Cortical specialisation to social stimuli from the first days to thesecond year of life: A rural Gambian cohort

Citation for this paper:

Lloyd-Fox, S., Begus, K., Halliday, D., Pirazzoli, L., Blasi, A., Papademetriou, M., …

Elwell, C.E. (2017). Cortical specialisation to social stimuli from the first days to

the second year of life: A rural Gambian cohort. Developmental Cognitive

Neuroscience, 25(June), 92-104.

https://doi.org/10.1016/j.dcn.2016.11.005

UVicSPACE: Research & Learning Repository

_____________________________________________________________

Faculty of Social Science

Faculty Publications

_____________________________________________________________

Cortical specialisation to social stimuli from the first days to thesecond year of life:

A rural Gambian cohort

S. Lloyd-Fox, K. Begus, D. Halliday, L. Pirazzoli, A. Blasi, M. Papademetriou, M.K.

Darboe, A.M. Prentice, M.H. Johnson, S.E. Moore, C.E. Elwell

August 2017

© 2017 The Authors. Published by Elsevier Ltd. This is an open access article under

the CC BY license (

http://creativecommons.org/licenses/by-nc-nd/4.0/

).

This article was originally published at:

ContentslistsavailableatScienceDirect

Developmental

Cognitive

Neuroscience

Cortical

specialisation

to

social

stimuli

from

the

first

days

to

the

second

year

of

life:

A

rural

Gambian

cohort

S.

Lloyd-Fox

_,

_K.

_Begus

_,

_D.

_Halliday

_,

_L.

_Pirazzoli

_,

_A.

_Blasi

_,

_M.

_{Papademetriou}

_,

M.K.

Darboe

_,

_A.M.

_Prentice

_,

_M.H.

_Johnson

_,

_S.E.

_Moore

_,

_C.E.

_Elwell

b_{DepartmentofMedicalPhysicsandBiomedicalEngineering,UniversityCollegeLondon,UK} c_{CognitiveDevelopmentCenter,CentralEuropeanUniversity,Hungary}

d_{DepartmentofPsychology,UniversityofVictoria,Canada} e_{MRCInternationalNutritionGroup,MRCUnit,Gambia} f_{MRCUnit,Banjul,Gambia}

g_{DivisionofWomen’sHealth,King’sCollegeLondon,UK}

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received12September2016 Receivedinrevisedform 17November2016 Accepted17November2016 Availableonline27November2016 Keywords:

fNIRS Infancy

Low-andmiddle-incomecountries Nutrition

Poverty Socialcognition

a

b

s

t

r

a

c

t

Brainandnervoussystemdevelopmentinhumaninfantsduringthefirst1000days(conceptiontotwo yearsofage)iscritical,andcompromiseddevelopmentduringthistime(suchasfromundernutritionor poverty)canhavelife-longeffectsonphysicalgrowthandcognitivefunction.Corticalmappingof cogni-tivefunctionduringinfancyispoorlyunderstoodinresource-poorsettingsduetothelackoftransportable andlow-costneuroimagingmethods.Havingestablishedasignaturecorticalresponsetosocialversus non-socialvisualandauditorystimuliininfantsfrom4to6monthsofageintheUK,hereweapplythis functionalNearInfraredSpectroscopy(fNIRS)paradigmtoinvestigatesocialresponsesininfantsfrom thefirstpostnataldaystothesecondyearoflifeintwocontrastingenvironments:ruralGambianand urbanUK.Resultsrevealrobust,localized,sociallyselectivebrainresponsesfrom9to24monthsoflife toboththevisualandauditorystimuli.Incontrastat0–2monthsofageinfantsexhibitnon-social audi-toryselectivity,aneffectthatpersistsuntil4–8monthswhenweobserveatransitiontogreatersocial stimulusselectivity.Thesefindingsrevealarobustdevelopmentalcurveofcorticalspecialisationover thefirsttwoyearsoflife.

©2016TheAuthors.PublishedbyElsevierLtd.ThisisanopenaccessarticleundertheCCBYlicense (http://creativecommons.org/licenses/by/4.0/).

1. Introduction

Infantsinresource-poorsettingsmay befrequentlyexposed toa rangeofsocial,environmental,nutritionalandpathological insults.Approximately1in2childrenarethoughttoliveinpoverty (CurrieandAlmond, 2011;UNICEF,2013),and165million chil-drenworldwideareundernourishedandstunted(UNICEF,2013), themajority ofwhomliveinSub-Saharan Africaor SouthAsia. Accordingtoa recentstudy,onethirdofchildren indeveloping countriesfailtoreachtheirdevelopmentalmilestones in cogni-tiveand/orsocio-emotionalgrowth,withthelargestnumber of affectedchildreninsub-SaharanAfrica(McCoyetal.,2016).This meansthat over80 millionchildren inlow andmiddle income

∗ Correspondingauthorat:CentreforBrainandCognitiveDevelopment,Birkbeck, UniversityofLondon,UK.

E-mailaddress:s.fox@bbk.ac.uk(S.Lloyd-Fox).

countries(LMICs) fail to develop a coreset of age-appropriate skillsthatallowthemtomaintainattention,understandand fol-lowsimple directions,communicateand cooperatewithothers, controlaggression,andsolvecomplexproblems.Theabsenceof theseskillshassignificantimpactontheiracademicachievement and mental health into adulthood,and as such theirpotential toleadfullandproductivelivesandsupportfuturegenerations. Whilemanystudiessuggestthatthepresenceoftheserisk fac-tors in infancy hasa lasting impactthroughout the life course (HackmanandFarah,2009;Martorelletal.,2010;Victoraetal., 2008),almostnothingisknownabouttheneuralbasesofthese earlydeficits.Thefirst1000daysoflifeareacriticalwindowfor brainandnervoussystemmaturation,andimpaireddevelopment duringthistimecanhavea significantimpactoncognitive out-come(CusickandGeorgieff,2012;MendezandAdair,1999;Powell etal.,1995).Toinforminterventionsthatmayreducetheimpact oftheseinsults,earlydetectionofatypicalneurocognitive func-tionisrequired.However,todatetherehasbeenalackofsuitable

http://dx.doi.org/10.1016/j.dcn.2016.11.005

S.Lloyd-Foxetal./DevelopmentalCognitiveNeuroscience25(2017)92–104 93

methodsforusefromearlyinfancy(Isaacs,2013).Investigationof thedevelopingbraininruralfieldsettingshasbeenbroadly lim-itedtobehaviouralassessments(Georgieff,2007;Sabanathanetal., 2015).However,measurementsofbehaviourcomewithsome lim-itations.Firstly,theycanonlybeusedtodetecteffectsoncethey reachthepointofobservablebehaviour,usuallyinthesecondyear of lifeor later. For example, whilstbehavioural measureshave beenunabletodistinguishbetweeninfantswithlowandhigh-risk ofdevelopingautism(definedbyafamilialdiagnosis)beforethe firstyearoflife,severalrecentneuroimagingstudieshave iden-tified differences in brainfunction in younginfants (Elsabbagh etal.,2012,2009;Foxetal.,2013;Guiraudetal.,2011;Lloyd-Fox etal.,2013;Luysteretal.,2011;McCleeryetal.,2009). Further-more,workontherelationshipbetweenfamilysocio-economic status(SES)andinfantbraindevelopmenthasevidencedatypical neuralactivityusingelectroencephalography(EEG)insixtonine montholdinfantsfromlowSESbackgroundsintheUK,highlighting theimportanceoftheearly-lifeenvironment onbrain develop-ment(Tomalskietal.,2013).Secondly, thereareissues relating tothe implementation, culturaladaptation and standardisation ofbehaviouralassessmentsbetweencontrastingpopulations.For example,manystandardisedassessmentmeasuresaredeveloped andnormed withina limitednumber ofhigh-income countries (Mullen, 1995).Thereforeresearchers needtodevelop country-specificnormsforthesemeasuresorcreateindependentmeasures andquestionnairesfortheirownpopulations(i.e.Abubakaretal., 2016;Kariukietal.,2016;Sabanathanetal.,2015).Adjustments tothesemeasurescanproducemorerobustandreliabledatasets withinpopulations,butcanalsohindercross-culturalcomparison duetoissueswithmeasurementequivalence.

Neuroimagingparadigmscanbedesignedtobeunbiased, objec-tiveandapplicableacrossdifferentpopulations.However,todate, therehasbeenalackofneuroimagingstudiesininfantsinLMICs.In manyinstancesthisisbecausetheengineeringandtechnicalissues associated withapplyingneuroimaging techniquesin resource-poorsettingshavenotbeenaddressed.Forexample,thehighcost andlow portabilityof neuroimagingmethodssuchasmagnetic resonanceimaging(MRI)hasexcludedtheiruseinresourcepoor settingsandfield-basedresearch.Directmeasuresofbrain activ-ityarepossibleinsuchsettingsusingEEGmethods,howevertheir usecanbelimitedbytestingconstraints(includingtheneedfor controlledtemperatureandhumiditylevelsinhotcountries, Kap-penmanand Luck, 2010), andmost studiesarenot undertaken untilinfantsreachtheageof18monthsofageorolder(Fernandes et al.,2014).Functional nearinfrared spectroscopy (fNIRS)is a non-invasiveopticalneuroimagingtechnique,whichcanmeasure corticalbrainfunction.Infantswearlightweightheadgearwhich facilitatesthedeliveryto,anddetectionofnearinfraredlightfrom thehead.Changesinnearinfraredlightintensityareacorrelateof changesinhaemodynamicsandoxygenationarisingfromlocalized neuronalactivityintheunderlyingcorticaltissue(Villringerand Chance,1997).fNIRSheadgearcanberapidlyadministeredandis welltoleratedbyyounginfantsfrombirth.Withoptimal position-ingofmeasurementchannels(pairsofsourcelightsanddetectors), fNIRSgenerallyallowsformorespecificspatiallocalizationof acti-vationwithrespecttoEEG.ThoughthedepthresolutionoffNIRS isdependentontheageoftheinfantandtheopticalpropertiesof thetissue(Fukuietal.,2003),anditofferslowerspatialresolution relative tofMRI,it is similarin that itmeasures the haemody-namicresponseresultingfromneuronalactivation.Researchfrom adultshasshownahighdegreeofcorrelationbetween simultane-ousrecordingsofhaemodynamicresponseswithfNIRSandfMRI (Satoetal.,2013;Steinbrinketal.,2006).WebelievefNIRScanbe widelyadoptedforfieldbasedresearchduetoitslowcost (rela-tivetootherneuroimagingmethodssuchasMRI),portability,ease ofusewithinfants(Gervainetal.,2011;Lloyd-Foxetal.,2010)

andclinicalpopulations(JacksonandKennedy,2013;Kolyvaetal., 2013), and suitabilityfor useinnaturalistic settings(Lloyd-Fox etal.,2015a).

In2013wetransportedanfNIRSneuroimagingsystemtoafield stationinruralGambiaanddemonstrateditsusetoacquiremapsof corticalfunctionfromyounginfants.Thefindingsfromourstudies in4–8monthGambianinfantswerepreviouslydescribedby Lloyd-Foxetal.(Lloyd-Foxetal.,2014a).Inthispaperwepresentdata fromGambianinfantsaged19daysto24monthsofage(seeFig.1). Theprimaryaimofthecurrentworkwastoassessthe speciali-sationofcorticalactivationinresponsetosocialcuesfromthefirst daysoflifetothesecondyear.Asecondarygoalwastocomparethe responsesfromtheruralGambiancohortswithknownresponsesin infantsfromanurbanUKpopulation.Weusetheterm‘social’inthis paperinthebroadestsense,i.e.theyarehuman-generatedsensory stimuliineitherthevisualorauditorydomain.Thisdoesnot nec-essarilyimplythatthesecuesareintendedtobecommunicative. Wechosetousethesestimuliforseveralreasons.Firstly, previ-ousresearchsuggeststhatinfantsareabletoidentify,andlearn from,voicesintheirsurroundingsfromaveryearlyage(Ockleford etal.,1988),makingthis anidealstimulustouseforour 0–24-monthagerange.Secondly,whilstthereisanextensiveandrich bodyoffNIRSresearchoninfantlanguageandspeechprocessing (Quaresimaetal.,2012),wechosetousenon-speechvocalisations inthecurrentparadigmbecauseasuitableparadigmwouldhave beenchallengingtodevelopforallofthelanguagesspokeninthe WestKiangdistrictofTheGambiawhereKenebaissituated.We thereforefelt thatthenon-speechvocalizationparadigmwould bemorewidelyapplicableacrossdifferentlanguagesandcultures. Thirdly,recentfNIRSresearchintypicallydevelopinginfantsfrom high-incomeurbanenvironmentshasshownrobustand consis-tentactivationtosocialvsnon-socialvisualandauditorystimuliin theinferiorfrontal,anteriortemporalandposteriorsuperior tem-poral–temporoparietaljunction(pSTS-TPJ)regionsofthecortex (Grossmannetal.,2010;Lloyd-Foxetal.,2012,2009; Minagawa-Kawaiet al.,2011).Therefore wewereconfidentthat wecould optimizethisparadigmacrossawideragerange.Havingpreviously establishedthissignatureresponseininfantsfrom4to8monthsof ageinruralGambia(Lloyd-Foxetal.,2014a),hereweappliedthis paradigmtotheinvestigationofsocialversusnon-socialcortical responsesininfantsfromthefirstdaysoflifetothesecondyear.

2. Materialsandmethods 2.1. Participants

Participantswererecruitedfromvillagesneighbouringafield stationinKeneba,WestKiangDistrictinTheGambia(see( Lloyd-Fox etal.,2014a)forfurtherdetails),identifiedusing theWest KiangDemographicSurveillanceSystem(http://www.ing.mrc.ac. uk/researchareas/westkiangdss.aspx).Allinfantswerebornfull term(37–42weeksgestation) andwithnormal birthweight.A combinationofprenatalgrowthretardation,poor-qualityandoften contaminatedfoodsandhighlevelsofinfectioncausemoderateto severegrowthfalteringinheightandweightgainfromaround3 monthsofageinthelocalpopulation(Lunnetal.,1991;Lunn,2000; vanderMerweetal.,2013).Thereforeexclusioncriteriaincluded weight-for-heightorheadcircumferencelessthan3z-scoresbelow WHOstandards.Growthmeasuresforthoseincludedinthestudy indicatetheinfantswereinthetypicalrangefor theirage(see

Table1).EthicalapprovalwasgivenbythejointGambia Govern-ment−MRCUnitEthicsCommittee,andwritteninformedconsent wasobtainedfromallparents/carerspriortoparticipation.

Theseparticipantswererecruitedintooneofthreecohorts(see

Fig.1.fNIRSheadgearonanewborn(Cohort1),6montholdinfantand13montholdinfant(Cohort2:longitudinal)anda2yearoldtoddler(Cohort3). Table1 ParticipantCharacteristics. Characteristics 0–2mths 4–8mths 9–13mths 12–16mths 18–24mths Sex(m/f) 14/4 10/14 15/10 9/10 7/9 Age(days) 41.0±14.32c _{174.4±40.7 348.4±37.8 428.2±34.2 631.2±76.0} Weight(kg) 4.59±1.04 6.91±0.75 8.12±1.48 8.24±0.84 9.42±1.44 Length(cm) 54.9±3.46 64.46±3.39 70.4±2.89 72.88±2.8 79.87±4.16 HeadcircumferenceH C(cm) 37.5±1.68 41.45±1.3 43.85±1.52 44.48±1.30 45.91±1.46 MUACa_{(cm) 12.3±0.12 13.8±0.59 14.1±7.04 13.9±0.55 14.1±0.85}

Growthanthropometricz-scoresb

Weight-for-age −0.41±1.2d _{−0.75±0.86 −1.28±1.50 −1.49±0.91 −1.51±1.15}

Length-for-age −0.43±1.08 −0.82±1.21 −1.65±1.08 −1.67±0.98 −1.37±1.04

H C-for-age(HCZ) −0.25±0.9 −0.89±0.99 −1.20±1.16 −1.13±0.93 −0.91±0.88

Weight-for-length(WHZ) −0.05±1.1 −0.21±0.82 −0.57±1.67 −1.67±0.86 −1.16±1.13

a_{Midupperarmcircumference.} b _{WiththeuseofWHOreferencecurves.} c _{Mean±SD(allsuchvalues).}

d _{zscore±SD(allsuchvaluesthereafter).}

S.Lloyd-Foxetal./DevelopmentalCognitiveNeuroscience25(2017)92–104 95

includedlongitudinaldatacollection,participatedat4–8monthsof age(seepreviouslypublishedresultsinLloyd-Foxetal.,2014a)and attwofurthersessions,six(aged9–13months)andninemonths (aged12–16months)later.Cohort3participatedat18–24months ofage.

Followingeachsession,infant’sdatacouldbeexcludedforthe followingreasons;dueto(1)motionartefactinthedata,(2)an insufficientnumberofvalidtrialsaccordingtolookingtime mea-sures,(3)tiredness/fussinessresultinginsession finishingearly, (4)experimenter error,(5)hairobscuring arrayand preventing measurements,or(5)theirweightforheight(WHZ)orhead cir-cumference(HCZ)z-scoresfallingbelow−3.ThereforeforCohort1, 18infantsparticipatedinthestudy(4female,meanage=41.0days, SD=14.32):afurthereightinfantsparticipatedbutwereexcluded fromgroupanalyses.ForCohort2(longitudinal),42infantswere initiallyrecruitedintothestudy.At4–8monthsofage(previously publishedin(Lloyd-Foxetal.,2014a)),24infantsparticipatedin thisstudy(10female,meanage=174.4days,SD=40.7)anda fur-ther18infantsparticipatedbutwereexcludedfromgroupanalyses. Sixmonthslaterparentsandinfantswereaskedtoreturnfora sec-ondsession.Nowaged9–13months,25infantsparticipatedinthis study(10female,meanage=348.4days,SD=37.8)andafurther 12infantsparticipatedbutwereexcludedfromfurtheranalyses. Inadditionfiveinfantsfromthefirstsessionat4–8monthsofage couldnotparticipateatthistimepointeitherbecausetheywere awayfromtheregionattime oftesting(4infants)or haddied sincethelastvisit(1infant).Threemonthslaterparentsandinfants wereaskedtoreturnforafinalsession.Nowaged12–16months,19 infantsparticipatedinthisstudy(10female,meanage=428.2days, SD=34.1)andafurther18infantsparticipatedbutwereexcluded fromfurtherdataanalyses.Inadditionthreeinfantsfromthefirst sessioncouldnotparticipateatthistimepointeitherbecausethey wereawayfromtheregionattimeoftesting(1infant)orhaddied sincethelastvisit(2infants).ForCohort3,16infantsparticipated inthestudy(9female,meanage=631.2days,SD=76.0)anda fur-ther8infantsparticipatedbutwereexcludedfromthestudy.Fig.2

providesdetailoftheparticipantstestedinthisstudyandreasons forexclusion.

2.2. Procedure

2.2.1. fNIRSmeasurements

Infantsworecustom-builtfNIRS-CBCDheadgearconsistingof anarrayovertherighthemisphere(Lloyd-Foxetal.,2010).Note thatmeasurementswererestrictedtotherighthemisphereas(1) ourfundingonlyallowedforarestrictednumberofsourcesand detectorswithrespecttotheNIRSsystemusedintheUK,and(2) welocalizedthechannelstoonehemispheretoensurewecould measuretheentiretemporallobearea.Thisarrayvariedinsize andconfigurationaccordingtotheagegrouptested(seeFig.3). Thearraycontaineduptoatotalof12channels(source-detector separations2cm),andinfantsweretestedwiththeUCLoptical topographysystem(Everdelletal.,2005).ForCohort1weuseda smallerarraytoadjustforthesmallerheadsizeatthisage employ-ing8channelstocoverthesameareaofthehead.ForCohort2 at4–8and12–16monthsof ageweusedthefullnumber of12 channels.For Cohort2at9–13monthsofageand Cohort3we employedadifferentshapedarraytoenableadditional measure-mentsovertheprefrontalcortex (whichwasdesigned totest a differentparadigmatthesamevisitwhenbothoftheseagegroups weretested).Howeverthisdesignresultedinsuboptimal place-ment of channels over the posterior temporal ROI (pSTS-TPJ – posteriorSuperiorTemporalSulcus–TemporoparietalJunction), whichmeantthatwemaynothavemeasuredthefullextentof thecorticalactivationforthevisualsocial/non-socialcontrastat theseagepoints(seeFig.3).Forthesourcedetectorseparations

usedinthisstudy,lighttransportmodelspredictlightpenetration depthsofuptoapproximately1cmfromtheskinsurface, poten-tiallyallowingmeasurementofboththegyriandpartsofthesulci nearthesurfaceofthecortex(Fukuietal.,2003;Richardsetal., 2012;Salamonetal.,1990).Withtheuseoftheco-registration MRI-fNIRSdatafromourrecentwork(Lloyd-Foxetal.,2014b)wecan approximatetheunderlyingcorticalanatomyofthefNIRS chan-nelsusedinthecurrentstudy.Therefore,weareconfidentthat wecanlocalizeourinvestigationtospecificregionsofthesocial brainnetworkanddrawcomparisonswithfindingsfromadult pop-ulations.Beforetheinfantsbeganthestudy,headmeasurements weretakentoaligntheheadgearwith10–20coordinates( Lloyd-Foxetal.,2014b).Theheadmeasurementsshowedthattheinfants’ headcircumferencedidnotchangeconsiderablybetween4and24 monthsofagesoweutilizedthesame2cmseparationsthroughout thecohorts.Theheadgearwasplacedwiththesourcelightoptode positionedbetweenchannel4and7onFig.3centredabovethe pre-auricularpoint(T4accordingtothe10–20system).

TheprotocolforCohorts2and3followedanidentical proce-duretothatoutlinedinourpreviouslypublishedstudy(Lloyd-Fox etal.,2014a).Forthereadersconveniencewerepeattextfromthis previouswork,butwithadditionalinformationprovidedaboutthe differenttestingprotocolfollowedforCohort1.OncethefNIRS headgearwasplacedontheirheads,theinfantssatontheir par-ent’slapin frontofascreen(forCohorts2&3)orasleepona mattressortheirparent’slapforCohort1.ForCohort1wewaited fortheswaddledinfantstofallasleep,thenwrappedtheheadband aroundtheirheadandwaitedforthemtosettlebeforebeginning thestudy.Weprioritizedplacingtheinfantsonacushioned mat-tressforthestudybutiftheywererestlesstheywereheldintheir parent’sarms,thereforewhilstsleepstatewasprioritizedsome infantsweretestedwhileinaquietalertstate.Fortwooftheinfants theybecamealertandthenfussyanditwasnecessarytofeedthem duringthestudyand trialsduringthissectionofthedatawere excludedfromdataanalysis.Forallparticipantstheparentwas instructedtorefrainfrominteractingwiththeinfantduringthe stimulipresentationunlesstheinfantbecamefussyorsoughttheir attention.Thesequenceofstimuluspresentationhasbeenusedin previousresearch(Lloyd-Foxetal.,2012,2013)andisillustratedin

Fig.3.Theconditionsalternatedoneaftertheother,withaperiod ofbaselinebetweeneach.Thethreetypesofconditions (visual-social(silent),auditorysocial,auditorynon-social)werepresented inthesameorderacrossinfantsinarepeatingloop(V-S,A-NS,A-S, V-S,A-S,A-NS)oftrials(singlepresentationofacondition)until theinfantsbecameboredorfussyasjudgedbytheexperimenter whowasmonitoringtheirbehaviour.Arestrictionofstudying audi-toryprocessinginawakeinfantsofthisageisthattheyneedto bepresentedwithconcurrentvisualstimulationtoreduceinfant movementandthusartefactinthesignal.Wechosetoemploythe samevisualstimuliduringthepresentationoftheauditorystimuli thatwecollecteddatafromwhenauditorystimulationwasabsent. ForCohort1weusedthesameparadigmandequipment,placing theTVmonitoratthesamedistancefromtheinfants’headasused forCohorts2&3.

2.3. Visualstimuli

Theseconsistedoffull-color,life-size(headandshouldersonly) socialvideosofadults(Gambiannationals)whoeithermovedtheir eyes leftorrightor performedhandgames—“Peek-a-boo”and “Incy-WincySpider.”Twovisualsocialvideoswerepresentedfor varyingduration overeach 9–12strialtoavoid inducing antic-ipatorybrain activity.To ensureinfants’ continuousattention– especiallysincethesocialvisualstimuliwasalsopresentedduring auditorytrials–thereweresixdifferentvisualsocialvideos(two actors;threetypesofsocialvideo),whileeachauditorycondition

Fig.3.Upperpanel–fNIRSheadgear:Cohort2at4–8and12–16monthsofageworethefullarray(blueandgreenchannels),Cohort2at9–13monthsandCohort3(18–24 months)worethepartialarray(green).Cohort1(0–2months)woreanarraycoveringchannels5,6,7,8,9,10,11&12.Sourcelightsareindicatedbyastaranddetectorsby acircle.Lowerpanel–Stimuluspresentation.

employedtwodifferentrecordings(twospeakers;onerecording each–seebelow).During thebaseline, visualstimuliwere dis-played,whichconsistedoffull-colorstillimagesofdifferenttypes oftransport(i.e.,carsandhelicopters)presentedrandomlyfora pseudorandomduration(1–3s)for9–12s(Lloyd-Foxetal.,2012). Dynamicnon-socialbaselinestimulihavealsobeenusedin previ-ousworkinvestigatingresponsestovisualsocialdynamicstimuli, andhavebeenfoundtoproducesimilareffectstothestatic non-socialbaselineusedinthecurrentstudy(Lloyd-Foxetal.,2011, 2009).Thesevisualstimuliweredisplayedona 24-inchplasma screenwithaviewingdistanceofapproximately100cm.

2.4. Auditorystimuli

Duringthepresentationofvisualstimulitheinfantswere pre-sented with auditory stimuli (see Fig. 3) at the onset of two outofeverythreeofthetrials.Thecontentanddurationofthe social videos(9–12s) werenot synchronizedwiththeauditory stimuli.Eachauditorystimuluspresentationlasted8sand con-sisted of four different sounds (of vocal or non-vocal stimuli) presentedfor0.37–2.92seach,interleavedbyshortsilence peri-ods(of0.16–0.24s).Thetwoauditoryconditionswereequivalent intermsofaveragesoundintensityandduration(p=0.65).Within theauditorysocial conditioninfantswerepresented with non-speechadultvocalizationsoftwospeakers(whocoughed,yawned, laughed,andcried).Withintheauditorynon-socialcondition,the infantswerepresentedwithcommonenvironmentalsounds(that werenothumanoranimalproduced,butwerelikelytobefamiliar

toinfantsofthisage;runningwater,bellsandrattles).Vocaland non-vocalstimuliwerechosenfromtheMontrealAffectiveVoices (formoredetail,see(Belinetal.,2000))andthestimuliofthevoice functionallocalizer (http://vnl.psy.gla.ac.uk/resourcesmain.php). Additionalnon-vocalstimuli(toysounds)werealsorecordedby theauthors(Blasietal.,2011).

2.5. Dataprocessingandanalysis

The NIRS system measuredthe light attenuation from each sourcedetectorpair.Theselightattenuationmeasureswereused to calculate changes in the concentration of oxy-haemoglobin (HbO2)anddeoxy-haemoglobin(HHb)in␮Molwhichwereused

ashaemodynamicindicatorsofcorticalneuralactivity(Obrigand Villringer,2003).Theanalysisprocedurefollowedasimilar proto-coltopreviousinfantresearch.Initially,therecordednearinfrared attenuationmeasurementsfor each infantwere analyzed sepa-rately.Trialswererejectedfromfurtheranalysisbasedonlooking timemeasures(codedofflinebyaresearcherunfamiliarwiththe study’saims,trialswith>60%triallookingwereconsideredvalid) andchannelswererejectedbasedonthequalityofthesignal,using artefactdetectionalgorithms(Lloyd-Foxetal.,2010,2009).Foreach infant,thelightattenuationsignalwaslow-passfiltered,usinga cutofffrequencyof1.8Hz.Thedatawasthendividedintoblocks consistingof4softhebaselinetrialpriortotheonsetofthe stim-ulus,theexperimentalstimulustrial,plusthefollowingbaseline. Thelightattenuationdatawasdetrendedwithalinearfitbetween thefirstandlast4sofeachblock.Thedatawerethenconverted

S. Lloyd-Fox et al. / Developmental Cognitive Neuroscience 25 (2017) 92–104 97 Table2

SignificantHbO2responsestotheSocial>Non-SocialVisualandAuditorycontrastsacrossthethreeCohorts.

4–8months(Cohort2) 12−16months(Cohort2)

VisualSocial>Non-Social AuditorySocial>Non-Social VisualSocial>Non-Social AuditorySocial>Non-Social

Ch TW t p df Ch TW t p df Ch TW t p df Ch TW t p df 8* 12–16 2.55 0.018 23 6* 12–16 2.22 0.04 18 5* 12–16 2.83 0.011 18 9 12–16 3.57 0.0016 23 9 8–12 3.63 0.002 17 6* 12–16 2.16 0.045 18 11* 8–12 2.56 0.018 23 9 12–16 3.55 0.0025 17 11 12–16 3.54 0.0017 23 11 8–12 3.67 0.0018 17 12 8–12 3.21 0.0004 23 11 12–16 3.39 0.0035 17 12–16 5.15 0.00003 23 12 8–12 4.63 0.0002 17 12 12–16 4.60 0.0003 17

9–13months(Cohort2) 18–24months(Cohort3)

VisualSocial>Non-Social AuditorySocial>Non-Social VisualSocial>Non-Social AuditorySocial>Non-Social

Ch TW t p df Ch TW t p df Ch TW t p df Ch TW t p df

9* 12–16 2.35 0.027 24 5* 12–16 2.74 0.011 24 9* 12–16 2.21 0.045 13 5 12–16 4.11 0.0009 15

0–2months(Cohort1) 0–2months(Cohort1)

AuditorySocial>Silence AuditoryNon-Social>Silence AuditoryNon-Social>Non-Social

Ch TW t p df Ch TW t p df Ch TW t p df 1* 8–12 2.46 0.025 17 1 8–12 4.90 0.00016 17 6* 8–12 2.45 0.025 17 1* 12–16 2.37 0.03 17 4 8–12 4.23 0.0006 17 2* 8–12 2.35 0.032 17 4* 12–16 3.30 0.0043 17 2* 12–16 2.39 0.029 17 5 8–12 4.77 0.00021 17 4 8–12 3.29 0.0044 17 6 8–12 3.38 0.0036 17 12–16 4.57 0.0003 17 6* 12–16 2.47 0.025 17 8* 8–12 2.14 0.047 17

intochangesinconcentration(␮Mol)inHbO2andHHbusingthe

modifiedBeerLambertlaw(Delpy etal.,1988)and assuminga differentialpathlengthfactorforinfants(5.13;basedon(Duncan etal.,1995)).Aminimumofthreevalidtrialsperconditionwas setasa thresholdforinclusionwithininfants.Inclusion criteria requiredeachchanneltocontainvaliddatainallthree experimen-talconditions.Foreachinfant,thetrialsandchannelsthatsurvived theserejectioncriteriawereenteredintofurtheranalyses. Follow-ingthis,validtrialsforeachcondition(visual-social(silent),auditory social,auditorynon-social)wereaveragedtogetherwithin chan-nelsforeachinfant,andatimecourseofthemeanconcentration changeinHbO2andHHbwascompiledforeachchannel.Twotime

windowswereselectedbetween8–12and12–16spost-stimulus onsetforeach trial.Thisperiodoftime wasselectedtoinclude therange of maximum concentration changes observed across infantsforHbO2andHHb,basedonvisualinspectionofthe

cur-rentdata,andinformedbydataanalysisapproachesusingthesame paradigminpreviouscohorts(Lloyd-Foxetal.,2013,2012,2009). Thetimewindowwassplitintotwoepochstoallowusto inves-tigatelatencyfurther,aswehavenoteddifferencesinthetiming andshapeoftheresponseacrossthedifferentvisualandauditory conditionsinpreviousresearch.Foreachchannel,statistical com-parisons(two-tailedt-tests)ofthemaximumchange(amplitude) (inHbO2 (increase in chromophoreconcentration) and/or HHb

(decreaseinchromophoreconcentration))wereperformedforthe (1)visual-social(silent)conditioncomparedtothenon-social base-linecondition(withsilence)andthe(2)directcomparisonofthe auditorysocialandnon-socialconditionsduringthespecifiedtime windows.InadditionforCohort1only,thecontrastbetweenthe twoauditoryconditionsandsilencewasalsoperformedtoexplore theagedependentspecialisationofthisresponsefurther.Eithera significantincreaseinHbO2concentration,orasignificantdecrease

inHHb,iscommonlyacceptedasanindicatorofcortical activa-tionininfantwork,howeverinaccordancewithpreviousresearch (Lloyd-Foxetal.,2010)wefoundthatthemajorityofthe signifi-canteffectswereinHbO2andsofocusedourresultsonthissignal.

ThesignificantHHbresponsesarereportedinSupplementaryData. Toresolvestatisticalproblemsofmultiplecomparisonsforthese groupanalysesweappliedthefalsediscoveryrate(FDR)correction (BenjaminiandHochberg,1995).Thechannelsthatdidnotsurvive thiscorrectionarehighlightedin Table2,however wechoseto reporttheresultsinfullintheResultssectionasthereplicationof effectacrossagesallaystheneedforastrictFDRcorrection.

3. Results

Thefindingsreveallocalizedpatternsofactivationinregions oftheposteriorsuperiortemporal,anteriortemporalandinferior frontalcortextothevisualandauditorysocialstimuli,in concor-dancewithpreviouscohortsofinfantsstudiedintheUK.

3.1. Visualsocialversusnon-social

Toassesstheresponsestothevisualsocialstimulithe visual-social(silent) conditionwas analysed relative to thenon-social baselinecondition(withsilence)(t-test,two-tailed)foreachcohort (ateachagepoint:seeTable2)intwotimeepochs(1)8–12sand(2) 12–16spoststimulusonset.NotethatCohort1didnotcontribute datatothiscontrastastheywereeitherasleep orinaposition wheretheywereunabletoviewthevisualstimuli.Thisanalysis revealedsignificantincreasesinHbO2centeredovertheposterior

areaofthearrays(seeFigs.4and5),correspondingtotheposterior STS-TPJregionofthecortex.

Asshown inFig.5thevisual-social>non-socialresponsewas localizedtothesameregionfrom4to24monthsofagewithone

channel(Channel 9)consistently revealingsignificantactivation acrossallfourcohorts.HaemodynamictimecoursesforChannel9 areshowninFig.6.Responsesat9–13and18–24monthswereless robustthanat4–8and12–16monthsofageduetothesub-optimal headgearplacement(thismayhavecontributedtothediminished responseseeninthetimecourseofCohort3(18–24months)in

Fig.5ascoveragemaynothavereachedtheareaofpeakactivation forthisagegroup).NosignificantHHbresponseswerefoundfor eitherthe8–12or12–16stimeepochsofanalysis.

3.2. Auditorysocialversusnon-Social

ForCohort1initialstatisticalanalysesofauditorysocialand non-socialresponsescomparedwithsilence(visualsocialonly)were conducted(seeFig.7)foreachtimeepoch(8–12and12–16spost stimulusonset).ThisanalysisrevealedamorewidespreadHbO2

responsetothenon-socialstimuli(5significantchannels) com-paredwiththesocialstimuli(3significantchannels).Oneofthese channelsalsorevealedasignificantHHbresponse.

Followingthis,pairedsamplechannel-by-channelt-tests (two-tailed)wereperformedtoassessthepresenceofauditorysocialand non-socialselectiveactivationacrosseachcohort(seeTable2).An agedependentresponsewasrevealedacross0–24monthsofage. ForCohort1(0–2months)andCohort2at4–8monthssignificant non-social>socialHbO2 selectivitywasevidenced ina posterior

temporalregionofthearray(seeFigs.4,5and7).Asreported pre-viously(Lloyd-Foxetal.,2014a),asocial>non-socialresponsewas alsoevidentinthe4–8montholdinfantsbutconfinedtoHHb(the analysesofHHbwiththenewtimeepochsarereportedin Supple-mentaryDataandconfirmthispreviousfinding).ForCohort2at 9–13and12–16monthsofage,andforCohort3(18–24months) significantsocial>non-socialHbO2selectivitywasevidencedinan

anteriortemporalregionofthearray(seeFig.4and5).These signif-icantresponseswerelocalizedwithinthesame3channelsacross 9–24monthsofage(seeFig.5&6).TherewerenosignificantHHb responsesinCohort2at9–13and12–16monthsofage,andCohort 3(18–24months).Haemodynamictimecoursesforthechannels showingsignificantsocialselectivityinCohort2at9–13and12–16 monthsofage,andCohort3(18–24months)andnon-social selec-tivityforCohort1(0–2months)andCohort2at4–8monthsare showninFig.8.

4. Discussion

Wehavesuccessfullyimplementedasocialversusnon-social fNIRSparadigmintwocontrastingenvironments:ruralGambian andurbanUK.ToourknowledgethisresearchinTheGambiais thefirstneuroimagingstudytoinvestigatecorticalspecialisation tosocialcuesacrosssuchawidespanofearlydevelopment,with participantsranging inage fromnewbornto toddlerhood. Fur-thermore,thestabilityoftheelicitedsocialresponsefrom9to24 months,andthetransitionfromnon-socialtosocialselectivityseen priortothisage,providesuswitharobustdevelopmentalcurveof specialisationtosocialcuesoverthefirsttwoyearsoflife.

WiththeuseoffNIRSwewereabletomeasurelocalizedbrain responsestovisualandauditorysocialcuesacrossthefirsttwo yearsoflife.For thevisualsocialversusnon-socialcontrast,the responsewasremarkablyconsistentacross4–24monthsofage. Inthetwomeasurementchannelsthatwereplacedinthesame positionoverallfourtimepoints(Cohort2at4–8,9–13,and12–16 monthsofageandCohort3at18–24monthsofage),asignificant responsewasfoundinthesamechannelateachage.Furthermore, forthelongitudinalcohorttheresponseappearstobecomemore rapidwithage,withafasterrisetopeakseenateachsubsequentage point.Usingco-registrationfNIRS–MRIdatafrompreviouscohorts

S.Lloyd-Foxetal./DevelopmentalCognitiveNeuroscience25(2017)92–104 99

Fig.4. SignificantHbO2haemodynamicresponsesforeachagegroupinthevisualsocial>non-socialcontrastandtheauditorysocialversusnon-socialcontrasts.Notethat fortheauditorycontrastforCohort2whentheinfantsareaged4–8monthstheyevidenceasignificantnon-social>socialresponse,whilstbythenextsessionwhentheyare aged9–13monthsandbeyond,theyshowasocial>non-socialresponse(similartoCohort3at18–24monthsofage).Noauditorynon-social>socialresponseswerefound between9and24monthsofage.Significantresultsareshownfortwotimewindows,8–12spoststimulusonset(grey)and12–16spoststimulusonset(black).

Fig.5.AsummaryofthelocationsofthesignificantHbO2haemodynamicresponsesforeachagegroupinthevisualandauditorysocialversusnon-socialcontrasts.This figurecombinesallsignificantresultsacrossthetimewindows8−12and12–16spoststimulusonset.NotethatforCohort2at4–8monthsofagetheinfantsdoexhibita significantauditorysocial>non-socialresponse,butonlyinHHbandisthereforenotillustratedhere(seeSupplementaryTable2fordetailsofsignificantHHbresponses).

ofUKinfantswewereabletoextrapolatetheapproximateposition ofthesechannelsoverunderlyinganatomy(Lloyd-Foxetal.,2014b; Lloyd-Foxetal.,2015b)andidentifythatthisregionofthearraywas positionedoverthepSTS-TPJ.Theseresponseslargelyreplicated ourpreviousfindingsfromtheUKin4–8montholdinfants( Lloyd-Foxetal.,2013,2012,2009).Furthermore,giventheinterestin comparingtheresultsfromtheGambiancohortswithage-matched datainotherparticipants,usinganidenticalprotocolwewereable tocollectdatafromacohortof12–16montholdsintheUK(seeBox 1).Despitethesmallersamplesize(N=12),wefoundverysimilar resultsinthiscohortwithsignificantresponseslocalizedwithinthe sameregionofpSTS-TPJchannels,supportingthecurrentfindings. Theauditorysocialversusnon-socialcontrast,whichwasused atallfiveagepointsfrom0to24monthsofage,revealedevidenceof adevelopmentalchangeinspecialisationforauditorysocial stim-uliacrossinfancy.Whilstthe0–2(Cohort1)and4–8(Cohort2) montholdsevidencedsignificantHbO2auditorynon-social

selec-tivity(non-social>social)inaregionofchannelslocalizedoverthe

pSTS-TPJ,Cohort2at9–13and12–16monthsofageandthe18–24 montholds(Cohort3)evidencedsignificantHbO2auditorysocial

selectivity(social>non-social) in a regionof channels localized overtheanteriortemporalcortex.Furthermore,thegroup aver-agedhaemodynamicresponsesoftheinfantsfrom9to24months ofagetothesocialandnon-socialauditorystimuliareremarkably similar,despitethecontributeddatabeingderivedfromdifferent combinationsofinfantsateachagepoint(withinthe longitudi-nalcohortitwasdependentonthoseavailableforfollowup,and who then contributedvalid data,and Cohort 3 compromiseda cross-sectionalsample).Incontrasttheaveragedhaemodynamic responsesofthe4–8montholds(Cohort2)suggestthatthelackof socialselectivityandevidentnon-socialselectivitymaybedueto adelayedresponsetothesocialstimuli.Thenon-socialselectivity isonlyapparentduringtheearliertimeepochandtheresponse tobothstimulibecomes equivalentatalater timepointduring stimulus presentation.Although speechperception in newborn infantsis well described (Dehaene-Lambertz and Spelke,2015;

Fig.6. HaemodynamictimecoursesoftheresponsetothevisualsocialstimuliinChannel9forCohort2(longitudinal)at4–8months(green),9–13months(orange)and 12–16months(red)andforCohort3(18–24months)(purple).NoteHbO2responsesareinfull,whileHHbresponsesaredashed,andthelocationofthisresponseisindicated ontheschematicofthehead.

Fig.7. SignificantHbO2haemodynamicresponsesforCohort1(0–2montholds):(1)auditorysocialvssilence,(2)non-socialvssilenceand(3)non-social>socialcontrasts. Significantresultsareshownfortwotimewindows,8–12spoststimulusonset(grey)and12–16spoststimulusonset(black).

Gervainet al.,2008; Penaet al.,2003; Vannasinget al., 2016), responsestohumanvocalisationsarelesswellknown.Toallow ustocomparetheseresponseswithaUKcohort,incollaboration withcolleaguesinCambridge,wewereabletousetheauditory paradigmwithnewborninfantsat thelocalmaternityhospital. ThefindingsfromtheUKcohortof 1–4dayoldnewbornswere consistentwiththecurrentfindings(ChuenWaiLee,TopunAustin etal.unpublishedresults).Widespreadnon-socialresponses (non-social>silence)werefound,whilstanisolatedregionoftheanterior temporalloberespondedtosocialsounds(social>silence). Fur-thermoreasignificantnon-social>social selectiveresponsewas localizedwithinthepSTS-TPJregionofthecortexinthese new-borns.Thesefindingsarealsoinlinewithotherrecentresearch (Cristiaetal.,2014)suggestingauditoryresponsestosuchcuesare lessspecializedatthisage.Thiscomparisonwithpreviousresearch at0–2monthsisalsoimportant,aswhilstintheseotherstudiesall infantswereasleep,inthecurrentstudysomeofourparticipants inCohort1wereinaquietalertstate.Thereforeitisimportantto seethattheresponseswerelargelyreplicatedacrosstheGambian andUKinfantsdespitethesedifferencesintheirstateofalertnessat

timeoftest.Previousresearchwith4–7montholdswouldsuggest thatauditorysocial>non-socialselectivityemergesfromthefirst monthsoflifebecomingmorerobustinthesecondhalfofthefirst yearoflife(Grossmannetal.,2010;Lloyd-Foxetal.,2012). Gross-mannandcolleaguesonlyfoundsocialselectivityoncetheinfants reached7monthsofage,andLloyd-Foxandcolleaguesreported thestrengthoftheresponseacross4–7monthsofagecorrelated withageofinfant.Inthecurrentwork,wedidfindevidenceofa social>non-socialresponseatthisagepointbutonlywithinthe HHbresponse.Interestingly,thiswasalsothecaseforoneofour previousstudiesinUKinfants(Lloyd-Foxetal.,2012),whilstin otherinfantswehavefoundasignificantHbO2selectiveresponse

(Lloyd-Foxetal.,2013).Collectively,thisresearchsupportstheview thatthisagerangemarksashiftinspecialisationtosocial over non-socialsoundsandfutureresearchshouldfocusonbothHbO2

and HHbresponsesinthis agegroup.In contrast,by9months ofageonwardsthesocial>non-socialselectivity(andabsenceof non-social>socialselectivity)becomesarobustresponselocalized withinaspecificanteriortemporalregionofthecortexwithall threeoftheolderagedcohortsshowingresponsesinthisarea.

Fur-S.Lloyd-Foxetal./DevelopmentalCognitiveNeuroscience25(2017)92–104 101

Fig.8. Haemodynamictimecoursesoftheresponsetotheauditorysocialstimuli(leftpanels)andnon-socialstimuli(rightpanels)inasociallyselectivechannel(onupper row)forCohort2(longitudinal)at4–8months(green),9–13months(orange)and12–16months(red)andforCohort3(18–24months)(purple)andtwonon-sociallyselective channels(onlowerrow)forCohort1(0–2months)(blue)andCohort2at4–8months(green).NoteHbO2responsesareinfull,whileHHbresponsesaredashed,andthe locationofthisresponseisindicatedontheschematicsofthehead(9–24montholds(black),0–2montholds(blue)and4–8montholds(green).

thermore,insupportofthesefindings,thelocationandselectivity ofthisresponsewasreplicatedinourUK12–16-month-oldcohort (seeBox1).

Thereareanumberofpotentialchallengesinperforming neu-roimaging studiesin a rural location in The Gambia, including transportationoftheinstrumentation,andintroductionof exper-imentalmethodstoa newcommunity andpopulationofAfrica

infants.TheportabilityofthefNIRStechnologyallowedusto trans-portittothevillageofKenebawithrelativeease.Thetechnology andexperimentalparadigmswerereadilyacceptedbyinfants, par-entsandlocalfieldstaff.Formanyoftheseinfantsthismayhave beenthefirsttimethattheyhavevieweda TVmonitor,andso weshouldbeconsciousofthisinourinterpretationoftheresults, neverthelesstheresponsesintheGambianinfantswere remark-ablyconsistentwiththeUKpopulationandinfantswerecalmand attentiveduringthestudy.Indeed,theattritionrateswerewithin thestandard rangefor infantfNIRS studieswithapproximately 20–30%excludedforinattentionorfussinessusingathree con-ditioncontrastdesign(previousresearchindicatesthatinawake infantsattritioncanincreasebyapproximately10%witheach addi-tionalexperimentalconditionemployed–seereviewby(Lloyd-Fox etal.,2010)).ForthefirstvisitofCohort2(4–8months)7outofthe 42infantswerealsoexcludedduetoexperimentererror.However thisisexpectedconsideringthatthisdatawascollectedatthefirst sessionoffNIRStestingintheGambia,andduetotimeandbudget constraints,webeganthestudywithoutanyleadintimetotrain withnewstaffandinthenewsetting.Theinfantsexcludeddue tomotionartifactorexperimentererrorinCohort1(0–2months) werealsomostlythoseinfantstestedatthebeginningwhenwe weregettingusedtoconductingfNIRSstudiesinthisagegroup (asthiswasthefirsttimetheteamhadworkedwithsuchyoung infants)and sotheinfantpositionand studysetupfordata col-lectionwasoptimizedovertime.Duringlatervisitstherewasno attritionduetothisfactor.ForCohort2at12–16monthsofage infants,theyseemtohavebeenmoreliabletolookawayorfuss outcomparedtotheotheragepoints,howeverthisisbehaviour consistentlyreportedatthisagepointindevelopmentalresearch. Furthermore,childrenwhowereexcludedduetofussinessor inat-tentionwerenotmorelikelytobethosewithverylowHCZorWHZ (only3infantsunderthisexclusioncategoryalsohadzscoresunder –3)indicatingthattheseinfantscanbemeasuredwithfNIRS.Other thangrowthfaltering,weusedthesameexclusioncriteriainThe GambiaaswewouldhaveintheUK,andsoaslongastheywere wellenoughtocontributeenoughvaliddatafor asession,they wereincludedinthestudy.Specifictestsforongoinginfectionsor neurodevelopmentalproblemswerenotadministeredontheday ofthevisitandfuturelargerscaleresearchprojectsshouldtakethis intoaccount.TheCBCDdesignedheadgearfitwelloneachageof infantandprovidedrobustsignals.Itmaybepertinenttonotehere thattherestrictiononfundingthatallowedustoonlymeasure therighthemispheremayhavebeenanadvantageasitallowed foroptimalfitoftheheadgearovertheregionofinterestwitha headbandlessliabletomoveontheheadandcauseartefactinthe data.Forexample,oncetheinfants’hairbecamethickeratolder agepointswesometimeshaddifficultymeasuringthroughbraids onthefemaleparticipantsbutfoundwecouldmeasurerobust sig-nalsbetweenthebraidsonthesameparticipant.Balancedagainst thisadvantageofasmallerheadgearprovidingmorerobustsignals, wastheissueoftryingtorunmorethanonecognitiveparadigm withasingleopticalarraydesign.Thelimitednumberofchannels intheopticalarraysonlyallowedustomeasureacertainnumber ofcorticalregions.Thereforeforonetestingtimepoint(whenwe werefollowingupourlongitudinalcohortat9–13montholdsand testing18–24montholds)inKenebawecompromiseddata col-lectiononthesocialparadigmbecauseweonlyhadtwochannels placedoverourROIforthevisualsocialresponseratherthanthe originalfourchannels.Inparticular,thismayhavecontributedto thediminishedresponseseeninthetimecourseat18–24months inFig.5.WhilstthepSTS-TPJROIwaswithinreachinthe9–13 monthswhoworethisheadgeardesign,coverageinthisoldestage groupmaynothavereachedtheareaofpeakactivation.Thoughwe hadcheckedthatheadcircumferencedidnotchangesignificantly between9and24monthsofageintheseCohorts,andtherefore

utilisedthesamesizedheadband,itmaybethatthepSTS-TPJhad extendedoutofreachofsomeofourmeasurementchannelsina ventro-dorsal,ratherthanananterior-posteriordirection,by18–24 months(Kabdebonetal.,2014;Lloyd-Foxetal.,2014b).In retro-spect,andwithinformationnowavailablefromthismorerecent research,wecanseethatfutureheadgeardesignsshouldbemore sensitivetothesechangesinanatomyoverthiswideragerangeby employingalargernumberofchannelsoverregionsofinterestto accountforindividualvariabilityinanatomy.

Webelievethattheseexperienceshaveallowedustooptimize datacollectionacross0–24monthsofageforresearchinfield set-tingssuchasKenebaand thisknowledgecanbe,and indeedis being,usedtoestablishfNIRSresearchinothersimilarsettings andforlargerscalestudies.Tooptimizehigherratesofvaliddata wewouldrecommend(1)abriefperiodoftrainingwithnewstaff alongsideanexperiencedresearcherwhiletheteamlearnshowto successfullyusefNIRSwithinfants/children;(2)headgearwhichis stable,optimizedforbrainregionsofinterest,andabletomeasure throughhair(our60-channelbilateralheadgeariscurrentlybeing successfullyusedwith36montholdsinBangladesh);(3) acknowl-edgementthatdataattritionwilllikelybehigherwithcertainage ranges(i.e.12–16monthsof age);(4) optimizedparadigms for successfuldatacollection(i.e.a2conditioncontrastdesignwould resultin<20%datadropout,andwehavepreviouslyconducteda 1conditiondesignwhichhadanattritionrateof4%−Lloyd-Fox etal.,2009);and(5)ideallymorethanonefNIRSsessionper indi-vidualandlargersamplesizestoallowonetotracedevelopmental change,whileaccountingforattritionofdata.Withthesefactors takenintoaccountweareconfidentthatfNIRScanbecomeavalid measureforlargerscaleresearchstudies.

We havenow demonstratedthat fNIRS canbeeasily imple-mentedina resource-poorruralsetting andusedfromthefirst fewdaysoflifethroughtotoddlerhoodtoprovidequantitativeand objectivemarkersofneurocognitivefunction.Furthermore,the sta-bilityoftheelicitedsocialresponsefrom9to24months,andthe transitionfromnon-socialtosocialselectivityseenpriortothisage, providesuswitharobustdevelopmentalcurveofspecialisationto socialcuesoverthefirsttwoyearsoflife.Anareathatmeritsfuture investigationistofurtherinterrogatethespecificsofthe neurovas-cularresponsebyusingamathematicalmodelofcerebralblood flowandmetabolism(Banajietal.,2008)tomorefullyinterpretthe hemodynamicresponsesacrossourfiveagepoints.Giventhe sta-bilityofthisdevelopmentalcurveofspecialisation,futureresearch ofcompromiseddevelopmentcouldusethis fNIRS paradigmto interrogateatypicalitiesinbrainfunctionandtheirassociationwith riskfactorssuchasundernutritionandpoverty.Forexample,in infantstestedwithfNIRSat4–6monthsofage–wholatergoon todevelopautismatthreeyears(whenbehaviouralatypicalities becomeevident)–itcanbeseen,usingthisparadigm,thatthe audi-torynon-sociallyselectiveresponseisofagreatermagnitudethan inlowriskagematchedinfants(Lloyd-Fox,Johnson,personal com-munication).Further,themagnitudeandlatencyoftheresponse appearstodifferforthevisualsocialstimuliinhighversuslowrisk infants(Lloyd-Foxetal.,2013).Thisresearchsuggeststhat atyp-icalbrainresponsesmaybemeasurablelongbeforebehavioural symptomsbecomeapparent(whichtypicallymanifestbetween2 and3yearsofage).Thesedevelopmentalhaemodynamicresponse curvemarkersoftimingandmagnitudeareidealcandidatesto fol-lowinfutureresearchininfantsatrisk(Aslin,2012;Vanderwert andNelson,2014).Thecurrentfindingsarehoweverlimitedtothe studyparadigmemployedandmaynotbedirectlyapplicableto theimpactofenvironmentalearlyliferiskfactors.Aslongitudinal prospectiveresearchcontinuesinthisfield,increasedsamplesizes andagepoints,and fNIRSparadigms thatinvestigateresponses acrossmultipleunderlyingcoreconstructs, ratherthan justthe socialdomain,willallowustointerrogateantecedentbiomarkers

S.Lloyd-Foxetal./DevelopmentalCognitiveNeuroscience25(2017)92–104 103

ofcompromiseddevelopment(suchaspovertyandunder nutri-tion)ingreaterdepth.Furthermore,thecollectionofthisdatafrom asearlyinlifeaspossibleshouldallowonetoidentifyhowdifferent factors(suchasmaternalunder-nutrition,familypoverty,orinfant under-nutrition,caregivingpractices)compound,orcompensate for,riskforcompromiseddevelopment.Importantly, neuroimag-ingmeasuressuchasfNIRSallowsustoidentifymarkersofatypical functionfromafarearlierage(i.e.frombirth)thanbehavioural assessments aretypically able to(from 2 to 3 years onwards). Largerscaleprospectivelongitudinalstudiescouldallowusto iden-tifyindividualsatgreatestrisk,targetadditionalfamilysupportto thosefamilieswiththegreatestneedandtargetinterventionsfrom anearlyagebeforecriticaldevelopmentalmilestoneshavebeen affected.Thelong-termaimofourresearchistoestablishfNIRSas auniversalassessmenttoolfortheinvestigationoftheimpactof adversityoncognitivedevelopmentininfantsirrespectiveofwhere thoseinfantsmighthavebeenborn.

References

Abubakar,A.,Kalu,R.B.,Katana,K.,Kabunda,B.,Hassan,A.S.,Newton,C.R.,Vande Vijver,F.,2016.Adaptationandlatentstructureoftheswahiliversionofbeck depressioninventory-IIinalowliteracypopulationinthecontextofHIV.PLoS One11,e0151030,http://dx.doi.org/10.1371/journal.pone.0151030. Aslin,R.N.,2012.Questioningthequestionsthathavebeenaskedabouttheinfant

brainusingnear-infraredspectroscopy.Cogn.Neuropsychol.29,7–33,http://

dx.doi.org/10.1080/02643294.2012.654773.

Banaji,M.,Mallet,A.,Elwell,C.E.,Nicholls,P.,Cooper,C.E.,2008.Amodelofbrain circulationandmetabolism:NIRSsignalchangesduringphysiological challenges.PLoSComput.Biol.4,e1000212,http://dx.doi.org/10.1371/journal.

pcbi.1000212.

Belin,P.,Zatorre,R.J.,Lafaille,P.,Ahad,P.,Pike,B.,2000.Voice-selectiveareasin

humanauditorycortex.Nature403,309–312.

Benjamini,Y.,Hochberg,Y.,1995.Controllingthefalsediscoveryrate:apractical

andpowerfulapproachtomultipletesting.J.R.Stat.Soc.Ser.BMethodol.57,

289–300.

Blasi,A.,Mercure,E.,Lloyd-Fox,S.,Thomson,A.,Brammer,M.,Sauter,D.,Deeley, Q.,Barker,G.J.,Renvall,V.,Deoni,S.,Gasston,D.,Williams,S.C.R.,Johnson,M.H., Simmons,A.,Murphy,D.G.M.,2011.Earlyspecializationforvoiceandemotion processingintheinfantbrain.Curr.Biol.21,1220–1224,http://dx.doi.org/10.

1016/j.cub.2011.06.009.

Cristia,A.,Minagawa,Y.,Dupoux,E.,2014.Responsestovocalizationsandauditory controlsinthehumannewbornbrain.PLoSOne9,e115162,http://dx.doi.org/

10.1371/journal.pone.0115162.

Currie,J.,Almond,D.,2011.Humancapitaldevelopmentbeforeagefive.Handb.

LaborEcon.4,1315–1486.

Cusick,S.E.,Georgieff,M.K.,2012.Nutrientsupplementationand neurodevelopment:timingisthekey.Arch.Pediatr.Adolesc.Med.166, 481–482,http://dx.doi.org/10.1001/archpediatrics.2012.199.

Dehaene-Lambertz,G.,Spelke,E.S.,2015.Theinfancyofthehumanbrain.Neuron 88,93–109,http://dx.doi.org/10.1016/j.neuron.2015.09.026.

Delpy,D.T.,Cope,M.,vanderZee,P.,Arridge,S.,Wray,S.,Wyatt,J.,1988.

Estimationofopticalpathlengththroughtissuefromdirecttimeofflight

measurement.Phys.Med.Biol.33,1433–1442.

Duncan,A.,Meek,J.H.,Clemence,M.,Elwell,C.E.,Tyszczuk,L.,Cope,M.,Delpy,D., 1995.Opticalpathlengthmeasurementsonadulthead,calfandforearmand

theheadofthenewborninfantusingphaseresolvedopticalspectroscopy.

Phys.Med.Biol.40,295.

Elsabbagh,M.,Volein,A.,Csibra,G.,Holmboe,K.,Garwood,H.,Tucker,L.,Krljes,S., Baroncohen,S.,Bolton,P.,Charman,T.,2009.Neuralcorrelatesofeyegaze processingintheinfantbroaderautismphenotype.Biol.Psychiatry65,31–38,

http://dx.doi.org/10.1016/j.biopsych.2008.09.034.

Elsabbagh,M.,Mercure,E.,Hudry,K.,Chandler,S.,Pasco,G.,Charman,T.,Pickles, A.,Baron-Cohen,S.,Bolton,P.,Johnson,M.H.,2012.Infantneuralsensitivityto dynamiceyegazeisassociatedwithlateremergingautism.Curr.Biol.22, 338–342,http://dx.doi.org/10.1016/j.cub.2011.12.056.

Everdell,N.L.,Gibson,A.P.,Tullis,I.D.C.,Vaithianathan,T.,Hebden,J.C.,Delpy,D.T., 2005.Afrequencymultiplexednear-infraredtopographysystemforimaging functionalactivationinthebrain.Rev.Sci.Instrum.76,93705,http://dx.doi.

org/10.1063/1.2038567.

Fernandes,M.,Stein,A.,Newton,C.R.,Cheikh-Ismail,L.,Kihara,M.,Wulff,K., Quintana,de,E.L.,Aranzeta,L.,Soria-Frisch,A.,Acedo,J.,Ibanez,D.,Abubakar, A.,Giuliani,F.,Lewis,T.,Kennedy,S.,Villar,J.,(INTERGROWTH-21st),fortheI.F, N.G.C.forthe21stC,2014.TheINTERGROWTH-21stproject

neurodevelopmentpackage:anovelmethodforthemulti-Dimensional assessmentofneurodevelopmentinpre-Schoolagechildren.PLoSOne9, e113360,http://dx.doi.org/10.1371/journal.pone.0113360.

Fox,S.E.,Wagner,J.B.,Shrock,C.L.,Tager-Flusberg,H.,Nelson,C.A.,2013.Neural processingoffacialidentityandemotionininfantsathigh-Riskforautism

spectrumdisorders.Front.Hum.Neurosci.,7,http://dx.doi.org/10.3389/

fnhum.2013.00089.

Fukui,Y.,Ajichi,Y.,Okada,E.,2003.MonteCarlopredictionofnear-infraredlight

propagationinrealisticadultandneonatalheadmodels.Appl.Opt.42,

2881–2887.

Georgieff,M.K.,2007.Nutritionandthedevelopingbrain:nutrientprioritiesand

measurement.Am.J.Clin.Nutr.85,614S–620S.

Gervain,J.,Macagno,F.,Cogoi,S.,Pe ˜na,M.,Mehler,J.,2008.Theneonatebrain detectsspeechstructure.Proc.Natl.Acad.Sci.105,14222–14227,http://dx.

doi.org/10.1073/pnas.0806530105.

Gervain,J.,Mehler,J.,Werker,J.F.,Nelson,C.A.,Csibra,G.,Lloyd-Fox,S.,Shuka,M., Aslin,R.A.,2011.Near-infraredspectroscopy:areportfromtheMcDonnell

infantmethodologyconsortium.Dev.Cogn.Neurosci.1,22–46.

Grossmann,T.,Oberecker,R.,Koch,S.P.,Friederici,A.D.,2010.Thedevelopmental originsofvoiceprocessinginthehumanbrain.Neuron65,852–858,http://dx.

doi.org/10.1016/j.neuron.2010.03.001.

Guiraud,J.A.,Kushnerenko,E.,Tomalski,P.,Davies,K.,Ribeiro,H.,Johnson,M.H., others,2011.Differentialhabituationtorepeatedsoundsininfantsathighrisk

forautism.Neuroreport22,845.

Hackman,D.A.,Farah,M.J.,2009.Socioeconomicstatusandthedevelopingbrain. TrendsCogn.Sci.13,65–73,http://dx.doi.org/10.1016/j.tics.2008.11.003. Isaacs,E.B.,2013.Neuroimaging,anewtoolforinvestigatingtheeffectsofearly

dietoncognitiveandbraindevelopment.Front.Hum.Neurosci.7,445,http://

dx.doi.org/10.3389/fnhum.2013.00445.

Jackson,P.A.,Kennedy,D.O.,2013.Theapplicationofnearinfraredspectroscopyin nutritionalinterventionstudies.Front.Hum.Neurosci.,7,http://dx.doi.org/10.

3389/fnhum.2013.00473.

Kabdebon,C.,Leroy,F.,Simmonet,H.,Perrot,M.,Dubois,J.,Dehaene-Lambertz,G., 2014.Anatomicalcorrelationsoftheinternational10–20sensorplacement systemininfants.Neuroimage99,342–356,http://dx.doi.org/10.1016/j.

neuroimage.2014.05.046.

Kariuki,S.M.,Abubakar,A.,Murray,E.,Stein,A.,Newton,C.R.J.C.,2016.Evaluation ofpsychometricpropertiesandfactorialstructureofthepre-schoolchild behaviourchecklistattheKenyanCoast.ChildAdolesc.PsychiatryMent. Health10,1,http://dx.doi.org/10.1186/s13034-015-0089-9.

Kolyva,C.,Kingston,H.,Tachtsidis,I.,Mohanty,S.,Mishra,S.,Patnaik,R.,Maude, R.J.,Dondorp,A.M.,Elwell,C.E.,2013.Oscillationsincerebralhaemodynamics inpatientswithfalciparummalaria.Adv.Exp.Med.Biol.765,101–107,http://

dx.doi.org/10.1007/978-1-4614-4989-815.

Lloyd-Fox,S.,Blasi,A.,Volein,A.,Everdell,N.,Elwell,C.E.,Johnson,M.H.,2009. Socialperceptionininfancy:anearinfraredspectroscopystudy.ChildDev.80, 986–999,http://dx.doi.org/10.1111/j.1467-8624.2009.01312.x.

Lloyd-Fox,S.,Blasi,A.,Elwell,C.E.,2010.Illuminatingthedevelopingbrain:the past,presentandfutureoffunctionalnearinfraredspectroscopy.Neurosci. Biobehav.Rev.34,269–284,http://dx.doi.org/10.1016/j.neubiorev.2009.07.

008.

Lloyd-Fox,S.,Blasi,A.,Everdell,N.,Elwell,C.E.,Johnson,M.H.,2011.Selective

corticalmappingofbiologicalmotionprocessinginyounginfants.J.Cogn.

Neurosci.23,2521–2532.

Lloyd-Fox,S.,Blasi,A.,Mercure,E.,Elwell,C.E.,Johnson,M.H.,2012.Theemergence ofcerebralspecializationforthehumanvoiceoverthefirstmonthsoflife.Soc. Neurosci.7,317–330,http://dx.doi.org/10.1080/17470919.2011.614696. Lloyd-Fox,S.,Blasi,A.,Elwell,C.E.,Charman,T.,Murphy,D.,Johnson,M.H.,2013.

Reducedneuralsensitivitytosocialstimuliininfantsatriskforautism.Proc.R. Soc.BBiol.Sci.,280,http://dx.doi.org/10.1098/rspb.2012.3026.

Lloyd-Fox,S.,Papademetriou,M.,Darboe,M.K.,Everdell,N.L.,Wegmuller,R., Prentice,A.M.,Moore,S.E.,Elwell,C.E.,2014a.Functionalnearinfrared spectroscopy(fNIRS)toassesscognitivefunctionininfantsinruralAfrica.Sci. Rep.4,4740,http://dx.doi.org/10.1038/srep04740.

Lloyd-Fox,S.,Richards,J.E.,Blasi,A.,Murphy,D.G.M.,Elwell,C.E.,Johnson,M.H., 2014b.Coregisteringfunctionalnear-infraredspectroscopywithunderlying corticalareasininfants.Neurophotonics1,http://dx.doi.org/10.1117/1.NPh.1.

2.025006.

Lloyd-Fox,S.,Széplaki-Köll"od,B.,Yin,J.,Csibra,G.,2015a.Areyoutalkingtome? Neuralactivationsin6-month-oldinfantsinresponsetobeingaddressed duringnaturalinteractions.CortexJ.DevotedStudyNerv.Syst.Behav.70, 35–48,http://dx.doi.org/10.1016/j.cortex.2015.02.005.

Lloyd-Fox,S.,Wu,R.,Richards,J.E.,Elwell,C.E.,Johnson,M.H.,2015b.Cortical activationtoactionperceptionisassociatedwithactionproductionabilitiesin younginfants.Cereb.CortexN.Y.NY25,289–297,http://dx.doi.org/10.1093/

cercor/bht207.

Lunn,P.,Northrop-Clewes,C.,Downes,R.,1991.Intestinalpermeability,mucosal injury,andgrowthfalteringinGambianinfants.Lancet338,907–910,http://

dx.doi.org/10.1016/0140-6736(91)91772-M.

Lunn,P.G.,2000.Theimpactofinfectionandnutritionongutfunctionandgrowth

inchildhood.Proc.Nutr.Soc.59,147–154.

Luyster,R.J.,Wagner,J.B.,Vogel-Farley,V.,Tager-Flusberg,H.,Nelson,C.A.,2011. Neuralcorrelatesoffamiliarandunfamiliarfaceprocessingininfantsatrisk forautismspectrumdisorders.BrainTopogr.24,220–228,http://dx.doi.org/

10.1007/s10548-011-0176-z.

Martorell,R.,Horta,B.L.,Adair,L.S.,Stein,A.D.,Richter,L.,Fall,C.H.D.,Bhargava, S.K.,Biswas,S.K.D.,Perez,L.,Barros,F.C.,Victora,C.G.,Group,C.onH.O.R.inT.S, 2010.Weightgaininthefirsttwoyearsoflifeisanimportantpredictorof schoolingoutcomesinpooledanalysesfromfivebirthcohortsfromlow-and middle-incomecountries.J.Nutr.140,348–354,http://dx.doi.org/10.3945/jn.

McCleery,J.P.,Akshoomoff,N.,Dobkins,K.R.,Carver,L.J.,2009.Atypicalfaceversus objectprocessingandhemisphericasymmetriesin10-Month-Oldinfantsat riskforautism.Biol.Psychiatry66,950–957,http://dx.doi.org/10.1016/j.

biopsych.2009.07.031.

McCoy,D.C.,Peet,E.D.,Ezzati,M.,Danaei,G.,Black,M.M.,Sudfeld,C.R.,Fawzi,W., Fink,G.,2016.Earlychildhooddevelopmentalstatusinlow-and

middle-incomecountries:national,regional,andglobalprevalenceestimates usingpredictivemodeling.PLoSMed.13,e1002034,http://dx.doi.org/10.1371/

journal.pmed.1002034.

Mendez,M.A.,Adair,L.S.,1999.Severityandtimingofstuntinginthefirsttwo

yearsoflifeaffectperformanceoncognitivetestsinlatechildhood.J.Nutr.

129,1555–1562.

Minagawa-Kawai,Y.,Lely,H.,vanderRamus,F.,Sato,Y.,Mazuka,R.,Dupoux,E., 2011.Opticalbrainimagingrevealsgeneralauditoryandlanguage-specific processinginearlyinfantdevelopment.Cereb.Cortex21,254–261,http://dx.

doi.org/10.1093/cercor/bhq082.

Mullen,E.M.,1995.Mullenscalesofearlylearning.Circ.PinesMNAm.Guid.Serv.

Obrig,H.,Villringer,A.,2003.Beyondthevisible—imagingthehumanbrainwith

light.J.Cereb.BloodFlowMetab.23,1–18.

Ockleford,E.M.,Vince,M.A.,Layton,C.,Reader,M.R.,1988.Responsesofneonates

toparents’andothers’voices.EarlyHum.Dev.18,27–36.

Pena,M.,Maki,A.,Kovac˘i ´c,D.,Dehaene-Lambertz,G.,Koizumi,H.,Bouquet,F., Mehler,J.,2003.Soundsandsilence:anopticaltopographystudyoflanguage

recognitionatbirth.Proc.Natl.Acad.Sci.U.S.A.100,11702.

Powell,C.A.,Walker,S.P.,Himes,J.H.,Fletcher,P.D.,Grantham-McGregor,S.M., 1995.Relationshipsbetweenphysicalgrowth,mentaldevelopmentand

nutritionalsupplementationinstuntedchildren:theJamaicanstudy.Acta

Paediatr.OsloNor.1992(84),22–29.

Quaresima,V.,Bisconti,S.,Ferrari,M.,2012.Abriefreviewontheuseoffunctional near-infraredspectroscopy(fNIRS)forlanguageimagingstudiesinhuman newbornsandadults.BrainLang.121,79–89,http://dx.doi.org/10.1016/j.

bandl.2011.03.009.

Richards,J.E.,Stevens,M.,Connington,A.,2012.AstereotaxicMRIbrainatlasfor

infantparticipants.Dev.Med.ChildNeurol.54,9–10.

Sabanathan,S.,Wills,B.,Gladstone,M.,2015.Childdevelopmentassessmenttools inlow-incomeandmiddle-incomecountries:howcanweusethemmore appropriately?Arch.Dis.Child.100,482–488,http://dx.doi.org/10.1136/

archdischild-2014-308114.

Salamon,G.,Raynaud,C.,Regis,J.,Rumeau,C.,1990.MagneticResonanceImaging

ofthePediatricBrain.RavenPress,NewYork,NY.

Sato,H.,Yahata,N.,Funane,T.,Takizawa,R.,Katura,T.,Atsumori,H.,Nishimura,Y., Kinoshita,A.,Kiguchi,M.,Koizumi,H.,2013.ANIRS–fMRIinvestigationof

prefrontalcortexactivityduringaworkingmemorytask.Neuroimage83,

158–173.

Steinbrink,J.,Villringer,A.,Kempf,F.,Haux,D.,Boden,S.,Obrig,H.,2006. IlluminatingtheBOLDsignal:combinedfMRI-fNIRSstudies.Magn.Reson. Imaging24,495–505,http://dx.doi.org/10.1016/j.mri.2005.12.034. Tomalski,P.,Moore,D.G.,Ribeiro,H.,Axelsson,E.L.,Murphy,E.,Karmiloff-Smith,

A.,Johnson,M.H.,Kushnerenko,E.,2013.Socioeconomicstatusandfunctional braindevelopment–associationsinearlyinfancy.Dev.Sci.16,676–687,

http://dx.doi.org/10.1111/desc.12079.

UNICEF,2013.Improvingchildnutrition:theachievableimperativeforglobal progress.

Vanderwert,R.E.,Nelson,C.A.,2014.Theuseofnear-infraredspectroscopyinthe studyoftypicalandatypicaldevelopment.Neuroimage85,264–271,http://dx.

doi.org/10.1016/j.neuroimage.2013.10.009.

Vannasing,P.,Florea,O.,González-Frankenberger,B.,Tremblay,J.,Paquette,N., Safi,D.,Wallois,F.,Lepore,F.,Béland,R.,Lassonde,M.,Gallagher,A.,2016. Distincthemisphericspecializationsfornativeandnon-nativelanguagesin one-day-oldnewbornsidentifiedbyfNIRS.Neuropsychologia84,63–69,

http://dx.doi.org/10.1016/j.neuropsychologia.2016.01.038.

Victora,C.G.,Adair,L.,Fall,C.,Hallal,P.C.,Martorell,R.,Richter,L.,Sachdev,H.S., 2008.Maternalandchildundernutrition:consequencesforadulthealthand

humancapital.Lancet371,340–357.

Villringer,A.,Chance,B.,1997.Non-invasiveopticalspectroscopyandimagingof humanbrainfunction.TrendsNeurosci.20,435–442,http://dx.doi.org/10.

1016/S0166-2236(97)01132-6.

vanderMerwe,L.F.,Moore,S.E.,Fulford,A.J.,Halliday,K.E.,Drammeh,S.,Young,S., Prentice,A.M.,2013.Long-chainPUFAsupplementationinruralAfrican infants:arandomizedcontrolledtrialofeffectsongutintegrity,growth,and cognitivedevelopment.Am.J.Clin.Nutr.97,45–57,http://dx.doi.org/10.3945/