1 23

Experimental Brain Research ISSN 0014-4819 Exp Brain ResDOI 10.1007/s00221-013-3548-7

Imitation behavior is sensitive to visualperspective of the model: an fMRI study

Rui Watanabe, Takahiro Higuchi &Yoshiaki Kikuchi

1 23

Your article is protected by copyright and

all rights are held exclusively by Springer-

Verlag Berlin Heidelberg. This e-offprint is

for personal use only and shall not be self-

archived in electronic repositories. If you wish

to self-archive your article, please use the

accepted manuscript version for posting on

your own website. You may further deposit

the accepted manuscript version in any

repository, provided it is only made publicly

available 12 months after official publication

or later and provided acknowledgement is

given to the original source of publication

and a link is inserted to the published article

on Springer's website. The link must be

accompanied by the following text: "The final

publication is available at link.springer.com”.

1 3

Exp Brain ResDOI 10.1007/s00221-013-3548-7

REsEaRch aRtIclE

Imitation behavior is sensitive to visual perspective of the model: an fMRI study

Rui Watanabe · Takahiro Higuchi · Yoshiaki Kikuchi

Received: 1 November 2012 / accepted: 25 april 2013 © springer-Verlag Berlin heidelberg 2013

be more effectively guided with the models filmed from the 1st person view.

Keywords Imitation · Perspective · Mirror neuron system · superior temporal sulcus · fMRI

Introduction

Recent advances in the understanding of the neural mecha-nisms of imitation suggest that there is a core circuitry of imitation comprising the superior temporal sulcus (sts) and the mirror neuron system (MNs), which consists of the posterior inferior frontal gyrus (IFG) and adjacent ven-tral premotor cortex (PMv), as well as the rostral inferior parietal lobe (IPl) (Iacoboni 2005). Iacoboni et al. (1999) reported significantly stronger brain activity in the infe-rior frontal and parietal areas during imitation than during execution of the same movements induced by geometric spa-tial cues. Based on the fact that these regions were homolo-gous to the monkey brain areas F5 and PE, which play a critical role in encoding the goal of imitation movements and kinesthetic information of the movements (Rizzolatti and arbib 1998), Iacoboni et al. concluded that similar types of information could have been encoded during human imitation.

During human imitation, the core imitation circuit also interacts with the dorsolateral prefrontal cortex (Ba46) and a set of motor preparation areas involving the dorsal pre-motor area (PMd), pre-supplementary motor area (sMa), and superior parietal lobule (sPl) (Iacoboni 2005). Grezes et al. (1999) showed significantly the stronger brain activ-ity in sPl and PMd during imitation of hand movements than during observation of the same movement. sPl is involved in integration of multimodal sensory information

Abstract Imitation behavior and accompanying brain activity can be affected by the perspective of the model adopted. the present study was designed to understand the effect of a model’s perspective in terms of the view (1st person vs. 3rd person) and the anatomical congruency of the limb between the model and the performer (congruent vs. incongruent). Eighteen young participants observed video clips of a model’s finger-lifting behavior and lifted the same finger on their right hand as quickly as possible. half of the video clips were filmed from the view of the participant (the 1st person view), whereas the other half were filmed from the perspective of facing a mirror (the 3rd person view). Each video clip depicted the finger lift-ing of the model’s right (congruent) or left (incongruent) hand. comparisons of the latency to imitate among the four perspective conditions showed significantly shorter latency for the 1st person-congruent and 3rd person-incongruent conditions. hemodynamic measurements with functional magnetic resonance imaging showed that shorter latency was explained with less involvement of the brain areas that are activated when a task is relatively complex. the brain areas considered to be a part of neural substrates of imita-tion were significantly activated under the 1st person view conditions regardless of the hand congruency. these find-ings suggest that, although the latency to imitate finger lift-ing was determined by the complexity of the task induced with the model’s perspective, imitation behavior seemed to

R. Watanabe · t. higuchi (*) health Promotion science, tokyo Metropolitan University, tokyo, Japane-mail: higuchit@tmu.ac.jp

Y. Kikuchi Frontier health science, tokyo Metropolitan University, tokyo, Japan

Author's personal copy

Exp Brain Res

1 3

and spatial information, while PMd plays a role in select-ing response movements (Grafton et al. 1992). Moreover, sPl projects heavily to PMd. With such literature in mind, Grezes et al. proposed that joint activation of sPl and PMd during imitation may reflect the integration of visually pre-sented movements for visuomotor transformations and pro-cessing of movement patterns into motor plans.

It is noteworthy that most studies regarding imitation have used video clips or direct demonstrations that depicted action from a viewpoint as if the model were facing the participants (e.g., schofield 1976; Gleissner et al. 2000; Koski et al. 2003; Buccino et al. 2004; Goldenberg and Karnath 2006; van Elk et al. 2011; Mengotti et al. 2012). this perspective is sometimes referred to as the 3rd person view (Jackson et al. 2006). there seems to be no adequate discussion regarding whether presenting model action from the 3rd person view is more effective to guide imitation behavior than doing so from another view, such as the 1st person view, from which an imitator observes model action from his own perspective.

to our knowledge, there are only a few brain imag-ing studies in which the effect of visual perspective of the model on the efficiency of imitation has been directly investigated (Jackson et al. 2006; Watanabe et al. 2011). Jackson et al. reported that several types of model hand and foot movement were videotaped from the 1st person and 3rd person views. Participants in their study tried to imi-tate the movement as quickly as possible using their right hand or foot. the models filmed from the 1st person view showed the movement of their right hand or foot, which was anatomically congruent with the participants’ hand or foot movement, while the models videotaped from the 3rd person view showed the movement of the their left hand was anatomically incongruent. the results showed a sig-nificantly shorter latency to imitate the model from the 1st person view. Moreover, the brain activity observed using functional magnetic resonance imaging (fMRI) was sig-nificantly stronger in the sensorimotor area spreading to the front–parietal network. these findings suggest the advan-tage of the model from the 1st person view.

Relevantly, whether the model’s hand was anatomically congruent with the imitator’s hand or not is likely to medi-ate the effect of the model’s view. Previous studies (scho-field 1976; Gleissner et al. 2000; Bekkering and Wohls-chlager 2000; Koski et al. 2003; Mengotti et al. 2012) demonstrated that when the models were presented from the 3rd person view, more efficient imitation behavior was guided when the model’s hand was anatomically incongru-ent (i.e., shown as if looking in a mirror) than when it was anatomically congruent. Koski et al. (2003) showed the stronger brain activity in the advanced motor areas involv-ing sMa and sPl when the model’s hand was anatomi-cally incongruent.

Koski et al. (2003) and other studies have only inves-tigated the effect of anatomical congruency between the model and imitator when the model was filmed from the 3rd person view. however, it seems to be natural to con-sider that, from the 1st person view, imitation would be more efficient when the model’s hand is anatomically con-gruent. Our previous findings (Watanabe et al. 2011) indi-rectly support this speculation by showing that, when the model’s behavior was presented from the 1st person view, stronger brain activity in the right posterior insula was observed for a presentation of the anatomically congruent right hand than for the incongruent left hand. however, we did not show behavioral evidence that participants’ imita-tion became more efficient. Relevantly, several studies regarding action observation showed that, when the model was filmed from the 1st person view, the stronger brain activity in a part of the front–parietal network was observed when the model was the anatomically congruent hand than when it was the anatomically incongruent hand (alaerts et al. 2009; Wakita and hiraishi 2011; Vingerhoets et al. 2012).

Based on such theoretical background, the present study was designed to test the hypothesis that the effect of visual perspective of the model would interact with the anatomical congruency of the limb between the model and an imitator; that is, imitation should be more efficient when the movement of the anatomically congruent hand is shown from the 1st person view or when the movement of the anatomically incongruent hand is shown from the 3rd person view. the experimental task was finger lifting, which has been used in previous studies regarding imita-tion behavior (Iacoboni et al. 1999; Iacoboni et al. 2001; Koski et al. 2003).

It is important to note that some of the previous findings regarding the effect of anatomical congruency (Koski et al. 2003; Jackson et al. 2006; Watanabe et al. 2011; Mengotti et al. 2012) can also be explained in line with spatial stimu-lus–response (s–R) compatibility of the finger arrangement between the model’s hand and the imitator’s hand. the response is generally easier when the spatial finger arrange-ment of the response hand corresponds to the spatial fin-ger arrangement of the model (Prinz 1997). that is, for the model of the 3rd person view, more efficient imitation behavior should be guided when the model’s hand is ana-tomically incongruent because of the compatibility of the finger arrangement. It is therefore necessary to remove the potential confound of spatial stimulus–response compat-ibility from the contribution of the anatomical congruence. In an effort to do this, the participant’s responding hand was misaligned with the model’s hand (i.e., both hands were not aligned on the same plane).

the task was categorized as a choice reaction-time task (cRt). the cRt includes the time required for both

Author's personal copy

Exp Brain Res

1 3

response selection of the correct finger and execution of the finger movement. to elucidate temporal behavioral charac-teristics of the response selection of the correct finger, we also asked participants to perform a simple reaction-time task (sRt) in which a predetermined finger was to be lifted as quickly as possible. Because the sRt mostly represents the time required for execution of the finger movement, the difference between the cRt and sRt, referred to as the differential Rt (DRt) (Ishihara et al. 2002; Ishihara and Imanaka 2008), was used to represent the time required for response selection of the correct finger. the shorter DRt was regarded to be a more efficient imitation.

We also measured the hemodynamic activity of the brain using fMRI. Based on the previous findings (Iaco-boni et al. 1999; Iacoboni 2005; Grezes et al. 1999), it was hypothesized that efficient imitation (i.e., shorter DRt) should be accompanied with stronger activation in the IFG, PMv, PMd, and post-parietal cortex (i.e., sPl and IPl), which were included in the MNs and the front–pari-etal network.

Method

Participants and ethics statement

Eighteen healthy subjects (nine females and nine males, age: 23.6 ± 5.5 years) were recruited. None of the partici-pants had any history of neurological or psychiatric illness. they were all right handed, as assessed by the Edinburgh handedness Inventory (Oldfield 1971). Each participant gave informed consent prior to the study. the experiment protocol was approved by the Institutional Ethics commit-tee of the tokyo Metropolitan University (approval Num-ber 10090). they provided their written informed consent to participate in the study. the tenets of the Declaration of helsinki were followed.

apparatus and task

the imitation models were presented using Presentation 14.0 (Neurobehavioral systems, Inc., Usa) and displayed using goggles equipped with the fMRI system (resolu-tion = 800 × 600 pixels, virtual image distance = 3 m, Resonance technology, Inc., Usa). a response box with 4 buttons (4 Button curve Right, current Designs, Inc., Usa) was used to measure the Rt. among them, 3 buttons from the left were used to respond with the index, middle, and ring fingers, respectively. the fMRI was conducted using a 3.0-tesla whole-body superconducting scanner system (achieva 3.0 tX, Philips, Inc., the Netherlands) equipped with a quadrature detection head coil of the birdcage type and an actively shielded gradient coil. the functional

scans were performed using a gradient-recalled echo, and echo-planar imaging MR sequence and t2*-weighted images were acquired. the BOlD-sensitive single-shot EPI sequence parameters were as follows: tR = 4,000 ms; tE = 90.5 ms; flip angle = 80°; matrix size = 128 × 128 pixels; FOV = 240 × 240 mm2; slice numbers = 20; and slice thickness = 6.0 mm.

the imitation models were a set of video clips of a sim-ple hand movement, namely, lifting the index, middle, or ring finger from a resting position on a response box with four buttons. the visual perspective of the model was manipulated so that (a) it was observed from either the 1st or 3rd person view and (b) it represented either the ana-tomically congruent hand (i.e., the right hand; congruent) or the mirror-facing hand (i.e., the left hand; incongruent). as a result, in each video clip, the finger-lifting movement was presented in one of the four views (Fig. 1). Each video clip lasted for 6 s; it consisted of a presentation of a fixation point for 1 s, a static view of the pronated hand resting on the response buttons for 2 s, movement of a finger for 1 s, and again a fixation point for 2 s.

the participants’ goal of the imitation task was to lift one of three fingers on their right hand, imitating the one lifted in the video stimuli as quickly as possible. During the imitation task, the participants held the response box with their right hand, while the right arm was aligned on the right side of the trunk with no pronation or supination of the forearm. With this alignment, the spatial arrangement of the fingers of the model and the participants was not

3rd person congruent 3rd person incongruent

1st person incongruent 1st person congruent

Fig. 1 Four visual perspectives of the model’s finger-lifting behavior

Author's personal copy

Exp Brain Res

1 3

horizontal. this manipulation was necessary to exclude the possibility that the Rt was determined simply by the s–R compatibility.

Procedure

at least 2 days prior to performing the main trials of the imi-tation task, the participants were involved in a pre-practice session. In this pre-practice session, the participants sat in chairs, instead of lying on a bed as they had done in the main session, and performed the imitation task. the alignment of their right arm while performing the task was identical to its alignment in the main session (i.e., the right arm was aligned on the right side of the trunk with no pronation or supination of the forearm). the participants performed a total of 336 tri-als, that is, four times the number of the cRt main trials. the pre-practice session consisted of four blocks. In each block, the four views for each model were included and presented in a randomized order. the purpose of so many practice trials prior to the main session was to reduce the concern that the misalignment of the finger arrangement between the models and the participants could result in a delayed response and/or a relatively large amount of response error.

In the main session, each participant assumed a supine position on a bed with his head fixed by firm foam pads. they were asked to keep their arms at their sides with the forearm in a neutral position. the main session was divided into two parts: an sR (i.e., simple reaction) and a cR (choice reaction). Identical video clips were used for each part. For the cR, the participants were asked to lift one of three fingers, which was identical to that lifted in the video stimuli, as quickly as possible. In contrast, for the sR, the finger to be lifted was predetermined; therefore, the participant’s task was simply to lift the predetermined fin-ger as quickly as the identical finger was lifted in the video clip. the Rts obtained in the sR were necessary to calcu-late the DRt, which was the subtraction of the Rts for the sR from those for the cR. the order of the two parts to be performed was counterbalanced among the participants. a 0.5-min rest period was scheduled between the parts.

Each part included 4 blocks. the perspective of the imi-tation model was fixed throughout each block and selected from the four perspectives. the order of the perspective to be selected was randomized among participants. Rest periods were not scheduled between blocks. Each block comprised 21 trials, in which each of three fingers was lifted for seven trials. the sR part was divided into three sub blocks. the finger to be lifted was constant under each sub block. the participants were precued before each sub block as to which finger was to be lifted. a total of 168 trials (i.e., two parts × four blocks × 21 trials) were presented for each participant. the main session lasted approximately 19 min.

Dependent measures and statistical analyses

For the behavioral analyses, the main dependent measure was the DRt. the DRt for each of the 12 stimuli (i.e., three fingers × four views) was calculated by subtracting the Rts obtained during the sR part from those during the cR part. the frequency of error in selecting the correct finger was also calculated based on the performance for the cR part. to statistically compare the DRts and the error rate among the four perspective conditions, a one-way aNOVa with repeated measures was used. a one-way aNOVa was also performed for the sRts to confirm whether any sig-nificant difference existed for the sRt. It is noteworthy that we used a one-way aNOVa instead of a two-way aNOVa (i.e., view × congruency) to statistically compare the pairs in the four conditions. this was because to statistically compare the pairs in the four conditions, that is, with a two-way aNOVa, we would be unable to compare 1st per-son congruent with 3rd person incongruent and 1st person incongruent with 3rd person congruent.

For the fMRI data, statistical analysis was performed using the sPM8 software package, implemented in Mat-laB 7.10.0 (Mathworks, Inc., sherborn, Usa). For each participant, the functional scans were realigned to cor-rect for subject motion. the scans were normalized to the Montreal Neurological Institute (MNI) template image and spatially smoothed by convolution with an 8 mm full-width at half-maximum (FWhM) Gaussian kernel. the smoothed data were analyzed using the general lin-ear model framework implemented in the sPM8. In the first level, for each participant, a general linear regression model was fitted to the data by modeling each of the event sequences of the four cR and four sR conditions to cal-culate the beta estimates of each visual perspective condi-tion. the onset of the event was defined as the moment when the finger of the models was raised. trials in which participants responded incorrectly were excluded from the analysis.

For each participant, the contrasts of interest were com-puted with regard to the following three respects. the first one was the effect of the 1st–3rd person views, which was calculated by both (a) [(1st person incongruent + 1st per-son congruent) − (3rd person incongruent + 3rd person congruent)] and (b) [(3rd person congruent + 3rd per-son incongruent) − (1st person congruent + 1st person incongruent)]. the second one was the effect of the hand congruency, which was calculated by (a) [(1st person congruent + 3rd person congruent) − (1st person incon-gruent + 3rd person incongruent)] and (b) [(1st person incongruent + 3rd person incongruent) − (1st person congruent + 3rd person congruent)]. the last one was the interaction between the view and hand congruency, which was calculated by (a) [(1st person incongruent − 1st

Author's personal copy

Exp Brain Res

1 3

person congruent) − (3rd person incongruent − 3rd per-son congruent)] and (b) [(3rd person incongruent − 3rd person congruent) − (1st person incongruent − 1st person congruent)].

For each calculation, the single-participant first-level con-trasts were introduced in the second-level random-effects analysis to allow for population inference. the main effects (i.e., the 1st–3rd person views and the hand congruency) and interactions between the view and hand congruency were computed using one-sample t test including all par-ticipants for each of the contrasts, which yielded a statistical parametric map of the t statistic (sPM t). areas of activa-tion were identified as significant if they passed a threshold of P < 0.001, uncorrected for multiple comparisons at the voxel level, and cluster size ≥8 voxels (alphasim/aFNI http://afni.nimh.nih.gov/afni/doc/manual/alphasim).

Results

Behavioral data

the means and standard deviations of the DRt, cRt, and sRt under each view condition are shown in Fig. 2, whereas the means and standard deviations of the error rate during the cR task under each visual perspective con-ditions are shown in Fig. 3. a one-way aNOVa for the DRt showed a significant main effect [F (3, 51) = 6.15, P < .005]. a follow-up test with the Ryan method showed that the DRt was shorter for the 1st person-congruent and 3rd person-incongruent conditions than for the 1st person-incongruent and 3rd person-congruent conditions (P < .05). there was no significant difference between the 1st per-son-congruent and 3rd person-incongruent conditions

tim

e (m

s) 1st-person×congruent

1st-person×incongruent

3rd-person×congruent

3rd-person×incongruent

tim

e (m

s)

tim

e (m

s)

(a)

(b) (c)

Fig. 2 Mean DRt (a), cRt (b), and sRt (c) for each of the four visual perspective conditions

Author's personal copy

Exp Brain Res

1 3

and between the 1st person-incongruent and 3rd person-congruent conditions. For the sRt, we statistically con-firmed that the sRts were not significantly different among the conditions [F (3, 51) = 0.50, ns]. We also confirmed

that there was no significant main effect for the error rate [F (3, 51) = 0.85, ns].

fMRI data

Brain regions that were significantly activated during the cR task in the 2nd-level random-effects analysis are shown in Figs. 4 and 5 and tables 1 and 2. the contrasts are shown as follows: the main effects of the 1st–3rd person view factors (Fig. 4; table 1), the main effects of the hand congruency factor, and interactions between the 1st–3rd person view factors and the hand congruency factor [(1st person incongruent − 1st person congruent) − (3rd person incongruent − 3rd person congruent)] (Fig. 5; table 2).

Effects of view and hand congruency

the effect of the 1st person view [i.e., (1st person con-gruent + 1st person incongruent) − (3rd person congru-ent + 3rd person incongruent)] induced significantly greater brain activity than the 3rd person view in MNs, in the front–parietal network and dorsal visual pathway and in the subcortical region. Brain regions that were more significantly activated in the 1st person view than in the

erro

r ra

te (

%)

Fig. 3 Mean error rates (+sD) for each of the four visual perspective conditions

Top Posterior Insulax = 36, y = -30, z = 8,

Right side Left side

Main effect: the 1st person view > the 3rd person viewFig. 4 Neural activations associated with the 1st person view > the 3rd person view are projected onto surface rendering obtained from an MNI-normalized single-subject brain. activation in the MNs and front–parietal network was observed. supra-threshold acti-vation involved in Rt posterior insula is displayed onto the sagittal section (x = 36)

Author's personal copy

Exp Brain Res

1 3

3rd person view were as follows: lt dorsal premotor area (PMd), Rt inferior frontal gyrus (IFG), Rt posterior supe-rior temporal sulcus (psts), lt superior parietal lobule (sPl), Rt posterior insula (PI), lt middle frontal gyrus, lt precuneus, Rt/lt cuneus, Rt/lt midbrain, lt thalamus, and Rt cerebellum (Fig. 4; table 1).

In contrast, no significantly activated regions of the brain were found with regard to the effect of the 3rd person view [i.e., (3rd person congruent + 3rd person incongru-ent) − (1st person congruent + 1st person incongruent)]. With regard to the effect of hand congruency both for the effect of congruent condition [i.e., (1st person congru-ent + 3rd person congruent) − (1st person incongru-ent + 3rd person incongruent)] and incongruent condition

[(1st person incongruent + 3rd person incongruent) − (1st person congruent + 3rd person congruent)], we found no brain regions that were significantly activated.

Interactions of two view factors

to compare the effect between the 1st–3rd person-view factors and hand congruency factor, we tested for the dif-ferential activity associated with two factors in interaction contrast [i.e., (1st person incongruent − 1st person congru-ent) − (3rd person incongruent − 3rd person congruent)]. Brain regions that were significantly activated were the Rt dorsal premotor area (PMd), lt anterior cingulate cor-tex (acc), and Rt fusiform gyrus (FG), indicating that the

Interaction effect: 1st-3rd * anatomical congruency

Right side ACCx = -12, y = 36, z = 18,

Fig. 5 Neural activations associated with the interac-tion effect between the 1st–3rd person views × anatomical congruency [i.e., (1st person incongruent − 1st person con-gruent) − (3rd person incongru-ent − 3rd person congruent)]. the left figure shows activation in PMd and FG, which is pro-jected onto surface rendering. the right figure demonstrates the activation found in acc, displayed onto the sagittal sec-tion (z = 18)

Table 1 the main effect: the 1st person view > the 3rd person view

showing brain regions with significant activation under (1st person incongruent + 1st person congruent) − (3rd person incongruent + 3rd person congruent)

P < 0.001 uncorrected, t value (t ≥ 4.05), voxel (kE ≥ 8)

l/R Region MNI coordinate t Voxels

x y z

l Dorsal premotor area −36 0 50 7.13 160

R Inferior frontal gyrus (Ba45) 54 26 10 4.47 24

l Middle frontal gyrus −44 44 −12 4.33 27

l superior parietal lobule −22 −50 56 4.27 34

R Posterior superior temporal sulcus 58 −54 10 4.22 79

l Middle frontal gyrus −50 22 38 4.22 10

R Posterior insula 36 −30 8 4.05 33

l Precuneus −14 −88 36 4.06 22

R cuneus 16 −96 8 4.26 52

l cuneus −12 −102 2 4.79 69

lR cuneus 0 −98 4 4.47

R Midbrain 6 −16 −8 6.45 55

l Midbrain −4 −16 −6 3.80

l thalamus −8 −22 10 4.17 30

R cerebellum 18 −46 −20 4.13 35

Author's personal copy

Exp Brain Res

1 3

models filmed from the 1st person-congruent and 3rd per-son-incongruent conditions, which showed shorter latency to imitate, induced less activity in these three regions than had been observed from the other perspectives (Fig. 5; table 2). No supra-threshold activation was associated with the opposite interaction contrast [(3rd person con-gruent − 3rd person incongruent) − (1st person congru-ent − 1st person incongruent)].

Discussion

as predicted, the behavioral results showed that the latency to imitate was significantly shorter for the 1st person-congruent and 3rd person-incongruent condi-tions. however, inconsistently with our hypothesis, the shorter latency under these conditions did not accom-pany significant supra-threshold activation in the brain areas that are included in the core imitation circuit, such as the MNs and sts. the hemodynamic activity under these two conditions were otherwise characterized as “lower” brain activity in Rt PMd, lt acc, and Rt FG (i.e., stronger activity were observed in these areas under the other two conditions; see Fig. 5 and table 2). On the other hand, stronger brain activity in the core imi-tation circuit was observed when participants observed the videos filmed from the 1st person view than from the 3rd person view. these findings may suggest that, regardless of the hand congruency, observation of the model’s behavior from the 1st person view induces the involvement of the core imitation circuit of the brain. the quickness of finger lifting in response to the model’s behavior may be explained by another reason, such as efficient cognitive processes in the brain or relative s–R compatibility of the finger alignment between the model and participants (Brass et al. 2000; Brass et al. 2001), although we did manipulate to reduce the potential com-pound of the effect of s–R compatibility.

shorter latency to initiate finger lifting under the 1st person-congruent and 3rd person-incongruent conditions

the hemodynamic analysis based on interaction con-trast [i.e., (1st person incongruent − 1st person congru-ent) − (3rd person incongruent − 3rd person congruent)] showed that the brain activity under the 1st person-congru-ent and 3rd person-incongruent conditions, which showed shorter latency to imitate, was characterized by less involvement of Rt PMd (i.e., the ipsilateral side of the par-ticipant’s moving hand), lt acc, and Rt FG than those under the 1st person-incongruent and 3rd person-congruent conditions. Previous studies have shown that these brain areas are likely to be involved when the task is relatively complex. For example, the brain activity in the ipsilateral PMd is greater for the control of precise or complicated movement (Ehrsson et al. 2000; sadato et al. 1996). acc is involved in monitoring conflict between competing responses, especially in a complex choice reaction task (Botvinick et al. 1999; carter et al. 1998; Magno et al. 2006). Previous studies also showed that the activation of FG was related to visual recognition of object alignment or shape (saito et al. 2003).

considering these previous studies, the participants may have performed the finger-lifting task with less involve-ment of Rt PMd, lt acc, and Rt FG under the 1st person-congruent and 3rd person-incongruent conditions because the perspective of the models under these conditions yields less complex and more efficient behavior. the relative s–R compatibility of the finger alignment between the model and participants under these conditions may have contrib-uted to the task being less complex, although we manipu-lated to reduce the potential compound of the effect of s–R compatibility with the misalignment of fingers between the model and participants. the alignment of the model filmed from the 1st person congruent condition was the most com-patible with the participant’s right hand, whereas the finger alignment of the model was consistent with the performer’s right-hand image projected onto a mirror under the 3rd per-son-incongruent condition.

Effects of the 1st person view

Imitation behavior with the model taken from the 1st person view was accompanied by significantly greater brain activ-ity in Rt sts and Rt IFG, which were involved in the core circuitry of imitation, and in lt sPl and lt PMd, which support imitation behavior (Iacoboni 2005). however, the models filmed from the 1st person view and the 3rd person view demonstrated no difference in behavioral character-istics. similarly, significantly stronger activation in Rt PI, which is considered to be engaged in body ownership (tsa-kiris et al. 2007; Karnath et al. 2005), was also found.

Table 2 the interaction effect: the 1st–3rd person views × the ana-tomical congruency of the limb

showing brain regions with significant activation under (1st per-son incongruent − 1st person congruent) − (3rd person incongru-ent − 3rd person congruent)

P < 0.001 uncorrected, t value (t ≥ 4.05), voxel (kE ≥ 8)

l/R Region MNI coordinate t Voxels

x y z

R Dorsal premotor area

18 −16 78 4.59 12

l anterior cingulate gyrus

−16 36 18 4.23 9

l Fusiform gyrus 30 −66 −10 5.47 13

Author's personal copy

Exp Brain Res

1 3

Greater activity in IFG and sts suggests that the mod-els filmed from the 1st person view likely elicited recog-nition of the models as finger movements relative to the video clips filmed from the 3rd person view. Previous studies have demonstrated that IFG plays a role in the rec-ognition of model movements, the encoding of the inten-tion of observed movements, and the understanding of the goal of the models (Iacoboni 2005; Iacoboni et al. 2005; Koski et al. 2002; Koski et al. 2003). Many imitation stud-ies have indicated the importance of lt IFG (Grezes and Decety 2001; Koski et al. 2003; tanaka and Inui 2002) and Rt IFG (Buccino et al. 2003; heiser et al. 2003; Johnson-Frey et al. 2003; Molnar-szakacs et al. 2005) for imita-tion behavior. Notably, several previous studies suggested that IFG would be potentially involved in cognitive control movement, such as selecting, conflicting, and inhibiting responses, rather than in imitation behavior (corbetta and shulman 2002; Brass and heys 2005; Molenberghs et al. 2009; Ocampo et al. 2011). Given that the response time was not delayed with the models filmed from the 1st per-son view, however, the stronger activation in the IFG under the 1st person conditions was unlikely to reflect the conflict and/or inhibition of responses.

several neuroimaging studies have demonstrated that sts is the main visual area involved in processing biologi-cal motion (allison et al. 2000; Jastorff and Orban 2009; Molenberghs et al. 2010). sts also provides a higher-order visual description of the imitation models to IFG, where the goal of the action and the motor specification are coded. after that, efferent copies of the motor imitative plan are sent from the MNs to sts, where the predicted sensory consequences of the planned imitative action and the visual description of the observed action are matched (Iacoboni et al. 2001; Molenberghs et al. 2010; Molnar-szakacs et al. 2005).

the significant brain activity in lt sPl and lt PMd observed in the present study suggest that the participants possibly used their multimodal sensory information to select the response for imitation. Previous studies proposed that sPl receives multimodal sensory input associated with body localization and movement control (Felician et al. 2004; Grezes et al. 1999; Parkinson et al. 2010). Further-more, lt PMd plays a role in response selection regard-less of the laterality of the response hand (schluter et al. 2001). there are distinct connections between sPl and PMd, which are associated with the selection of subsequent movements and motor planning (Parkinson et al. 2010; Grezes et al. 1999). the Rt PI activity suggests the possi-bility that the participants considered the models as their own bodies regardless of hand congruency. Previous stud-ies have reported that Rt PI shows greater activity when we experience a sense of body ownership and self-awareness (tsakiris et al. 2007; Karnath et al. 2005).

taken together with these previous findings, the present findings indicate that the models filmed from the 1st person view are likely to more effectively guide imitation behav-ior than those from the 3rd person view. Under the 1st per-son view, the brain areas of Rt sts and Rt IFG were more activated, probably because the participants perceived the kinesthetic information of imitation models efficiently and recognized the models as finger lifting. Moreover, lt PMd and lt sPl were more activated, probably because the par-ticipants selected their own finger position on the basis of their somatosensory schema and discriminated the lifted finger of the models to plan their response.

We found no significantly stronger brain activity in PMv and IPl, which are also included in the core imita-tion circuitry (Iacoboni 2005; Iacoboni and Dapretto 2006). In the present study, comparisons were made between the experimental conditions, all of which included the model’s behavior. Previous studies which have conducted similar comparisons also showed no stronger brain activity in PMv and IPl (chaminade et al. 2005; Williams et al. 2007). In contrast, other previous studies have made comparisons between an experimental condition that included the mod-el’s behavior and an experimental condition that included a neutral stimulus, such as dot (Iacoboni et al. 1999; Koski et al. 2003). considering these results, it seems likely that PMv and IPl play roles in representing hand movement during imitation behavior; therefore, no stronger brain activity may have been observed when comparing experi-mental conditions, both of which included the model’s behavior.

Effect of the 3rd person view

the present study failed to show the brain areas which were more significantly activated under the 3rd person-view condition than under the 1st person-view condition. this was inconsistent with previous findings in which sev-eral brain areas included in the core imitation circuit were activated (Goldenberg and Karnath 2006; Iacoboni et al. 1999; Iacoboni et al.2005; Koski et al. 2003). although the reason for such discrepancy is unclear, the task character-istics of finger lifting in response to model’s behavior may have affected such results. In other words, the participants may have performed this task simply as the task of choice reaction time rather than as imitation of model’s behav-ior. a future study should address the validity of these explanations.

Conclusion

We tentatively conclude that presenting the model’s action from the 1st person view is more adequate to facilitate

Author's personal copy

Exp Brain Res

1 3

imitation behavior than presenting it from the 3rd person view. this is because the models filmed from the 1st person view induced brain activity in the core circuit of imitation and front–parietal network. however, the latency to imitate finger lifting was not explained by the involvement of the core circuit of imitation in the brain; rather, the latency was significantly shorter when the finger alignment between the model and participants was spatially compatible and, thus, the brain areas which are involved when the task is relatively complex are less active. Future studies need to address whether such a tentative conclusion is task-specific or not.

References

alaerts K, heremans E, swinnen sP, Wenderoth N (2009) how are observed actions mapped to the observer’s motor system? Influence of posture and perspective. Neuropsychologia 47: 415–422

allison t, Puce a, Mccarthy G (2000) social perception from visual cues: role of the sts region. trend cogn sci 4:267–278

Bekkering h, Wohlschlager a (2000) Imitation of gestures in children is goal–directed. Q J Exp Psychol a 53a(1):153–164

Botvinick M, Nystrom lE, Fissell K, carter cs, cohen JD (1999) conflict monitoring versus selection-for-action in anterior cingu-late cortex. Nature 402:179–181

Brass M, heys c (2005) Imitation: is cognitive neuroscience solving the correspondence problem? trends cog sci 9(10):489–495

Brass M, Bekkering h, Wohlschlager a, Prinz W (2000) compatibil-ity between observed and executed finger movements: comparing symbolic, spatial, and imitative cues. Brain cogn 44(2):124–143

Brass M, Bekkering h, Prinz W (2001) Movement observation affects movement execution in a simple response task. acta Psychol 106(1–2):3–22

Buccino G, Binfoski F, lucia R (2003) the mirror neuron system and action recognition. Brain lang 89:370–376

Buccino G, Vogt s, Ritzl a, Fink GR, Zilles K, Freund hJ, Rizzolatti G (2004) Neural circuits underlying imitation learning of hand actions: an event–related fMRI study. Neuron 42(2):323–334

carter cs, Braver ts, Barch DM, Botvinick MM, Noll D, cohen JD (1998) anterior cingulate cortex, error detection, and the online monitoring of performance. science 280:747–749

chaminade t, Meltzoff aN, Decety J (2005) an fMRI study of imita-tion: action representation and body schema. Neuropsychologia 43:115–127

corbetta M, shulman Gl (2002) control of goal-directed and stimu-lus-driven attention in the brain. Nat Rev Neurosci 3(3):201–215

Ehrsson h, Fagergren a, Jonsson t, Westling G, Johansson Rs, For-ssberg h (2000) cortical activity in precision-versus power-grip tasks: an fMRI study. J Neurophysiol 83(1):528–536

Felician O, Romaiguere P, anton Jl, Nazarian B, Roth M, Poncet M, Roll JP (2004) the role of human left superior parietal lobule in body part localization. ann Neurol 55(5):749–751

Gleissner B, Meltzoff aN, Bekkering h (2000) children’s coding of human action: cognitive factors influencing imitation in 3-year-olds. Dev sci 3:405–414

Goldenberg G, Karnath h (2006) the neural basis of imitation is body part specific. J Neurosci 26:6282–6287

Grafton s, Mazziotta J, Woods R, Phelps M (1992) human func-tional anatomy of visually guided finger movements. Brain 115:565–587

Grezes J, Decety J (2001) Functional anatomy of execution, mental simulation, observation, and verb generation of actions: a meta-analysis. hum Brain Mapp 12(1):1–19

Grezes J, costes N, Decety J (1999) the effects of learning and inten-tion on the neural network involved in the perception of meaning-less actions. Brain 122:1875–1877

heiser M, Iacoboni M, Maeda F, Marcus J, Mazziotta Jc (2003) the essential role of Broca’s area in imitation. Eur J Neurosci 17(5):1123–1128

Iacoboni M, Dapretto M (2006) the mirror neuron system and the consequences of its dysfunction. Nat Rev Neurosci 7(12):942–951

Iacoboni M, Woods RP, Brass M, Bekkering h, Mazziotta Jc, Riz-zolatti G (1999) cortical mechanisms of human imitation. sci-ence 286:2526–2528

Iacoboni M, Koski l, Brass M, Bekkering M, Woods h, Dubeau RP, Mazziotta Mc, Rizzolatti G (2001) Reafferent copies of imitated actions in the right superior temporal cortex. Proc Natl acad sci Usa 98(24):13995–13999

Iacoboni M, Molnar-szakacs I, Gallese V, Buccino G, Mazziotta Jc, Rizzolatti G (2005) Grasping the intentions of others with one’s own mirror neuron system. Plos Biol 3(3):e79

Iacoboni M (2005) Neural mechanisms of imitation. curr Opin Neu-robio 15(6):632–637

Ishihara M, Imanaka K (2008) Visual perception and motor prepara-tion of manual aiming: a review of behavioral studies and neural correlates. In: Nilsson Il, lindberg WV (eds) Visual perception: new research. Nova science Publishers, hauppauge, pp 1–48

Ishihara M, Imanaka K, Mori s (2002) lateralized effects of target location on reaction times when preparing for manual aiming at a visual target. hum Mov sci 21(5–6):563–582

Jackson Pl, Meltzoff aN, Decety J (2006) Neural circuits involved in imitation and perspective-taking. Neuroimage 31(1):429–439

Jastorff J, Orban Ga (2009) human functional magnetic resonance imaging reveals separation and integration of shape and motion cues in biological motion processing. J Neurosci 29(22):7315–7329

Johnson-Frey sh, Maloof FR, Newman-Norlund R, Farrer c, Inati s, Grafton st (2003) actions or hand-object interactions? human inferior frontal cortex and action observation. Neuron 39(6):1053–1058

Karnath hO, Baier B, Nagele t (2005) awareness of the functioning of one’s own limbs mediated by the insular cortex? J Neurosci 25(31):7134–7138

Koski l, Wohlschlager a, Bekkering h, Woods RP, Dubeau M-c, Mazziotta J, Iacoboni M (2002) Modulation of motor and premo-tor activity during imitation of target-directed actions. cereb cor-tex 12:847–855

Koski l, Iacoboni M, Dubeau Mc, Woods RP, Mazziotta Jc (2003) Modulation of cortical activity during different imitative behav-iors. J Neurophysiol 89(1):460–471

Magno E, Foxe J, Mloholm s, Robertson I, Garavan h (2006) the anterior cingulate and error avoidance. J Neurosci 26(18):4679–4773

Mengotti P, corradi-Dell’acqua c, Rumiati RI (2012) Imitation components in the human brain: an fMRI study. Neuroimage 59(2):1622–1630

Molenberghs P, cunnington R, Mattingley JB (2009) Is the mirror neuron system involved in imitation? a short review and meta-analysis. Neurosci Biobehav Rev 33(7):975–980

Molenberghs P, Brander c, Mattingly JB, cunnington R (2010) the role of superior temporal sulcus and the mirror neuron system in imitation. hum Brain Map 31(9):1316–1326

Molnar-szakacs I, Iacoboni M, Koski l, Mazziotta Jc (2005) Func-tional segregation within pars opercularis of the inferior frontal gyrus: evidence from fMRI studies of imitation and action obser-vation. cereb cortex 15(7):986–994

Author's personal copy

Exp Brain Res

1 3

Ocampo B, Kritikos a, cunnington R (2011) how frontoparietal brain regions mediate imitative and complementary actions: an fMRI study. Plos ONE 6(10):e26496

Oldfield Rc (1971) the assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia 9:97–113

Parkinson a, condon l, Jackson sR (2010) Parietal cortex coding of limb posture: in search of the body-schema. Neuropsychologia 48:3228–3234

Prinz W (1997) Perception and action planning. Eur J cogn Psychol 9(2):129–154

Rizzolatti G, arbib M (1998) language within our grasp. trends Neurosci 21(5):188–194

sadato N, campbell G, Ibanez V, Deiber M, hallett M (1996) com-plexity affects regional cerebral blood flow change during sequential finger movements. J Neurosci 16(8):2691–2700

saito DN, Okada t, Morita Y, Yonekura Y, sadato N (2003) tactile-visual cross-modal shape matching: a functional MRI study. Brain Res cogn Brain Res 17(1):14–25

schluter ND, Krams M, Rushworth MF, Passingham RE (2001) cer-ebral dominance for action in the human brain: the selection of actions. Neuropsychologia 39(2):105–113

schofield WN (1976) Do children find movements which cross the body midline difficult? Q J Exp Psychol 28:571–582

tanaka s, Inui t (2002) cortical involvement for action imitation of hand/arm postures versus finger configurations: an fMRI study. NeuroReport 13(13):1599–1602

tsakiris M, hesse MD, Boy c, haggard P, Fink GR (2007) Neural signatures of body ownership: a sensory network for bodily self-consciousness. cereb cortex 17(10):2235–2244

van Elk M, van schie ht, Bekkering h (2011) Imitation of hand and tool action is effector-independent. Exp Brain Res 214(4):539–547

Vingerhoets G, stevens l, Meesdom M, honore P, Vandemaele P, achten E (2012) Influence of perspective on the neural correlates of motor resonance during natural action observation. Neuropsy-chol Rehabil 22(5):752–767

Wakita M, hiraishi h (2011) Effects of handedness and viewing per-spective on Broca’s area activity. NeuroReport 22(7):331–336

Watanabe R, Watanabe s, Kuruma h, Murakami Y, senoo a, Mat-suda t (2011) Neural activation during imitation of movements presented from four different perspectives: a functional magnetic resonance imaging study. Neurosci lett 503(2):100–104

Williams Jh, Whiten a, Waiter GD, Pechey s, Perrett DI (2007) cor-tical and subcortical mechanisms at the core of imitation. soc Neurosci 2(1):66–78

Author's personal copy