NeuroImage 103 (2014) 476–484

Contents lists available at ScienceDirect

NeuroImage

j ourna l homepage: www.e lsev ie r .com/ locate /yn img

“Mom called me!” Behavioral and prefrontal responses of infants toself-names spoken by their mothers

Masahiro Imafuku a,b, Yoko Hakuno a, Mariko Uchida-Ota c, Jun-ichi Yamamoto a, Yasuyo Minagawa a,d,⁎a Department of Psychology, Faculty of Letters, Keio University, 2-15-45 Mita, Minato-ku, Tokyo 108-8345, Japanb Graduate School of Education, Kyoto University, Yoshida-honmachi, Sakyo-ku, Kyoto 606-8501, Japanc Graduate School of Human Relations, Keio University, 2-15-45 Mita, Minato-ku, Tokyo 108-8345, Japand Institut d'étude de la cognition, Ecole Normale Supérieure, 29, rue d'Ulm, 75005 Paris, France

⁎ Corresponding author at: Department of PsychoUniversity, 2-15-45 Mita, Minato-ku, Tokyo 108-8345, Jap

E-mail address: myasuyo@bea.hi-ho.ne.jp (Y. Minagaw

http://dx.doi.org/10.1016/j.neuroimage.2014.08.0341053-8119/© 2014 Elsevier Inc. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Accepted 20 August 2014Available online 29 August 2014

Keywords:Self-nameInfantsMother's voiceNear-infrared spectroscopyThe medial prefrontal cortex (mPFC)

Development of a sense of self is a fundamental process needed for human social interaction. Although functionalneuroimaging studies have revealed the importance of medial prefrontal cortex (mPFC) in self-referencing, howthis function develops in infancy remains poorly understood. Todetermine the cerebral basis underlying process-ing of self-related stimuli, we used behavioralmeasures and functionalmulti-channel near-infrared spectroscopy(fNIRS) tomeasure prefrontal cortical responses in 6-month-old infants hearing their own names.We also inves-tigated the influence of a mother's voice on name perception in infants— an ability that plays a crucial role in therecognition of social signals. Experiment 1 measured the behavioral preferences of infants for their own namesand for other names, spoken either by their mothers or by strangers. Results showed that infants significantlypreferred their own name to other names, regardless of speaker type. Experiment 2 examined hemodynamic re-sponses to the same four conditions in the prefrontal cortex. Compared with other names, hearing their ownnames, especiallywhen spoken by theirmother, elicited greater activity in the infant's dorsalmPFC. Furthermore,the magnitude of the cerebral response correlatedwith the degree of behavioral preference onlywhen involvingtheir mother's voice. These findings suggest that, particularly in the context of their mothers' voice, the dorsalmPFC of infants is already sensitive to social signals related to self at 6 months. At the same time, familiarityand affection related processing are also discussed as possible factors modulating dorsal mPFC activation atthis age.

© 2014 Elsevier Inc. All rights reserved.

Introduction

Humans develop their communication skills from birth throughinteraction with others. To detect messages conveyed by othersintended for interaction, one must understand signals directed at theself. Humans are extremely sensitive to ostensive signals like name call-ing that indicate someone's communicative intention toward them(Csibra, 2010). An infant's name is usually spoken in infant-directedspeech (‘motherese’) by caregivers and gradually becomes an ostensivesignal. Infants learn their own name as an auditory signal about the selfand can recognize names at around 4.5 months of age (Mandel et al.,1995). From this age, infants will selectively turn their head whentheir name is called, showing that they interpret their own name as avocative.

The underlying mechanism behind name recognition is thought tobe “mentalizing” — a term that refers to an automatic cognitive processthat underlies our ability to attribute intentions to ourselves or to others

logy, Faculty of letters, Keioan. Fax: +81 3 5427 1209.a).

(Amodio and Frith, 2006). This involves understanding communicativeintentions toward ourselves by interpreting audio signals as our names.The brain regions that adults employ in processing these communica-tive signals are localized specifically in the rightmedial prefrontal cortex(mPFC) and the left temporal pole (Kampe et al., 2003). Recently,Grossmann et al. (2010) examined whether 5-month-old infants aresensitive to eye contact and their own names as communicative signalsby measuring PFC responses using functional near-infrared spectrosco-py (fNIRS). They showed that infants automatically process communi-cative signals directed at them, as the left dorsal prefrontal cortex wasactivated in response to their own names. Although this difference inlocalization patterns between 5-month-old infants and adults may bedue to changes during development, supportive evidence is sparse,particularly for infants. The present study aims to reveal further detailsof the neuralmechanismunderlying early name recognition by focusingon prefrontal activation, particularly themPFC, and the effect of speakerfamiliarity. In their daily lives, infants usually hear their names spokenby principal caregivers (e.g. mothers), and these familiar voices mighthave a large impact on developmentally early processing of socialsignals toward self. However, the influence of familiar voices on infantname recognition has not been explored in previous studies. Thus, our

477M. Imafuku et al. / NeuroImage 103 (2014) 476–484

study has an additive effect on the recognition of one's own name as anostensive signal. Namely, we examine whether mother's voice wouldenhance the association between names and self.

Each infants' response to their own name influences social-communicative development. Previous studies have focused on therole of an infant's own name in early language acquisition, includingobject learning. Because infants focus their attention in response tohearing their own names they may be more likely to associate eventswith their labels. For example, 6-month-old infants recognize wordsthat follow their own name, but not words that follow other names(Bortfeld et al., 2005). Using event-related potentials (ERP), Pariseet al. (2010) showed that ERP amplitude following the Nc peak waslarger for objects preceded by the infant's own name. This suggeststhat infants use their own name as a social cue to focus their attentionon events, and suggests that infants can easily absorb a great deal ofinformation by effectively responding to their own name.

As many behavioral studies have shown, establishment of themother–infant bond is also crucial for social-communicative develop-ment. An experiment using a measure of sucking behavior showedthat at birth neonates already prefer their own mother's voice to thatof a stranger (DeCasper and Fifer, 1980). Infants at 3.5 months of agecan only recognize their mothers' facial expressions, while they canonly recognize facial expressions of other people four months later(Montague and Walker-Andrews, 2002). Barker and Newman (2004)showed that word learning in 6- to 8-month-old infants was facilitatedwhen words were spoken by the infant's own mother. These studiessuggest that mothers play a crucial role in infant acquisition of basicsocial-communicative abilities in the first year of life.

Indeed, mothers spend a large amount of time with their infants,thus, it is assumed that the mother contributes to their infant's neuro-logical development. Using fNIRS, Minagawa-Kawai et al. (2009) mea-sured cerebral responses in infants to video-clips that showed theirown mothers and unfamiliar mothers engaged in smiling behavior,and found that infants about 1 year in age showed activation in the an-terior part of the orbitofrontal cortex (OFC) specifically when watchingtheir own mothers' smiles. This suggests that infant prefrontal cortexmight play a role in regulating and encoding affective cues from theprimary caregiver. Using ERP, Purhonen et al. (2005) found that themother's voice has special importance and enhances involuntary atten-tion; this indicates a clear memory template of the own mother's voiceat a very early age. Furthermore, it has been shown that the mother'svoice elicits stronger activations in the anterior prefrontal cortex andthe left posterior temporal region than an unknown voice, suggestingthat the mother's voice plays a role in the early shaping of the humanauditory system (Dehaene-Lambertz et al., 2009). As shown by thesestudies, the mother's voice has a strong impact on both behavioral andneural responses in young infants. A previous study using ERP in adulthumans showed that a participant's own name elicited larger responseamplitudes in the late phase of novelty P3 in the fronto-parietal regionwhen the namewas produced by a familiar voice thanwhen it was pro-duced by an unfamiliar voice (Holeckova et al., 2006). This indicates thatfamiliar voices may induce deeper processing of self-related stimulithan unfamiliar voices even in infants. Hence, it is likely that an infant'sname being called by themother plays a special role in the developmentof the mechanism of mentalizing, and this should be studied byassessing the brain of the infant.

As discussed above, principal caregivers play a crucial role in thesocial-communicative development of infants. Although a previous be-havioral study has shown that infants recognize their own name calledby strangers as a familiar word form (Mandel et al., 1995), it has not yetbeen shownwhethermothers' voices affect the ability of their infants torecognize communicative signals directed at the self. Furthermore, onlya few infant studies have assessed behavioral and neurophysiologicalresponses simultaneously. Using both approaches provides rich infor-mation on the interpretation of neural basis about behaviors (Sethet al., 2008). Therefore, the present study aimed to assess the responses

of 6-month-old infants to names (theirs or others), spoken by women(their mothers or strangers), using both behavioral and neuroimagingmeasures. In Experiment 1, we examined infant behavioral responsesto names called by either mothers or strangers using a modified head-turn preference procedure. In Experiment 2, we used fNIRS to examinechanges in infant cerebral activation patterns in response to hearingnames spoken in different voices. To this end, we measured activationin frontal brain regions that have been implicated in similar studiesusing adults. These regions include part of the medial prefrontal area,which is crucial for processing self-related stimuli. We expected tofind higher levels of activation in the prefrontal brain regions in re-sponse to one's own name (self-name) than to others (other-names).Furthermore, we predicted that a self-name spoken by one's ownmother is the most effective self-referencing stimulus for activatingthe mPFC.

Experiment 1

Using a modified head-turn preference paradigm, we investigatedwhether mothers' voices or strangers' voices differentially affect howinfants respond to hearing their names (Nelson et al., 1995). Previousstudies have suggested that infants can perceptually discriminatebetween self-names and other-names at 4.5 months of age. However,no studies have been performed to assess whether the mother's voiceaffects the recognition of self-name as a familiar word form in infants.In this experiment, we investigated self-name recognition in infantsby varying speaker familiarity. We hypothesized that infants wouldshow a preference for self-names over other-names, and that theywould show more preference to self-names when spoken by theirown mother than when spoken by strangers.

Materials and methods

ParticipantsSeventeen 6-month-old infants (10 boys and 7 girls; mean age,

181.5 ± 22.8 days) and their mothers participated in the experiment.Three participants were excluded from final analysis because of fussi-ness (n = 2) and technical failure (n = 1). All infants were screenedfor typical cognitive functioning using the Kyoto Scale of PsychologicalDevelopment (KSPD) (Ikuzawa et al., 2002) at the time of the experi-ment and all infants had been full term at birth and had no history ofserious illnesses or disorders in both Experiments 1 and 2. The studywas approved by the ethics committee of Keio University, Faculty ofLiterature (No. 11047). Informed consent was obtained from the par-ents prior to the study. On the basis of our interaction with themothersand the results of KSPD, all mothers were regarded to be healthy withno cognitive deficits.

StimuliAudio recordings of the infants' first names spoken by their mothers

were used as stimuli. The mothers' voices were recorded (PMD671,Marantz) before the behavioral study. We checked in advance whatkind of names infants usually hear such as infants' sibling names andchose a name for other names condition. Mothers were instructed tosay the name of their own child and those of the other infants usinginfant-directed speech. The experimenter demonstrated an exampleutterance before the recording. The audio stimuli were digitized at asampling rate of 44.1 kHz and a resolution of 16 bits. All infant nameswere controlled to have the same number mora which is a sub-syllabicstructure serving a basic unit in Japanese (e.g. Ma-sa-hi-ro, 4 morae;Ryo-o-he-i, 4 morae). The auditory stimuli comprised four conditions:self-name/mother's voice, self-name/stranger's voice, other-name/mother's voice, and other-name/stranger's voice. The name stimuli ofeach condition consisted of 40 s which included repeated one namewith an inter-stimulus interval (ISI) of 500 ms and the number of onename was in each stimuli. Stimuli used in the self-name condition for

478 M. Imafuku et al. / NeuroImage 103 (2014) 476–484

one participant were used once in the other-name condition for theother participant. Likewise, stimuli used in the mother's voice conditionfor one participant were used once in the stranger's voice conditionfor the other participant. For example, infant A heard self- and other-names spoken by mother A as the mother's voice condition and thosespoken bymother B as the stranger's voice condition; conversely, infantB heard self and other-names spoken bymother B as themother's voicecondition and those spoken by mother A as the stranger's voice condi-tion. All auditory stimuli were edited using audio editing software(Sound Forge Pro 10.0), and the root mean square (RMS) value of thesound amplitude was approximately equal among stimuli.

ProcedureInfants were seated on the lap of a parent with their eyes approxi-

mately 40 cm from the monitor in a sound-attenuated room. Motherswore headphones so that they would not hear the stimuli during theexperiment. Stimuli were presented on a 19-inch monitor controlledby a computer. Two loudspeakers were mounted on either side of themonitor at about the height of the infant's head. A video camera wasmounted behind the monitor to record the infant's eye gaze duringthe experiment. An experimenter observed the infant's behaviorthrough another monitor connected to the video camera.

Responses to the nameswere assessed using themodified head-turnpreference paradigm (Nelson et al., 1995), which has previously beenused in infant studies (e.g., Hakuno et al., 2012). We used the paradigmto assess preferences toward auditory stimuli bymeasuring the amountof time infants looked at visual stimuli thatwere concurrently presentedon a computer monitor. Thus, the longer infants looked at a visual stim-ulus, the greater their preference for the accompanying name wasthought to be.

Before the start of the trial, the attention getter (fireworks; visualangle= approximately 17°) was presented at the center of themonitorand a sound was played through both speakers to grasp the infant's at-tention, as shown in Fig. 1. The trial started when the infant looked atthe attention getter. Each trial began with a visual stimulus (a movieof puppet; visual angle = approximately 13°) shown at either the leftor the right side on the monitor with auditory stimuli (one of fourname conditions). In the test phase,when the infant looked at the visualstimulus, the auditory stimulus was presented through the speaker onthe same side of the visual stimulus of the monitor at approximately65 dB SPL. The trial lasted until the infant looked away from the visualstimulus for more than 2 s or continued to look at the stimulus for lon-ger than 40 s. The inter-trial interval was 1 s. Each trial comprised about

Fig. 1. The timeline of one

(mean± SEM) 26.1 ± 4.0 name stimuli in average within 40 s of max-imum presentation. If the infant returned attention to the movie in the2-s interval, the trial continued. One session comprised 4 trials (one ofeach name condition type) and each infant experienced two sessions.Thus, each conditionwas repeated twice for each infant. This procedurecan reliably evaluate the degree to which infants recognize familiarsounds (e.g., Lebedeva and Kuhl, 2010). The presentation order of thefour conditions was counterbalanced across participants. Presentationof visual and auditory stimuli was controlled using codewritten in Visu-al Basic 6.0.

Data analysisFixation durations toward audio-visual stimuli (henceforth listening

duration) were coded by two trained individuals (coders) at 100ms in-tervals using behavioral coding software (GenobsX) that judged eye po-sition and movements. To check the reliability of the first coder, asecond coder who had no knowledge of the experimental conditionsalso coded roughly half the data. The correlation between the codedvalues of the two observers was statistically significant with a correla-tion value of 0.98 (Pearson correlation, P b 0.001) confirming the reli-ability of the first coder. Thus, we used the behavioral data coded bythe first observer for analysis.

Results

As shown in Fig. 2, the mean listening durations were (mean ±SEM) 12.3 ± 7.3 s for the self-name/mother condition, (mean ± SEM)10.3 ± 7.6 s for other-name/mother, (mean ± SEM) 12.9 ± 6.7 s forself-name/stranger, and (mean ± SEM) 10.3 ± 5.1 s for other-name/stranger. A two-way analysis of variance (ANOVA) with name (self-name or other-name) and speaker (mother or stranger) as within-subject factors revealed a significant main effect of name, indicatingthat listening duration to self-names was significantly longer thanto other-names (F1,16 = 6.00, P = 0.026, partial η2 = 0.27). Therewas no statistically significant main effect for speaker (F1,16 = 0.03,P = 0.857, partial η2 = 0.00) or any interactions between name andspeaker (F1,16 = 0.03, P= 0.862, partial η2 = 0.00), indicating that in-fants preferred their own name regardless of voice familiarity.

Discussion

We found that 6-month-old infants preferred their own name toother names. Our finding is consistentwith those of previous behavioral

trial in Experiment 1.

Fig. 2.Mean listening times to auditory stimuli in Experiment 1 (N=17). Error bars indi-cate 1 standard error of the mean (N = 17). *P b 0.05.

479M. Imafuku et al. / NeuroImage 103 (2014) 476–484

studies showing responsiveness of infants to their own name (Mandelet al., 1995; Nadig et al., 2007). Although no previous studies have in-vestigated the impact of caretakers' voice on name perception, wefound that speaker's voice familiarity did not affect the degree of self-name preference. The results of Experiment 1 showed that 6-month-old Japanese infants discriminate their own name from other names re-gardless of speaker familiarity. These results suggest that 6-month-oldinfants have already been aware of the phonological word form oftheir name and that such familiar phonological structure can grabtheir attention regardless of the voice familiarity. However, at thispoint we cannot conclude which factor played a crucial role in gettinginfants' attention. The factor that catches infants' attention could beeither familiarity of the word form or self-referencing induced by thename. This can be alternatively rephrased as the awareness of auditorystimulus at the perceptual level or recognition of the self at the cognitivelevel. To further investigate these possibilities, we plan to conductNIRS study using a similar experimental paradigm to examine bothbehavioral and neurophysiological data together.

Experiment 2

Although the behavioral study did not show any effect of speakertype on name preference, brain activity associated with self-namesmay still differ depending on voice familiarity. Therefore, to investigatewhether cerebral responses to self-names and other-names differ in6-month-olds, we measured hemodynamic responses while infantsheard names (self or other) spoken by either their own mother or bya stranger.

Materials and methods

ParticipantsSeventeen 6-month-old infants (6 boys and 11 girls; mean age,

185.6 ± 21.9 days) participated in Experiment 2. Four other infantswere tested in our experiment but were excluded from the analysis be-cause of fussiness. Ten of the 17 infants had also participated in Experi-ment 1. For those 10 infants, we performed the two experiments atleast 1 week apart to avoid any habituation bias. As in Experiment 1,all mothers were judged to be healthy without any cognitive deficits.All mothers participated as paid volunteers.

StimuliAuditory stimuli were prepared in the same manner as in Experi-

ment 1. All infants completed two 3.8-min sessions, one with themother's voice and one with a stranger's voice. We divided the sessioninto two because 7.6 min in total is too long to keep their attention.There was a few minutes rest period between the two sessions andthe session orderwas counterbalanced across participants.We followedthe procedures used in Naoi et al. (2012). For each session, infants expe-rienced six self-name trials and six other-name trials, which were

randomly presented. We used a block design such that each trialconsisted of a silent period as the baseline block (10 s) and a name stim-ulation period as the target block (8 s). During the target block, a namestimulus corresponding to one of the four conditions was repeated fivetimes with an ISI of approximately 300 ms. Length of ISI was a littleshorter than that of the behavioral study in Experiment 1 because dura-tion of one block in such functional imaging study is desired to be nottoo long but concise with enough information. However, the differencedid not seem to affect the results, becausewe aimed to compare the fourconditions in each experiment.

ProcedureInfants watched a silent puppet show to keep calm throughout the

fNIRS session (both baseline and target blocks). Their behavior was re-corded on a DVD throughout the experiment to assess movements.Using a multichannel fNIRS system (ETG-7000, Hitachi Medical Co.,Japan), we measured changes in hemoglobin (Hb) concentration andHb oxygenation levels in the frontal brain region (Fig. 3A). We used asilicon probe pad in which eight emission and seven detection probeswere arranged in a 3 × 5 rectangular lattice, with probes separatedby 20 mm. The estimated spatial resolution was 15–30 mm. Near-infrared light penetrates deeper into the brain in infants comparedwith adults because infant brains contain less reflective white matter(Fukui et al., 2003). Thus, the 20-mm distance between the sourcesand detectors in our study was able to capture brain activity in deeperregions of infant PFC. The estimated depth of near-infrared light wasaround 20–35 mm from skin surface. The probe was placed on the par-ticipants' frontal cortical areas and the resultant number of recordingchannels was 22, as shown in Fig. 3A. The lowest lines of the probeswere positioned in a direction parallel to the T3–Fp1–Fp2–T4 line inthe international 10–20 system and the center of the channels wasplaced across thenasion–inion line (Jasper, 1958). These procedures en-abled us to use the spatial estimation of the brain based on the virtualregistration method (Tsuzuki et al., 2007). After probe placement, theexperimenter checked if each probe was adequately in contact withthe scalp. If the attached check was verified, then we started NIRSrecording.

Data analysisNIRS data were analyzed with Platform for Optical Topography

Analysis Tools (POTATo) (Research & Development Group, Hitachi,Ltd.)withinMatLab 7.7 (TheMathWorks, Inc., Natick,MA,USA).We cal-culated changes in the concentrations of oxy-Hb and deoxy-Hb, whichwere estimated based on the change in absorbance by using laserbeams of approximately 780 and 830 nm, sampled at 10 Hz. The rawdata of oxy- and deoxy-Hb in each channel were high-pass filtered at0.0143 Hz and low-pass filtered at 0.8Hz to remove components arisingfrom physiological signals (i.e., respiratory and cardiac activity) andbody movements (Naoi et al., 2012; Taga et al., 2003). We excludedblocks with motion artifacts (signal variations greater than 2 standarddeviations from the mean over 0.2 s). After the removal of inappropri-ately fitted channels and blocks including motion artifacts, all datawere averaged. Baselines of all NIRS signals were corrected using linearfitting (Penã et al., 2003; Sato et al., 2007). Consequently, we removedthe data of nine lower channels because the amplitudes of the changein the concentration of oxy-Hb and deoxy-Hb were too large andchanged together in the same direction, indicating loose attachment ofprobes in the lowest line. This was caused by unfitted silicon probepad against the lower part of the forehead.

The mean numbers of trials for the 17 infants were 5.94 for the self-name by mother's voice condition, 5.53 for the self-name by stranger'svoice condition, 5.94 for the other-names by mother's voice condition,and 5.88 for the other-names by stranger's voice condition. Since hemo-dynamic peak latency was 5.1 s on average across all participants andchannels, we defined the analysis time window as between 3 and 7 safter the onset of the stimulus. Mean concentrations of oxy-Hb and

480 M. Imafuku et al. / NeuroImage 103 (2014) 476–484

deoxy-Hb in each time window were calculated for all channels andeach subject.

To determine brain activity during each condition, changes betweenhemodynamic responses during name stimulation and baseline werecompared during 4-s analysis windows for each channel and each in-fant. To find out which brain regions were activated, the differences inthe changes in Hb concentration between baseline and target blockswere assessed using a two-tailed t-test for each channel under all fourconditions. We used the false discovery rate (FDR) to correct for multi-ple comparisons across all channels. The region of interest (ROI) wasbased on this result as well as our hypothesis. In the next analysis, weapplied a two-way ANOVA with name (self-names or other-names)and speaker (mothers or strangers) as a within-subject factors on theresponsemagnitude of oxy-Hb changes (changes in Hbduring the aver-aged 4 s baseline block were subtracted from the averaged 4 s targetblock) for the ROIs. Effect sizes were also calculated as indexed bypartial eta-squared (ηp2). Our ROIs resulted in one area in the prefrontalcortex that combined two proximate channels (CH16 and CH21). Thisbrain region corresponds with the dorsal–medial prefrontal cortex(dmPFC), based on Automatic Anatomical Labeling (AAN) (left superiormPFC= 73.8%, right superiormPFC= 26.1% for CH16 and right superi-or mPFC= 51.1%, frontal superior cortex= 48.9% for CH21), accordingto the spatial registration for NIRS channels (Okamoto et al., 2004;Okamoto and Dan, 2005; Tsuzuki et al., 2007). This method enablesthe placement of a virtual probe holder on the scalp by simulating theholder's deformation and by registering probes and channels onto thecanonical brain template in the standard stereotaxic coordinate system.This probabilistic registration is based on MRI anatomical database andcan be used with a scalp-based positioning system, such as the 10–20system. This method has been reliably used in the fNIRS literature(e.g. Dan et al., 2013, Minagawa-Kawai et al., 2009). These ROIs coverthe same area (themedial surface of the dorsal frontal gyrus) identifiedin previous adult literatures. Although this is the estimation for adulthead/brain, this is also applicable for infant's head. This is because weuse the probe separation of 20 mm instead of 30 mm employed by thespatial registration for adults. As infant circumference of head is roughlyone third of adults, the aforementioned modification resulted in similarrelationship between the probe positions and brain areas of infants andadults, because the relative scalp position between infants and adults isnot different (Barkovich et al., 1988; Minagawa-Kawai et al., 2009).Furthermore, examining infant's MRI anatomical images of the brain,Matsui et al. (2014) recently reported that macroanatomical structureswere generally comparable between adult and infant atlases, indicatingthat the virtual registration employed for fNIRS can be also reliablyapplied to infant brain. Finally, to examine the correlation coefficients be-tween the infants' behavioral responses (listening duration) in Experi-ment 1 and the neural responses measured with NIRS in Experiment 2,we analyzed the correlations between the behavioral results and the Hbchanges with a Spearman rank for 10 infants participating in both Exper-iments 1 and 2.

Results

Increases in oxy-Hb concentration were observed in prefrontalareas in response to all name stimuli, regardless of condition. How-ever, compared with the other three conditions, a larger increasewas observed when infants heard their own name spoken by theirown mother (Fig. 3B). While statistically significant differencesin the changes in oxy-Hb for channel (CH) 21 (P = 0.0011, t = 3.22)

Fig. 3. (A) Grand averaged time courses of changes in oxy- and deoxy-Hb in response toself-name being called by their own mother, and larger responses in CH16 and CH21,which was estimated to be around the dmPFC. (B) Hemodynamic responses to auditorystimuli different in name and voice conditions (CH16 and CH21). The red line showsthe change of oxy-Hb and the blue line does the change of deoxy-Hb. (C) Changes inoxy-Hb vs. baseline in response to four different conditions during the time windows.Error bars indicate 1 standard error of the mean (N = 17). *P b 0.05.

481M. Imafuku et al. / NeuroImage 103 (2014) 476–484

and CH16 (P = 0.0032, t = 3.01) were observed during this condition(self-name/mother), changes observed in the other three conditionsdid not survive the correction for multiple comparisons.

As mPFC was the targeted area for our study, and because the CH-based analysis revealed some critical regions in relation to nameprocessingwithin themPFC, we defined our ROI by combining two prox-imate channels from two brain regions (CH16 and CH21; dorsalmPFC). Atwo-way ANOVA with factors of name and speaker revealed statisticallysignificant main effects for both factors in dmPFC (Fig. 3C; name:F1,16 = 7.65, P = 0.014, ηp2 = 0.32; speaker: F1,16 = 5.82, P = 0.028,ηp2 = 0.27). Post-hoc tests (simple main effect tests) demonstrateda greater response to the mother's voice than to a stranger's voice(F1,16 = 4.46, P = 0.043) and a greater response to the self-namethan to the other-name condition (F1,16= 5.59, P= 0.024). These resultsshow that self-names elicited larger hemodynamic responses than other-names. Furthermore, greater activation was observed when names werespoken in a mother's voice than in a stranger's voice. These activatedchannels particularly cover the anterior part of dmPFC, taking intoaccount the NIRS light propagation. There was no statistically significantinteractions (F1,16 = 0.20, P= 0.658, ηp2 = 0.01).

Further, we found a significant positive correlation between listen-ing duration (in Experiment 1) and changes in oxy-Hb in the anteriordmPFC (in Experiment 2) during the combined mother conditions(Pearson correlation; r8 = 0.55, P = 0.012; Fig. 4), indicating that be-havioral preferences for mothers' voices was accompanied by largerbrain activation in the anterior dmPFC. In contrast, no statistically signif-icant correlations were found between behavioral preferences and oxy-Hb changes for the stranger conditions (r8 = 0.24, P = 0.316).

Discussion

Experiment 2 showed that both name and speaker influenced theamount of activation in the anterior part of dmPFC in 6-month-oldinfants. Specifically, hearing their own name elicited larger cerebral re-sponses than other names, and this increase was enhanced when thenames were spoken by their own mothers. Grossmann et al. (2010)showed more activation in the left dorsolateral prefrontal cortex of5-month-old infants after they heard their own names than afterother names. Our finding was consistent with Grossmann's study inthat the dorsal prefrontal area was activated more in response to self-names. However, in our study the activation was more medial, andstronger in the right hemisphere. These differences may have resultedfrom our use of a familiar voices, which possibly stimulated recognitionof intentional signals toward self. Mother gaze-shifts had a larger influ-ence on infant object processing than stranger gaze-shifts did in4-month-old infants, suggesting that perceptual familiarity and previ-ous interactions facilitate gaze processing (Hoehla et al., 2011). There-fore, processing social signals from mothers and strangers may differ

Fig. 4. Correlation between listening time and prefrontal cortex (dmPFC; CH16 and CH21) respoparticipating both in Experiments 1 and 2 (n= 10). Inmother's voice condition (left sides in thsure (looking time) (x-axis) were significantly correlated with oxy Hb changes (y-axis) in the dn.s., not significant.

in infants when the signals are directed toward self, and this mayexplain the difference we observed in activation area.

The present study shows that 6-month-old infants are sensitiveto differences in social signals as exemplified by self-names and other-names. Furthermore, brain activation in response to names was stron-ger when the names were spoken by the infants' mothers. Because nostrong activation was observed for self-names uttered by strangers,this voice factor seems to be particularly important for eliciting cerebralresponses related to social cognition. Rather than in lateral regions,the present study shows brain activation in response to self-namesin the medial prefrontal area related to self-referential processing(e.g., Kampe et al., 2003). It could be that at this age with a small socialworld, infants, when hearing their name, only associate the mother'svoice with social intention. Further support for this interpretationcomes from the fact that a significant correlation between dmPFC activ-ity and name preference in our behavioral results was observed for themother condition. In summary, the anterior part of the dorsal–medialfrontal cortex is involved in the processing of self-names in 6-month-old infants, and this may be facilitated by a mother's voice. However,there are still other possible interpretations about the present data.Such interpretations involve familiarity effect and affection relatedbrain activation. Familiarity and affection could serve as strong candi-dates that explain the dmPFC activations. Thus we will detail themin the general discussion by taking into account both behavioral andfNIRS data.

General discussion

We employed behavioral and neuroimaging methods to examinewhether voice familiarity affects responses to self-names in 6-month-old infants; there has been no study so far addressing this topic. Previ-ous studies in infants using these two methods have also been limited.Our study showed that the dorsal–medial prefrontal area is engagedin name recognition, and that a mother's voice has a strong impact onname processing in this region at this age. The results of the behavioralstudy showed that 6-month-old infants significantly preferred theirownname to other names, regardless of the speaker, suggestingpercep-tual sensitivity for their own name. Our findings are in agreement withthose that have shown first-name recognition in 4.5-month-old infants(Mandel et al., 1995). To further investigate the underlying mechanismbehind the recognition of social signals related to self, we combinedanalysis of behavioral and fNIRS data and further revealed a significantcorrelation between the dmPFC responses and listening durations toself-name stimuli. Our result suggests a possible role of processingself-related stimuli for this brain region. However, there are other pos-sible factors that contribute to elicit dmPFC activations. Thus this sectiondiscusses such possibilities. We will first focus on the self-referencingfunction and then we will examine familiarity and affective factorsthat may be involved in the dmPFC activations.

nse to self-name spoken by their ownmother's or a stranger's voice in 6-month-old infantse figure), therewere 10 plots in each name condition and in sum 20 plots. Behavioralmea-mPFC in the mother's voice condition, but not in the stranger's voice condition. *P b 0.05;

482 M. Imafuku et al. / NeuroImage 103 (2014) 476–484

Two previous neurophysiological studies in adults have investigatedthe effects of both name and speaker on name processing. Using ERP,Holeckova et al. (2006) showed that the prefrontal regionwasmore ac-tive in response to a person's own name spoken by a familiar voice thanwhen spoken by an unfamiliar voice. In a combined ERP and PET study,Holeckova et al. (2008) further examined how familiar and unfamiliarvoices affect cerebral blood flow in adult mPFC. Again, they reported asignificant effect of speaker familiarity in name processing, implyingspecific higher-level processing of names spoken in a familiar voicecompared with those spoken in an unfamiliar voice. In adults, it hasbeen shown that the right mPFC plays an important role in processingone's own name (Kampe et al., 2003). Furthermore, by using ERP andPET, Perrin et al. (2005) showed that the mPFC is activated when hear-ing one's own name, and that mPFC activation as obtained by PET mea-surement is more correlated with the amplitude of the P3 componentobtained for the self-name than with that obtained for other-names.This suggests that the mPFC plays the most prominent role in name-processing. Additionally, the right mPFC is also activated in responseto other stimuli relevant to the notion of self, such as a person's ownface and voice (Keenan et al., 2001; Nakamura et al., 2001; Uddinet al., 2005).

In the present study, significant activity was found at CH16 andCH21, which according to the spatial registration for NIRS in adults,correspond to mPFC and frontal superior cortex, respectively (Tsuzukiet al., 2007). Taking into account that the NIRS light propagation isaround the white matter in infancy (Fukui et al., 2003), these areasmay include anterior part of the right mPFC. Additionally, fMRI studiesindicate that the surfaces of Brodmann areas 9 and 10 (which includemPFC) are involved in the detection of first names (Kampe et al.,2003; Perrin et al., 2005). Consequently, our results suggest that6-month-old infants recruit a similar neural region as adults whentheir own name is called by their own mother. We speculate that thisregion,which has limited function at a young infant age, is the early pre-cursor of self-referencing. Self-referencing is thefirst candidate among afew to interpret the dmPFC activation. The following paragraphs willdiscuss the dmPFC role either by supporting the interpretation associat-ed with self-referencing from literature or limiting its function at thisage.

According to Neisser (1988), interaction with others is essential forthe development of the concept of self and that the interpersonal selfcan be viewed as self-consciousness that is triggered by interactionwith others. Our findings may be interpreted to mean that in the firsthalf year of life, social interaction is chiefly initiated by the mother,resulting in specific stimuli (e.g. the infant's name) triggering self-consciousness. Plenty of evidence supports the special role of themater-nal stimuli — particularly in a communicative context (DeCasperand Fifer, 1980; Montague and Walker-Andrews, 2002). A meta-analysis of imaging studies regarding the self has shown activationin mPFC while viewing self-related stimuli such as autobiographicalmemory and self-body images, including faces (Northoff et al., 2006).Another meta-analysis of social cognition (Van Overwalle, 2009) in-dicates that the activation foci corresponding with self-referencingare located inferiorly, and that the dmPFC is more engaged in pro-cessing “traits of others”. Thus, the activation of the anterior dmPFCin this study may simply represent a response to the external stimuliof the mother's voice rather than an indication of self-referencing.However, this is unlikely because several studies have reportedthat dmPFC is particularly related to mentalizing used in self-referential tasks (Fossati et al., 2003; Amodio and Frith, 2006).An fMRI study on name perception in adults that is comparablewith our study (Kampe et al., 2003) also found activation in thedmPFC. Furthermore, Sassa et al. (2007) showed stronger responsesin the polar part of themPFC to communicative speech produced by afamiliar actor than by an unfamiliar actor, suggesting that activity inanterior mPFC is related to understanding social interactions in thecontext of autobiographical information, and that familiarity with

the speaker is integrated into the communicative context. On thebasis of these previous studies and in the activation we observed inthe anterior part of dmPFC in response to names, we speculate that6-month-old infants have internalized their names as self-relatedstimuli that are exclusively produced by their own mother. Here,we should again emphasize that this activation may not reflectrepresentation of self as observed in adults but very early formof self-referencing which was developed through mother–infantinteraction.

Ontogenetically, the capacity to understand other people's actions asgoal-directed and infer others' intention emerges at around 6 months ofage (Kanakogi and Itakura, 2011; Kovács et al., 2010;Woodward, 1998).Accordingly, brain activation in the dmPFC found here may be relatedto detecting intentions toward one's self from another person,which in-cludes recognizing self-names in the context of communication. Suchsocial signals that are directed to the self are reported to be hard to rec-ognize for individuals with autism spectrum disorders (ASD) (Frith,2001). Thus, infants at risk for autism show weak behavioral responsesto their own name even at the age of 6 and 12 months (Nadig et al.,2007; Osterling et al., 2002). The investigation of the neural responseto self-name in ASD child would be valuable. Nonetheless, the “self”presumed to be related to dmPFC activity in this study cannot be themature form of self which involves the extender self, the private self,and the concept of self that is defined by Neisser (1988). The represen-tation of self that can be observed here may only involve what Neisserrefers to as the interpersonal self.

One of the significant findings of the present study is that dmPFCactivation was higher in response to the self-name being spoken bythe mother than by a stranger. Research in developmental psychologyhas shown that themother plays a crucial role in the social and cognitivedevelopment of infants. Neonates prefer their own mothers' voices,suggesting that the maternal voice is significant even before birth(DeCasper and Fifer, 1980). A recent ERP study showed that a mother'svoice elicits stronger responses in neonate language-relevant brainareas than a stranger's voice does (Beauchemin et al., 2010). Asfor the social cognitive domain, a behavioral study has reportedthat because infants spend most of their time with their mothers, at3.5 months of age infants can match the emotional voice and face oftheir mother, but not of their father or of strangers (Montague andWalker-Andrews, 2002). These findings suggest that human infantsperceive their mother as a special agent very early in life, and socialcognitive abilities are acquired through interaction with the mother.

Here, we found that the anterior dmPFC was activated when theinfants' own names were called by their mothers; this may be becauseinfants at this stage recognize communicative intentions toward theself from their mothers, but do not yet recognize the same intentionsfrom others. Although infants behaviorally responded to the wordform of their own name, they may not yet have been fully enabled todetect intentions from non-mothers. We interpret the dissociation inresponses that we observed in the behavioral and neurophysiologicaldata to derive from this factor. Specifically, behavioral preference inExperiment 1 may chiefly reflect perceptual attention to familiar wordform, whereas dmPFC activation in Experiment 2 may derive from cog-nitive factor of self-referencing at the very primitive form. We assumethat perceptual attention to familiar word form wasn't observed asHb responses in PFC, because such external attention often deacti-vate mPFC in general (Gusnard and Raichle, 2001). Even in neonates,familiarity probably without self and affective factors deactivatedPFC (Uchida-Ota et al., 2011). It may be that the familiar word form(i.e. name) spoken by strangers may have drawn external attention forsome infants causing deactivation, but elicited weak self-referencing forsome infants at the same time, resulting in weak Hb changes in average.Itmay be also possible that external attention evoke cerebral responses inthe temporal area, which we didn't measure this time. Because this isonly a tentative interpretation, we should further examine how percep-tual level of attention differentially activates infant brains in the future.

483M. Imafuku et al. / NeuroImage 103 (2014) 476–484

Further studies should also measure older infants using fNIRS to deter-mine how this ability changes with age, and how infants learn to applythis ability to strangers.

Finally, here we discuss a couple of alternative interpretations of thepresent data pertaining to affect related activation and familiarity effect.As stated by Grossmann (2013), activation of mPFC can be associatedwith processing affective stimuli. Because self-name spoken by mother'svoice may involve the most affection among the four conditions in thisstudy, the affection factor should be taken into consideration to explainthe results. As affective related activation is not specific to the dmPFCbut observed in different areas such as in the anterior orbital part of PFC(Minagawa-Kawai et al., 2009) and left side of dorsal PFC (Naoi et al.,2012) in the case ofmaternal affection, we assume that affection is a sub-sidiary interpretation to the dmPFC activation. Furthermore, a significantmain effect of name type in Experiment 2 seems not to relate to affectivefactor much because difference in self- and other-name is generally notdirectly related to affective factors.

In fact, familiarity factor more plausibly explains the results than theaffective factor. For both voice and name conditions, familiarity is stron-ger in self-name and mother-voice conditions than other-name andstranger-voice conditions. This is consistent with our NIRS experimentresults showing stronger responses to the formers. Although the samelogic cannot be flippantly applied to the behavioral data, behavioraldata can also be explained by familiarity factor if we hypothesize differ-ent type of familiarity exclusively for the behavioral data. This differenttype of familiarity is specifically the familiarity of the word form and itsphonetic representation. In case of behavioral preference, the infants'gaze preference may have influenced only by sound form of names.Nevertheless, there are some facts that weaken the familiarity interpre-tation. First, mPFC is not specific area to process familiarity (Skinner andFernandes, 2007). Second, familiarity may trigger one's attention to ex-ternal stimuli and such attention modulation often induce decreasedbrain activations around mPFC (Gusnard and Raichle, 2001) as ex-plained in the previous paragraph. This is contrary to the present resultsof increased Hb changes. On the other hand, we can expect increasedbrain activities in mPFC, if the familiarity factor induces attention to in-ternal representations such as infants' self-related factors or autobio-graphic memories based on one's experience. In this sense, familiarityis crucially related to the sense of self-referencing. It is true that familiar-ity plays a significant role in recognizing oneself in a social communica-tive context or evoking autobiographic memories in a neuroimagingliterature (e.g. Sassa et al., 2007). If we take the familiarity of this partic-ular sense, we can hypothesize that the familiarity factor cruciallyevoked dmPFC activation in this study. However, teasing apart self-referencing function and familiarity triggering attention to internal rep-resentation is difficult in this experiment.

In conclusion, the most likely explanation of dmPFC activation in6-month-olds for the self-name andmother-voice condition is the func-tion of self-referencing, which is related to familiarity factor triggeringinternal attention. Evidence supporting this explanation includes, thecorrelation of the dmPFC activation and behavioral preferences inmother's voice, function of brain region (i.e. dmPFC) fromneuroimagingliterature and a pattern of oxy-Hb response (i.e. increased pattern).However, since this set of evidence is not strong enough to completelyconclude self-referencing function in 6-month-olds, this issue shouldbe still open to further examination.

Conclusion

Using the fNIRS and behavioral measures, we demonstrated thatthe anterior dmPFC in 6-month-old infants is more sensitive to theirown names and to their own mothers' voices than other names andstranger's voices. As the adult dmPFC is critically related to the process-ing of social stimuli related to the sense of self and the activation inthis area was significantly correlated with the behavioral measureonly for the mother's voice condition, the activation seen here may be

interpreted as a precursor to self-referencing facilitated by a familiarvoice. Although dmPFC activation more likely signals self-referencingin a social communicative context, other factors such as familiarity areassumed to be indirectly involved in this process. Thus further examina-tion to tease apart those factors is desired in the future studies to pre-cisely determine the onset of self-referencing in human infants. Still,our findings demonstrate neuro cognitive bases for earlier social abili-ties than previously expected — particularly for mothers' communica-tive context.

Acknowledgments

This work was supported by the Global Center of Excellence (GCOE)program at Keio University, Grant-in-Aid for Scientific Research(KAKENHI) (B) (Project No. 24118508) and Grant-in-Aid for ScientificResearch on Innovation Areas (Project No. 24300105) to YM. Part ofthis work was the undergraduate thesis of the first author. We wouldlike to thank all the children and their parents who participated in thisresearch at Keio Baby Lab. We also thank S. Ishii for assistance withthe experiments, M. Myowa-Yamakoshi and lab members of KyotoUniversity for the helpful suggestions, I. Imahori for proof reading andR. Wu for helpful comments on earlier drafts of this paper.

References

Amodio, D.M., Frith, C.D., 2006. Meeting of minds: the medial frontal cortex and socialcognition. Nat. Rev. Neurosci. 7, 268–277.

Barker, B.A., Newman, R.S., 2004. Listen to your mother! The role of talker familiarity ininfant streaming. Cognition 94, 45–53.

Barkovich, A.J., Kjos, B.O., Jackson Jr., D.E., Norman, D., 1988. Normal maturation of theneonatal and infant brain: MR imaging at 1.5 T. Radiology 166, 173–180.

Beauchemin, M., González-Frankenberger, B., Tremblay, J., Vannasing, P., Martinez-Montes, E., Belin, P., et al., 2010. Mother and stranger: an electrophysiological studyof voice processing in newborns. Cereb. Cortex 21, 1705–1711.

Bortfeld, H., Morgan, J.L., Golinkoff, R.M., Rathbun, K., 2005. Mommy and me: familiarnames help launch babies into speech-stream segmentation. Psychol. Sci. 16, 298–304.

Csibra, G., 2010. Recognizing communicative intentions in infancy. Mind Lang. 25,141–168.

Dan, H., Dan, I., Sano, T., Kyutoku, Y., Oguro, K., Yokota, H., Tsuzuki, D., Watanabe, E., 2013.Language-specific cortical activation patterns for verbal fluency tasks in Japaneseas assessed by multichannel functional near-infrared spectroscopy. Brain Lang. 126,208–216.

DeCasper, A.J., Fifer, W.P., 1980. Of human bonding: newborns prefer their mothers'voices. Science 208, 1174–1176.

Dehaene-Lambertz, G.,Montavont, A., Jobert, A., Allirol, L., Dubois, J., Hertz-Pannier, L., et al.,2009. Language ormusic, mother orMozart? Structural and environmental influenceson infants' language networks. Brain Lang. 114, 53–65.

Fossati, P., Hevenor, S.J., Graham, S.J., Grady, C., Keightly, M.L., Craik, F., et al., 2003. Insearch of the emotional self: an fMRI study using positive and negative emotionalwords. Am. J. Psychiatry 60, 1938–1945.

Frith, U., 2001. Mind blindness and the brain in autism. Neuron 32, 969–979.Fukui, Y., Ajichi, Y., Okada, E., 2003. Monte Carlo prediction of near-infrared light propa-

gation in realistic adult and neonatal head models. Appl. Opt. 42, 2881–2887.Grossmann, T., 2013. The role of medial prefrontal cortex in early social cognition. Front.

Hum. Neurosci. 7, 1–6.Grossmann, T., Parise, E., Friederici, A.D., 2010. The detection of communicative signals

directed at the self ininfant prefrontal cortex. Front. Hum. Neurosci. 4, 201–205.Gusnard, D.A., Raichle, M.E., 2001. Searching for a baseline: functional imaging and the

resting human brain. Nat. Rev. Neurosci. 2, 685–694.Hakuno, Y., Omori, T., Yamamoto, J., Minagawa-Kawai, Y., 2012. A preliminary study on

infants' learning of word–object relations: the role of social interaction. In:Watanabe, S. (Ed.), CARLS Series of Advanced Research. vol. 5. Keio UniversityPress, Tokyo, pp. 177–182.

Hoehla, S., Wahla, S., Michela, C., Striano, T., 2011. Effects of eye gaze cues provided by thecaregiver compared to a stranger on infants' object processing. Dev. Cogn. Neurosci.2, 81–89.

Holeckova, I., Fisher, C., Giard, M.H., Delpuech, C., Morlet, F., 2006. Brain responses to asubject's own name uttered by a familiar voice. Brain Res. 1082, 142–152.

Holeckova, I., Fisher, C., Morlet, D., Delpuech, C., Costes, N., Morlet, F., 2008. Subject's ownname as a novel in a MMN design: a combined ERP and PET study. Brain Res. 1189,152–165.

Ikuzawa, M., Matsushita, Y., Nakase, A., 2002. Kyoto Scale of Psychological Development2001. Kyoto International Social Welfare Exchange Centre, Kyoto, (in Japanese).

Jasper, H.H., 1958. The ten-twenty electrode system of the international federation.Electroencephalogr. Clin. Neurophysiol. 10, 367–380.

Kampe, K.W., Frith, D., Frith, U., 2003. “Hey John”: signals conveying communicativeintention toward the self activate brain regions associated with “mentalizing”,regardless of modality. J. Neurosci. 23, 5258–5263.

484 M. Imafuku et al. / NeuroImage 103 (2014) 476–484

Kanakogi, Y., Itakura, S., 2011. Developmental correspondence between action predictionand motor ability in early infancy. Nat. Commun. 2, 1–6.

Keenan, J.P., Nelson, A.M., Pascual-Leone, A., 2001. Self-recognition and the right hemi-sphere. Nature 409, 305.

Kovács, Á.M., Téglás, E., Endress, A.D., 2010. The social sense: susceptibility to others'beliefs in human infants and adults. Science 330, 1830–1834.

Lebedeva, G.C., Kuhl, P.K., 2010. Sing that tune: infants' perception of melody and lyricsand the facilitation of phonetic recognition in songs. Infant Behav. Dev. 33, 419–430.

Mandel, D.R., Jusczyk, P.W., Pisoni, D.V., 1995. Infant's recognition of sound patterns oftheir own names. Psychol. Sci. 6, 314–317.

Matsui,M., Homae, F., Tsuzuki, D.,Watanabe, H., Katagiri,M., Uda, S., et al., 2014. Referentialframework for transcranial anatomical correspondence for fNIRS based on manuallytraced sulci and gyri of an infant brain. Neurosci. Lett. 88, 55–68.

Minagawa-Kawai, Y., Matsuoka, S., Dan, I., Naoi, N., Nakamura, K., Kojima, S., 2009.Prefrontal activation associated with social attachment: facial-emotion recognitionin mothers and infants. Cereb. Cortex 19, 284–292.

Montague, D.P.F., Walker-Andrews, A.S., 2002. Mothers, fathers, and infants: the role offamiliarity and parental involvement in infant's perception of emotion expressions.Child Dev. 73, 1339–1352.

Nadig, A., Ozonoff, S., Young, G.S., Rozga, A., Sigman, M., Rogers, S., 2007. A prospectivestudy of response to name in infants at risk for autism. Arch. Pediatr. Adolesc. Med.161, 378–383.

Nakamura, K., Kawashima, R., Sugiura, M., Kato, T., Nakamura, A., Hatano, K., et al., 2001.Neural substrates for recognition of familiar voices: a PET study. Neuropsychologia39, 1047–1054.

Naoi, N., Minagawa-Kawai, Y., Kobayashi, A., Takeuchi, K., Nakamura, K., Yamamoto, J., et al.,2012. Cerebral responses to infant-directed speech and the effect of talker familiarity.NeuroImage 59, 1735–1744.

Neisser, U., 1988. Five kinds of self-knowledge. Philos. Psychol. 1, 35–59.Nelson, D.G., Jusczyk, P.W.,Mandel, D.R., Myers, J., Turk, A., Gerken, L., 1995. The head-turn

preference procedure for testing auditory perception. Infant Behav. Dev. 18, 111–116.Northoff, G., Heinzel, A., de Greck, M., Bermpohl, F., Dobrowolny, H., Panksepp, J., 2006.

Self-referential processing in our brain—a meta-analysis of imaging studies on theself. NeuroImage 31, 440–457.

Okamoto, M., Dan, I., 2005. Automated cortical projection of head-surface locations fortranscranial functional brain mapping. NeuroImage 15, 18–28.

Okamoto, M., Dan, H., Sakamoto, K., Takeo, K., Shimizu, K., Kohno, S., et al., 2004. Three-dimensional probabilistic anatomical cranio-cerebral correlation via the international10–20 system oriented for transcranial functional brain mapping. NeuroImage 21,99–111.

Osterling, J., Dawson, G., Munson, J., 2002. Early recognition of first year old infants withautism spectrum disorder versus mental retardation: a study of first birthday partyhome videotapes. Dev. Psychol. 14, 239–252.

Parise, E., Friedrici, A.D., Striano, T., 2010. “Did you call me?” 5-month-old infants ownname guides their attention. PLoS ONE 5, 1–9.

Penã, M., Maki, A., Kovac, D., Dehaene-Lambertz, G., Koizumi, H., Bouquet, F., et al., 2003.Sounds and silence: an optical topography study of language recognition at birth.Proc. Natl. Acad. Sci. U. S. A. 100, 11702–11705.

Perrin, F., Maquet, P., Pergneux, P., Ruby, P., Degueldre, C., Balteau, E., et al., 2005. Neuralmechanisms involved in the detection of our first name: a combined ERPs and PETstudy. Neuropsychologia 43, 12–19.

Purhonen, M., Kilpeläinen-Leesb, R., Valkonen-Korhonena, M., Karhub, J., Lehtonena, J.,2005. Four-month-old infants process own mother's voice faster than unfamiliarvoices—electrical signs of sensitization in infant brain. Cogn. Brain Res. 24, 627–633.

Sassa, Y., Sugiura, M., Jeong, H., Horie, K., Sato, S., Kawashima, R., 2007. Cortical mecha-nism of communicative speech production. NeuroImage 37, 985–992.

Sato, Y., Sogabe, Y., Mazuka, R., 2007. Brain responses in the processing of lexical pitch–ac-cent by Japanese speakers. Neuroreport 18, 2001–2004.

Seth, A.K., Dienes, Z., Cleeremans, A., Overgaard, M., Pessoa, L., 2008. Measuring con-sciousness: relating behavioural and neurophysiological approaches. Trends Cogn.Sci. 12, 314–321.

Skinner, E.I., Fernandes, M.A., 2007. Neural correlates of recollection and familiarity: a re-view of neuroimaging and patient data. Neuropsychologia 45, 2163–2179.

Taga, G., Asakawa, K., Maki, A., Konishi, Y., Koizumi, H., 2003. Brain imaging in awake infantby near-infrared optical topography. Proc. Natl. Acad. Sci. U. S. A. 100, 10722–10727.

Tsuzuki, D., Jurcak, V., Singh, A.K., Okamoto, M., Watanabe, E., Dan, I., 2007. Virtual spatialregistration of stand-alone fNIRS data to MNI space. NeuroImage 34, 1506–1518.

Uchida-Ota, M., Arimitsu, T., Yagihashi, T., Kojima, S., Ikeda, K., Takahashi, T., et al., 2011.Preliminary report on differences in cerebral hemodynamic responses to mother'svoice between preterm and term neonates. In: Watanabe, S. (Ed.), CARLS Series ofAdvanced Research. vol. 5. Keio University Press, Tokyo, pp. 137–146.

Uddin, L.Q., Kaplan, J.T., Molnar-Szakacs, I., Zaidel, E., Iacoboni, M., 2005. Self-face recogni-tion activates a frontoparietal “mirror” network in the right hemisphere: an event-related fMRI study. NeuroImage 25, 926–935.

Van Overwalle, F., 2009. Social cognition and the brain: a meta-analysis. Hum. BrainMapp. 30, 829–858.

Woodward, A.L., 1998. Infants selectively encode the goal object of an actor's reach.Cognition 69, 1–34.