Using eye-tracking technology with babies to address socio ...€¦ · Using eye-tracking technology with babies to address socio-economic status associated educational disadvantages

Usingeye-trackingtechnologywithbabiestoaddresssocio-economicstatusassociatededucationaldisadvantages

Afinalreportonprojectsexploringearlyscreeningandtrainingofinfantsinchildren’scentresinEastLondon

DerekMoore

HaikoBallieux,PrzemekTomalski,ElenaKushnerenko,DeirdreBirtles,SamWass,AnnetteKarmiloff-Smith,MarkJohnson

SchoolofPsychology,UniversityofSurrey

InstituteforResearchinChildDevelopment,UniversityofEastLondonCentreforBrainandCognitiveDevelopment,BirkbeckCollege

2

Tableofcontents1. Abouttheauthors....................................................................................................................32. ExecutiveSummary..................................................................................................................43. Acknowledgements..................................................................................................................6 Funding...................................................................................................................................6 Partners..................................................................................................................................64. Rationale..................................................................................................................................75. Whatiseye-tracking?...............................................................................................................76. Primaryaimsofthestudies......................................................................................................8 Screeningstudy.......................................................................................................................8 Trainingstudy..........................................................................................................................8 Studydesign...........................................................................................................................97. WorkingwithChildren’sCentres............................................................................................11 Trainingstafftouseeye-trackers.........................................................................................12 Engagingstaffwiththeassessmentofinfants.....................................................................128. Recruitingandworkingwithparentsandinfants..................................................................13 Recruitment..........................................................................................................................13 Demographicdetailsofparticipantsrecruited.....................................................................15 DetailsoftestingfacilitiesacrossCCs...................................................................................18 UserfeedbackontestingsessionsinScreenstudy..............................................................199. Eye-trackingtasks..................................................................................................................21 Screeningstudy....................................................................................................................21 Trainingstudy.......................................................................................................................24 Qualityofeye-trackingdata.................................................................................................2810.Screeningstudy-Infantperformanceonscreeningeye-trackingtasks................................2911.Screeningstudy(cont.)-Associationsoftasksperformancewithsocio-economicstatus....3212.Screeningstudy(cont.)-Associationswithlanguageoutcomesintoddlers.........................3313.Trainingstudy-Effectivenessoftrainingtasks......................................................................3514.Conclusions............................................................................................................................38 Usingeye-trackingforscreening..........................................................................................38 Usingeye-trackingfortraining.............................................................................................3915.Publicationsandpresentationsfromthestudy.....................................................................4016.References..............................................................................................................................41

1. AbouttheauthorsProfessor Derek Moore is Pro-Vice-Chancellor Education and Health at the University ofGreenwich.PreviouslyhewasProfessorofPsychologyandHeadofSchoolofPsychologyattheUniversityofSurrey,before thathewasDirectorof the Institute forResearch inChildDevelopmentattheUniversityofEastLondon.ProfessorMarkJohnsonisHeadoftheDepartmentofPsychologyatCambridgeUniversity.Before that, andAssociateDirectorof theCentre forBrain andCognitiveDevelopment atBirkbeckCollege.In2008hewasawardedtheBritishPsychologicalSocietyPresident'sAwardfordistinguishedcontributionstopsychologicalknowledge,andin2011,hewaselectedasFellowoftheBritishAcademy(FBA).ProfessorAnnetteKarmiloff-SmithwasaprofessorialresearchfellowattheDevelopmentalNeurocognition Lab at Birkbeck College. Before moving to Birbeck, she was Head of theNeurocognitiveDevelopmentUnit at Institute of ChildHealth,University College, London.ProfessorKarmiloff-Smithpassedawayin2016.DrPrzemekTomalskiisaLecturerinPsychologyattheUniversityofWarsaw.PreviouslyhewasaResearchFellowattheInstituteforResearchinChildDevelopmentattheUniversityofEastLondon,andbeforethathewasattheCentreforBrainandCognitiveDevelopmentatBirkbeckCollege.DrHaikoBallieux is a Senior Lecturer in Psychology at theUniversity ofWestminster. HepreviouslywasaResearchFellowattheInstituteforResearchinChildDevelopmentattheUniversityofEastLondon,andbeforethatattheCentreforBrainandCognitiveDevelopmentatBirkbeckCollege.DrSamWassisaSeniorLecturerandESRCFutureResearchLeaderFellowinPsychologyattheUniversityofEastLondon,andpreviouslywasaBritishAcademyPostdoctoralFellowattheMRCCognitionandBrainSciencesUnitinCambridge.DrElenaKushnerenkowasaResearchFellowattheUniversityofEastLondon,UELBabylabLeaderandEEGLabCoordinator.Previously,shewasaResearchFellowattheCentreforBrainandCognitiveDevelopmentatBirkbeckCollege.

4

2. ExecutiveSummaryThisreportsummarisesfindingsfromasetofstudiesundertakeninEastLondonChildren’sCentres (CCs) that have important implications for future approaches to screening andinterventionofchildrenfromdisadvantagedareaswhoaremostatriskforpooreducationaloutcomes.Alargescreeningstudyperformedwithafinalsampleof174infantsaimedtoexaminethefeasibilityofusingeye-trackingassessmentsincommunitysettings.Thespecificaimwastoimproveourunderstandingofearlycognitivedevelopmentininfantsgrowingupinlowsocio-economicstatus(SES)families,aswellastoencouragethesefamiliestoparticipateandtolearnmoreabouttheirbabies.Asmallertrainingstudywithafinalsampleof23infantsaimedto examine the feasibility of attention training in community settings using eye-trackingtechnology.Inthesestudies,wedemonstratedthat:

• itispossibletouseeye-trackingtechnologytoscreeninfantsforpossibledeficitsinearlyattention

• it is possible to run training sessions targeting attention control development ininfantsatelevatedriskofatypicaldevelopmentwithacommunitysetting

• that children’s centre staff are willing and capable of being trained to use thistechnology

• thatparents found theuseofeye-trackinganengagingmethod for themto learnabouttheirinfant’semergingabilitiestocontroltheirattention

Theresultsweobtainedsuggestedthat:

• differencesininfants’processingofvisualandauditoryinformationcontingentonSESarealreadydiscernibleby6/7monthsof age.However, the sizeof thesedifferenceswasrelativelysmall.

• thetrainingeffects thatobtainedwithour low-SESsamplewerehighlycomparable tothoseobservedinpreviousstudieswithhigh-SESinfants.

Wealsoconducted,underseparatefunding,afollow-upassessmentinwhichtheinfantsthatwehadoriginallyassessedaspartofthescreeningstudyinthisprojectweretestedagainwhentheywere 2 years of age. These analyses identified a number of significant associations betweeninfants’performanceontheeye-trackingtasksat6/7monthsandtheirlanguageoutcomesat2years. This suggests that the CC assessments that we administered can successfully predictchildren’ssubsequentlanguagedevelopmentoutcomes.

Thereportincludespreviouslypublisheddata(seeReferencessection,below)aswellasotherpreviouslyunpublisheddata.

6

3. Acknowledgements

FundingThetwolinkedstudiesreportedhereweresupportedbygrantsfromtheNuffieldFoundation.TheNuffieldFoundationisanendowedcharitabletrustthataimstoimprovesocialwell-beinginthewidestsense. Itsupportsresearchandinnovationineducationandsocialpolicyandalsoworkstobuildcapacity ineducation,scienceandsocialscienceresearch.TheNuffieldFoundationhasfundedthisproject,buttheviewsexpressedarethoseoftheauthorsandnotnecessarily those of the Foundation (more information available atwww.nuffieldfoundation.org).ThisworkwasalsosupportedbyaBritishAcademyPostdoctoralFellowshiptoSW,andweacknowledgeadditionalsupportoftheWellcomeTrust(098330/Z/12/Z,AKS),theUKMedicalResearchCouncil(G0701484,MJ),andtheUniversityofEastLondonSchoolofPsychology.

PartnersWewouldliketothankallparticipatingfamiliesfortheircontribution,aswellasmanagementandstaffinCCsinTowerHamletsandNewham(London,UK).Wewouldespecially like to thankSallyParkinson,HeadofCommissioning inNewham, forhelpingwithsettinguppartnershipswithCCs,andMonicaFortyandherteamforsupportinchildren'sservicesinTowerHamlets.WewouldalsoliketoacknowledgethehelpandsupportofAcuitywhoareprovidersofTobiieye-trackingequipmentintheUKwhoprovidedtrainingmaterialforCCstaff.

4. RationaleOneofthekeychallengesforsocietytodayisidentifying,asearlyaspossible,precursorstocognitive,languageandcommunicationdifficultiessothatproblemscanbeaddressedatasearlyastageindevelopmentasispossible.Thisisparticularlyimportantforinfantrearedinfamiliesfacingeconomichardshipwhereresearchshowsthatearlydifficultiesarefarmorelikelytoleadtonegativelong-termconsequencesforavarietyofmentalhealthandwellbeingoutcomes (Noble et al., 2005, 2007). The importance of early intervention is increasinglyrecognised by governments, and there is cross-party political momentum for developingtargetedearlysocial-emotionalinterventionsforlowsocio-economicstatus(SES)groups.Thework presentedherewas intended to developneweye-tracking technology currentlyused within laboratory settings in the UK and the US, and to assess its potentialtransformationalimpactonearlyassessmentandinterventionwithincommunitysettings.Wealsowishedtoassessthefurtherpotentialforthistechnologytoengageparentswiththeirinfants’cognitivedevelopment,andtodirectlytrainearlyattentionaldifficultiesthatcouldunderpinbothsocialandcognitiveoutcomes.Ultimately,eye-trackingcouldbeusedinthreeways: first, as part of a ‘toolkit’ for identifying potential attention, language and socialperception problems in infants even before language production emerges. Second, as anintervention tool for directly training attentional processes in infants. Third, as a tool tointervenewith,andengage,parentsintheirinfant’sdevelopment.Weanticipatethat,inthelongterm,thisworkwillcontributesignificantlytothequalityoflifeofpeoplefromlow-SESbackgroundsandwillhaveassociatedeconomicbenefitsforsociety;upto7%ofchildrenenteringschoolhavesomeformoflearningorbehaviouraldifficultyandmany of these may be alleviated if identified earlier. Alleviating potential educationalproblemsbeforechildrenentertheschoolsystemwillimprovethequalityoflifeforchildren,their families and communities. If we can engage early-years professionals to use thistechnologyaspartoftheirassessments,then,astheparadigmsandmethodsimprove,thosechildrenwithpotentialdifficultiescouldbeidentifiedevenbeforetheyspeak.Thiswouldalsohelpusindevelopingearlyinterventionstargetedattheseinfantsandtheirmothers.Inthisway, we could have a significant positive impact by reducing the proportion of childrenpresentingwithproblemsatentrytoschool.Consequently, the potential longer-term social and economic impact of a successfulprogrammeofscreeningandtraining interventionsusing thesenewtechnologiescouldbelarge.USandUKreviewssuggestthattodatereturnsonsuccessfulearlyinterventionmaybeten-foldforcostexpendedintermsoflong-termeducationalandhealthoutcomes.Theyalsosuggestthatthisreturnincreasestheearliertheinterventionsstart,andthemoretargetedtheinterventionsare.IfthistechnologycanberapidlyandeffectivelyintegratedintoCCs,thenthiscouldhavesignificantlong-termhealthandeducationalimplications.

8

5. Whatiseye-tracking?Eye-trackingisatermusedtorefertotheuseofremotescanningequipmentthatcantellpreciselywheresomeoneislooking.Eye-tracking equipment uses low-level infra-redlightwhichsimultaneouslyreflectsoffthecorneaandtheretinaatthebackoftheeye(Gervainetal., 2011).The reflection from the retina revealstheoutlineofthepupil.Adetectorpicksup thereflectionof thecorneaandcomputes therelativepositionof thecornealreflectiontotheoutlineofthepupil.Assomeonemovestheireyetolookindifferentplacestherelativelocationofthetworeflectionschangesandthiscanbeusedtoworkout

wheresomeoneislooking.An initial procedure is used to calibratethese angles against a set of referencepointsdisplayedonavideoscreen.Withbabies, attractive animated attention-

grabbersareusedascalibrationreferencepoints,asweknowinfantswilllookatthese.Oneoftheissueswitholdergenerationeye-trackerswasthat,sincethecalibrationwasonlyperformed tomatcheyemovementswhen theeyewaswithinaparticular2-Dplane, thecalibrationwouldrapidlyloseaccuracywhentheparticipantmovedeithercloserto,orfurtherawayfrom,theeye-tracker.Morerecenteye-trackers, includingtheTobiiT120eye-trackerusedinthisstudy,alsotrackthepositionoftheparticipant’sheadin3-Dspace.Thismakesitpossibletocorrect,online,forthemovementofthehead,thussubstantiallyimprovingtheaccuracyoftracking.The uses of eye-tracking for studying infants has increased rapidly over the last 10 years.Increasingly,eye-trackingisusedfordetailedstudiestoestablishwhichpartsofstimuliaremostsalienttoinfantsandhowtheycontroltheirattentionasstimulichange,aswellastoseehowinfantattentionchangeswithage.TheTobiiT120iscurrentlyemployedinanumberof leading infant research centres around theworld, but to our knowledge had yet to beemployedinsituinthecommunityasascreeningandtrainingtool.

6. Primaryaimsofthestudies

In partnership with local CCs we carried out two studies using Tobii T120 eye-trackingequipment.TheScreeningstudyinvolved1746-month-oldinfantsandtheirparents.Theaimofthisstudywastoexplorethefeasibilityofdeliveringeye-trackingassessmentswithinCCs,andtoexplorethepotentialofeye-trackingtoengagestaffandparents fromlowersocio-economicstatustoparticipateintheseactivities.

TheTrainingstudy,specificallyexaminedthefeasibilityofdeliveringtargetedinterventioninCCs.Thetrainingwasdeliveredoverfiveweeklysessionsandusedcomputerisedparadigmsto train attention control in infants. 33 12-month-old infantswere recruited, 23 ofwhomcompletedthetraining.

Screeningstudy1Theprimaryaimsofthisstudywere:

1. Todeterminethefeasibilityofdeliveringeye-trackingassessmentsthathadpreviouslybeenperformedinthelabwithin‘realworld’,communitysettings.

2. Tousethesetechniquestoengageparentsfromlowersocio-economicclassgroupstohelpthembettertounderstandtheirinfants’development.

3. To build partnerships with Early Years providers and to engage them with thesetechniques, therebyfostering interest in incorporatingthesetechniques intargetedscreeningandinterventionprogrammes.

4. To examine the longitudinal relationships between eye-tracking performance andlanguageoutcomesand,iffound,toexaminehowtheseassociationsweremediatedbyfactorsassociatedwitheconomic,socialandcommunitydeprivation.

5. Toinvestigatewherefurtherrefinementoftheseparadigmsisrequiredifassociationsarenotclear.

Trainingstudy2Theaimsofthetrainingstudywere:

1. To investigate the feasibility of using eye-tracking to deliver targeted earlyinterventionsinCCs.

1Titleoffundedproject:Usingnewtechnologiestoengageparentsfromdisadvantagedareasintheassessmentofthelanguageandsocialdevelopmentoftheirbabies(TakeaLookBabyStudy,TALBY). 2Titleoffundedproject:Traininginfantattentionalcontrolinacommunitysetting:apilotstudyusingeye-trackingtechnologytodelivergaze-contingencytraininginchildren’scentres(iSTARs)

10

2. Toinvestigatewhetherattentiontrainingcouldbeshowntoleadtoimprovementsinattentional control in infants from families with varied SES background, and tocompare the improvements observed with those documented in previous studiestargetinghigh-SESinfants.

3. Todemonstratethatparentalengagementandattendancecouldbemaintainedacrossa5-week-longtrainingprogramme.

StudydesignThescreeningstudyconsistedofthreeelements:1) Working with the staff in six CCs in East London to train them to use eye-tracking

equipment.2) Creatinganddeliveringneweye-trackingassessmentstothebabies,assessingthequality

ofthedataproduced,anddeterminingtherangeofdifferencesininfantattention.3) Assessingthelevelofengagementofparentswiththesemethods.Inaddition,underseparatefunding,weconductedafollow-upassessmentinwhichasubsetoftheinfantsfromthepresentstudyweretestedagainwhentheywere2yearsofage.Theaimofthisstudywastodeterminewhetherthedatacollectedatage6-7monthsisabletopredictemergingdifferencesinlanguageandcommunication.The intention was to recruit 200 infants for the initial training study. We slightly under-recruitedrelativetotheseinitialaims:thefinalsampleconstituted174infantsat6-7months.For the follow-up assessment, 45 of the children who participated in the original studyattendedasubsequenttestingsessionat24monthsofage.

FortheTrainingstudy,ouraimwastorecruitaseparatecohortof30infantsaged11to12months.Allinfantsparticipatinginthetrainingstudyweretocompleteaninitialassessmentof their attention control capacities (‘pre-test’). Then, over thenext fiveweeks, halfwereallocatedtoreceiveattentiontraining,andtheremaininghalfwereallocatedtoamatched,activecontrolgroup.Oncompletionofthetrainingprogram,all infantsrepeatedtheinitialassessmentoftheirattentioncontrolcapacitiesaspresented(‘post-test’).Weslightlyunder-recruitedrelativetoourinitialaims:althoughweinitiallyrecruited33infants,only23infantscompletedthefullprogramandthepost-testassessment.

Post-testn=23

TraininggroupWeeklycontingenttraining

over5weeksN=10

ControlgroupN=13

Pre-testn=23

12

7. WorkingwithChildren’sCentres(CCs)Thisstudywasundertaken

in partnership with six CCs (seecoloured locations on the map,with the red pin indicating thelocationofUEL).TheseCCsweresupportedbyChildren’sServicesand local Health Services in theLondon boroughs of Newhamand Tower Hamlets. Localauthorities in the UK followdifferent models of CCmanagement,andtheCCsinourstudyreflectthisdiversity.

InNewham, CCs are semi-independent and often set up by existing schools and nurseries responsible for theirmanagement and budget.Managers and staff in three CCs in Newhamwere approachedseparately,allrespondingpositivelytotheideaofparticipationintheproject.Theyindicatedthattheyvaluedtheproject’sgoalsandcouldseehowitcouldpotentiallybenefitfamiliesintheircentres.CommitmenttotheprojectintheboroughofTowerHamletswasalsogood.Themanagement of CCs in this boroughwasmore centralised,with a clearmanagementstructureoverseeingallgovernance,research,andexternalcollaborations.AttheearlystagesoftheprojectweapproachedandreceivedformalsupportfromtheHeadofEarlyYearsinbothboroughsanddeliveredformalpresentationsattheirCCmanagers’meetingandtoeachcentreindividually.Whenapplyingforfundsfortheprojectwereceivedacommitmentinkindinstafftimetoallowstafftotakepartintrainingandassessments.

These two boroughs

chosen for the study are inthe top five for social andeconomic deprivation inEngland (see IMD densitymap DCLG, 2010), and havearound42%ofchildrenlivingbelow the poverty line(Aldridgeetal.,2013).

Trainingstafftouseeye-trackersOneaimofthisprojectwastoestablishthefeasibilityofengagingCCstaffwitheye-

tracking technology, in order to ascertain whether they would be comfortable withundertaking assessments. To facilitate this process, we worked closely with eye-trackingtechnologyspecialists(AcuityLtd.)todevelopashorttrainingprogramme,andcreatedstudymaterialstoeducateandengagethestaff.

Thetraininglasted2-3hours,andusuallytookplaceintheCC.Itcomprisedaseminar

oninfantcognitiveresearchtocontextualisethework,followedbybasictrainingofpracticalskillsonloadingandrunningtheeye-trackerparadigms.Thisincludedinformationonhowtosetupandtoundertakeeye-trackingrecording,aswellashowtoreplaytoparentsthevideooftheinfanteye-trackingtrace.

StaffmembersinallbutoneCC(duetotimerestrictions)participatedinthetraining

sessionsatthebeginningoftheproject.Intotal,weconductedfivetrainingsessionsattendedby16CCstaffmembers.Weaskedallparticipantstogivefeedbackbyansweringquestionsonthequalityofthetraining,itsabilitytoholdattentionanditsusefulness.Eachquestionwascoded1to5(poortoexcellent).Ourfeedbacksuggestedthat:

• 15/16staffreportedthelevelofcontentoftraininggood,verygoodorexcellent• 16/16foundtheabilityofthetrainingtoholdtheirattentiongood,verygoodorexcellent• 16/16foundtheusefulnessofthetrainingfortheirworktobeverygoodorexcellent.Overall,thesefindingsindicatethatstaffthoughtthatthetrainingwasgenerallygoodqualityandwasrelevanttotheirwork.

Engagingstaffwiththeassessmentofinfants Onthewhole,engagementofstaffmembersandtheirmanagerswashigh.Mostmanagerswereinterestedintheprojectandwerekeentoletstaffmemberstaketimetoconducttheeye-tracking sessions with infants. The participation of staff members who took part intraininginassessmentswasgood,with75%(n=12)sittinginon(i.e.,observing)and/orpartlyrunning(i.e.,activelydelivering)atleastonesession.InoneCC,staffwereveryengagedandsetuparotaforstaffmemberstoparticipateinthesessionsinagivenweek.IntwootherCCs,wehadonestaffmemberparticipate.Staffmembersincludednurseryandteachingstaff,withagesrangingfrom20to47years.However,inthreeCCs(includingtheonementionedearlierthatdidnotparticipateintrainingsessionsduetotimerestrictions)staffmembersdidnotparticipateinthesessionsdeliveredbytheexperimenter.Thiswasnotduetolackofinterest,but rather reflected increased workload in Tower Hamlets in 2010-11 as a result of re-organisationcausedbynationalbudgetcuts.

14

8. RecruitingandworkingwithparentsandinfantsRecruitment

BothstudiesreceivedclearancefromtheUniversityofEastLondonethicscommittee,

andadditionalclearancewasobtainedfromtheResearchGovernanceDirectorateofTowerHamlets.

For the Screening studyall participatingCCs advertisedour ‘LearnAboutYourBaby’

sessions(seeadvertisement,below).Thesesessionswereadvertisedasapotentiallearningexperienceforparents.SessionswerescheduledandadvertisedinCCquarterlyactivities

calendars for parents alongside other baby-targeted activities (e.g. baby yoga, baby club,parentandtoddlergroup).

WeprovidedtheCCswithourrequiredage-rangeandinclusioncriteria:agerange,6

months0daysto7months30days;nopre-terminfants;nomajormedicalcondition;andnomajordeliverycomplications.Then,theCCsaccessedtheirowndatabaseandsenttheflyeranda study information sheet toallparentswith infants fitting thesecriteria. Inaddition,flyersandpostersweredistributedintheCCreceptionareas.

The informationmaterials were written in English except for the ‘calling all babies’

phraseontheflyerandposter3.ManyCCstaffmemberswereabletospeakotherlanguages,andwerethereforeabletoexplainthestudyinmoredetailtoparentswhohadqueriesafterseeingtheflyer.ParentswhowishedtotakepartcontactedtheCC,orresearcherdirectly,tobookanappointment.SincetheparentsoftenalreadyknewthestaffmembersworkingattheCCs,thismayhavemadethemmoreinclinedtojointhesessions.

TheCCmanagersestimatedthataround50%ofthetotalnumberoftheparentson

theirdatabaseswhomtheycontactedactuallytookpartinthestudy(thisestimatevariedfrom33%to65%acrossthecentres).Allparentswerebriefedpriortotakingpartinthestudythatthiswasaresearchprojectandtheresultscouldnotbeusedindiagnosinganydifficultiesofindividualbabiesbeforethesemethodshadbeenvalidated.

Onedisadvantageofbeingpartofthetimetablewasthatthisrestrictedustothesame

sloteachweekwhenwecouldtestparticipantsinagivencentre.Ifthisonce-weeklyslotwasnotconvenientforparents,thentheycouldnotalwaysbetested.Inafewcases,weassessedinfantsinanotherCCifthetimingofthesessionandlocationweremoreconvenient.

Groupsofsessionsweretimetabledeitherforamorningorafternoon,orinsomecases

allday.Aswewereusingonesetofequipment,andoneteamofresearchers,itwasessentialtocarefullycoordinatethetimetablingofsessionsthroughouttheweekacrossCCs.Overall,this allowed us to comfortably assess on average three infants a week in each CC, orapproximately20infantsperweekacrossthesixCCs.

The testing session itself consisted of an introduction by the experimenter,

administrationofthefiveeye-trackingtasks,eachlasting5minutes,ascriptedplaybackofvideos of the infant performance, the completion of parent questionnaires and a sessionevaluation.Parentsweregivenacertificateofparticipation for theirbaby,a£10shopping

3GiventhediversityoflanguagesinEastLondonitwasnotpossibletocreatematerialstranslatedforall.However,infuturestudiesitwouldbepossibletotargetspecificlanguagegroupswithtranslatedmaterialsifthiswerethefocusofthestudy.

16

voucher,andachildren’sbook.Inaddition,withpermission,wetookapictureofthebabyandsentanA4printoutonphotopapertotheparent’shome.

Demographicdetailsofparticipantsrecruited

ScreeningstudyOurinitialtargetwastorecruit200infantsoversevenmonths.Werecruited12whotookpartininitialpiloting,afurther9whowereassessedbutlaterexcludedfromanalysisastheydidnotmeet age and/or health criteria, and a further 174who completed a full assessmentsession,andfromwhomeye-trackingdatawerecollected.Allparticipantsincludedinthefinalsample were born full-term (36-42 weeks gestational age). A comparable proportion ofparticipantscamefromeachofthetwoboroughs(Newham54.5%;TowerHamlets45.5%).Thetablebelowshowskeydemographicdetailsforourscreeningstudyparticipants:Demographicdetailsforourscreeningstudyparticipants.Resultsshowmeanvaluesforkeydemographic details for our screening study participants. Results in brackets show thestandarddeviations,indicatingthevarianceinthesevaluesacrossthepopulationsampled.Measure Measure

Age(days) 209.3(19.7)

Gender FemaleMale

39.1%60.9%

Gestationalage(weeks)

39.5(1.5)

Birthweight(g)

3229.1(501.5)

InfantEthnicity WhiteBritishOtherwhiteAfro-CaribbeanAsian,Indian&PakistaniAsianBangladeshiMixed/Other

15.3%11.2%11.8%21.2%14.0%26.5%

MotherEthnicity WhiteBritishOtherwhiteAfro-CaribbeanAsianIndian&PakistaniAsianBangladeshiMixed/Other

19.0%14.9%11.5%23.6%14.9%16.1%

Mother’sageatbirth(years)

30.1(4.9)

Timeatpresentresidenceinmonths

42.1(40.5)

No of oldersiblings

01234

57.5%29.3%6.9%5.7%0.6%

Residence Ownhouse/flatRentedhouse/flatCouncilhousingRentedroomOther

28.7%39.7%25.9%2.3%3.4%

Average grosshouseholdincome

£49,487(65,456)

Averagegrossincomeperfamilymember

£15,226(21,774)

MotherSOC* (1)(2)(3)(4)(5)(6)

32.8%8.0%1.1%2.3%11.5%44.3%

FatherSOC*

(1)(2)(3)(4)(5)(6)

47.2%8.1%4.3%6.2%21.1%13.0%

18

Mother’seducationlevel**

(1)(2)(3)(4)(5)(6)

20.1%35.1%14.4%10.3%14.4%5.7%

Father’seducationlevel**

(1)(2)(3)(4)(5)(6)

19.0%31.6%11.5%12.1%13.8%4.6%

Breast-feeding

0-3months4-6monthsCurrently

84.5%70.1%58.0%

Age introducedsolidfood

3months4months5months6months

2.2%21.8%24.7%47.2%

*Social Occupation Classification (UK Office for National Statistics, 2010) as: (1) higher managerial orprofessional,(2)intermediateoccupations(3)smallemployersandownaccountworkers(4)lowersupervisoryandtechnicaloccupations(5)routine/semi-routineand(6)long-termunemployed.**Educationbands:(1)postgraduatedegree(2)degree(university)(3)F.E.qualification(4)A-levels,(5)GCSE(6)noformaleducationThetablebelowshowstheprimaryandsecondlanguagesoftheresearchparticipantswhoparticipatedinthefinalcohort:

Primaryandsecondarylanguageofparticipants

PrimaryLanguage

EnglishBengaliHindi/UrduGujarati/PunjabiPolishMandarin/CantoneseRomanianBulgarianTamilLithuanian/Latvian

Other

58.6%8.0%6.9%4.0%1.7%1.7%1.1%1.1%1.1%1.1%

14.7%

SecondaryLanguage

NosecondlanguageEnglishBengaliGujarati/PunjabiMandarin/CantoneseHindi/UrduFrenchSpanishGermanArabic

Other

28.7%30.5%5.7%6.9%4.6%2.9%2.9%2.9%1.1%1.1%

12.7%

Thefigurebelowshowsabreak-downofparentethnicityacrossthesixCCsweused:

ParentethnicityacrossthesixCCsrecruited.

This informationshowsthatallCCsattractedahighlydiversesample–butthat,as

eachCCwaslocatedwithinadifferentcommunity,theyrecruitedsub-sampleswithdifferentprofilesofethnicities.Thisreflectsthemanydifferentpopulation‘pockets’ofethnicityfoundacrossthesetwoboroughs inEastLondonandindicatesthatthestudymanagedtoreflectwellthelocaldiversityofeachCC4.

TrainingstudyInaddition,wealsorecruitedafurtherseparatecohortof3311-to12-month-oldinfantsfortheTrainingstudy.Ofthese33enrolledinfants,8droppedoutafteronesessionandafurther2after2sessions.Reasonsfordrop-outincludedequipmentfailure,sicknessinthefamily,andlackofparentalengagementintheprogramme.Thefinalsample,therefore,constituted23infants.Thetablebelowshowsthedemographicprofileofthefinalsample.Ofthe23infantscompletedtheTrainingstudy,10wereinthetraininggroupand13inthecontrolgroup.Genderratioswere5male/5femaleforthetraininggroup,and5male/8femaleforthecontrolgroup.Theaverageageswere347(SD=14.3)and362(SD=17.6)daysforthetrainedandcontrolgroupsrespectively.4ThedataisavailableatthelatestCensus2011dataforindividualwards.CensusdataforTowerHamletscanbefoundherehttp://www.towerhamlets.gov.uk/Documents/Borough_statistics/Ward_profiles/Census-2011/RB-Census2011-Ethnicity-2013-01.pdfAndforNewhamhttp://www.newham.info/dataviews/view?viewId=576

20

Demographic details for our training studyparticipants. Results showmean values for keydemographic details for our screening study participants. Results in brackets show thestandarddeviations,indicatingthevarianceinthesevaluesacrossthepopulationsampled.

Measure

N 23Ageindays 352(14.8) Gender(%female/male) 57%/43%

Infantethnicity(%white.non-white) 21.7%/71.3% Meanfamilyincome £22,911(27838) Medianfamilyincome £10,702 Rangeofincome £0-10,6250 Gestationalageinweeks 38.7(1.7)Birthweightingrams 3100.5(573.8)Mother’seducationlevela 1:9.1%

2:36.4%3:27.3%4:0%5:13.6%6:13.6%

Father’seducationlevela 1:5.0% 2:55.0%

3:25.0%4:5.0%5:10.0%6:0%

_____________________________________________________________________aParentaleducation:(1)–post-graduate;(2)–highereducationdegree;(3)–furthereducation;(4)-highschoolA-levels;(5)–GCSE;(6)noqualification.

DetailsoftestingfacilitiesacrossCCsOne of the potential challenges of testing in CCs is the potential variability of availablefacilities. Testing was conducted in rooms varying in size, background noise, lighting,temperature,andvisualdistraction.Theroomswerenormallyemployedforgroupactivities(minimum4x5meters),orasconsultingroomsbyhealthvisitorsormidwives(onaverage3x3meters).Thebackgroundnoiseinmostroomswaslow,apartfromonecentrelocatednexttoabusyroad.Staffmembersunderstoodtheimportanceofnoiselevelsandwereverycooperativeinkeepingthemtoaminimuminthecorridorduringtestingsessions.Wemeasuredambientsound levels in 7%of sessions at a positionequivalent to thatof the infant’s head to the

screen.Theaverageoverallsoundlevelinthetestingroomswas49.5dB,rangingfrom40.2to55.1dB.Anyobviousvisualdistractions(colourfulpostersetc.)weremovedoutofsight.InthreeCCs,theamountofdaylightenteringtheroomwasnoteasytocontrol.Lightingconditionswerethereforeslightlydifferentineachcentre.Thefigurebelowcontainstwophotosillustratingtherangeoftestingfacilitiesavailableatthevarioussites.

Picturesillustratingthetestingset-upsusedattwoCCs

UserfeedbackontestingsessionsintheScreeningstudy

Following completion of their participation in the Screening study, all participatingparentswereaskedtogivefeedbackonthesession.Theiranswerstothequestionspresentedareshowninthetablebelow.

Theseresults,documentedabove,wereconsistentwithourexpectations:wedidnot

expectparentstoradicallychangehowtheythoughtabouttheirbabymerelyfollowinga1/2-hour assessment session, but we did expect them to change to some extent theirunderstandingofwhatkindsofthingstheirbabywasinterestedinandcapableofdoing,evenatsuchayoungage(e.g.attentionshifting,decidingwhattoobserve).ParticularlynoteworthyisthefactthatparentsoflowerSESconsideredthesessionmoreimportantandinfluentialontheirknowledgeoftheirbaby.Unemployedmothersorthosewithmanualjobsmoreoftenreportedthatitverymuchchangedtheirunderstandingcomparedtoemployedmotherswithmoreskilledjobs(Chi-squared=26.89,p=.001).Similardifferencesemergedforfatherswithroutinejobsor long-termunemployedcomparedtoworkingfathers(Chi-squared=18.3,p=.019).

22

Resultsofquestionnaireexaminingparents’satisfactionwithparticipationintheScreeningstudy,brokendownbymaternalSES

%ofparentsresponding(n=174)

1–notatall;2–alittle;3–neutral;4–quitealot;5-very

1 2 3 4 5

Q1.How interesting didyoufindthissession?

Overall 0 0.6 2.3 26.6 70.5SESgrp1&2 0 0 1.4 31.5 67.1SESgrp3 0 1.0 3.0 23.0 73.0

1 2 3 4 5

Q2. How much did youenjoythesession?

Overall 0 0.6 7.5 37.6 54.3SESgrp1&2 0 0 9.6 38.4 52.0SESgrp3 0 1.0 6.0 37.0 56.0

Q3. How much has thissession changed the wayyouunderstandyourbaby?

1 2 3 4 5

SESgrp1&2 5.4 6.8 37.0 48.0 2.8SESgrp3 2.0 14.0 24.0 40.0 20.0

1 2 3 4 5

Q4. How much has thissession changed the wayyouthinkaboutyourbaby?

Overall 36.4 11.0 23.7 21.4 7.7SESgrp1&2 39.6 9.6 27.4 22.0 1.4SESgrp3 34.0 12.0 21.0 21.0 12.0

SECclassifications(1)–highermanagerialandprofessionaloccupations;(2)–intermediateoccupations;(3)–routineandmanualoccupationsorlong-termunemployed.

9. Eye-trackingtasks

For theScreeningstudy,weusedadaptedversionsof tasks thatarewidely inuse inotherprojectsaroundtheworldthattrackearlycognitivedevelopmentininfants.Weselectedthesefive ‘candidate’ tasks inordertoassesswhethertheymightbeuseful for identifyingpotential individual differences in early social attention and communication in non-labsettings. SES has been associated with differences in the quality of early mother-infantinteractions,whichmayalsobeassociatedwithdifferencesinattentiontofaces(Gligaetal.,2009),and in theemergenceofdifferences in jointattentionbehaviours,particularlygazemonitoring(Elsabbaghetal.,2012),aswellasdifferencesinaudio-visualspeechintegrationandauditorydiscrimination(Guiraudetal,2012).Thefollowingtasksmeasureparametersofsocial attention such as attention to faces, eye gaze, joint attention behaviours, and gazemonitoring.Theyalsoexamineaudio-visual speech integrationandauditorydiscriminationwhichareprecursorsforsuccessfullanguagedevelopment.

Screeningstudy

Assessingearlysensitivitytosocialcues.

Task1:Focusonfaces(‘Pop-out’)Inthistask,infantsarepresentedwitharraysofimagesofdifferentclassesofobjectsarrangedinacircle.Inhalfthetrialsoneoftheimagesisaface. The aim is to assess the extent towhichinfants show a preference for looking at facesoverotherclassesofobjects(birds,cars,shoes,etc.).Measuresofindividualdifferencesinface‘pop-out’includethenumberoflooks,durationoffirstlook,andtotallookingtothefacecomparedtoobjectsovertrials.Groupdifferenceshavebeenfoundinmeasuresof pop-out between infants with and withoutautism(Gligaetal.,2009;seealsoFrank,Amso&Johnson,2014;andFrank,Vul&Johnson,2009).Inourversion,weadaptedthetaskforusewithadiversepopulationbyincludingawidervarietyofethnicitiesoffaces(Ballieuxetal.,2013).

Ascreenshotfromthe‘pop-out’task.Purpleareasshowthegazelocationsof

theinfant.

24

Task2:Eye-GazefollowingFromtheageof6months, infants increasinglyusethedirectionofaperson’sgazeasacueforlooking towards objects of attention (Senju &Csibra,2008),andtreatgazereferencesascuesfor learning new words (e.g. Gliga & Csibra,2009; see also Gillespie-Lynch et al., 2013).Differencesininfantsensitivitytodynamiceyegazehavebeenassociatedwithlateremergingautism(Elsabbaghetal.,2012).In this task, we measured the frequency oforientingresponsestowardsthelocationofanobject congruent with the direction of gaze,eithertotheleftorright,shownbyafaceonavideoscreen.Again,weadaptedthistasktobeappropriateformorediversesamplesbyincludingarangeofethnicitiesoffaces.

Task3:ViewingasocialinteractionShort (30 seconds) video clips showing talking andinteracting people were presented to measureinfants’ orienting to social cues in naturalisticsettingsaswellastrackingofinteractionpartners.The analysis of fixation distributions (that is, howlongandofteneyegazeisfixedondifferentpartsoftheimageorvideo)allowsustostudytheallocationofattentiontotheeyesandmouthandothersourcesofinformationinadisplay(Võetal.,2012).Individualdifferences in allocation of attention to thesenaturalistic scenes can predict later social andcognitivedifficulties(Wass&Smith,2014).

Ascreenshotfromthegazefollowingtask.Redshowsthegazelocationsoftheinfant.

Ascreenshotfromthemovie‘Charliebitmyfinger’,usedinTask3

Assessingearlysensitivitytospeechandphonemes

Task4:Audio-visualspeechintegration(AVSI)This task assessed infants’ expectations of the relationship between pronounced speechsoundsandexpectedlipmovements.Weusedaneye-trackingversion(Tomalskietal.,2012)ofanaudio-visualspeechintegrationtaskdevelopedbyKushnerenkoetal.(2008).Weadaptedthistaskformorediversesamplesby including a range of ethnicities of faces(Mooreetal.,2014). Infantswerepresentedwithfourtypesofvideos:2congruentvideoswhere the auditory trackmatches the seenarticulatory lip movements (using sounds/ba/ and /ga/) and 2 incongruent videos(visual /ba/ dubbed onto auditory /ga/ andvice versa) along with a silent face controlcondition.Lookingtimestotheeyesandmouthat6-9months of age in this task have been shown to be predictors of receptive languagedevelopmentintoddlersandfoundtobeassociatedwithdistinctivepatternsofbrainactivityduring AVSI (Kushnerenko et al., 2013a,b; also see Lewkowicz & Hansen-Tift, 2012). Inaddition,infantsatriskfordevelopingautismshowatypicalfacescanningduringaudio-visualintegration(Guiraudal.,2012).

Task5:VoweldiscriminationtaskApreferentiallisteningprocedure(modelledafterPolkaetal.,2008;Mattocketal.,2008)wasusedtotestthediscriminationoftwovowelsembeddedinaword(‘dog’vs.‘dug’and‘bet’vs.‘bat’),whileinfantssawapictureofamobilephoneonthescreen.Infantswerefamiliarisedwithonewordpresentedrepeatedlyfor30secondsandthenwerepresentedwithtwotesttrials – onewith the familiarisedword only, and anotherwith the novel word presentedbetween instancesof the familiarisedone.Preference (longer looking times) for thenovelvowel/wordwastreatedasanindicationofvoweldiscrimination.Thereasonforincludingthistaskwastoassesswhetherdelaysintheemergenceofvoweldiscriminationmaypredictlaterlanguagedifficulties.

Left-ascreenshotfromtheAVSItask.Right–theareasusedfortheAreaof

Interestcoding

26

TrainingStudy

TrainingtasksThetrainingstudyusedanidenticalsetoftrainingstimulitothoseusedinpreviouslypublishedresearchwithhigh-SESchildren(Wassetal.,2011).Thefollowingfourtrainingtaskswereincluded:

Screenshotsfromthefourtrainingtasksused.a)ButterflyTask;b)StarsTask;c)WindowsTask;d)SuspectsTask.

Task1:ButterflyThebutterfly(indicatedinred)scrolledfromlefttorightaslongasthechildlookeddirectlyatit,withstaticandmoving(indicated inblue)distractorspresented inthechild’speripheralvisualfield.Ifthechildlookedtoanyofthedistractors,allofthedistractorsdisappeared,andthescrollingstopped.Asindicatedbytherectangles,alldistractorsonscreenweremoving(inthedirectionindicatedbythearrows)withtheexceptionofthewallandstars,whichwerestatic.

Task2:StarsAtarget (indicatedred)waspresentedon-screenalongwithstaticandmovingdistractors.Movingdistractorsarehighlightedinblue;intheimageshownonlyonedistractorismoving,butothertrialsfeatureduptoeightmovingdistractors.Ifthechildlookedtothetargetwithina timewindow, theyreceivedanaudio-visual reward.Bothtargetanddistractorschangedbetweentrials.

Task3:WindowsAtarget(indicatedred)waspresentedinonelocationonthescreen.Allfourwindowsthenclosedandfixationtarget(theredflower)appearedforavariableinter-stimulusinterval.Afterthe fixation targetdisappeared, a lookback to the cuedwindow triggeredanaudio-visualreward(acartoonanimation).

Task4:SuspectsAtarget(indicatedred)waspresentedalongwitharangeofdistractors.Ifthechildlookedtothetargetwithinatimewindow,theyreceivedanaudio-visualreward.Onceperblockof12trialsthetargetchanged.Targetsfromthepreviousblock(indicatedyellow)werepresentedconcurrentlywiththecurrenttarget,asdistractors.

Pre-andpost-testassessmenttasksIn order to assess transfer of training effects, the following tasks were presented to allparticipating infantsbeforeandafter their completionof the intervention.The taskswerepresentedinterleavedwithoneanotherinorder, inabatterythatlastedapproximately20minutesintotal.Inordertomaintainengagementduringtesting,anumberofshortclipsfromTVprogrammeswerealsopresentedbetweenexperimentalblocks.

Task1:SustainedAttentionFourdifferentstillimageswerepresented:twoperblockintwoblocks.Eachblockcontainedoneoftwo‘interesting’images(i.e.attractive,detailedimagesofflowersandfish)andoneoftwo‘boring’images(i.e.low-detail,monochromeoutlinesofadiamondandacross).Trialscommenced once the baby had fixated a central target. Trials endedwhen the baby hadlookedawayfromthescreenfor1second,asjudgedbyanexperimenter,orwhen15secondshadelapsed.Attheendofeachtrial,afixationtargetandbriefauditorystimulus(<1second)werepresented. If thebaby fixatedthetarget, thenext trial started immediately; ifnot,asequenceofdifferent fixation targetsandauditoryattentiongetterswas repeated.Stimuliwerere-presenteduntil:i)twosuccessivelookswerelessthan50%ofthelongestunbrokenlooksofar, ii)eightsuccessivelookshadtakenplacewithoutreachingtheaforementionedcriterion,oriii)thetotalpresentationlengthexceeded120seconds.

28

Screenshots fromthe fourpre-andpost-testassessment tasksadministered in thetrainingstudy.a)ExamplesofthestimuliintheSustainedAttentiontask:‘Boring’(top)and‘Interesting’(bottom)stimuli.b)SchematicshowingthescreenlayoutforatrialintheShort-TermMemorytask.c)SchematicshowingtheCognitiveControltask.d)IllustrationofscreenlayoutfortheoverlapconditionGap-Overlaptask.Inthebaselinecondition,thecentraltargetdisappearedasthelateraltargetwaspresented.

Task2:Short-termmemoryShort-termmemory(STM)wasassessedintwoblocks,eachconsistingofseventrials.Ineachtrial, two targets (each subtending 5˚) were presented for 6500 milliseconds (ms). Twoseparateoccludersthenappearedandcoveredtheobjectsfor2500ms.Theoccludersthenrevealed the objects; one of the objects had changed colour. The two objectswere thenpresentedfor7000ms.Thedependentvariablewaswhetherthefirst lookwastothesidewhere thecolourof the targethadchanged following theocclusionperiod,or to the sidewhere the targetwas the samecolourasprior to theocclusion (followingKaldy&Blaser,2013).Thelocationofthechangesidevariedbetweentrials.

Task3:CognitivecontrolThistaskwaspresentedintwoblocks,eachlasting18trials.Afterfixatingacentraltarget(acartoonflowersubtending4.5˚ofvisualanglefromtheperspectiveoftheinfant),thetrialcommencedfollowinga300msdelay.Twoblankrectangles(10.8˚x9˚ofvisualangle)werepresentedleftandright,concurrentlywithanauditorystimulusfor2000ms(theanticipatorywindow).Avisualrewardanimation(lasting4000ms)thenappearedononeside(ineithertheleftorrightrectangle)forninetrialsinarow(thepre-switchphase)beforeswappingsidesfor thenextnine trials (thepost-switchphase). If theparticipant correctly anticipated thepresentationofthereward(definedasasaccadebeginningbetween300and2300msaftertrialonsetandsubjecttoaminimumlookdurationof400ms),thenthevisualrewardstimulusappeared immediately. The outcomemeasure was the proportion of correct anticipatorylooks.

Task4:Gap-overlapThis task was presented in three blocks. The first two blocks lasted 20 trials; the thirdcontinueduntileitherenoughvalidtrialshadbeencollected(12usabletrialspercondition)or80trialshadbeenpresented,orthechildbecameinattentive.Afterfixatingacentraltarget(CT, a cartoon flower, 4.5˚ of visual angle from theperspective of the infant), following avariableInter-StimulusInterval(seebelow)alateraltarget(LT,acartooncloud,3˚ofvisualangle from the perspective of the infant) was presented to the left or right; when theparticipantfixatedtheLTtheyreceivedabriefanimationasanaudio-visualreward.Threeconditionswerepresented:Gap–CTdisappears200msbefore LTappears;Baseline–CTdisappears concurrently with LT appearance; Overlap – CT remains onscreen with LTappearance.Theorderoftrialswasrandomisedbetweenconditions.Thereactiontime(RT)wasthetimeelapsedbetweenLTappearanceandthereportedpositionofgazeleavingthecentralfixationarea(a9˚boxaroundtheCT).Disengagementlatencieswerecalculatedastheparticipant’saverage reaction time in theoverlapcondition subtracted fromtheiraveragereactiontimeinthebaselinecondition(followingElsabbaghetal.,2008).

30

Qualityofeye-trackingdataOneofthekey issuesfordeterminingtheeffectivenessofusingeye-trackingforscreeningpurposesinCCsistheextenttowhichthedataproducedisneartothequalityobtainedinalabsetting.Toevaluatethequalityofeye-trackingdataobtained,wecomparedthedataontheAVSItaskfromtheScreeningstudy,forwhichwehadalsocollecteddatainalaboratorysituationaspartofanearlierstudyandontheAVSItaskfromtheScreeningstudy,carriedoutinCCs.Thesameequipmentandexperimentalparadigmwereemployedinbothstudies(seeTomalskietal.,2013b;Ballieuxetal.,2015).

ComparisonofresultsinlaboratoryandCCsettings___________________________________________________________________________ ELASlaboratorystudy

CCScreeningstudy

Proportionofparticipantscompletingtask

84.2% 82.8%

Proportionofvalidtrialsperparticipant(SD)

94.4%(9.39)

88.8%(16.43)

Proportionoftimepointsatwhichtheeye-trackercollectedvalidgazedata(SD)

76.5%(16.39)

60.5%(21.61)

Comparingperformanceonthesameaudio-visualspeechintegration(AVSI)taskintwokindsofsettings,labversusCC,didnotleadtoanincreaseinparticipantdropout,nortoareductionintheproportionofvalidtrialsperparticipant.Bothdatasetswerecomparableinthisregard.However,onedifferencewasthattestinginCCsledtoalowerproportionoftimepointsatwhichtheeye-trackerrecordedvalidgazedata(theTobiiT120eye-trackerwassamplinggazepositionat120Hz;proportionofvalidsamplesM=60.5%versusELASstudyM=76.5%),andacorresponding increasedvariabilitywithin thesample (theELASstudySD=16.39, range26-97%;TALBYstudySD=21.61andrange0-99%).WeconcludethattestinginaCCsettingdoesnotsignificantlyalterinfantengagementwiththe eye-tracking tasks nor completion rates, but that variable conditions (e.g. light levels,noise, general distractions) and greater sample variability may somewhat reduce dataintegrity.

10. Screeningstudy-infantperformanceoneye-trackingtasks

Facepop-outeffectBasedonpreviouswork(discussedabove)wehypothesisedthat,forthemajorityofinfants,thefirstfixationwouldbemorelikelytobetowardstheface,butthatsomeinfantsmayshowdifferencesinthetotalamountoftimelookingattheface,mayshowreducedorincreasedfixationdurations,anddifferentvisitation-foragingpatternsthatcouldrevealsub-groupsofinfantsat risk for laterdifficulties.Thebasic facepop-outeffectwasclearlyevident inourcohort (see figure below) indicating that the tasks obtained the expected result whenconducted in CCs. In those slides where only objects were presented, we found a smallpreferenceforlookingtowardsanimageofaclock.

VisitdurationtoeachAreaofInterest(AOI)insecondsfora)objectonlyslidesandb)faceandobjectslides.Although all infants looked longer at the face than other objects in the arrays, the taskrevealedarangeofindividualdifferencesinmeasuresthatmaybeearlyindicatorsofpotentialproblems.Indicatorsofpotentialdifferencesaretheaveragedurationofeachfixationorvisittothefaceandthetotalnumberoffixationsandvisits.Previousstudieshavesuggestedthatthoseinfantswhomakemanyshorterorlongervisitsorfixationscouldbeatrisk.Smallsub-groupsofinfantscanbeidentifiedinoursample(seefigure,below)thatmadelongeraveragevisitsandthosewhomade largernumbersof fixations.These infantsmaybeatrisk inthelongerterm.

32

ScatterplotshowingtherelationshipbetweenaveragevisitdurationperAreaofInterest(AOI)andthenumberoffixationsperAOI.Thisfigureillustratesthevariabilityobservedacross

differentindividualinfantsthatwetested.

Audio-visualspeechintegrationConsistentwithpreviousstudies,wefoundthatinfantsshowedstrongpatternsoflookingtothemouthandtheeyesindynamicspeakingfacesacrossthedifferentconditionsoftheAVSItask.Wealsosuccessfullyreplicatedothereffectsobservedinpreviousstudies,suchasthefact that infants’ lookingbehaviordiscriminatedbetweencongruentofauditoryandvisualcueswhilstlookingatapersonspeakingandincongruentmatches,aphenomenonknownastheMcGurkeffect(McGurk&MacDonald,1976).

Eye-GazefollowingThe condition in the gaze following task proved tobedifficult for infants of this agewithinfantsfollowinggazeonlyaround25%oftrials(seefigure,below).Infantsweremarginallymore likely to follow a gaze if precededby eye contact thannot. This revealed individualdifferencesthatrelatetoSESandoutcomesatage2(seesection12,EyeGazeTask).

Proportionofvalidtrialsbyconditionforthegazefollowingtask.Bluedotsindicate

datapointsthatwereidentifiedasoutliersfollowingthestandardcriteria(asdescribedinBarnett&Lewis,1974)

34

11. Screeningstudy(cont.)-associationsoftasksperformancewithsocio-economicstatus

Facepop-putOuranalysesrevealedthatdifferencesarefoundinthepop-outtaskinperformancebySES(seefigures,below).High-SESgroupsshowingareducedrangeinproportionsoffirstlooksonface,andlowSESgroupsshowingawiderrangeofresponses.Thissuggeststhatthetaskmaybeusefulforidentifyingsub-groupswithinat-riskpopulationsoflower-SESgroups.Significanteffectsof father’seducationwerealsonoted.TheseresultssuggestthateffectsofSESarediscernible,evenbytheagerangeof6-7monthstestedinthestudy,oninfants’processingofsocialstimuli.

Proportionofinfants’firstlookstofacesubdividedbymaternalsocialclass(left)andfather’s

education(right)

Audio-visualspeechintegrationTheAVSItaskalsoidentifiedsmallassociationsbetweenSESmeasuresandtheratiooflookingtimeobservedforcongruousrelativetoincongruousvisualstimuli,forbothhouseholdincome(R2=-.15)andmaternaloccupationallevel(R2=-.18).Thissuggeststhat,evenbytheagerangeof 6-7 months tested in the study, there are small effects of SES discernible in infants’processingoflanguageandspeech.

12. Screeningstudy(cont.)-associationswithlanguageoutcomesintoddlers

Wehave,inaseparateprojectnototherwisedescribedinthisreport,followeduptheinfantswhoparticipatedintheTALBYprojectataseparateassessment,conductedwhentheyweretwoyearsofage.Thedatawehavecollectedhasidentifiedsignificantassociationsbetweeninfants’performanceontheeye-trackingtasksdescribedabove,administeredat6-7monthsof age, and children’s’ emerging language and communication outcomes, measured withPreschool LanguageScalesPLS-4 inventoryandCommunicativeDevelopmental Inventories(CDI) respectively at two years. While these effects are small, they do suggest that thisapproachmayoffersomeadditionaldiagnostictoolforthosemostatrisk.

AVSItaskA significant relationship was observed between infants’ ability to differentiate betweencongruentandincongruentauditoryandvisualcueswhilstlookingatapersonspeakingat6-7monthsandtheirperformanceonPre-schoolLanguageScalesatage2(R2=0.29,p=.03)(seefigure).Thissuggeststhataspectsofinfants’earlyabilitytoprocessspeechinformationduringinfancycanpredicttheirlaterlanguageoutcomes.

Scatterplotshowingtherelationshipbetweenprocessingofspeechsoundsat6/7months,as

assessedintheTALBYstudy,andchildren’s’laterlanguageperformanceat2years.

Eye-GazefollowingtaskTheeye-gazefollowingtask,administeredaspartoftheTALBYprojectat6-7months,alsosignificantlypredictedinfants’subsequentlanguageoutcomeswhenassessedatafollow-upassessmentattwoyears(seefigure,below).Infants’likelihoodoffollowingthegazeoftheactor during infancy correlated with their productive vocabulary on the Communicative

36

Development Inventory as assessed at two years (R2=0.30, p=.03). Infants’ likelihood offollowinggazeafternoeye-contactalsopredictedchildren’s’subsequentscoresonthePre-SchoolLanguagescales(R2=0.29,p=.05).Theseresultsfurtherconfirmourconclusionsthatthe CC assessments administered to infants as part of the TALBY project can successfullypredictchildren’ssubsequentlanguagedevelopmentoutcomes.

Scatterplotillustratingtherelationshipobservedbetweeninfants’Eye-Gazefollowing

observedat6/7monthsandlaterproductivelanguagedevelopment

Scatterplotillustratingafurtherrelationshipobservedbetweeninfants’Eye-Gazefollowing

observedat6/7monthsandlaterlanguagedevelopment

13. Trainingstudy-effectivenessoftrainingtasks

Within-taskimprovementsduringthetrainingphaseFirst,weassessedwhetherchangesinperformanceonthetrainingtaskswereobservedacrossthe four training sessions conducted. The figure below demonstrates how training taskperformance changed across the four visits in the current study. In order to assess thesignificance of these changes, linear regression lines were calculated based on change inperformanceacrossthefourtrainingsessions.Thegradientsoftheselineswerepositiveforeightoutofteninfants,suggestingthattheyimprovedacrossthetrainingsessions.At-testanalysissuggestedthatthisresultdifferedsignificantlyfromchance(t(9)=2.25,p=.025,one-tailed).Thischangewasconsistentwiththatpredicted,basedonpreviousresearch(Wass,etal.,2011).Thisfindingestablishedameasurableeffectoftraining,andsuggeststhatinfants’attentioncanbetrainingwithinthecontextofCCvisitswithinfantsfromthisdemographicpopulation.

Changeinperformanceacrossthefourtrainingvisits.Theperformanceateachtrainingtaskwascomparedtothatindividual’saverageperformanceacrosstheentiretrainingphasebycalculatingaz-score.Inthisway,weidentifiedsignificantimprovementsintaskperformance

acrossthefourvisits.

38

Impactoftrainingonpre-andpost-testsInordertoexaminewhetherthetrainingledtoimprovementsinotheraspectsofattentioncontrolthathadnotbeendirectlytrained,anumberofotherassessmentsofinfants’capacityfor voluntary, self-directed control of attention were administered before and after theintervention phase to all participating infants (see section 9, above). The results of theseassessmentsarepresentedbelow.

Bar charts show change (Δ) scores on pre-post assessments, calculated from themarginalmeans.a)Sustainedattention.b)Gap-overlaptask.(Becausethevalenceofthepredictedandobserved change was negative, -Δ scores are presented for this measure, for ease ofcomparison.) c) Cognitive control task. d) Short-term memory task. Stars indicate thesignificanceofthetrainingeffectsobserved.*-p<.05,(*)–p<.10.Wefoundthat,aftertraining,thetrainedgroupshowedasignificantincreaserelativetothecontrolgroupinsustainedattentionto'Interesting'targets(seeFigurea),above).However,weobservednochangeaftertraininginlookingtimeto'Boring'targets.Thissuggeststhattheeffectoftrainingwasnotsimplyanoverallincreaseinlookingtimetothescreen.Rather,itis

consistent with a model suggesting that training attention leads to increases in infants’capacitytoexercisevoluntary,selectivecontroloverthewhattheypayattentionto.Wealsofoundthataveragereactiontimeonthegap-overlaptaskwassignificantlyimprovedfollowingtraining(seeFigureb),above).Bycontrast,nochangeswerefoundondisengagingvisualattention.However,thismayhavebeentodowiththehighratesofdatadrop-outonthis task,dueto loweye-trackingdataquality.Ofnote,substantially longerreactiontimeswereobtainedonthistaskcomparedtolab-basedversionsofthesametask,apatternthatispredictedbylowerdataquality(Leppänen,etal.,2014;Wass.,etal.,2014).We also observed improvements on a task assessing anticipatory saccades during rule-learning(seeFigurec).Changesinperformancewereobservedduringtheinitialrulelearningphaseonly;thiswasslightlyunexpected,aspreviousfindings(Wassetal.,2011)hadsuggestedthat training improvements would be observed on both phases of the task. Finally, weobservedtrend-level improvementsona taskassessingshort-termmemory (seeFigured).Thisisincontrasttothepreviousstudy(Wass,2011),whichusedadifferentassessment,inwhichnoimprovementswerefound.Inseparateinvestigations(Wass,2011)wefoundtest-retestreliabilityofthetaskthatwepreviouslyusedtobeverylow,anditmaybethatmemorytrainingeffectscanbedetectedwhenamoresensitivemeasurementparadigmisused.In conclusion, these results areencouraging, anddemonstrate that early attention canbetrainedinhigh-riskinfantsfromlowSESbackgrounds,workingwithCCsettings.Theresultsobtainedareencouraginglysimilartothoseobtainedinpreviouslab-basedstudies,suggestingthereplicabilityofthiswork.

40

14. Conclusions

Usingeye-trackingtoscreeninfantsinCCsettingsOurkeyfindingsfromtheScreeningstudywereasfollows:

• CCs offer an effective context for recruiting and assessing large and highly diversesamplesofinfantsfromareasofhighdeprivation.

• CC staff are receptive to learning to use eye-tracking equipment and delivering

assessments,butonlyifstaffinglevelsandtimearesufficient.

• Playing back a video of an infant’s looking pattern is an effectiveway of engagingparentswiththeassessmentoftheirinfants,andofopeningupadiscussionabouttheimportanceofattentionininfants’development.

• Thesessionswereeffectiveinchanginghowparentsreportedwhattheyunderstand

andthoughtabouttheirbabies.Thiswasmosteffectiveforparentswithlowersocio-economicoccupations.

• The eye-tracking technique proved to be convenient and effective in collecting

researchdatainCCs.Wewereabletocollectdatathatisofreasonablequalityandcomparabletodataobtainedinlaboratorysettings.

• Our eye-tracking assessments provided some indicators that differences in infants’

processingofvisualandauditoryinformationcontingentonSESarealreadydiscernibleby6/7monthsofage.However,thesizeofthesedifferenceswasrelativelysmall.

• Inaddition,weconductedaseparateassessment,inwhichwefolloweduptheinfants

assessedintheTALBYstudywhentheywere2yearsofage.Ouranalysesidentifiedanumberofsignificantassociationsbetweeninfants’performanceontheeye-trackingtasksat6/7monthsandtheirlanguageoutcomesat2years.ThissuggeststhattheCCassessments that we administered can successfully predict children’s subsequentlanguagedevelopmentoutcomes.

Asaresultofthesefindings,wecanconcludethateye-trackingoffersafeasibleandeffectivemethod formeasuringearly infantcognitivedevelopmentwithincommunitysettings.Eye-tracking research has considerable potential for use as an early screening tool to identifyinfantsatelevatedriskofsubsequentatypicaldevelopment.Particularlyencouragingarethefindingsfromourfollow-upstudywhichsuggestedthatinfants’performanceontheseearlyinfant eye-trackingmeasures significantly predicted later language outcomes during earlychildhood.

From a policy perspective, early identification of infants who are developing atypically isconsideredvital,insofarasitallowsforinterventionstobeadministeredatasearlystageaspossible. Early interventions are widely considered to be a more cost-effective way ofremediating problems than interventions administered later in development (Heckman,2006). Investigating inearly screeningand interventionmay, therefore,offeramorecost-effectivewayofremediatingproblemsthanwaitinguntilproblemshavedevelopedgreaterseveritylateron,andattemptingtoremediatethemthen.Otherresearchhasalsosuggestedthat this type of early screening work should be concentrated on infants from low-SESbackgrounds,forwhomtheriskofatypicaldevelopmentintermsbothofacademicandlong-termmentalhealthoutcomesismostsevere(Businelle,2013).

Usingeye-trackingtotraininfants’attentioninCCsettingsOurkeyfindingsfromtheTrainingstudywereasfollows:

• Eye-tracking canbeused to train infants’ attentionwithinCC settings,with infantsfromlow-SESbackgrounds.

• Thetrainingeffectsobservedwereconsistentwiththoseobservedonprevious,lab-

basedstudieswithhigh-SESinfants.

• However,participantdrop-outwasaproblem,with10ofthe33familiesrecruitedforthestudyfailingtocomplete.Futureworkshouldinvestigatewhetheritispossibletomakethisinterventionlessonerousforparticipatingparents,suchasbyconductingtrainingsessionswithinparticipants’homesasacomplementtoCCvisits.

Overall,our resultsprovidedpromisingevidence for the feasibilityandefficacyofanearlyinterventionthattargetsinfants’capacitytoexercisevoluntarycontroloverthefocusoftheirattention.Infuture,thisworkoffersapotentialnewearlyinterventiontooltotargetinfantsidentified as particularly high-risk via early screening. These twomethods can be used tocomplement one another, to provide more sensitive and cost-effective means forunderstandingearlydevelopmentininfantsfromlow-SESbackgrounds,whoareknowntobeatthemostelevatedriskofadverselong-termoutcomes.

42

15. Publicationsandpresentationsfromthestudytodate

Ballieux, H., Tomalski, P., Kushnerenko, E., Karmiloff-Smith, A., Johnson, M.H., &Moore, D.G.(2013).Eye-tracking in theCommunity:RelationsBetween InfantAttentiontoSocialCuesand their Socio-economic, Ethnic, and Language Background. Poster presented at theBiennialmeetingoftheSocietyforResearchinChildDevelopment,Seattle,US.

Ballieux,H.,Tomalski,P.,Kushnerneko,E.,Johnson,M.H.,Karmiloff-Smith,A.,andMoore,D.G.(2015), Feasibility of Undertaking Off-Site Infant Eye-Tracking Assessments of Neuro-CognitiveFunctioninginEarly-InterventionCentres.InfantandChildDevelopment 25(1),95-113doi:10.1002/icd.1914.

Ballieux, H.,Wass, S. V., Tomalski, P., Kushnerenko, E., Karmiloff-Smith, A., Johnson,M. H., &Moore,D.G.(2016).Applyinggaze-contingenttrainingwithincommunitysettingstoinfantsfromdiverseSESbackgrounds.JournalofAppliedDevelopmentalPsychology,43,8-17.

Ballieux, H., Tomalski, P, Karmiloff-Smith, A., Johnson,M., Kushnerenko,E. & Moore, D.G.(manuscriptprepared) Individualdifferences in6-7month-old infants’facepreferencesaspotentialpredictorsofemerginglanguageandsocialcommunicationoutcomesat2years.

Haiko Ballieux, Derek Moore, Annette Karmiloff-Smith, Mark Johnson, Deirdre Birtles, ElenaKushnerenko,PrzemyslawTomalski(2014) IndividualDifferences inSocialOrienting in6-month-oldsasaLongitudinalPredictorofLanguageandSocialCommunicationOutcomesatAgeTwo,andTheirRelationshiptoSESInternationalConferenceonInfantStudies,Berlin.

Moore,D.G.Tomalski,P.Ballieux,H.Birtles,D.Johnson,M.H.Karmiloff-Smith,A.&Kushnerenko,E.(2014)Audio-visualSpeechIntegrationatSixMonthsasaPredictorofLanguageOutcomesatAgeTwoinInfantsfromFamiliesVaryinginIncome,Occupation,EthnicityandLanguage.Paperpresentedat:InternationalConferenceonInfantStudies,Berlin.

MooreD.G.,KushnerenkoE.,BallieuxH.,Karmiloff-SmithA.,JohnsonM.H.&TomalskiP.(2015)Adaptingeye-trackingparadigmstoassessSES-related individualdifferences insix-month-oldinfants,SocietyforResearchinChildDevelopmentBiennialMeeting.Philadelphia,USA.

Tomalski,P.,Moore,D.G.,Ballieux,H.,Kushnerenko,E.V.,Johnson,M.H.,&Karmiloff-Smith,A.(2016).Separatingtheeffectsofethnicityandsocio-economicstatusonsleeppracticesof6-to7-month-oldinfants.LearningandIndividualDifferences46(2016)64–69

TomalskiP.,BallieuxH.,Karmiloff-SmithA.,JohnsonM.H.,KushnerenkoE.,MooreD.G.(2015).Earlypredictorsoflanguagedevelopmentininfantsfromlow-SESfamilies:amulti-ethniceye-trackingstudyinthecommunity.SocietyforResearchinChildDevelopmentBiennialMeeting.Philadelphia,USA.

16. References

Aldridge,H.;Bushe,S.;Kenway,P.;MacInnes,T.;Tinson,A. London’spovertyprofile.PublishedbyTrust for London & New Policy Institute; 2013. ISBN: 978-1-901373-75-4.www.londonspovertyprofile.org.uk/LPP_2013_Report_Web.pdf

Barnett,Vic,andTobyLewis.Outliersinstatisticaldata.Wiley,1974.

Bayer,J.K.,Hiscock,H.,Hampton,A.,&Wake,M.(2007).Sleepproblemsinyounginfantsandmaternalmentalandphysicalhealth.JournalofPaediatricsandChildHealth,43(1-2),66–73.doi:10.1111/j.1440-1754.2007.01005.x

Bedford, R., Elsabbagh, M., Gliga, T., Pickles, A., Senju, A., Charman, T., … BASIS team. (2012).Precursorstosocialandcommunicationdifficulties in infantsat-riskforautism:gazefollowingandattentionalengagement.JournalofAutismandDevelopmentalDisorders,42(10),2208–18.http://doi.org/10.1007/s10803-012-1450-y

Businelle,M.S.,Mills,B.A.,Chartier,K.G.,Kendzor,D.E.,Reingle,J.M.,&Shuval,K.(2013).Dostressfuleventsaccountforthelinkbetweensocioeconomicstatusandmentalhealth?JournalofPublicHealth,36(2),205-212.

DepartmentforCommunitiesandLocalGovernment.Englishindicesofdeprivation.2010.ISBN:978-1-4098-2924-9.www.gov.uk/government/publications/english-indices-of-deprivation-2010

Gervain,J.,Mehler,J.,Werker,J.F.,Nelson,C.a,Csibra,G.,Lloyd-Fox,S.,…Aslin,R.N.(2011).Near-infraredspectroscopy:areportfromtheMcDonnellinfantmethodologyconsortium.DCN,1(1),22–46.http://doi.org/10.1016/j.dcn.2010.07.004

Guiraud, J. a, Tomalski, P., Kushnerenko, E., Ribeiro,H.,Davies, K., Charman, T.,… Johnson,M.H.(2012). Atypical audiovisual speech integration in infants at risk for autism. PloS One, 7(5),e36428.http://doi.org/10.1371/journal.pone.0036428

Hackman, D. A., Farah, M. J., &Meaney, M. J. C.-P. (2010). Socioeconomic status and the brain:mechanistic insights from human and animal research. Nat Rev Neurosci, 11, 651–9 ST –Socioeconomicstatusandthebrain:me.http://doi.org/nrn2897[pii]10.1038/nrn2897

Heckman,J.J.(2006).Skillformationandtheeconomicsofinvestingindisadvantagedchildren.Science,312(5782),1900-1902.

Kushnerenko,E.,Teinonen,T.,Volein,A.,&Csibra,G.(2008).Electrophysiologicalevidenceofillusoryaudiovisual speech percept in human infants. Proc Natl Acad Sci U S A, 105(32), 11442–5.http://doi.org/10.1073/pnas.0804275105

Kushnerenko,E.,Tomalski,P.,Ballieux,H.,Potton,A.,Birtles,D.,Frostick,C.,&Moore,D.G.(2013a).Brain responses and looking behavior during audiovisual speech integration in infants predictauditory speech comprehension in the second year of life. Frontiers in Psychology, 4,10.3389/fpsyg.2013.00432.http://doi.org/10.3389/fpsyg.2013.00432

Kushnerenko, E., Tomalski, P., Ballieux, H., Ribeiro, H., Potton, A., Axelsson, E. L., …Moore, D. G.(2013b).Brainresponsestoaudiovisualspeechmismatchininfantsareassociatedwithindividualdifferences in looking behaviour. Eur J Neurosci, doi:10.1111/ejn.12317.http://doi.org/10.1111/ejn.12317

44

Lewkowicz,D.J.,&Hansen-Tift,A.M.C.-P.(2012).Infantsdeployselectiveattentiontothemouthofa talking face when learning speech. Proc Natl Acad Sci U S A, 109(5), 1431–6.http://doi.org/1114783109[pii]10.1073/pnas.1114783109

Noble, K. G.,McCandliss, B. D., & Farah,M. J. (2007). Socioeconomic gradients predict individualdifferencesinneurocognitiveabilities.DevSci,10,464–80ST–Socioeconomicgradientspredictindiv.http://doi.org/DESC600[pii]10.1111/j.1467-7687.2007.00600.x

Noble,K.G.,Norman,M.F.,&Farah,M.J.(2005).Neurocognitivecorrelatesofsocioeconomicstatusin kindergarten children. Dev Sci, 8, 74–87 ST – Neurocognitive correlates of socioecon.http://doi.org/DESC394[pii]10.1111/j.1467-7687.2005.00394.x

Mattock,K.,Molnar,M.,Polka,L.,&Burnham,D.(2008).Thedevelopmentalcourseof lexicaltoneperception in the first year of life. Cognition, 106(3), 1367–81.http://doi.org/10.1016/j.cognition.2007.07.002

McGurk,H.,&MacDonald,J.(1976).Hearinglipsandseeingvoices.Nature,264,746–748.

Polka,L.,Rvachew,S.,&Molnar,M.(2008).SpeechPerceptionby6-to8-Month-OldsinthePresenceofDistractingSounds.Infancy,13(5),421–439.http://doi.org/10.1080/15250000802329297

Tomalski,P.,Moore,D.G.,Ribeiro,H.,Axelsson,E.L.,Murphy,El.,Karmiloff-Smith,A.,…Kushnerenko,E. (2013a). Socio-economic status and functional brain development – associations in earlyinfancy.DevSci,16(5),676–87.http://doi.org/10.1111/desc.12079

Tomalski, P., Ribeiro,H., Ballieux,H., Axelsson, E. L.,Murphy, E.,Moore,D.G.,&Kushnerenko, E.(2013b).Exploringearlydevelopmentalchangesinfacescanningpatternsduringtheperceptionofaudiovisualmismatchofspeechcues.EuropeanJournalofDevelopmentalPsychology,10(5).http://doi.org/10.1080/17405629.2012.728076

Tomalski, P., Ribeiro,H., Ballieux,H., Axelsson, E. L.,Murphy, E.,Moore,D.G.,&Kushnerenko, E.(2012).Exploringearlydevelopmentalchangesinfacescanningpatternsduringtheperceptionofaudiovisualmismatchofspeechcues.EurJDevPsychol,DOI:10.1080/17405629.2012.728076.http://doi.org/10.1080/17405629.2012.728076

Tomalski, P.,Moore, D. G., Ballieux, H., Kushnerenko, E. V., Johnson,M. H., & Karmiloff-Smith, A.(2016).Separatingtheeffectsofethnicityandsocio-economicstatusonsleeppracticesof6-to7-month-oldinfants.LearningandIndividualDifferences46(2016)64–69

TomalskiP.,BallieuxH.,Karmiloff-SmithA.,JohnsonM.H.,KushnerenkoE.,MooreD.G.(2015).Earlypredictorsoflanguagedevelopmentininfantsfromlow-SESfamilies:amulti-ethniceye-trackingstudyinthecommunity.SocietyforResearchinChildDevelopmentBiennialMeeting.Philadelfia,USA.

UK Office for National Statistics. (2010). Standard Occupational Classification 2010. Volume3: TheNationalStatisticsSocioeconomicClassificationUserManual.Basingstoke:PalgraveMacmillan.

Wass,S.,Porayska-Pomsta,K.,&Johnson,M.H.(2011).Trainingattentionalcontrolininfancy.CurrentBiology :CB,21(18),1543–7.http://doi.org/10.1016/j.cub.2011.08.004

Wass, S. V., & Smith, T. J. (2014). Individual differences in infant oculomotor behavior during theviewingofcomplexnaturalisticscenes.Infancy,19(4),352-384.

Zimmerman,I.,Steiner,V.,&Pond,R.(2002).PreschoolLanguageScale(4thed.).SanAntonio,TX:ThePsychologicalCorporation.