWT narrow-band, and
pulse hierarchy. The narrow-
band width was computed based
on the inter-beat-interval (IBI)
calculated for each individual
take/piece/duo
±π
π2
0
−π
2
In phaseOut of phase
Pianist is leading
Singer is leading
Singer is leading
Pianist is leading
Phase differences and their interpretation
Fig. 3 Scheme to interpret phase differences between the pianist and
singer in a duo, referring to the period in question. The period is rep-
resented by the full circle. Phase difference is measured as angles in
the interval from
−𝜋
to
𝜋
. Zero phase difference indicates that the two
corresponding velocity series are perfectly in sync; while the series
move in perfect anti-phase for values of
±𝜋
. The relation of the two
series is still called in-phase or out-of-phase when shifting away from
these two extremes by less than a quadrant, i.e.less than a quarter
of the period. Phase difference values in the intervals
[0, 𝜋∕2]
and
[−𝜋,−𝜋∕2]
indicate that the pianist is leading and the singer is fol-
lowing; conversely, values in the intervals
[𝜋∕2, 𝜋]
and
[−𝜋∕2, 0]
show that the singer is leading and the pianist is following. The R
package “WaveletComp” (Roesch & Schmidbauer, 2018) used for the
analysis received the pianist’s time series as the first input, andthe
singer’s as second input; the roles of “leader” and “follower” swap
with the input order.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
2021; Ragert etal., 2013; Timmers etal., 2014) because,
unlike position, its values do not depend on the specific
orientation of the performer in the motion capture volume,
and it is less affected by noise than are other derivatives of
position (e.g., acceleration and jerk; see van DorpSkogstad
(2014) pp 30–31 for a mathematical explanation).
Next, velocity trajectories were subject to cross-wavelet
transformation (CWT) for each duo and take (i.e., take 0,
before rehearsal, and takes 2 and 3, after rehearsal), using
the R package “WaveletComp” (Roesch & Schmidbauer,
2018) with the complex-valued Morlet wavelet as mother
wavelet. The range of periods to be considered was set in
line with the phrase structure of the pieces, and ranged from
about one half bar to 4 bars. The upper value of this range
was calculated using the slowest mean IBI identified across
duos and takes, whilst the lower value of the period range
was calculated using the fastest mean IBI identified across
duos and takes. Thus, the range for the Fauré piece was from
1 to 26s, and the range for the Schumann piece was from 0.2
to 6.2s. From now on, this broad range of frequencies will
be referred to as “broad-band”.
From the cross-wavelet transformation of 3D veloci-
ties across broad-band frequencies, the power spectrum
was extracted as a measure of the strength of coordina-
tion between musicians in head motion in the time-period
domain (step 2c).3 Phase differences between musicians’
head velocity oscillations were also extracted as a measure
of leading and lagging.This information is given in terms of
angles in a cycle, as shown schematically in Fig.3.
Step 3: Analysis ofmusicians’ coordination strength
andleading/lagging behaviour acrossphrase levels
Using the power and phase information computed for broad-
band frequencies, five non-overlapping “narrow-bands” were
then defined. These narrow bands included a range of fre-
quencies with periods corresponding to approximately one
half bar, one bar, two bars, three bars, and four bars (step
3a). Narrow bands were computed individually for each
duo, piece, and take, using the corresponding average IBI.
A scaled width for each narrow-band was defined for the two
pieces, as shown in Fig.2.
Then, for each period within the narrow-bands, time-
series data of power and phase difference per observation
across the duration of the performance were extracted, and
grand averages were computed across timestamps and period
range for each narrow-band, duo, and take. Two linear mixed
models were implemented to compare CWT power between
narrow-bands (phrase levels) and identify the most dominant
phrase level(s) for each piece and take across the 24 duos,
i.e. the phrase level with the highest CWT power (step 3b).
Eventually, several different models were set up to study
the impact of piece, take, piece section, and the musicians’
empathic profile on CWT power and phase difference, as
measures of interperformer coordination (steps 3c–d).
Robustness checks
Model residuals were found to pass the test for homoskedas-
ticity in all cases, using the R package “DHARMa” (Har-
tig, 2021). However, the results of CWT analysis presented
above are based on a series of decisions (e.g., choice of the
mean rather than the median), which might have given rise
to several issues; these issues are addressed in the Supple-
mentary material(see Supplementary Material Analysis)
with the aim of validating the robustness of the results.
Robustness checks covered alternative velocity data process-
ing and outliers sensitivity of results. In summary, all checks
supported our findings.
Results
This section reports the results of cross-wavelet transform
(CWT) analysis by phrase levels (step 3), aiming at: (1)
analysing CWT power and phase difference by phrase level
(steps 3a-b), (2) analysing CWT output by piece/take/sec-
tion (step 3c), and (3) investigating the impact of the musi-
cians’ empathic perspective taking (EPT) on CWT power
and phase difference (step 3d). The section concludes with
the analysis of the overall quantity of motion of the per-
formance, providing an additional feature of interperformer
coordination, and a report on the outcome of various robust-
ness checks, aiming at validating the results.
Coordination strength andleading/lagging
byphrase level
Step 3a: In order to investigate interperformer coordina-
tion, the CWT power spectrum of the front-head mark-
ers’ velocity trajectories was computed for each duo, take
and piece, within the chosen broad-band of 1–26s for
the Fauré piece and 0.2–6.2s for the Schumann piece.
Figure4 provides an example image plot of the power
spectrum in the time-period domain, as extracted for one
performance of the Schumann piece, together with the per-
formers’ velocity trajectories. While the horizontal axis
monitors the time elapsed (in s), the vertical axis (on a log-
arithmic scale) refers to the chosen broad-band of potential
periodic components in the two trajectories. The color bar
3 For the purpose of this study, we used cross-wavelet power instead
of wavelet coherence, which, in analogy to the concept of correlation,
is normalized to give a quantity between 0 and 1, but requires further
smoothing.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
reveals the gradient of power with which a period occurs
jointly, white contour lines delineate areas of significance.
The plot also features information about the performers’
phase difference in the time-period domain, and thus their
tendency to lead/lag, depicted by arrows according to the
scheme in Fig.3
For each period within the chosen broad-band, CWT
power was averaged across time and analysed in relation
to the individual phrase levels (i.e., half bar, one bar, two
bars, three bars and four bars, computed based on the
respective duo’s own IBI), as shown in the example in
Fig.5. An analysis of the number of occurrences of the
average power peaks (at least out of the three highest)
that fell inside the duration range of one of the individual
phrase levels was done for each duo/take/piece. As shown
in Table2, most of the peaks fell into the half-bar and one-
bar range in case of the Fauré piece, while two bars were
found most prominent in case of the Schumann piece. Few
peaks were also found within the remaining phrase levels.
These findings suggest that interperformer coordination
occurs predominantly within the musical phrase levels.
Step 3b: Next, two linear mixed-effects models (i.e.,
one for each piece) were fitted (estimated using Restricted
Maximum Likelihood, REML, and nloptwrap optimizer)
to measure the impact of phrase levels on power data. This
also identified the phrase level with the strongest CWT
power (i.e., the most dominant phrase level). The strength
of the CWT power might be related to the tempo of the
overall performance, andfor this reason we decided to run
a model for piece, as they differed in average tempo. CWT
power as response variable was aggregated by phrase level
and take per duo. Take number was nested under phrase
level and duo number was entered as random effect. The
power model formula used for each piece was:
For the Fauré piece, the model’s total explanatory power
was substantial (conditional
R2=0.55
) and the part
power ∼ phrase_level∕take.
Fig. 4 Example heat plot of the
CWT power spectrum (bottom)
as computed from the front-
head marker velocities (top) for
a duo performing the Schumann
piece for the third time after a
rehearsal session. Arrows indi-
cate the CWT phase difference,
as described in Fig.3. Vertical
lines identify piece section
boundaries, as listed in Table1.
The segmentation of the CWT
data in line with the score was
mostly based on the MIDI data,
providing note-by-note time
stamps. Unfortunately, eight out
72 MIDI recordings were cor-
rupted; the segmentation in this
case was manually computed
(by the first author) by extract-
ing the relevant time-stamps
from the room microphones
using Audacity (Audacity,
2021)
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
related to the fixed effects alone (marginal
R2
) was 0.43.
Results demonstrate that, regardless of the take num-
ber, CWT power was stronger at the half bar level
(𝛽=9.5, 95% CI [6.9, 12.2],t(343)=7.1, p<0.001)
and
the one bar level
(𝛽=9, 95% CI [6.3, 11.6],
t
(343)=6.6,
p
<
0.001)
compared to the four bar level, set as baseline.
Interestingly, post-hoc comparisons using the Tukey test
indicated that the half bar level was most prominent for take
0 and 2, whilst for take 3 it was the 1 bar level(see Fig.6A,
C, E).
For the Schumann piece, also, the model’s total explan-
atory power was substantial (conditional
R2=0.49
) and
the part related to the fixed effects alone (marginal
R2
)
was 0.37. Results demonstrate that, regardless of the take
number, CWT power was significantly stronger at the two
bar level
(
𝛽
=4.1, 95% CI [2.2, 6],t(343)=4.3, p
<
0.001)
and the three bar level
(𝛽=2.2, 95% CI [0.3, 4.1]
,
t(343)=2.3, p<0.05)
compared to the four bar level,
set as baseline(see Fig.6B, D, F). Post-hoc Tukey com-
parisons demonstrate that for take 3, the two bar level was
more prominent even than the three bar level (see Fig.6F).
Figure6 provides a summary of the impact of phrase
level on CWT power for the Fauré piece (Fig.6A, C, E)
and the Schumann piece (Fig.6B, D, F). Table3 presents
the period range pertaining to the greatest grand mean
power for each take/piece, which from now on will be
referred to as the most dominant phrase level. Taken
together, these results demonstrate that the phrase level
had an impact on the CWT power, which was strongest at
half bar (takes 0 and 2) and one bar (take 3) in the Fauré
piece, and at about two bars in the Schumann piece (all
takes).
To test the effect of phrase level on phase differences
(musicians’ leading/following behaviour), two other linear
mixed models (i.e., one per piece) were fitted. CWT phase
difference data aggregated by phrase level and take per
duo were entered as response variables, with take number
nested under phrase level as fixed effect. The phase differ-
ence model formula was as follows:
For the Fauré piece, the model’s total explanatory power
was weak (conditional
R2=0.13
) and the part related to
the fixed effects alone (marginal
R2
) was 0.02. For the
Schumann piece, the model’s total explanatory power was
weak (conditional
R2=0.02
) and the part related to the
fixed effects alone (marginal
R2
) was 0.02. Results show no
significant effect of phrase level on CWT phase difference,
neither for the Schumann piece nor the Fauré piece. Musi-
cians’ head motion was, on average, in-phase in all takes and
both pieces, as shown in Fig.7 (produced using R package
“circlize”; Gu etal., 2014).
Impact ofpiece, take andsection
Step 3c: Two models tested the impact of piece and take on
mean CWT power and phase difference, respectively. For
CWT power, a linear mixed-effects model was fitted (esti-
mated using Restricted Maximum Likelihood, REML, and
phase_difference ∼ phrase_level∕take.
Fig. 5 Example of a CWT power (red line) averaged across time per
period in the Schumann piece, take 3, plotted against the non-over-
lapping phrase levels of a given duo (i.e. half bar, one bar, two bars,
three bars and four bars duration), estimated based on the individual
inter-beat interval (IBI). Visual inspection of this graph reveals a ten-
dency for CWT power peaks to correspond to the phrase levels, and
for mean CWT power to lower with phrase levels longer than two
bars
Table 2 Occurrence of one of the first three CWT power peaks by
phrase level. The occurrence is expressed as percentage of the total
number of peaks falling in one of the phrase levels
Piece Half bar One bar Two bars Three bars Four bars
Fauré 46 43.5 3.5 3.5 3.5
Schumann 14 23 44 17 4
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
Fig. 6 Grand mean power by
different phrase levels (i.e., half
bar, one bar, two bars, three
bars, and four bars) for the three
takes (i.e., take 0, take 2 and
take 3) of the Fauré piece (A, C,
E) and the Schumann piece (B,
D, F). Horizontal lines display
the significant outcomes from
the Tukey test of equal means;
significance codes:
∗∗∗
0.1%,
∗∗
1%,
∗
5%
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
nloptwrap optimizer), using the R package “lme4” (Bates
etal., 2015). The power model formula was as follows:
The model included CWT power at the dominant period
range identified in Step 3b, aggregated per duo by piece
and take, as the response variable. This allowed us to enter
piece as a fixed effect. To measure the impact of repeated
performances of the same piece, take was also entered nested
power ∼ piece∕take.
under piece. Finally, duo number was entered in the model
as random effect. The model’s total explanatory power was
substantial (conditional
R2=0.32
) and the part related to the
fixed effects alone (marginal
R2
) was 0.03. Neither piece nor
take was a significant predictor of CWT power (Fig.8). In
other words, no significant difference in the strength of the
interperformer coordination was detected across pieces nor
across repeated performances within the same piece.
Table 3 Summary of the
most dominant phrase level
computed for each take and
piece, featuring related phrase
level, and mean CWT power,
mean (SD) signed CWT phase
difference (transformed from
radians into ms), leadership, and
phase direction
Piece Take Most dominant
phrase level
Mean
CWT
power
Mean (SD) CWT
phase difference (ms)
Leadership Phase direction
Fauré 0 Half bar 23.88 396 (539) Pianist In phase
2 Half bar 23.22
−
128.3 (873) Singer In phase
3 One bar 22.65
−
1050 (2175) Singer In phase
Schumann 0 Two bars 22
−
108 (753) Singer In phase
2 Two bars 21.3 87 (719) Pianist In phase
3 Two bars 22.4
−
136 (683) Singer In phase
four
three
two
one
half
±π
π2
0
−π 2
Fauré Take 0
four
three
two
one
half
±π
π2
0
−π 2
Fauré Take 2
four
three
two
one
half
±π
π2
0
−π 2
Fauré Take 3
four
three
two
one
half
±π
π2
0
−π
2
Schumann Take 0
four
three
two
one
half
±π
π2
0
−π
2
Schumann Take 2
four
three
two
one
half
±π
π2
0
−π
2
Schumann Take 3
Fig. 7 Grand mean CWT phase differences between pairs of musicians’ head velocity trajectories
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
For phase difference, a linear model was fitted (estimated
using Ordinary Least Squares, OLS), to predict the effect of
take, nested within piece, on CWT phase difference at the
dominant phrase level (i.e., CWT power), aggregated per duo
by piece and take. The phase difference model formula was
as follows:
In this model, duo number was initially entered as random
effect, but dropped because the variance was negligible
(
<0.001
), and including it in the model generated a singular
error. Thus, a linear model was chosen and run using the lm
function from the “stats” package, which is basic R function-
ality. The model explained a moderate proportion of variance
(R2=0.10, F(5, 138)=3.13, p=.010, adj. R2=0.07)
.
The tendency for the singer to lead was weaker in
the Schumann piece
(𝛽=−0.14, M=−0.1, 95% CI
[−0.28, −0.03],t(138)=−2.12, p
<
0.5)
, than in the
Fauré piece
(M=−0.3, 95% CI [−1.14, −0.04])
. Take
number predicted CWT phase difference depending on the
piece being performed: it was not a significant predictor
in the Schumann piece, but it was a significant predictor
in the Fauré piece. As shown in Fig.8, the tendency for
the singer to lead and pianist to follow was higher in the
last take
(𝛽=−0.25, M=−0.15, 95% CI [−0.39, −0.12],
t(138)=−3.74,
p
adj
<
0.001
), than in the first take
(M=0.1, 95% CI [−1.59, −0.49])
. Thus, the leader-follower
relationships between co-performers differed depending on
the piece being performed and across repetitions. Singers
tended to lead and pianists tended to follow more pronounc-
edly in the Fauré piece than the Schumann piece, and more
pronouncedly in take 3 than take 0 of the Fauré piece.
phase_differ ence ∼ piece∕take.
In addition to the last two analyses conducted on the
take as a whole, CWT power and phase difference were
also analysed as a function of the solo versus together sec-
tions for each piece/take at the dominant phrase level, using
a model for each CWT measure. The power and phase dif-
ference model formulae were, respectively, as follows:
Solo versus together sections predicted CWT
power in all takes of the Fauré piece, and CWT
power was higher in the together sections (take 0:
𝛽=6, M=26, 95% CI [1.9, 11.8],t(140)=2.8, p<0.01
;
take 2:
𝛽=11.1,
M
=27, 95% CI [7.1, 15.2],
t
(140)=5.4,
p
<
0.001
; take 3:
𝛽=9.4, M=25.76, 95% CI [5.2, 13.6],
t(140)=4.4, p
<
0.001
) than the solo parts (take 0:
M=19.8, 95% CI [0.1, 0.8]
; take 2:
M=15.9, 95% CI
[0.5, 1.1]
; take 3:
M=16.2, 95% CI [0.4, 1]
). However, solo
versus together parts did not predict CWT power in any of
the Schumann takes. Interestingly, results demonstrate that
phase difference did not change across solo/together sec-
tions, regardless of the piece and take, demonstrating that
the leader–follower relationships between musicians were
unchanged during the piano solo sections.
In summary, these results indicate that the tendency to
lead/lag a co-performer was impacted by the piece being
performed and the repeated performances. The tendency for
the singer to lead and the pianist to follow was stronger in
the Fauré piece (more rhythmically stable), and leadership
patterns were stronger at the ending of the laboratory session
than before rehearsing in the case of the Fauré piece. The
strength of interperformer coordination, as measured by the
power ∼ solo_together
phase_differ ence ∼ solo_together
Fig. 8 Estimates of piece and
take effects (the latter nested
under piece) related to mean
CWT power (left side plot)
and phase difference (right
side plot). Piece estimates are
given above with reference to
the specified base level of the
factor (i.e., Schumann versus
Fauré, take 2 versus take 0
and take 3 versus take 0). The
Tukey method has been used for
adjusted p values.Significance
codes:
∗∗∗
0.1%,
∗∗
1%,
∗
5%
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
CWT power, did not change across piece or take, but was
stronger in the together sections than in the piano solos of
the three takes of the Fauré piece.
Impact ofempathy oninterperformer coordination
Whole piece
Step 3d: Having shown that coordination strength (as meas-
ured by CWT power) and leader-follower relationships (as
measured by CWT phase difference) depend on the cho-
sen piece and take, generalized linear models were imple-
mented to investigate the impact of empathic perspective
taking (EPT) profile of the musicians on motion coordina-
tion. A total of 24 models were fitted: 1 per take/piece (thus
6 in total), per CWT measure (i.e., power and phase dif-
ference) and EPT measure (i.e., EPT group and musician’s
own EPT score). We decided to investigate the effects of
empathy on each take and piece (rather than nesting take
under piece) since each performance might be musically
unique, and therefore the effects of empathy might change
between pieces and takes. Average CWT power and phase
difference of the most dominant phrase level were entered
in the models as response variables, whilst EPT group (i.e.,
LL, HH, LH, and HL) and musician’s own EPT score (i.e.,
singer’s EPT score and pianist’s EPT—discrete variables)
were entered, once at a time, as explanatory variables. A
Bonferroni correction was implemented for multiple linear
models, dividing the critical p value level (0.05) by the num-
ber of comparisons being made, 24, corresponding to the
total number of models developed for the impact of empathy
on interperformer coordination. For this reason, a p value
threshold was set at 0.002083. The model formulae related
to power and EPT and musician’s own EPT score were,
respectively, as follows:
The model formulae related to phase difference and EPT and
musician’s own EPT score were, respectively, as follows:
Results demonstrate that there was no significant relation-
ship between CWT power and the different EPT measures
(i.e., EPT group and musician’s own EPT score), regardless
of the piece being performed or the take number. Results
also demonstrate that EPT group was not a significant pre-
dictor, regardless of the piece performed and the take num-
ber. Interestingly, singer’s EPT score had an impact on CWT
phase difference of the last repeated performance of the
Fauré piece: the higher the empathy score of the singer, the
higher the tendency was for the singer to lead and pianist to
follow
(𝛽=−0.05, 95% CI [−0.08, −0.02],
t
(21)=−3.95,
p
<
.001)
. In other words, in the last take of the Fauré piece,
we found a significant relationship between singers’ empa-
thy profile and the tendency to lead/follow, with a 0.05 unit
increase in the tendency for the singer to lead and pianist
to follow per unit of increase in singer’s EPT, as shown in
Fig.9. The model explains a substantial proportion of vari-
ance
(R2=0.44, F(2, 21)=8.30, adj.R2=0.39)
.
This analysis of the impact of the musicians’ own EPT
score on CWT power and phase difference was based on
the most dominant phrase level. As explained above, results
demonstrate that empathy did not predict CWT power, but
did predict CWT phase difference in case of the Fauré piece,
take 3. But, an analysis of the occurrences of the first three
power peaks by phrase level (see Table2 and Supplemen-
tary materialanalysis, Fig.1) also highlighted that some
peaks fell within less dominant phrase levels. It was there-
fore of interest to investigate the impact of musicians’ EPT
score also based on the less strong phrase levels. For this
reason, step-wise linear models were implemented (one for
each piece and take) to investigate the impact of empathy on
CWT power and phase difference data related to each phrase
level. As shown in Table4, empathy did not predict CWT
power ∼ EPT_group
power ∼ Pianist_EPT + Singer _EPT
phase_differ ence ∼ EPT_group
phase_differ ence ∼ Pianist_EPT + Singer _EPT
Fig. 9 Tendency to lead/lag as measured by the CWT phase differ-
ence in the Fauré piece, take 3, plotted as a function of the musician’s
own empathic score, based on a multiple linear regression model. The
blue line shows the association between the explanatory variables
(i.e., pianist’s EPT score on the left side plot and singer’s EPT score
on right side plot) and the response variable (mean CWT phase dif-
ference), while holding the value of the other predictor variables con-
stant (singer’s EPT and pianist’s EPT, respectively). The slope angle
represents the
𝛽
coefficient. As we can see, results demonstrate that
the higher the EPT score of the singer, the higher the tendency was
for the singer to lead and pianist to follow; in contrast, the EPT score
of the pianist did not impact the leader-follower relationships
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
power of any of the less dominant phrase levels, in line with
the results for the most dominant phrase level. Interestingly,
the singer’s EPT score did predict CWT phase difference
at three bar level in the Fauré piece, take 2: the higher the
singer EPT, the lower the tendency for the singer to lead
and pianist to follow. These results are in contrast with the
direction of the empathy impact that was found in Fauré,
take 3, in which a higher singer empathy was associated
with a higher tendency for the singer to lead and the pianist
to follow.
Musical sections
Step 3d: Having found an impact of the singer’s empathy
on the CWT phase difference in the case of the Fauré piece,
take 3, a follow-up analysis was run on each take and piece
as a function of the musical sections forming the piece (i.e.,
Table 4 Impact of musician’s EPT score on CWT power (top) and phase difference (bottom) for each take/piece/phrase level, as measured by
sign and level of significance of the
𝛽
coefficients resulting from the linear regression models implemented
Power
PieceTake Musician Half bar One bar Two bars Three barsFour bars
Fauré
0 Pianist −+−− −
Singer ++−− +
2 Pianist −+−− −
Singer −++ ++
3 Pianist −−− +−
Singer +−− ++
Schumann
0 Pianist ++
++ +
Singer ++
++ +
2 Pianist ++++ +
Singer −−
−− −
3 Pianist +−++ −
Singer −−
+−+
Phase difference
PieceTake Musician Half bar One bar Two bars Three barsFour bars
Fauré
0 Pianist −−++ +
Singer +−+−−
2 Pianist −−−− −
Singer +−++
∗∗∗ +
3 Pianist −−− −−
Singer +−
∗∗∗
−− −
Schumann
0 Pianist +−+−+
Singer ++
++ +
2 Pianist −+ + −−
Singer −+−++
3 Pianist +− − ++
Singer ++ ++
−
+
indicates a positive relationship between empathy and CWT power/phase difference, whilst − indicates a negative relationship. Grey cells refer
to the most dominant phrase level for a given piece and take. Significance code:
∗∗∗0.1%
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
A, B and A’ sections for the Fauré piece and A, A’, B, A”,
C, A”’ for the Schumann piece; form analysis reported in
Table1a and b, respectively). This was done to investigate
whether the impact of empathy on CWT phase difference
would change during the course of the piece.4 For this pur-
pose, six linear nested models were fitted to predict the
impact of the musical section nested under the singer’s EPT
score on mean CWT phase difference. CWT phase differ-
ence data, corresponding to the most dominant phrase level
identified above, were entered in the models as mean values,
aggregated per sections. The phase difference model formula
was as follows:
As shown in Fig.10, the direction of the influence of
the singer’s EPT in take 3 of the Fauré piece was the
same across the three sections: the higher the singer’s
EPT, the higher the tendency for the singer to lead and
pianist to follow in sections A, B and A’. A similar ten-
dency was found above when considering the whole take.
Interestingly, this tendency was greater in the first sec-
tion (i.e., A)
(95% CI [−0.18, 0.61]
, than its repetition
(i.e., A’)
(𝛽=0.02, 95% CI [8.55
e
−03, 0.04],
t
(138)=3.21,
p
<
.01)
. The impact of the singer’s EPT on CWT phase dif-
ference did not differ significantly between sections A and
B. The model explains a moderate proportion of variance
(R2=0.14, F(3, 138)=7.19, p<.001, adj.R2=0.12)
.
Results from the models run on the remaining takes
demonstrate that musicians’ own EPT did not predict CWT
phase difference, regardless of the musical sections. These
results are also consistent with the findings resulting from
the analysis of the whole piece.
Furthermore, in light of the presence of solo moments
focused on piano performance and contrasting with together
sections, the analysis of the impact of empathy was con-
ducted also as a function of the solo/together moments.
To this end, six linear nested models (one per take/piece)
were fitted to predict the impact of the solo/together section
nested under the singer’s EPT score on mean. The CWT
phase difference model formula was as follows:
Results demonstrate that the impact of singer’s EPT as
function of the type of performance (i.e., solo vs together)
on the phase difference, i.e. the tendency to lead/lag, was
consistent. In particular, for the Fauré piece, take 3, the ten-
dency for the singer to lead and pianist to follow with more
empathic singers did not change significantly between solo
versus together moments, as shown in Fig.11.
phase_differ ence ∼ Singer_EPT∕sect ion.
phase_differ ence ∼ Singer_EPT∕solo_t ogether
Fig. 10 Scatter plot of the tendency to lead/lag as measured by the
CWT phase difference in the three musical sections (i.e., A, B, and
A’) of the Fauré piece, take 3, plotted as a function of the singer’s
empathic score. The slope represents the
𝛽
coefficient. Interestingly,
results demonstrate that the tendency for the singer to lead and pianist
to follow in presence of more empathic singers was greater in the first
section of the piece (i.e., section A), than after its repetition after the
B section
Fig. 11 Scatter plot of the tendency to lead/lag as measured by the
CWT phase difference in sections where the pianist played solo ver-
sus sections where they played together. The graph relates to the third
performance of the Fauré piece and CWT phase difference is plotted
as a function of the singer’s empathic score. The slope represents the
𝛽
coefficient with error bars plotted over
4 The analysis of the impact of EPT on CWT power by section was
not conducted, since there was no impact of individual EPT score nor
of EPT group on CWT power in the whole take in any of the pieces
and takes.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
Quantity ofmotion
In addition to the analysis of CWT power and phase differ-
ence of musicians’ front-head velocity trajectories, we also
measured the overall quantity of motion (QoM) in perfor-
mances across the 14 markers that were placed on musi-
cians’ head and upper body. The extra four markers placed
on the singer’s hip were not included in this analysis. QoM
was computed as sum of 3D velocities5 across markers and
performers per second for each duo, and then compared
across pieces and repeated performances.
Two models were implemented to test the fixed effects of
EPT measures (EPT group and musicians’ own EPT score,
tested in separate models), take nested under piece on QoM
(the response variable). Duo number was entered in both
models as random effects to account for repeated measures.
QoM values were aggregated by performance. The QoM
model formulae were as follows:
The analysis of changes of QoM by piece and take
based on linear mixed-modelling showed that take
and piece predicted QoM, which was higher in Schu-
mann than Fauré
(𝛽=148.7, 95% CI [140.8, 171.9]
,
p<.001)
, and increased across repeated performances.
For the Schumann piece, QoM was higher in take 2
(𝛽=85.7, 95% CI [49.4, 122],p<.001)
and take 3
(𝛽=87.8, 95% CI [51.5, 124.2],p<.001)
than in take 0,
demonstrating that body motion increases with more prac-
tice. For the Fauré piece, QoM was also higher in take 2
(𝛽=75.4, 95% CI [39.1, 111.7],p<.001)
and take 3
(𝛽=69.2, 95% CI [32.8, 105.5],p<.001)
than in take 0,
demonstrating that body motion increases with more practice
also in the Fauré piece. Conversely, EPT group and musi-
cians’ own EPT did not predict QoM, demonstrating that
EPT is irrelevant for the overall energy of the performance.
Discussion
This study focused on how head motion of piano-singing
duos relates to the phrase and formal structure of the piece
being performed, and musicians’ empathic perspective tak-
ing (EPT). Musicians’ head motion was investigated while
QoM ∼ Pianist_EPT + Singer_EPT + piece∕take
QoM ∼ EPT_group.
performing two pieces (one from Schumann and one from
Fauré) that contrasted in terms of tempo and structural fea-
tures, including the clarity of the distinction between melody
and accompaniment, before and after rehearsal. By applying
cross-wavelet transform (CWT) analysis, we quantified the
degree of similarity and phase difference in common perio-
dicities between head velocity trajectories, as measures of
the power of the interperformer coordination and the ten-
dency to lead/lag the co-performer, respectively.
The main results can be summarised as follows:
• Peaks of interperformer coordination power predomi-
nantly corresponded to the phrase structure of the piece,
demonstrating that musicians tend to jointly align their
periodic head motion with a subdivision of phrase length.
• The singer’s EPT score was positively or negatively
related to the singer’s tendency to lead and the pianist’s
tendency to follow after a rehearsal session, depending
on the take, and only for the Fauré piece (the slower and
structurally simpler piece), revealing the complexity of a
direct link between empathy and musicians’ head motion.
Interperformer coordination bypiece andtake
Research has recently analysed how musicians’ movements
during performance relate to the piece being performed
(Clayton etal., 2019; Demos & Chaffin, 2017; Demos
etal., 2017; Eerola etal., 2018; Thompson & Luck, 2011),
by establishing a link between movements and music struc-
ture. Through the application of CWT theory to motion cap-
ture data, we provided a novel contribution to the field of
music performance science, by quantifying the relationship
between head motion and the phrase structure of the piece
for two selections of Romantic classical repertoire. We found
that a majority of the three most salient CWT power peaks
fell within one of the chosen phrase levels, demonstrating, as
expected (see RQ1), that interpersonal coordination power
of head motion corresponds to the phrase structure of the
piece.
We also found that the coordination power did not change
across pieces. However, future investigations replicating this
study are needed to test these results across a broad selection
of pieces and verify whether advanced musicians can com-
pensate for differences between pieces and maintain a simi-
lar strength of interactions, regardless of the piece features.
Interestingly, the averaged quantity of body motion
increased with successive takes, but the strength of inter-
action, contrary to our expectations, did not differ across
takes. This means that musicians moved more across takes
but the way they periodically related to each other remained
the same, which also demonstrates that quantity of body
motion and periodic coordination of head motion are two
different factors of interpersonal coordination. Furthermore,
5 The raw position data of the 14 markers applied to the head and
upper body of each musician were smoothed using a Savitzky–Golay
filter and then velocity derived, as was done for the calculation of
CWT power and phase difference of the front head-marker velocity
series.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
these results suggest that the time musicians spent together
rehearsing and negotiating a common interpretation might
be irrelevant for the similarity of head motion. There might
still be differences in terms of sound synchronization accu-
racy as result of practicing together, as studies on ensemble
synchronization suggest (D’Amario etal., 2018a); future
investigations might analyse note-to-note synchronization
in parallel to head and body motion analysis to shed more
light in this respect.
Our results demonstrate that leadership, as manifested
in musicians’ head motion, is shaped in large part by piece
structure, regardless of musicians’ empathy (RQ 2). The
overall tendency for the singer to lead and the pianist to
follow that was found in the study is in line with the more
traditional distribution of leader-follower roles expected
in a Lied duo. But, variations in leadership intensity and
direction observed between pieces, takes and musical sec-
tions within pieces do provide evidence that leadership in
ensembles represents a dynamic concept rather than a more
straightforward role division. This view is very much in line
with studies analysing leadership in string quartets, showing
complex and very well differentiated patterns of dependen-
cies between musicians (Timmers etal., 2014).
A stronger tendency for the singer to lead and the pianist
to follow in the Fauré piece than in the Schumann piece
could have been induced by the pieces’ features: The piano
part is more clearly presented as an accompaniment in the
Fauré piece. Conversely, the pianist in the Schumann piece
plays also a leader role during solo sections and when per-
forming in unison segments of the singer’s melody. Interest-
ingly, the strength of leadership tendency did not change
within pieces during solo and together sections regardless of
the piece and take, contrary to our expectations (see RQ2),
whilst the interaction power was, as expected, stronger in
together sections than piano solos. These results suggest that
singers can play a leader role even when not required to sing
notes during the short prelude and interludes of the piece.
This can be understood by the role that these solo sections
play: they remain part of the duo piece, implying a continu-
ous performing and acting of both musicians, including that
of singers during prolonged rests.
We observed stronger leadership in the last take of the
Fauré piece after rehearsal compared with that of the first
take before rehearsal (RQ3). These results corroborate
evidence of a longitudinal study observing changes in
the leader-follower relationships emerging spontaneously
between members of a semi-professional singing quin-
tet during five rehearsal sessions across a term of study
(D’Amario etal., 2018a). Interestingly, D’Amario etal.
(2018a) found evidence of a more democratic approach to
leadership by the last rehearsal session, in which leadership
became equally shared among the five singers, compared
with earlier sessions. Conversely, we observed a stronger
tendency for the singer to lead and the pianist to follow in
Fauré take 3 than take 0. These apparently contrasting results
could be understood in light of differences in piece features:
D’Amario etal. (2018a) made use of two Bach’s chorales
arranged for the study so no clear leader-follower roles were
markedly induced by the pieces and leadership could emerge
spontaneously during rehearsals; we, in the current study,
used a Fauré piece with a clear definition of leader-follower
roles. Taken together, these results suggest that leadership in
ensembles evolves in the course of practicing in the direction
the music piece implies. Future studies are needed to test this
hypothesis across a broad selection of pieces.
Ultimately, by means of CWT phase difference, we quan-
tified the accuracy of interperformer coordination of head
motion, observing that signed asynchrony between musi-
cians’ head motion was on average between about 400 ms
and 1s for the slower Fauré piece, and between 90 and 140
ms for the faster Schumann piece (see 3). The standard devi-
ation of the signed asynchronies was between 500ms and
2s for the Fauré piece, and about 700 ms for the Schumann
piece. Research analysing interperformer synchronization
in Western Classical music at low-order note-to-note coor-
dination reports typical standard deviations of signed asyn-
chronies (and mean absolute asynchrony) between 30 and
50 ms, and mean signed asynchronies mostly close to 0ms
(Bishop & Goebl, 2015; D’Amario etal., 2018b; Goebl &
Palmer, 2009; Timmers etal., 2014). Taken together, these
results imply that the magnitude of interperformer note-to-
note coordination tends to be smaller than that of the head
motion coordination. This is not surprising in light of the
fact that note synchronization in Western Classical music is
highly intentional, whilst head motion is not.
Singer’s empathy impact oncoordination power
andleadership
A growing body of research on empathy and music is high-
lighting how joint musical activities facilitate empathy
(Cho, 2019; Cho & Han, 2022; Rabinowitch etal., 2012;
Schellenberg etal., 2015). More recently, Novembre etal.
(2019) expanded on this by demonstrating how EPT scores
of untrained musicians can promote accuracy in synchro-
nizing their actions during a joint music-making task, in
which participants were required to play synchronously two
different streams of musical sounds by rotating an electronic
music-box.
Our study expands the work of Novembre etal. (2019),
by analysing the impact of empathy in the context of natu-
ralistic music ensemble performances, where the complex-
ity of the music and performers’ social and musical roles
in the ensemble may affect how relevant empathic abilities
are, and what effect they have on ensemble interactions. We
found no evidence of a relationship between empathy and
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
interperformer coordination power of head motion trajec-
tories, regardless of the piece performed and the take num-
ber (RQ4). Participants in our study were advanced trained
musicians, with extensive experience in ensemble settings
working with numerous other musicians with, presumably,
different personalities. Therefore, they might have been
skilled in adapting to the personality and cognitive differ-
ences that they might find amongst their co-performer(s).
Overall, these results suggest that empathy can impact
interperformer synchronization skills manifested at low-
order note-to-note temporal accuracy, as found in Novembre
etal. (2019), but might not reflect higher-order interper-
former synchronization power of head motion. Future stud-
ies might corroborate these results by analysing simultane-
ously the impact of EPT on interpersonal synchronization of
sounds and body motion between musicians in ensembles.
We also found that the singer’s EPT impact on leader-
ship depended on the piece being performed: empathy influ-
enced leadership in the Fauré piece, but not in the Schu-
mann piece (RQ4). The piano part of the Fauré piece plays
a clear accompaniment role featuring a constant ternary pat-
tern supporting the melodic line of the singer. Conversely,
the piano part of the Schumann piece is less repetitive and
includes instances doubling the melodic lines of the singer
(see Supplementary material Fig.2, bar 17). This more dis-
tinct distribution of leader and follower roles induced by
the Fauré piece compared with the Schumann piece might
have promoted the stronger tendency for the singer to lead
and the pianist to follow in the Fauré piece as argued above,
and might also explain differences in the empathy impact
on leadership between pieces. These results suggest that the
impact of empathy on leader–follower relationships might
depend on the overall leadership distribution induced by the
piece. Future investigations on the impact of empathy as a
function of the intensity of the leader–follower roles induced
by pieces, in which leadership is manipulated accordingly,
might shed more light in this respect. Furthermore, the fact
that in Lied-duos singers mostly perform the melodic line
might explain the reason why the empathy impact on lead-
ership was promoted by the singer and not the pianist, who
plays and is expected to act as the accompaniment at stage
(Frăţilă, 2021).
Interestingly, we found that the direction of the singer’s
empathy impact on leadership differed between takes: the
higher the singer empathy score, the higher the tendency
for the singer to lead and the pianist to follow in take 3, and
lower the tendency for the singer to lead and the pianist
to following in take 2. These two takes took place after a
rehearsal session when musicians were required to give their
post-rehearsal performances as though on a concert stage.
Music performances are thought to be unique in nature in
terms of motivation and energy within and between musi-
cians (Chaffin etal., 2007; Devaney, 2015); this uniqueness
of music performance might explain differences in empa-
thy impact across takes. Furthermore, when musicians
were asked to identify their best take of each piece after
the rehearsal session, 58% of the duos chose take 3, 35%
take 2 and only 6% take 0. Musicians were very reluctant
to pick their best performance as they often reported that
all post-rehearsal performances were of high-level, but very
distinctive quality. Musicians usually do not just propose a
replication across performances, but they usually strive to
provide a unique performance(Chaffin etal., 2007).
Limitations andfuture directions
An important aspect to consider is that our results are related
to two specific pieces from a certain repertoire, and we found
some differences between pieces and consecutive takes in
terms of interperformer coordination. This suggests poten-
tially high variability across musical material and perfor-
mance situations. Nevertheless, generalisability might be a
difficult goal to attain, since music performance is highly
situation-dependent. In addition, laboratory-based experi-
ments such as these are highly peculiar, since they require a
specific population of (semi-)professional musicians work-
ing for long periods of time. Each of our lab sessions, for
example, that involved a period of setup, some rehearsal,
and three repeated performances of two different pieces,
lasted about 2h in total—not to mention the time that par-
ticipants spent learning the music at home before arriving
for the session. An additional piece or pieces would have
induced fatigue in the participants, degrading the quality of
our results, and limited the number of musicians who were
willing to participate in the study.
In this study, we conceptualized musical leadership as
manifested in musicians’ head-motion. Such a tendency to
anticipate or lag somebody else’s head motion is certainly
not a comprehensive view of leadership, which is studied
often in terms of social roles (Garrido & Requena, 2015;
King, 2006; Lim, 2014; Page-Shipp etal., 2018) rather than
body motion’s phase synchronicity. Future mixed studies
might further advance our understanding of leadership in
ensembles, by analysing patterns in verbal social interac-
tions during rehearsals in combination with sound and body
motion synchronization.
Furthermore, we analysed the impact of empathy in piano-
singing duos. Larger ensembles with and without an assigned
leader (i.e., a conductor) might provide more variability in
the leader-follower relationships and the interperformer coor-
dination, and this might also open up to an even more com-
plex nature of the impact of empathy in ensemble playing.
Notably, the range of the EPT scores of our participants (from
13 to 28) spanned roughly the top half of the possible range
(from 0 to 28). Stronger effects of empathy might be under-
mined by insufficient heterogeneity in empathy scores. The
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
mean EPT score of our participants is also higher than that
observed by Davis (1980) among US university students. The
mean EPT of our female participants, in fact, was 20.6 (SD =
4.1) and that of males was 19.2 (SD = 3.5); whilst that found
by Davis (1980) was 17.96 (SD = 4.8) for the females and
16.78 (SD = 4.7) for the males. This might be a characteris-
tics of the musician sample, since training and performance
contexts can potentially promote empathy at various degrees
(Sevdalis & Raab, 2014). Some self-selection issues may
have affected our sample as well, since people with greater
empathy might be more likely to succeed at music studies.
Future studies are needed to show how the distribution of
empathic profiles among high-level musicians compares to
that of the general population.
Finally, we focused the analysis of interperformer coor-
dination on musicians’ head motion, since this can signal
the emotionally expressive intentions that performers aim to
transmit. Musicians in ensembles often use also breathing as
a visual/auditory cue to synchronize. The simultaneous anal-
ysis of body motion and breathing coordination might shed
some light on the factors influencing interperformer coor-
dination in music ensembles. In addition to head motion,
pianists’ shoulders also exhibit more movements than other
body parts (i.e., hip, torso, neck, elbows, wrists, middle fin-
gers) and can reflect different expressive intentions during
solo performances (Thompson & Luck, 2011). Therefore,
pianists might use shoulder motion as a visual cue to facili-
tate coordination in ensembles, similarly to head motion.
Conclusion
The present study provides a novel contribution to research
on interpersonal coordination in music ensembles by show-
ing that musicians’ head motion reflects the phrase structure
of the piece, and the singer’s empathy score can contribute
to the establishment of leader-follower relationships between
musicians during piano-singing duo performances, depend-
ing on the piece being performed. These findings provide
evidence, for the first time to our knowledge, of an asso-
ciation between empathy and musicians’ body motion in
the final performances of a certain piece, by revealing how
the singer’s role as leader in piano-singing duos might be
emphasised or diffused by extra-musical factors like empa-
thy. These results provide a better understanding of the
mechanisms underpinning performance science and inter-
personal coordination in music ensembles and during joint
action activities.
Supplementary Information The online version contains supplemen-
tary material available at https:// doi. org/ 10. 1007/ s00426- 023- 01818-8.
Acknowledgements We thank the participants for taking part in the
study.
Author contributions All authors contributed to the conception and
design of the study. SD collected, analysed and interpreted the data,
and drafted and critically revised all versions of the manuscript. LB
contributed to the data collection, analysis and interpretation, and to
the article revisions. HS and AR made a valuable contribution to the
data analysis and interpretation, and to the revision of the article. All
authors approved the submitted version.
Funding Open access funding provided by Austrian Science Fund
(FWF). This research was funded by the Austrian Science Fund (FWF)
project P 32642 and the Research Council of Norway through its Cen-
tres of Excellence scheme (project number 262762).
Data Availability The datasets, including participants’ EPT scores and
raw MoCap data, can be found at https:// doi. org/ 10. 5281/ zenodo. 63670
15.
Declarations
Public significance statements This study demonstrates the importance
of the phrase structure of the music and the empathic profile of the
singer in piano-singing duos in establishing interperformer coordina-
tion power and leader-follower relationships between musicians dur-
ing ensemble performances, respectively, as manifested in musicians’
body motion. These findings provide a novel contribution to research
on interaction in music ensembles and performance science, by quan-
tifying the relationships between body motion and phrase levels of the
piece and revealing for the first time to our knowledge an association
between body motion and empathy, in an ecologically valid situation.
Conflict of interest We have no conflicts of interest to disclose.
Ethical approval The experiment was conducted in accordance with
the Declaration of Helsinki, and ethical approval for the study, with
reference EK Nr: 05/2020, was obtained from the Ethics Committee
at mdw—University of Music and Performing Arts Vienna (Austria).
Open Access This article is licensed under a Creative Commons Attri-
bution 4.0 International License, which permits use, sharing, adapta-
tion, distribution and reproduction in any medium or format, as long
as you give appropriate credit to the original author(s) and the source,
provide a link to the Creative Commons licence, and indicate if changes
were made. The images or other third party material in this article are
included in the article's Creative Commons licence, unless indicated
otherwise in a credit line to the material. If material is not included in
the article's Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will
need to obtain permission directly from the copyright holder. To view a
copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.
References
Audacity. (2021). Audacity(R): Free audio editor and recorder [Com-
puter application]. [Computer software manual]. https:// audac ityte
am. org/ (Version 3.0.0).
Bates, D., Mächler, M., Bolker, B., & Walker, S. (2015). Fitting linear
mixed-effects models using lme4. Journal of Statistical Software.
https:// doi. org/ 10. 18637/ jss. v067. i01
Bishop, L., Cancino-Chacón, C., & Goebl, W. (2019). Moving to com-
municate, moving to interact. Music Perception, 37(1), 1–25.
https:// doi. org/ 10. 1525/ mp. 2019. 37.1.1
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
Bishop, L., & Goebl, W. (2015). When they listen and when they
watch: Pianists’ use of nonverbal audio and visual cues during
duet performance. Musicae Scientiae, 19(1), 84–110. https:// doi.
org/ 10. 1177/ 10298 64915 570355
Bishop, L., & Goebl, W. (2017). Communication for coordination:
Gesture kinematics and conventionality affect synchronization
success in piano duos. Psychological Research, 82, 1177–1194.
https:// doi. org/ 10. 1007/ s00426- 017- 0893-3
Bishop, L., Jensenius, A., & Laeng, B. (2021). Musical and bodily
predictors of mental effort in string quartet music: An ecological
pupillometry study of performers and listeners. Frontiers in Psy-
chology, 12, 653021. https:// doi. org/ 10. 3389/ fpsyg. 2021. 653021
Castellano, G., Mortillaro, M., Camurri, A., Volpe, G., & Scherer,
K. (2008). Automated analysis of body movement in emotion-
ally expressive piano performances. Music Perception, 26(2),
103–119. https:// doi. org/ 10. 1525/ mp. 2008. 26.2. 103
Chaffin, R., Lemieux, A. F., & Chen, C. (2007). “It is different each
time I play”: Variability in highly prepared musical performance.
Music Perception, 24(5), 455–472. https:// doi. org/ 10. 1525/ mp.
2007. 24.5. 455
Chang, A., Kragness, H. E., Livingstone, S. R., Bosnyak, D. J., &
Trainor, L. J. (2019). Body sway reflects joint emotional expres-
sion in music ensemble performance. Scientific Reports, 9,215.
https:// doi. org/ 10. 1038/ s41598- 018- 36358-4
Chang, A., Livingstone, S. R., Bosnyak, D. J., & Trainor, L. J. (2017).
Body sway reflects leadership in joint music performance. Pro-
ceedings of the National Academy of Sciences,114, E4134–
E4141. https:// doi. org/ 10. 1073/ pnas. 16176 57114
Cho, E. (2019). The relationship between small music ensemble expe-
rience and empathy skill: A survey study. Psychology of Music,
49(3), 600–614. https:// doi. org/ 10. 1177/ 03057 35619 887226
Cho, E., & Han, J. Y. (2022). Small music ensemble and empathy: A
replication study in a south Korean music student sample. Psy-
chology of Music,50(4), 1121–1135. https:// doi. org/ 10. 1177/
03057 35621 10316 63
Clayton, M., Jakubowski, K., & Eerola, T. (2019). Interpersonal
entrainment in Indian instrumental music performance: Synchro-
nization and movement coordination relate to tempo, dynamics,
metrical and cadential structure. Musicae Scientiae, 23(3), 304–
331. https:// doi. org/ 10. 1177/ 10298 64919 844809
Dahl, S., & Friberg, A. (2007). Visual perception of expressiveness in
musicians body movements. Music Perception, 24(5), 433–454.
https:// doi. org/ 10. 1525/ mp. 2007. 24.5. 433
D’Amario, S., Daffern, H., & Bailes, F. (2018a). A longitudinal study
investigating synchronization in a singing quintet. Journal of
Voice. https:// doi. org/ 10. 1016/j. jvoice. 2018. 06. 011
D’Amario, S., Daffern, H., & Bailes, F. (2018b). Synchronization in
singing duo performances: The roles of visual contact and leader-
ship instruction. Frontiers in Psychology, 9, 1208. https:// doi. org/
10. 3389/ fp- syg. 2018. 01208
Davis, M. H. (1980). A multidimensional approach to individual dif-
ferences in empathy. JSAS Catalog of Selected Documents in
Psychology, 10, 85.
Davis, M. H. (1983). Measuring individual differences in empathy:
Evidence for a multidimensional approach. Journal of Personality
and Social Psychology, 44(1), 113–126. https:// doi. org/ 10. 1037/
0022- 3514. 44.1. 113
De Carli, F., Nobili, L., Beelke, M., Watanabe, T., Parrino, L., Terzano,
M. G., & Ferrillo, F. (2004). Quantitative analysis of sleep EEG
microstructure in the time? Frequency domain. Brain Research
Bulletin, 63, 399–405. https:// doi. org/ 10. 1016/ s0361- 9230(04)
00092-9
Dean, R. T., & Dunsmuir, W. T. M. (2016). Dangers and uses of
cross-correlation in analyzing time series in perception, per-
formance, movement, and neuroscience: The importance of
constructing transfer function autoregressive models. Behavior
Research Methods, 48(2), 783–802. https:// doi. org/ 10. 3758/
s13428- 015- 0611-2
Demos, A. P., & Chaffin, R. (2017). Removing obstacles to the analysis
of movement in musical performance: Recurrence, mixed models,
and surrogates. In M. Lesaffre, P.-J. Maes, & M. Leman (Eds.),
The Routledge Companion to Embodied Music Interaction (pp.
341–349). Routledge.
Demos, A. P., Chaffin, R., & Logan, T. (2017). Musicians body sway
embodies musical structure and expression: A recurrence-based
approach. Musicae Scientiae, 22(2), 244–263. https:// doi. org/ 10.
1177/ 10298 64916 685928
Devaney, J. (2015). Evaluating singer consistency and uniqueness in
vocal performances. In T. Collins,D. Meredith, A. Volk. (Eds),
Mathematics and Computation in Music. MCM 2015. Lecture
Notes in Computer Science, 9110. Springer.https:// doi. org/ 10.
1007/ 978-3- 319- 20603-5_ 17
Dotov, D., Bosnyak, D., & Trainor, L. J. (2021). Collective music
listening: Movement energy is enhanced by groove and visual
social cues. Quarterly Journal of Experimental Psychology, 74(6),
1037–1053. https:// doi. org/ 10. 1177/ 17470 21821 991793
Eerola, T., Jakubowski, K., Moran, N., Keller, P. E., & Clayton, M.
(2018). Shared periodic performer movements coordinate inter-
actions in duo improvisations. Royal Society Open Science, 5(2),
171520. https:// doi. org/ 10. 1098/ rsos. 171520
Frăţilă, L. (2021). Technical strategies for improving the accompanying
skills in the piano-voice duo. Bulletin of the Transylvania Univer-
sity of Braşov Series VIII Performing Arts, 14(63), 41–50. https://
doi. org/ 10. 31926/ but. pa. 2021. 14. 63.1.5
Garrido, D. C., & Requena, S. O. (2015). Music and leadership: The
role of the conductor. International Journal of Music and Per-
forming Arts, 3(1), 84–88. https:// doi. org/ 10. 15640/ ijmpa. v3n1a8
Glowinski, D., Badino, L., Ausilio, A., Camurri, A., & Fadiga, L.
(2012). Analysis of leadership in a string quartet. In Proceedings
of the Third International Workshop on Social Behaviour in Music
at ACM ICMI (pp. 763–774). Santa Monica.
Glowinski, D., Mancini, M., Cowie, R., Camurri, A., Chiorri, C.,
& Doherty, C. (2013). The movements made by performers in
a skilled quartet: A distinctive pattern, and the function that it
serves. Frontiers in Psychology, 4, 841.
Goebl, W., & Palmer, C. (2009). Synchronization of timing and
motion among performing musicians. Music Perception, 26(5),
427–438. https:// doi. org/ 10. 1525/ mp. 2009. 26.5. 427
Grachten, M., Cancino-Chacón, C., & Gadermaier, T. (2019). partitura:
A python package for handling symbolic musical data. In Late
Breaking/demo session at the 20th International Society for Music
Information Retrieval Conference, Delft, The Netherlands.
Grevenstein, D. (2020). Factorial validity and measurement invari-
ance across gender groups of the German version of the inter-
personal reactivity index. Measurement Instruments for the
Social Sciences, 2, 8.
Grinsted, A., Moore, J. C., & Jevrejeva, S. (2004). Application of
the cross wavelet transform and wavelet coherence to geophysi-
cal time series. Nonlinear Processes in Geophysics, 11(5/6),
561–566. https:// doi. org/ 10. 5194/ npg- 11- 561- 2004
Gu, Z., Gu, L., Eils, R., Schlesner, M., & Brors, B. (2014). Circlize
implements and enhances circular visualization in R. Bioinfor-
matics, 30(19), 2811–2812. https:// doi. org/ 10. 1093/ bioin forma
tics/ btu393
Hartig, F. (2021). DHARMa: Residual Diagnostics for Hierarchical
(Multi-Level/Mixed) Regression Models [Computer software
manual]. http:// flori anhar tig. git hub. io/ DHARMa/ (R package
version 0.4.4).
Issartel, J., Bardainne, T., Gaillot, P., & Marin, L. (2015). The rel-
evance of the cross-wavelet transform in the analysis of human
interaction: A tutorial. Frontiers in Psychology, 5, 1566. https://
doi. org/ 10. 3389/ fpsyg. 2014. 01566
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Psychological Research
1 3
Issartel, J., Marin, L., Gaillot, P., Bardainne, T., & Cadopi, M.
(2006). A practical guide to time-frequency analysis in the
study of human motor behavior: The contribution of wavelet
transform. Journal of Motor Behavior, 38(2), 139–159. https://
doi. org/ 10. 3200/ jmbr. 38.2. 139- 159
Jakubowski, K., Eerola, T., Ximenes, A. B., Ma, W. K., Clayton,
M., & Keller, P. E. (2020). Multimodal perception of interper-
sonal synchrony: Evidence from global and continuous ratings
of improvised musical duo performances. Psychomusicology:
Music, Mind, and Brain, 30(4), 159–177. https:// doi. org/ 10.
1037/ pmu00 00264
Jensenius, A. R., Wanderley, M. M., Godøy, R., & Leman, M.
(2010). Musical gestures: Concepts and methods in research.
In: R. Godøy & M. Leman (Eds.), Musical Gestures: Sound,
Movement, and Meaning (pp. 12–35). Routledge.
Keller, P. E. (2014). Ensemble performance: Interpersonal align-
ment of musical expression.In: D. Fabian, R. Timmers, & E.
Schubert (Eds.), Expressiveness in music performance: Empiri-
calapproaches across styles and cultures (pp. 260–282). Oxford
University Press.
Keller, P. E., & Appel, M. (2010). Individual differences, auditory
imagery, and the coordination of body movements and sounds
in musical ensembles. Music Perception, 28(1), 27–46. https://
doi. org/ 10. 1525/ mp. 2010. 28.1. 27
Keller, P. E., Novembre, G., & Hove, M. J. (2014). Rhythm in joint
action: Psychological and neurophysiological mechanisms for
real-time interpersonal coordination. Philosophical Transac-
tions of the Royal Society B: Biological Sciences, 369(1658),
20130394. https:// doi. org/ 10. 1098/ rstb. 2013. 0394
King, E. C. (2006). The roles of student musicians in quartet rehears-
als. Psychology of Music, 34(2), 262–282. https:// doi. org/ 10.
1177/ 03057 35606 061855
Lim, M. C. (2014). In pursuit of harmony: The social and organi-
sational factors in a professional vocal ensemble. Psychology
of Music, 42(3), 307–324. https:// doi. org/ 10. 1177/ 03057 35612
469674
Nakamura, E., Yoshii, K., & Katayose, H. (2017). Performance error
detection and post-processing for fast and accurate symbolic
music alignment. In Proceedings of the International Society
for Music Information Retrieval Conference (ISMIR), Suzhou,
China, October 23-27.
Novembre, G., Mitsopoulos, Z., & Keller, P. E. (2019). Empathic
perspective taking promotes interpersonal coordination through
music. Scientific Reports, 9, 12255. https:// doi. org/ 10. 1038/
s41598- 019- 48556-9
Novembre, G., Ticini, L. F., Schutz-Bosbach, S., & Keller, P. E.
(2012). Distinguishing self and other in joint action. Evidence
from a musical paradigm. Cerebral Cortex, 22(12), 2894–2903.
https:// doi. org/ 10. 1093/ cercor/ bhr364
Novembre, G., Ticini, L. F., Schütz-Bosbach, S., & Keller, P. E.
(2014). Motor simulation and the coordination of self and other
in real-time joint action. Social Cognitive and Affective Neuro-
science, 9(8), 1062–1068. https:// doi. org/ 10. 1093/ scan/ nst086
Page-Shipp, R., Joseph, D., & van Niekerk, C. (2018). Conductorless
singing group: A particular kind of self-managed team? Team
Performance Management: An International Journal, 24(5/6),
331–346. https:// doi. org/ 10. 1108/ tpm- 09- 2016- 0041
Paulus, C.M. (2009). The saarbruecken personality questionnaire
on empathy: Psychometric evaluation of the german version
of the interpersonal reactivity index. http:// hdl. handle. net/ 20.
500. 11780/ 3343
R Core Team (2013). R: A Language and Environment for Statistical
Computing. R Foundation for Statistical Computing, Vienna,
Austria. http:// www.R- proje ct. org/ (ISBN 3-900051-07-0).
Rabinowitch, T.-C., Cross, I., & Burnard, P. (2012). Long-term musi-
cal group interaction has a positive influence on empathy in
children. Psychology of Music, 41(4), 484–498. https:// doi. or g/
10. 1177/ 03057 35612 440609
Ragert, M., Schroeder, T., & Keller, P. (2013). Knowing too little or
too much: The effects of familiarity with a co-performer’s part
on interpersonal coordination in musical ensembles. Frontiers
in Psychology, 4, 368. https:// doi. org/ 10. 3389/ fpsy g. 2013. 00368
Roesch, A., & Schmidbauer, H. (2018). Waveletcomp: Computa-
tional wavelet analysis. R package version 1.1. https:// cran.r-
proje ct. org/ web/ packa ges/ Wavel etComp/ index. html
Schellenberg, E. G., Corrigall, K. A., Dys, S. P., & Malti, T. (2015).
Group music training and children’s prosocial skills. PLoS One.
https:// doi. org/ 10. 1371/ journ al. pone. 01414 49
Schmidt, R. C., Morr, S., Fitzpatrick, P., & Richardson, M. J. (2012).
Measuring the dynamics of interactional synchrony. Journal of
Nonverbal Behavior, 36(4), 263–279. https:// doi. org/ 10. 1007/
s10919- 012- 0138-5
Schmidt, R. C., Nie, L., Franco, A., & Richardson, M. J. (2014).
Bodily synchronization underlying joke telling. Frontiers in
Human Neuroscience, 8, 633. https:// doi. org/ 10. 3389/ fnhum.
2014. 00633
Sevdalis, V., & Raab, M. (2014). Empathy in sports, exercise, and
the performing arts. Psychology of Sport and Exercise, 15(2),
173–179. https:// doi. org/ 10. 1016/j. psych sport. 2013. 10. 013
Stevens, A., & Ramirez-Lopez, L. (2021). An introduction to the
prospectr package [Computer software manual]. (R package
version 0.2.2).
Thompson, M. R., & Luck, G. (2011). Exploring relationships
between pianists’ body movements, their expressive intentions,
and structural elements of the music. Musicae Scientiae, 16(1),
19–40. https:// doi. org/ 10. 1177/ 10298 64911 423457
Timmers, R., Endo, S., Bradbury, A., & Wing, A. M. (2014). Syn-
chronization and leadership in string quartet performance: A
case study of auditory and visual cues. Frontiers in Psychology,
5, 645. https:// doi. org/ 10. 3389/ fpsyg. 2014. 00645
Titchener, E. B. (1909).Lectures on the experimental psychology of
the thought processes. The MacMillan Company, New York.
Torrence, C., & Compo, G. P. (1998). A practical guide to wave-
let analysis. Bulletin of the American Meteorological Society,
79(1), 61–78. https:// doi. org/ 10. 1175/ 1520- 0477(1998) 079<
0061: APGTW A>2. 0. CO;2
van DorpSkogstad, S.A. (2014). Methods and technologies for using
body motion for real-time musical interaction. Dissertation, Uni-
versity of Oslo.
Volpe, G., Dausilio, A., Badino, L., Camurri, A., & Fadiga, L. (2016).
Measuring social interaction in music ensembles. Philosophi-
cal Transactions of the Royal Society B: Biological Sciences,
371(1693), 20150377. https:// doi. org/ 10. 1098/ rstb. 2015. 0377
Walton, A. E., Richardson, M. J., Langland-Hassan, P., & Chemero,
A. (2015). Improvisation and the self-organization of multiple
musical bodies. Frontiers in Psychology, 6, 313. https:// doi. org/
10. 3389/ fp- syg. 2015. 00313
Washburn, A., Demarco, M., Vries, S. D., Ariyabuddhiphongs, K.,
Schmidt, R. C., Richardson, M. J., & Riley, M. A. (2014). Danc-
ers entrain more effectively than non-dancers to another actor’s
movements. Frontiers in Human Neuroscience, 8, 800. https:// doi.
org/ 10. 3389/ fnhum. 2014. 00800
Wood, E. A., Chang, A., Bosnyak, D., Klein, L., Baraku, E., Dotov,
D., & Trainor, L. J. (2022). Creating a shared musical interpreta-
tion: Changes in coordination dynamics while learning unfamiliar
music together. Annals of the New York Academy of Sciences,
1516, 106–113.
Publisher's Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
Content courtesy of Springer Nature, terms of use apply. Rights reserved.

1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com