R. Dai et al. 1143
Fig. 1. (A) Underlying individual cognitive processes of two-person synchronization. Subscripts i and j represent two persons. (B) A conceptual sketch of ‘two-in-one’
system from holistic view that shows holistic cognitive processes in two-person synchronization. Subscript H means ‘Holistic’.
performed a joint-tapping task in unidirectional and bidi-
rectional conditions (Konvalinka et al., 2010), and their brain
activities were simultaneously recorded using functional near-
infrared spectroscopy (fNIRS) hyperscanning. Previous studies
demonstrated that synchronization performance in bidirec-
tional condition was better than in unidirectional condition. We
hypothesized that the reason for this performance difference
was that holistic error correction was more sufcient in
bidirectional condition than in unidirectional condition. We also
hypothesized that different performances of dyads in the same
condition would be explained by integrated effects of the dyad’s
all holistic cognitive processes. To explore the effect of each
holistic cognitive process on synchronization performance,
each holistic behavioral index would be compared between
conditions; and correlation of holistic cognitive processes and
empathy trait/holistic neural processes would be calculated
because interpersonal SMS, as a basic form of social interaction,
requires mentalizing ability that can be measured by empathy
scale and inter-brain couplings (Cummins et al., 2005; Babiloni et
al., 2012;Cuiet al., 2012). At neural level, we hypothesized that
inter-brain neural coupling (especially coupling of right PFC)
would be higher in bidirectional condition than in unidirectional
condition because bidirectional condition was more cooperative
than unidirectional condition (Cui et al., 2012;Jianget al., 2012).
Methods
Denition and quantitative measurements of holistic
cognitive and neural processes
We used the simplest interpersonal SMS form, two-person
synchronization without tempo change, as an example.
Holistic cognitive processes. According to linear phase cor-
rection model that was proposed by Vorberg (Vorberg and
Wing, 1996; Vorberg and Schulze, 2002) to describe single
person synchronizing to metronome and extend to two-
person synchronization (Vorberg, 2005), the individual cognitive
processes in two-person synchronization mainly contain
internal timekeeper, phase correction and motor execution
(Figure 1A). To achieve and maintain synchronization at
a given tempo, each individual needs an internal time-
keeper to produce predened tempo and predict the onset
of next tap, and then the predicted tap is executed by motor
system. The actual taps inevitably have synchronization
errors due to noise presented in the timekeeper and motor
systems. Thus, each individual must utilize the feedback
of actual taps to adjust his internal timekeeper to reduce
synchronization errors. The parameters of individual cognitive
processes can be estimated using linear phase correction
model
t
i,n+1
= t
i,n
+T
i,n
+M
i,n+1
−M
i,n
−α
i
t
i,n
−t
j,n
t
j,n+1
= t
j,n
+T
j,n
+M
j,n+1
−M
j,n
−α
j
t
j,n
−t
i,n
,(1)
where the variables t, T and M represent the tap onset
event time, timekeeper interval and motor delay, respectively
(Figure 2A). The phase correction gain α is a xed proportion
of the last asynchrony representing the degree of correction
(Figure 2B). The subscripts i and j represent two persons, and
n indicates the number of taps. We used this model to rst
estimate individuals’ phase correction gains (α
i
, α
j
), individuals’
Downloaded from https://academic.oup.com/scan/article-abstract/13/11/1141/5131634 by Beijing Normal University Library user on 05 January 2019