synchrony and specificity in the ... - home | openscholar · new research synchrony and...

NEW RESEARCH

Synchrony and Specificity in the Maternaland the Paternal Brain: Relations to

Oxytocin and VasopressinShir Atzil, M.A., Talma Hendler, M.D., Ph.D., Orna Zagoory-Sharon, Ph.D.,

Yonatan Winetraub, B.A., Ruth Feldman, Ph.D.

Objective: Research on the neurobiology of parenting has defined biobehavioral synchrony,the coordination of biological and behavioral responses between parent and child, as acentral process underpinning mammalian bond formation. Bi-parental rearing, typicallyobserved in monogamous species, is similarly thought to draw on mechanisms of mother–father synchrony. Method: We examined synchrony in mothers’ and fathers’ brainresponse to ecologically valid infant cues. Thirty mothers and fathers, comprising 15couples parenting 4- to 6-month-old infants, were visited at home, and infant play wasvideotaped. Parents then underwent functional magnetic resonance imaging scanningwhile observing own-infant compared with standard-infant videos. Coordination in brainresponse between mothers and fathers was assessed using a voxel-by-voxel algorithm, andgender-specific activations were also tested. Plasma oxytocin and arginine vasopressin,neuropeptides implicated in female and male bonding, were examined as correlates. Re-sults: Online coordination in maternal and paternal brain activations emerged in social–cognitive networks implicated in empathy and social cognition. Mothers showed higheramygdala activations and correlations between amygdala response and oxytocin. Fathersshowed greater activations in social– cognitive circuits, which correlated with vasopres-sin. Conclusions: Parents coordinate online activity in social–cognitive networks that supportintuitive understanding of infant signals and planning of adequate caregiving, whereasmotivational–limbic activations may be gender specific. Although preliminary, thesefindings demonstrate synchrony in the brain response of two individuals within anattachment relationship, and may suggest that human attachment develops within thematrix of biological attunement and brain-to-brain synchrony between attachmentpartners. J. Am. Acad. Child Adolesc. Psychiatry, 2012;51(8):798 – 811. Key words: neu-roimaging, parental brain, oxytocin, vasopressin, intersubject correlation

mhatOthF

B io-behavioral synchrony, the coordinationof biological and behavioral signals betweenparent and child, has been described as a

central process underlying the formation of mam-malian bonding.1 For instance, the amount of fe-male rats’ parenting behavior was found to shapethe receptor distribution of their offspring’s oxyto-cin (OT), a nine–amino acid neuropeptide synthe-sized in the hypothalamus that plays a key role insocialization and affiliation throughout life. Humanmothers and infants similarly coordinate biology

Clinical guidance is available at the end of this article.
t
JOURN

798 www.jaacap.org

and behavior. For example, mothers and infantswere found to synchronize their heart rhythmsonline during social interactions, and the extent ofsuch biological coupling was determined by thedegree of interactive synchrony between motherand child.3 Similarly, mothers who engaged in

ore synchronous interactions showed more co-erent activations of the amygdala and nucleusccumbens (NAcc) to their infant’s stimuli, andhese activations correlated with maternal plasmaT.4 In addition, synchrony has been found be-

ween mothers’ and fathers’ physiological and be-avioral response in the context of infant cues.ollowing triadic interactions between parents and
heir 6-month-old infants, maternal and paternal
AL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY

VOLUME 51 NUMBER 8 AUGUST 2012

tllpf

sd

hc

SYNCHRONY AND SPECIFICITY OF MATERNAL AND PATERNAL BRAIN

plasma OT was found to be interrelated and corre-lated with the degree of affect synchrony betweenparents. Such mother–father synchrony is thoughtto underpin the evolution of family units in mam-mals,5 to facilitate the emergence of co-parenting,and to consolidate the mother–father attachment inthe context of family formation.2

Becoming a parent involves a major neuro-hormonal reorganization that prepares for theexpression of adequate caregiving.6 Across mam-malian species, pregnancy and childbirth areassociated with marked changes in maternalbrain areas implicated in motivation, nurturance,and attention.6 Research in bi-parental speciespoints to similar alterations in fathers’ brains, con-tingent upon exposure to infant cues.7 In humans,imaging studies suggest the involvement of twoglobal networks underlying maternal care, a moti-vational–emotional limbic network that includesthe amygdala, NAcc, and anterior-cingulate cortex(ACC), and a social-attention network implicatedin social cognition and empathy, including themedial-prefrontal-cortex (mPFC), superior-temporal-gyrus (STG), insula, inferior-parietal lob-ule (IPL), and inferior-frontal gyrus (IFG).4,8-10 Thissocial-cognition and emotional-modulation cir-cuitry underlies the parent’s ability to read theinfant’s nonverbal signals and to provide synchro-nous parenting. However, because little researchhas addressed the neurobiology of human father-hood, it is unclear whether fathering involves sim-ilar integration of limbic and cortical networks andis mediated by processes related to pair bonding asin other biparental mammals. Of particular interestis whether brain mechanisms related to mother–father synchrony support the development of fa-therhood, a question that bears important implica-tions for the study of human attachment.

From an evolutionary perspective, the mater-nal and paternal roles are distinct,6 yet they mayshare underlying physiological mechanisms.11

Paternal care is observed in 3% to 5% of mam-malian species, and these species are notablymonogamous,12 suggesting that pair-bonding sup-ports the development of mammalian fathering.Animal monogamy is the result of co-evolution ofmothering and fathering expressed in facultativefathering, i.e., paternal care that increases infantendurance in the context of mothering.13 Humanpaternal investment contributes to children’s cog-nitive, social, and emotional development,14

whereas father absence increases psychopathology
and antisocial conduct.15,16 Taken together, these
JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY


findings suggest that the development of co-parenting may be supported by bio-behavioralmechanisms through which parents synchronizetheir efforts to jointly raise their child.

In addition to brain areas that may showmother–father synchrony, neurohormonal changesduring pregnancy, parturition, and lactation mayresult in distinct profiles of brain activations inmothers and fathers.17 Brain and behavioral sys-ems undergo changes with motherhood, particu-arly those implicated in motivation and vigi-ance.6 An increase in gray matter volume in therefrontal cortex, parietal lobe, and midbrain was

ound in mothers during the postpartum period,17

suggesting that activations in motivational systemsmay be more pronounced in mothers. Studies onpostpartum maternal motivation in rats differenti-ate early from late-onset motivation. Hormonalevents associated with parturition initially activateneural circuitry implicated in maternal immediateresponsiveness to pups, possibly through neuralconnectivity to motivational–dopaminergic brainareas, such as the NAcc,17 and human researchimilarly pointed to the centrality of mesolimbicopamine/reward pathways for mothering.18,19

Thereafter, continued sensory experiences throughmother–pup interactions establish long-term ma-ternal responsiveness. Late post-partum motiva-tion for infant care requires learning and memoryand involves functional reorganization of brainstructures and their connections.20 It is possiblethat the unconditional maternal motivation ini-tiated by the biology of pregnancy would bemore pronounced in mothers, expressed ingreater motivation–limbic activations. Thelater, learning-based care may be shared bymothers and fathers, leading not only to com-parable responsiveness in social– cognitive net-works but a possible synchrony of such activa-tions based on the parents’ growing experiencein co-parenting their child.

Both OT and arginine vasopressin (AVP) areneuropeptides that support mammalian socialbehavior in females and males21; yet research indi-cates dimorphism in their effects on social, affilia-tive, and sexual behavior.22 OT has been linkedwith maternal–infant bonding,1,2,10 whereas AVP

as been associated with male bonding by in-reasing aggressive and territorial behavior.22,23

AVP is thought to underlie mammalian fatheringby increasing vigilance to selective protection.24

Recent studies assessing OT and AVP in relation to
brain activations showed that whereas maternal
799www.jaacap.org

TB

ATZIL et al.

OT levels correlated with left NAcc and rightamygdala response to infant cues,4 intranasal AVPincreased men’s response in left temporo-parietaljunction (TPJ), a key theory-of-mind node.25 Thesestudies indicate that AVP may modulate social-emotional responses at the cortical level, particu-larly circuits related to emotional regulation andsocial cognition, whereas OT may be more closelylinked with motivational–limbic structures.22

In the current study, we examined whethermothers and fathers would synchronize theirbrain responses to own-infant cues. Mothers andfathers underwent functional magnetic reso-nance imaging (fMRI) scanning while observingvideos of their infant’s solitary play, and thedegree of coordination in their brain reactivitywas assessed. As we are aware of no studyassessing the time-locked coordination of brainactivations among attachment partners, suchfindings, albeit preliminary, may have significantimplication for the study of human attachment.Consistent with research on the co-evolution ofmothering and fathering, we expected that brainareas implicated in empathy and social cognition,which are based on representing the other’s statewithin one’s physiology, will synchronize be-tween the maternal and paternal brains. How-ever, the strong effects of childbirth on motiva-tional structures, including the amygdala andNAcc, were expected to result in higher activa-tions of these structures among mothers. In ad-dition, we tested correlations between plasmaAVP and OT with activations in key parentingareas. AVP was hypothesized to correlate withpaternal brain activations, whereas OT correlatedwith the maternal brain. By assessing both brainsynchrony and network specificity and their neu-rohormonal correlates, we hoped to shed furtherlight on the biobehavioral processes underlyingthe development of co-parenting in humans.

METHODParticipantsThirty parents (15 married couples) of 4- to 6-month-old, singleton, full-term infants participated, includingseven first-time and 12 veteran parents (12 breast-feeding mothers). Thus, this study addressed the firststages of parental bonding, not the transition to par-enthood. Participants were 22 through 37 years old(mean � 29.3, SD � 3.45), had completed 12 to 21 yearsof education (mean � 15.76, SD � 1.85), and had nohistory of major physical or mental illness. Parents wererecruited through advertisement in the community. Data
from some mothers, but not fathers, were also included
JOURN

800 www.jaacap.org

in a previous report.4 The study was approved by theel-Aviv Sourasky Medical Center Institutional Reviewoard, and all participants signed informed consent.

ProcedureThe study included three sessions. In the first session,families were visited at home when the infant was 4 to6 months old and were videotaped. Films were used asfMRI stimuli and included mother–infant and father–infant interactions and infant solitary play, each lasting2 minutes. In the second session, several days after thehome visit, each parent underwent brain scanning.Blood samples were collected on a third visit, to avoidthe potential impact of blood draw on brain responseor vice versa, and this decision is supported by thehigh individual stability in plasma neuropeptide levelsover several months.26

fMRI ParadigmThe fMRI paradigm was individually tailored andincluded eight consecutive, 2-minute, infant-relatedvideos with rest between stimuli. The various clipsincluded vignettes of own infant during solitary playand unfamiliar-infant during solitary play standardacross participants, and these two vignettes were usedin the current report. Parents also viewed clips of ownand standard parent–infant interactions. Stimuli orderwas counterbalanced and presented in five differentorders across participants. Clips were previewed byrest with fixation period of 1 minute. Rest with fixationperiods of 15 to 18 seconds was presented betweenfilms. All infant videos were micro-coded for infantaffective state on a computerized system (Noldus,Wageningen, the Netherlands) using a validated cod-ing scheme by trained coders to ensure that infant statedid not differ between own-infant and standard-infantvideos. In all films, infants were in neutral or positiveaffective states. The stimulus presented in the own-infant condition was individually tailored for each cou-ple, and the group analyses combines 15 different infantfilms as one own-infant condition. It is thus assumed thatbrain responses to the own-infant in parents, at a grouplevel, are beyond the infant’s specific state.

fMRI AcquisitionImaging was performed on a GE-3T Sigma Horizonecho-speed scanner with a resonant gradient echopla-nar imaging system. Functional images were acquiredusing a single-shot echo-planar T2*-weighted se-quence. The following parameters were used: 128 �128 matrix; field-of-view of 20 � 20 cm; 39 slices with3-mm thickness and no gap. TR/TE 3000/35; flip angle90°, acquisition orientation of the fourth ventricle plane.In addition, each functional scan was accompanied by athree-dimensional (3D) anatomical scan using anatomi-cal 3D sequence spoiled gradient (SPGR) echo sequences
that were obtained with high-resolution of 1 � 1 � 1 mm.


2lcea

cttobbSaiwmasleDcbd


fMRI Preprocessing and AnalysesBrainVoyager QX version 2.1 was used to analyzefMRI data. The first six functional volumes, beforesignal stabilization, were excluded from analysis.

Preprocessing3D motion correction was conducted, using trilinearinterpolation, linear trend removal, and high-pass fil-tering. A 4-mm, full-width at half maximum Gaussiansmoothing was used to overcome differences in inter-subject localization. Functional maps were manuallycoregistered with 1 � 1 � 1 triliniar-interpolatedanatomical maps, which were normalized into Ta-lairach space. Three-dimensional first-level statisticalparametric maps were calculated separately for eachsubject using a generalized linear model (GLM) inwhich all stimuli conditions were positive predictorswith a hemodynamic lag of 6 seconds. For second-levelanalyses, we used a statistical threshold of p � .007 anda voxelwise extent threshold of � � 3 functional voxels(3 � 27mm3 � 81mm3) to identify significant clusters.The threshold was determined a priori based on ourprevious work4 and was designed for a preliminaryexploratory investigation. We attempted to optimizethe ratio between type I and type II errors,27 whileconsidering that our natural stimuli may be a source ofintergroup variance.

AnalysesWhole-Brain Intercouple Correlations. We evaluatedthe parental brain response by using a BOLD-fMRI–based analysis that tracked the voxel-by-voxel syn-chronization between mothers’ and fathers’ brainsacross the 2-minute video of own-infant. To search forbrain regions that would work synchronously acrossthe film,28 we developed an algorithm that first trans-formed all brains to Talairach space and then pre-formed a voxel-by-voxel Pearson correlation betweenthe mother’s and father’s brains during observation ofsame clip, resulting in 15 correlation values for eachvoxel, one for each couple. A cross-subject brain syn-chrony analysis was conducted by Hasson et al.28;however, unlike their study, in which all subjectswatched the same film, our correlation scope ad-dressed the intercouple matrix, the focus of this re-search. The correlation was computed for the voxelsignals of each mother and father of the same infant.The matrices included 51 time-points of voxel signalfor each parent (120-second film proceeded and fol-lowed by 15- and 18-second rest periods, TR � 3), andyielded a single r value per voxel. We then normalizedthese correlations using a Fisher z transformation andassessed whether the 15 samples had a mean correla-tion value that was significantly greater (or smaller)than 0 using the t test. To evaluate the brain areas that
correlates specifically to the own-infant film and not to


infants in general, we conducted the same procedure(i.e., 15 whole-brain correlations and t tests) for thestandard-infant control film and subtracted the stan-dard-infant multi-subject correlation map from theown-infant multi-subject correlation map in a second-level t test of normalized correlation values. The testrevealed the brain areas that were differentially corre-lated in the mother’s and father’s brain responsespecifically to their own infant at the level of thegroup. We refer to such intersubject correlation as“mother–father synchrony,” a term addressing thecoordinated changes in BOLD signals in specific brainareas in mother’s and father’s brain in relation to thesame infant. The intersubject correlation measurescoordinated changes of activation in two brainssummed across the film (yielding the r value for eachvoxel of each pair of subjects).Whole-Brain GLM. A second-level, multi-subject,repeated-measures GLM was computed for mothersand fathers, in which the various infant-related filmswere defined as distinct block predictors. Total num-ber of blocks included two “own stimuli” (own infant,own parent–infant interaction) and four control stimuli(two unfamiliar infant films, two unfamiliar mother–infant interaction, and additional gender-relatedparent–infant interaction, male for fathers, female formothers). A 2�2 analysis of variance (ANOVA) wasconducted for all subjects and all conditions (F8,224, p ��10�14). A “between-subjects factor” contained twoevels (mothers�fathers), and a “within-subject factor”ontained two levels (own-infant�standard infant). Tovaluate the differences between mothers and fathers,second-level post hoc contrast was calculated.

Hormonal Analysis. Blood was drawn from the ante-ubital vein of the parents into 9-mL, chilled Vacuetteubes (Grenier Bio-One, Kresmunster, Austria) con-aining lithium–heparin supplemented with 400 KIUf Trasylol (Bayer, Leverkusen, Germany) per 1 mL oflood. Samples were kept ice chilled for up to 2 hoursefore centrifugation at 1,000 g for 15 minutes at 4°C.upernatants were collected and stored at �70°C untilssayed. Parents were asked to refrain from foodntake 30 minutes before blood draw. Maternal blood

as drawn at least 30 minutes after nursing and 30inutes before nursing. OT and AVP assays were

vailable for 13 mothers and 11 fathers. Previoustudies2,26 showed no differences between plasma OTevels in breast-feeding and non–breast-feeding moth-rs when OT was not measured around breast-feeding.eterminations of hormones were performed using a

ommercial OT and AVP enzyme-linked-immunosor-ent assay kit (Assay Design, Ann Arbor, MI), asescribed in earlier studies.2,29 Measurements were

performed in duplicate, and the concentrations ofsamples were calculated by MATLAB-7 (MathWorks,Natick, MA) according to relevant standard curves.The intra-assay and interassay coefficients for OT were
less than 7% and 15.8%, respectively. The intra-assay
801www.jaacap.org

cl

ATZIL et al.

and interassay coefficients for AVP were less than 3.9%and 16.9%, respectively.

Whole-Brain Hormone CorrelationsThese analyses examined which brain areas selective tothe own-infant�standard-infant contrast correlate withplasma OT and AVP. We used whole-brain analysis ofcovariance (ANCOVA) with each hormone plasma levelsas covariates, separately for fathers and mothers.

Region-of-Interest AnalysisThe region-of-interest (ROI) analysis was conductedusing the right amygdala. This region was selectedbased on previous research that marks the maternalright amygdala as a limbic structure responding to the

FIGURE 1 Mother–father brain synchronization duringshow interparent correlations among 15 fathers and 15 mcortex, pre-motor and motor cortices, cerebellum, inferiorX � 47, Z � 12.

own-infant contrast in relation to caregiving.4,30 In s

JOURN

802 www.jaacap.org

addition, the right amygdala differentiated betweenmothers and fathers and was defined from theANOVA (in the contrast mothers versus fathers � ownversus unfamiliar infant) as a functional ROI. The ROIincluded 108 voxels, peak activation in Talairach coor-dinates: X � 15, Y � �2, Z � �12, t � 3.456168, p �.001289. Mean GLM � values of the 108 voxels in theontrast own-infant�standard-infant was then corre-ated to OT and AVP.

RESULTSSynchrony and Specificity in Mothers’ andFathers’ Brains: Whole-brain IntersubjectCorrelationsFigure 1 and Table 1 present brain areas that

rvation of own-infant video. Note: Brain areas thatrs in response to their own-infant video include visualetal-lobule (IPL), inferior frontal gyrus (IFG), and insula.

obseothe

-pari

howed concurrent activations in mothers and



pt

erior fr


fathers during observation of own-infant solitaryplay. Results demonstrate synchrony betweenmothers’ and fathers’ brains in associative visualcortex, lentiform nucleus, cerebellum, ACC, pre-motor and motor cortices, IFG, cuneus, IPL, me-dial and lateral prefrontal cortex (PFC), temporalcortex, and insula when responding to theirinfant.

Whole-Brain GLMFigure 2 and Table 2 present differences betweenmothers’ and fathers’ brains activations whenresponding to their infant in the contrast owninfant�standard infant. As seen, mothers showedhigher activations in the right superior temporalgyrus, right post-central gyrus, right fusiform,right amygdala, right lentiform nucleus, righttemporal pole, and right caudate. Fathersshowed higher activations in superior-occipitaland temporal gyri, the left medial PFC, left

TABLE 1 Brain Areas Showing Correlations Between MoOwn-Infant Video

Peak X Peak Y

Left lingual gyrus �15 �91Left lentiform nucleus �18 5Right occipital gyrus 6 �91Right cerebellum 3 �43Left dorsal ACC �9 23Right pre-central gyrus 54 �4Right IFG 48 11Right posterior cingulate 12 �37Right cuneus 18 �88Left IPL �42 �55Left IPL �63 �40Right pre-central gyrus 42 �16Left middle frontal gyrus �45 2Right dlPFC 48 44Right middle temporal gyrus 63 �46Right IFG 27 17Right insula 33 5Left superior parietal lobule �39 �58Right dmPFC 3 38Left superior frontal gyrus �12 20Right IPL 60 �31Right superior frontal gyrus 15 38Left middle frontal gyrus �18 53

Note: Mother–father correlation analysis. X, Y, Z represent Talairach coothreshold min � 4.15. K � number of voxels in significant cluster; Apre-frontal cortex; dmPFC � dorso-medial pre-frontal cortex; IFG � inf

precuneus, and left inferior parietal gyri.



Oxytocin and Vasopressin. Consistent with previ-ous studies,2,26 no mean-level differences in

lasma OT emerged between mothers and fa-hers (F1,21 � 1.4029, p � .2495; mothers’ OT

mean � 592.94 pM, SD � 219; fathers OT mean �1042.47 pM, SD � 836.09). Similarly, no mother–father differences were found in AVP levels(F1,21 � 0.7605, p � .3935, mothers’ AVP mean �207.1 pM, SD � 60.87; fathers’ AVP mean �263.38 pM, SD � 218.831).

Whole-Brain Hormone CorrelationsWhen considering the brain activations in thecontrast own infant�standard infant with OT orAVP as a covariate, differences were found be-tween mothers and fathers.Oxytocin. Figure 3 and Table 3 present correla-tions between brain areas that responded to thecontrast own infant�standard infant and OTlevels separately for mothers and fathers (ran-

’ and Fathers’ Brain Response During Presentation of

Peak Z BA t p K

1 17 8.987 .000 1321 6.866 .000 210

28 19 6.642 .000 176�29 6.485 .000 203

31 32 6.451 .000 9346 4 6.372 .000 10622 44 6.063 .000 4819 23 6.032 .000 10637 19 5.914 .000 13334 40 5.814 .000 26737 40 5.664 .000 9346 4 5.395 .000 20849 6 5.287 .000 14716 46 5.216 .000 322�5 21 5.174 .000 86�8 47 5.150 .000 21313 13 5.100 .000 15758 7 5.037 .000 12328 9 4.965 .000 37549 6 4.873 .000 9222 40 4.670 .000 8343 8 4.380 .001 11710 10 4.158 .001 112

es. N � 30 subjects, 15 pairs. Cluster size threshold � 3. Confidenceanterior cingulate cortex; BA � Brodmann area; dlPFC � dorso-lateral

ontal Gyrus; IPL � inferior parietal lobule.

thers

rdinatCC �

dom effect, uncorrected, p � .0071, cluster size

803www.jaacap.org

.fH

ATZIL et al.

�3; Figures 3A to 3B and Table 3). Mothers’ OTpositively correlated with the GLM � values inthe left insula, left IPL, left and right temporalcortices, left ventral ACC, and left NAcc. Fathers’OT negatively correlated with the left inferiorand superior frontal gurus, left primary motorcortex, medial PFC, and left ACC.Vasopressin. Figure 4 and Table 4 present corre-lations between brain areas that responded to thecontrast own infant�standard infant and plasmaAVP levels. Mothers’ brain responses to owninfant were negatively correlated with AVP inthe bilateral superior-frontal gyrus, right pre-central gyrus, right-medial-frontal gyrus, andright middle-temporal gyrus. In fathers, AVPcorrelated negatively with GLM � values in theright inferior-parietal lobule, right temporal pole,right IFG, right-medial-frontal gyrus, and leftinsula (random effect, uncorrected, p � .0071,

FIGURE 2 Differences between mothers’ and fathers’covariance (ANCOVA) of blood oxygen level–dependestatistical map of the “between subjects” factor contrasand the “within-subject” factor contrasts between “owngreater positive activations of right amygdala and tempFathers showed greater positive activations of the dorsIFG � inferior frontal gyrus; Y � �3, X � �20.

cluster size � 3) (Figure 4, Table 4). p

JOURN

804 www.jaacap.org

ROI Analysis of Right AmygdalaWe extracted the signal from the right amygdalaregion and used it as a seed region for hormonalcorrelation.Hormonal Correlation With Right Amygdala. Figure5 present the correlations between the rightamygdala ROI GLM � values in the contrastown-infant�standard-infant and plasma OT andAVP. We extracted the signal from the rightamygdala that yielded the analysis of covariancebetween mothers’ and fathers’ brain activationsin the contrast own-infant�standard-infant (Ta-ble 2), and correlated this signal with plasma OTand AVP in mothers and fathers separately. Inmothers, the right-amygdala was positively cor-related with maternal OT levels (r � 0.653, p �0154, Figure 5), but such a correlation was notound in fathers (r � 0.3045, p � 0.362134).owever, in fathers only, there was a trend for

n responses to own-infant video. Note: Analysis ofOLD) functional magnetic resonance imaging (fMRI)tween 15 mothers (in pink) and 15 fathers (in blue)nt” and “standard-infant” videos. Mothers showedl, occipital, and parietal gyri compared with fathers.e-frontal cortex (dPFC) compared with mothers.

braint (B

ts be-infaora

al-pr

ositive right-amygdala correlation with plasma




AVP (r � 0.686, p � .01972). However, becausethe correlation was influenced by a single subject,it was not possible to determine an AVP–amygdala correlation in fathers (Figure 5).

DISCUSSIONThe current results suggest that mothers andfathers coordinate their brain responses to theirown infant stimuli and point to the differentialassociations of mothers’ and fathers’ brain acti-vations with neuropeptides supporting femaleand male bonding. Findings indicate that syn-chrony between mothers’ and fathers’ brain acti-vations emerged in social– cortical networksassociated with mentalizing and empathy,8 in-cluding the insula, IPL, dmPFC, and IFG, sug-gesting that parents may share in real timetheir intuitive understanding of the infant’sstate and signals. Thus, the findings may pointto potential links between mechanisms in-volved in pair bonding and those implicated inhuman parenting, and may suggest that co-parenting evolved on the basis of the highermammals’ capacity for online neural coordina-tion with a social partner and the ability torepresent the other’s state in one’s physiol-

TABLE 2 Brain Activations in Mothers vs. Fathers in the

Peak X Peak Y

Mothers�FathersRight superior temporal gyrus 42 11Right post-central gyrus 18 �34Right fusiform gyrus 39 �1Right middle temporal gyrus 42 �70Right amygdala 12 �7Right lentiform nucleus 15 8Right cuneus 15 �85Right superior temporal gyrus 45 8Right caudate 9 17

Fathers�MothersLeft superior occipital gyrus �36 �91Left medial frontal gyrus �21 32Right superior temporal gyrus 48 �22Left medial frontal gyrus �3 65Left precuneus �30 �76Right middle temporal gyrus 63 �37Left inferior parietal gyrus �46 �67Left inferior parietal gyrus �42 �52

Note: X, Y, Z represent Talairach coordinates. N � 30 subjects. Clustercluster. BA � Brodmann area.

ogy.31 However, because of our small sample b



size and the exploratory nature of our analysis,the current findings are preliminary and re-quire much further research. If validated, suchfindings may have important implications forunderstanding the neurobiology of fatherhoodand may extend evolutionary models on theemergence of co-parenting, the formation ofsocial family units in humans, and the devel-opment of the infant’s social brain within thematrix of neurobehavioral synchrony.

In addition to synchrony, results may alsopoint to specificity in the brain networks thatsupport mothering and fathering and suggestthat, overall, mothering is marked by greaterlimbic responsiveness. Mothers exhibitedgreater amygdala activations to own-infantcues, consistent with findings for other mam-mals. Second, plasma OT, a key neurohormoneunderlying bond formation, correlated withlimbic activations only among mothers, includ-ing the NAcc, amygdala, and ventral ACC.Such enhanced limbic–motivational activity inmothers may point to the deeply rooted, phy-logenetically ancient role of mothering,whereas fathers’ enhanced social– cognitive ac-tivations may reflect the more culturally facul-tative role of fathering.5,11 It is possible that,

rast Own-Infant vs. Standard-Infant Stimuli

Peak Z BA t p K

�8 13 4.791 .000 16864 3 4.186 .000 411

�20 20 4.119 .000 22613 39 4.111 .000 1023

�17 34 4.019 .000 386�11 4.011 .000 128

19 18 3.687 .001 98�23 38 3.552 .001 226

4 3.546 .001 126

25 19 �4.182 .000 35028 9 �4.147 .000 239�8 22 �3.955 .000 105

4 10 �3.878 .001 12356 7 �3.839 .001 1004 22 �3.802 .001 126

49 7 �3.726 .001 17858 40 �3.693 .001 99

3. Confidence threshold min � 2.79. K-Number of voxels in significant

Cont

size �

ecause of its critical importance for infant

805www.jaacap.org

sm

ATZIL et al.

survival, the evolution of mothering has beenmore tightly coupled with heightened respon-siveness to infant cues and elevated amygdalaresponse. It is important to note, however, thatmothers also showed cortical responses andfathers showed limbic activations. The evolu-tion of fathering and co-parenting requiresmuch further research, and it is still unknownwhether there are distinct patterns of “mater-nal” and “paternal” brain response or whetherthey reflect individual variability in emotionalresponsivity or caregiving opportunities.

As seen, the insula may play an importantrole in mother–father synchrony in response totheir infant. The insula showed synchronizedactivity in the inter-couple correlation analysis

FIGURE 3 Correlations of plasma oxytocin levels with wmothers, plasma oxytocin levels correlated positively withleft nucleus accumbens (NAcc), inferior parietal lobule (IPAmong fathers, plasma oxytocin levels correlated negativ(dmPFC), insula, inferior frontal gyrus (IFG), parietal corte

(right middle insula), as well as correlations

JOURN

806 www.jaacap.org

with OT in both mothers and fathers (leftinsula). These findings are consistent with re-search showing middle insula activations whensubjects viewed their romantic partners com-pared with friends.32 Along with the insula, inthe IPL, IFG, and motor areas, there was evi-dence for mother–father brain synchrony.These regions constitute a social network in-volved in mentalization, action representation,simulation, mirroring, and attention func-tions.9,33 These networks provide the basis forocial interactions among kin and non-kinembers of society,8 and correlate with the

minute-by-minute interactive synchrony be-tween mother and infant.4 Such social brainresponsivity included higher-order cognitive

e-brain activations in mothers and fathers. Note: Amongations in the ventral anterior cingulate cortex (vACC),nd temporal and frontal gyri (A) (X � �6, Y � 5).ith activations in the dorsal medial prefrontal cortex

nd cerebellum (B) (X � 6, Z � 12).

holactivL), aely wx, a

functions implicated in mentalization and em-



signific


pathy, as well as more basic functions of per-ception (lingual gyrus) and action representation(motor areas). Possibly, when perception–actionregions are co-activated with mentalization re-gions, they have a social-attentive function thatenhances the salience of the social context andthe planning of adequate action. These uniformand synchronized parental responses in so-cially relevant brain areas likely represent a gen-der-independent parental orientation, whichmay provide the foundation for alloparentingand the formation of attachment bonds through-out life; however, these hypotheses require muchfurther study.

The brain–OT correlations may provide ad-ditional support to the notion that mothering isguided by greater motivational– emotional fo-cus. While keeping in mind that the hormonalanalyses used an extremely small sample sizeof 13 mothers and 11 fathers, the preliminaryresults suggest that plasma OT levels signifi-cantly correlated with limbic– emotional brainareas among mothers, including the left NAcc,right amygdala, anterior insula, temporal gyri,

TABLE 3 Oxytocin Correlations to Brain Areas in the Co

Peak X Peak Y

MothersLeft insula �33 2Left inferior parietal lobule �45 �55Right middle temporal gyrus 54 �43Left inferior parietal lobule �45 �55Left superior temporal gyrus �66 �43Left middle temporal gyrus �54 �31Left anterior cingulate gyrus �9 35Left superior frontal gyrus 0 59Right anterior cingulate gyrus 6 41Left NAcc �16 7Left inferior parietal lobule �55 �55

FathersLeft inferior frontal gyrus �45 8Left superior frontal gyrus �18 41Left pre-central gyrus �51 �4Left medial frontal gyrus �3 38Right cerebellum 21 �73Left inferior parietal lobule �48 �28Right medial frontal gyrus 3 35Left cingulate gyrus 0 17Left medial frontal gyrus �3 59

Note: Brain regions of mothers and fathers showing significant correlationrepresent Talairach coordinates. Fathers(N) � 11, cluster size � 3, p �

p � .0071, min correlation threshold � 0.69. K-number of voxels in

and ventral-ACC. Among fathers, OT corre-



lated with higher activations in cognitive areas,including the dorso-lateral PFC, dorsal ACC,IPC, and pre-central gyrus. Interestingly, thelimbic–OT correlations in mothers were posi-tive, whereas the cortical–OT correlations infathers were negative. The dorso-medial PFCplays a regulatory role in organizing behav-ioral response to emotional stimuli, and higherpaternal OT may have somewhat decreasedthis modulatory response. Oxytocin is centralfor the formation of social bonds in general andparenting in particular, and is critical for ma-ternal behavior2; human studies have pointedto its role in social competencies, includingtrust, “mind-reading,” and empathy.34 Despitecomparability in baseline plasma levels, whichis consistent with previous research,2 oxytocinwas differentially related to limbic and corticalactivations in mothers and fathers. It is possiblethat oxytocin acts differently on the develop-ment of mothering and fathering, through mo-tivation enhancement or cognitive modulation.Yet, the underlying mechanisms for genderspecificity in human parenting in relation to

t Own-Infant vs. Standard-Infant Stimuli

Peak Z BA r p K

13 13 0.868 .000 24855 40 0.865 .000 91

�14 20 0.861 .000 44740 40 0.855 .000 1297 22 0.851 .000 207

�2 21 0.832 .000 145�2 32 0.821 .001 47428 9 0.820 .001 113�2 32 0.801 .001 112�6 0.800 .001 47449 40 0.799 .001 83

13 13 �0.933 .000 110531 9 �0.930 .000 122840 6 �0.928 .000 56340 8 �0.923 .000 126

�23 �0.903 .000 49728 40 �0.902 .000 27631 9 �0.866 .001 18840 32 �0.847 .001 16319 10 �0.823 .002 104

plasma oxytocin levels in the contrast own-infant�standard-infant. X, Y, Z1, min correlation threshold � 0.75. Mothers(N) �13, cluster size � 3,ant cluster. BA � Brodmann area; NAcc � nucleus accumbens.

ntras

s with.007

oxytocin brain receptors remain unclear.

807www.jaacap.org

fprlapaaaap

ATZIL et al.

AVP has been implicated in male bondingby supporting defensive behaviors in mam-mals35 and in social cognition and face percep-tion in humans.36 The current findings, the firstto examine plasma AVP in relation to humanfathering, support the special role of AVP inmale bonding by showing its correlations withbrain activations only among fathers. AVP–brain correlations emerged in fathers in the IFGand insula, highlighting the link between pa-ternal affect and social cognition. The IFG and

FIGURE 4 Correlations of plasma vasopressin levels withfathers, plasma vasopressin levels correlated negatively withtemporal pole (B) (X � 31), and inferior parietal lobule (IPL)correlated negatively with activations in superior and mediapre-frontal cortex.

insula process mirror functions that are crucial l

JOURN

808 www.jaacap.org

or human social bonding, specifically betweenarents and infants, and the father’s socialesponse to the infant may be related to circu-ating levels of plasma AVP. This AVP– brainssociation may represent elevated AVP-de-endent vigilance, and support the father’sbility to read the intention of others to defendnd respond to the young. However, it shouldlso be noted that in small sample sizes, there isgreater chance for type II errors, and it is notossible to ascertain whether OT–limbic corre-

le-brain activations in mothers and fathers. Note: Amongations in the inferior frontal gyrus (IFG) (A) (X � 53),X � 45) (A). Among mothers, plasma vasopressin levelstal gyri and middle temporal gyrus (B). mPFC � medial

whoactiv(C) (

l fron

ations are unique to mothers and AVP– brain



in sign


correlations are unique to fathers. Larger sam-ples are thus required to validate the associa-tions between OT and AVP with gender-spe-cific patterns of parental brain activationsdescribed here.

The findings that synchrony between moth-ers’ and fathers’ brain activations in areasimplicated in social cognition and empathymay suggest that brain activity can synchro-

TABLE 4 Vasopressin Correlations to Brain Areas in the

Peak X Pea

MothersLeft superior frontal gyrus �9Right precentral gyrus 27 �

Left medial frontal gyrus �3Right middle temporal gyrus 63 �

Left medial frontal gyrus �12 �

Right superior frontal gyrus 3Fathers

Right inferior parietal lobule 39 �

Left middle frontal gyrus �33Right superior temporal gyrus 33Right inferior frontal gyrus 48Right medial superior frontal gyrus 9Left insula �39

Note: Brain regions of mothers and fathers showing significant correlationZ represent Talairach coordinates. Fathers(N) � 11, cluster size � 3, p �

p � .0071, mean correlation threshold � 0.70. K-Number of voxels

FIGURE 5 Region-of-interest (ROI) analysis. Note: (A)amygdala response to own-infant versus standard-infantonly in fathers. (B) Amygdala–oxytocin correlation in mostandard-infant contrast correlated positively with plasma

0.3

0.6 r= 0.653, p= 0.0154

A

-0.6

-0.3

0

0 500 1000Oxytocin pM Right amOxytocin pM g



nize in real time between attachment partnerswithin an affiliation-related context. Thesefindings are consistent with current notions of“embodiment,”31 which suggest that when in-dividuals are required to perceive and to un-derstand social signals, the interpretation ofthese signals is based on changes in theseindividuals’ own physiology. This line of re-search was taken to suggest that bio-behavioral

trast Own Infant vs. Standard Infant Stimuli

Peak Z BA r p K

55 6 �0.929 .000 11152 4 �0.868 .000 25037 6 �0.859 .000 83

�14 21 �0.855 .000 12058 6 �0.837 .000 73055 6 �0.822 .001 81

22 40 �0.927 .000 27916 10 �0.894 .000 52

�36 38 �0.888 .000 1427 45 �0.858 .001 122

37 8 �0.855 .001 8110 13 �0.800 .003 87

lasma vasopressin levels in the contrast own-infant�standard-infant. X, Y,1, mean correlation threshold � 0.75. Mothers(N) � 13, cluster size � 3,ificant cluster. BA � Brodmann area.

dala–vasopressin correlation in fathers. The rightast correlated positively with plasma vasopressin levels. The right amygdala responses to own-infant versusocin levels only in mothers.

0.3

0.4 r=0.686, p=0.01972

B

0

0.1

0.2

0 500 1000Vasopressin pMala ROI p p

Con

k Y

292838252829

283814235011

s to p.007

Amygcontrthersoxyt

ygdyg

809www.jaacap.org

ATZIL et al.

dyadic mechanisms provide the basis for thehuman capacity to understand complex emo-tional expressions in others. Combined withthe present findings and the bio-behavioralsynchrony conceptual model,31 it is possiblethat the formation of human attachment in-cludes a finely tuned adaptation of the parentand infant’s brains in areas implicated in em-bodied simulation.

Limitations of the study include the rela-tively small sample size and homogeneous SES,the less restricted statistical thresholds, and theinability to measure OT and AVP in the brain.It is a preliminary study that used an explor-atory approach at a whole-brain level withrelatively liberal thresholds for statistical sig-nificance complemented with a priori regionsof interests. In addition, we did not includenon-infant stimuli in our paradigm. In themother–father correlation analysis, we did nottrack the timely changes in synchronization orevents that might influence the intercouplecorrelation; rather, the analysis correlated theactivations across the signal time-course dur-ing the film and yielded a single r value foreach maternal and paternal voxel.

Our findings are among the first to addressthe brain basis of fatherhood, and may suggestthat human attachment involves the onlinecoordination between the neurobiological pro-cesses of two individuals within an attachmentrelationship. Although the findings are prelim-inary and require further validation, they mayshed light on process of brain-to-brain syn-chrony in the formation of human attachment,and contribute to the study of coparenting andthe ways in which family units are constructedfrom the joint efforts of parents raising theirinfants. As fathers’ role in childcare is increas-ing, the neurobiology of fatherhood is amongthe key question for future research. Finally,future research on conditions associated withdisruptions to social functioning or parent–infant bonding is required in order to constructmore specific interventions based on the coor-dination of brain mechanisms and behavioral
expression.
Oxytocin-dopamine interactions mediate variations in maternalbehavior in the rat. Endocrinology. 2010;151:2276-2286.

JOURN

810 www.jaacap.org

Clinical Guidance

• Research has demonstrated that attachment bondsare formed through processes of bio-behavioralsynchrony—the coordination of physiology andbehavior between attachment partners. Here weexamined whether mothers and fatherssynchronize their brain response to their owninfant’s cues.

• Mothers and fathers synchronized brain activationsonline to own-infant video in brain areas implicatedin social cognition, theory of mind, and empathy.Such brain synchrony may help parents to jointlyread their infant’s nonverbal communications and toplan adequate parenting.

• Mothers showed greater activations in limbic areas,and fathers showed greater activations in social–cognitive cortical areas. These were differentiallycorrelated with oxytocin and vasopressin,neuropeptides associated with female and malebonding.

• These findings are the first to demonstrate brainsynchrony between attachment partners in empathy-related networks. Conditions associated with earlysocial dysfunctions, such as autism, are linkedwith less optimal functioning of these networks,and the findings may suggest disruptions in thecapacity for brain-to-brain synchrony.

Accepted June 4, 2012.

This article was reviewed under and accepted by Deputy Editor EllenLeibenluft, M.D.

Ms. Atzil and Dr. Zagoory-Sharon are with Bar-Ilan University. Dr.Hendler and Mr. Winetraub are with the Functional Brain Center,Tel-Aviv Sourasky Medical Center. Dr. Feldman is with the GondaBrain Sciences Center, Bar-Ilan University.

Research at Dr. Feldman’s laboratory is supported by the IsraeliScience Foundation, the Brain and Behavior Research Foundation,and the Katz Family Foundation.

Disclosure: Ms. Atzil, Drs. Hendler, Zagoory-Sharon, and Feldman,and Mr. Winetraub report no biomedical financial interests or poten-tial conflicts of interest.

Correspondence to Ruth Feldman, Ph.D., Gonda Brain SciencesCenter, Bar-Ilan University, Ramat-Gan, Israel 52900; e-mail:[email protected]

0890-8567/$36.00/©2012 American Academy of Child andAdolescent Psychiatry

http://dx.doi.org/10.1016/j.jaac.2012.06.008

REFERENCES1. Feldman R. Oxytocin and social affiliation in humans [published

online ahead of print January 20 2012]. Hormones Behav. 2012;61:380-3891.

2. Shahrokh DK, Zhang TY, Diorio J, Gratton A, Meaney MJ.

3. Feldman R, Magori-Cohen R, Galili G, Singer M, Louzoun Y.Mother and infant coordinate heart rhythms through episodesof interaction synchrony. Infant Behav Dev. 2011;34:569-577.

4. Atzil S, Hendler T, Feldman R. Specifying the neurobiological
basis of human attachment: brain, hormones, and behavior in


mailto:[email protected]

http://dx.doi.org/10.1016/j.jaac.2012.06.008


synchronous and intrusive mothers. Neuropsychopharmacology.2011;36:2603-2615.

5. Geary DC. Evolution and proximate expression of human pater-nal investment. Psychol Bull. 2000;126:55-77.

6. Kinsley CH, Amory-Meyer E. Why the maternal brain? J Neu-roendocrinol. 2011;23:974-983.

7. Franssen CL, Bardi M, Shea EA, et al. Fatherhood alters behav-ioural and neural responsiveness in a spatial task. J Neuroendo-crinol. 2011;23:1177-1187.

8. Decety J. The neuroevolution of empathy. Ann N Y Acad Sci.2011;1231:35-45.

9. Iacoboni M, Dapretto M. The mirror neuron system and the conse-quences of its dysfunction. Nat Rev Neurosci. 2006;7:942-951.

10. Strathearn L. Maternal neglect: oxytocin, dopamine and theneurobiology of attachment. J Neuroendocrinol. 2011;23:1054-1065.

11. Storey A, Walsh C. How fathers evolve: a functional analysis offathering behavior biosocial foundations of family processes. In:Booth A, McHale SM, Landale NS, eds. Biosocial Foundations ofFamily Processes. National Symposium on Family Issues, 2011,Part 1. New York: Springer; 2011:35-47.

12. Insel TR, Young LJ. The neurobiology of attachment. Nat RevNeurosci. 2001;2:129-136.

13. Garfield CF, Isacco A. Fathers and the well-child visit. Pediatrics.2006;117:e637-e645.

14. Lamb ME. The Role of the Father in Child Development. Hobo-ken, NJ: Wiley; 2010.

15. Williams E, Radin N. Effects of father participation in childrearing: twenty-year follow-up. Am J Orthopsychiatry. 1999;69:328-336.

16. Luo J, Wang LG, Gao WB. The influence of the absence of fathersand the timing of separation on anxiety and self-esteem ofadolescents: a cross-sectional survey [published online ahead ofprint September 9, 2011]. Child Care Health Dev. 2011;DOI:10.1111/j.1365-2214.2011.01304.x

17. Kim P, Leckman JF, Mayes LC, Feldman R, Wang X, Swain JE.The plasticity of human maternal brain: longitudinal changes inbrain anatomy during the early postpartum period. Behav Neu-rosci. 2010;124:695-700.

18. Landi N, Montoya J, Kober H, et al. Maternal neural responses toinfant cries and faces: relationships with substance use. FrontPsychiatry. 2011;2:32.

19. Strathearn L, Mayes LC. Cocaine addiction in mothers. Ann N YAcad Sci. 2010;1187:172-183.

20. Pereira M, Morrell JI. Functional mapping of the neural circuitryof rat maternal motivation: effects of site-specific transient neuralinactivation. J Neuroendocrinol. 2011;23:1020-1035.

21. Cho MM, DeVries AC, Williams JR, Carter CS. The effects ofoxytocin and vasopressin on partner preferences in male and



female prairie voles (Microtus ochrogaster). Behav Neurosci.1999;113:1071-1079.

22. Meyer-Lindenberg A, Domes G, Kirsch P, Heinrichs M. Oxytocinand vasopressin in the human brain: social neuropeptides fortranslational medicine. Nat Rev Neurosci. 2011;12:524-538.

23. Thompson RR, George K, Walton JC, Orr SP, Benson J. Sex-specific influences of vasopressin on human social communica-tion. Proc Natl Acad Sci U S A. 2006;103:7889-7894.

24. Carter CS. Neuroendocrine perspectives on social attachment andlove. Psychoneuroendocrinology. 1998;23:779-818.

25. Zink CF, Stein JL, Kempf L, Hakimi S, Meyer-Lindenberg A.Vasopressin modulates medial prefrontal cortex-amygdala cir-cuitry during emotion processing in humans. The Journal ofNeuroscience. 2010;30:7017-7022.

26. Gordon I, Zagoory-Sharon O, Leckman JF, Feldman R. Oxytocinand the development of parenting in humans. Biol Psychiatry.2010;68:377-382.

27. Lieberman MD, Cunningham WA. Type I and type II errorconcerns in fMRI research: re-balancing the scale. Soc Cogn AffectNeurosci. 2009;4:423-428.

28. Hasson U, Nir Y, Levy I, Fuhrmann G, Malach R. Intersubjectsynchronization of cortical activity during natural vision. Science.2004;303:1634-1640.

29. Kramer KM, Cushing BS, Carter CS, Wu J, Ottinger MA. Sex andspecies differences in plasma oxytocin using an enzyme immu-noassay. Can J Zool. 2004;82:1194-1200.

30. Swain JE, Lorberbaum JP, Kose S, Strathearn L. Brain basis ofearly parent-infant interactions: psychology, physiology, and invivo functional neuroimaging studies. J Child Psychol Psychiatry.2007;48:262-287.

31. Niedenthal PM, Mermillod M, Maringer M, Hess U. The simula-tion of smiles (SIMS) model: embodied simulation and themeaning of facial expression. Behav Brain Sci. 2010;33:417-433;discussion 433-480.

32. Bartels A, Zeki S. The neural basis of romantic love. Neuroreport.2000;11:3829-3834.

33. Dodell-Feder D, Koster-Hale J, Bedny M, Saxe R. fMRI itemanalysis in a theory of mind task. Neuroimage. 2011;55:705-712.

34. Bartz JA, Zaki J, Bolger N, Ochsner KN. Social effects of oxytocinin humans: context and person matter. Trends Cogn Sci. 2011;15:301-309.

35. Bielsky IF, Hu SB, Young LJ. Sexual dimorphism in the vasopres-sin system: lack of an altered behavioral phenotype in female V1areceptor knockout mice. Behav Brain Res. 2005;164:132-136.

36. Guastella AJ, Kenyon AR, Alvares GA, Carson DS, Hickie IB.Intranasal arginine vasopressin enhances the encoding of
happy and angry faces in humans. Biol Psychiatry. 2010;67:1220-1222.
811www.jaacap.org

synchrony and specificity in the ... - home | openscholar · new research synchrony and...

Documents