Developmental social neuroscience meets public health challenge: A new system of healthcare delivery for infants and toddlers with autism spectrum disorder National Autism Conference, State College, PA August 1, 17 Ami Klin, PhD Director,, Children s Healthcare of Atlanta Georgia Research Alliance Eminent Scholar Professor & Chief, Division of Autism, Department of Pediatrics, Emory University School of Medicine Emory Center for Translational Social Neuroscience Thank You The children and families who support our clinical and research activities Warren Jones and many wonderful colleagues and students The National Institute of Mental Health The Marcus Foundation The JB Whitehead and Woodruff Foundations The Children s Healthcare of Atlanta Foundation The Georgia Research Alliance as well as The National Institute of Child Health and Human Development The Autism Science Foundation Conflicts of Interest No conflicts of interest associated with this presentation 3 4
Translation Impact Clinical Resources at a glance Excellence Science Faculty Advancement Research Resources >, unique patients/yr >3, in the community Tx: set protocols (x visits) > 6% on Medicaid ~ 3% minorities/under-served Clinical Assessment/Diagnosis Treatment Programs Center/Home/School/Community Care Coordination Program Educational Outreach Program Autism and other developmental delays are a Public Health Challenge Prevalence: 1:68 autism; 1:1 developmental delays Autism Societal Cost/Year in the US: $ 136 billion Autism Lifetime Cost of Care Per Child: $.4-3.6 million Importance of early diagnosis and intervention for lifelong outcome and cost of care American Academy of Pediatrics recommends screening for autism at 18 and 4 months Autism Median age of diagnosis in US: 4-6 to.7 years % of primary care providers who routinely screen not known CENTER-BASED MODEL PROGRAM COMMUNITY-VIABLE OUTREACH MODEL CLINICAL TRAINING RESEARCH ADVOCACY <% of children identified before age 3 years The Science of Clinical Care CDC, 14; Peacock et al., 1; Cidav et al., 1; Mandell et al., 1; 9; 13; 14; Wang et al., 13; Buescher et al., 14; Wiggins et al., 6; Shattuck et al., 9; Honigfeld et al., 1; Heidgerken et al., ; Dosreis et al., 6; Johnson & Myers, 7; is an affiliate of Children s Healthcare of Atlanta. 1 Children s Healthcare of Atlanta Inc. All rights reserved. 6 ASD symptoms RESULT from deviations from normative socialization Sociality: the evolutionary roots of our social brain GENETIC MECHANISMS OF SOCIALIZATION BEHAVIORAL LIABILITY SYMPTOMS CONCEPTS Jones et al. (8). Arch Gen Psy, 6(8), 946-4; Klin et al. (9). Nature, 49, 7-61; Jones & Klin (9). J Am Acad of Child Psy, 48(): 471-3; Jones & Klin (13). Nature, 4, 47-431; Klin et al. (14). Neurosci Biobehav Rev. 8
Universal Principle: the Platform for Development of Social Brain Born to Socially Orient The co-creation of social experience Reciprocal Social Interaction MH Johnson PhD Neuroplasticity WHITE MATTER DEVELOPMENT H-J Park PhD 9 1 Anthropomorphizing
Anthropomorphizing
Attention to Biological Motion not significantly different from chance, p Non-verbal mental-age Verbal mental-age match Klin A & Jones matched W. (8). control Dev Science, 1: 4-46. Attention to Biological Motion Klin A & Jones W. (8). Dev Science, 1: 4-46.
Preference in Autism Only When Clapping Happens Typically- Developing Children Developmentally- Delayed Children Children with Autism Typically- Developing Children Developmentally- Delayed Children Children with Autism Klin, Lin, Gorrindo, Ramsay, & Jones, Nature, 9. Klin, Lin, Gorrindo, Ramsay, & Jones, Nature, 9. Physical, rather than social, cues guide looking in toddlers with autism Audiovisual Synchrony Quantification Change in Motion * Change in Sound = Audiovisual Synchrony CHANGE IN MOTION CHANGE IN SOUND AUDIOVISUAL SYNCHRONY Klin, Lin, Gorrindo, Ramsay, & Jones, Nature, 9. Time
Cumulative Audiovisual Synchrony Relative Audio-Visual Synchrony = Normalized Peak Difference Clap Location Max Synchrony Pat-a-cake Inverted Upright Feeding Inverted Upright No Synchrony
Patterns of visual fixation to approaching caregiver How do -year-olds with autism watch the face of a caregiver? Eye: F,63 = 1.87, p<.1 d = 1.6 Mouth: F,63 =.99, p<.6 d = 1.4 Jones, Carr, Klin (8). Archives of General Psychiatry. 6(8):946-4. Jones, Carr, Klin (8). Arch Gen Psychiatry. 6(8):946-4. Watching a Face But Seeing Physical Contingencies? Looking at Eyes and Mouth As a Function of Audiovisual Synchrony TD ASD Caregiver Audiovisual Synchrony
Social Interaction is the Platform for Brain Development Brain size doubles in the 1st year of a baby s life, synaptic density quadruples. (Gilmore et al, 7; Pfefferbaum et al, 1994; Huttenlocher, 1979; Petanjek et al, 11) 33
Derivation of Attentional Funnel
Objective, Quantitative Measures The majority of typically-developing -year-olds fixate on the same locations, at the same moments, during 8% of viewing time. Experimental Presses
The accrual of missed opportunities for social learning Scenes of Social Action % of Total TD Permutations TD, N=44 ASD, N= Divergence Expected by Chance (permutation testing) Number of Divergences Observed Typically- Developing Toddlers with Autism Scenes of Social Interaction Scenes of Social Interaction Toddlers with Autism Toddlers with Autism Typically- Developing Typically- Developing
8 Time 1 Time Probability Ratio, % 6 4 Probability of looking at the same locations at the same times as typical control children is significantly correlated with levels of autistic social disability, both contemporaneously and 1. years after initial testing. Quantitative Indices for Assessing Presence of ASD Presence of ASD: Diagnostic Accuracy All ages - training reference + - N=36 N=18 training N=18 testing True Positive Fraction 1.8.6.4. Entire Sample Training AUC =.94..4.6.8 1 + 87 11 98 device - 1 11 1 97 11 18 estimate lower upper 89.7% 81.9% 94.% se sp 9.9% 84.3% 9.% ppv 88.8% 8.8% 93.8% npv 91.7% 8.% 9.6% accuracy 9.4% 8.7% 93.7% kappa.8.736.891 TD normative funnels = Empirical ROC curve Fitted ROC curve +/- 9% CI True Positive Fraction 1.8.6.4. False Positive Fraction Entire Sample Testing AUC =.871..4.6.8 1 False Positive Fraction All ages - external validation reference + - + 37 1 49 device - 9 9 46 6 18 estimate lower upper 8.4% 66.6% 89.6% se sp 8.6% 69.% 88.7% ppv 7.% 61.8% 8.% npv 84.7% 73.3% 9.% accuracy 8.6% 7.% 87.% kappa.66.4.77 ASD comparison scanpaths =
Quantitative Indices for Assessing Severity of ASD TD normative funnels = ASD comparison scanpaths = INFANT SIBLING STUDY Translational Opportunities A National Institutes of Health Autism Center of Excellence 4-Mon 18-Month-Olds 4-Month-Olds 18-Month-Olds 4-Month-Olds 1 1 1 6 r = -.6 (-.7, -.769), p =.11 1 r = -. (-.11, -.7), p =.18 r = -.73 (-.444, -.886), p =.1 1 4 CCC = -.619 (-.444, -.749), p <.1 CCC = -.47 (-.319, -.71), p <.1 CCC = -.776 (-.637, -.866), p <.1 CCC = -.9 Social (-.9, Ability Index -.747), p =.6 CCC = -.436 (-.39, -.714), p =.334 CCC = -.63 (-.6, -.831), p =.6 ADOS Severity: r = -.13 (-.77, -.691), p =.1 r = -. (-.11, -.7), p =.18 ADOS Severity: r = -.79 (-.67, -.87), p <.1 Assessment of Cognitive Functioning CCC = -.47 (-.319, -.71), p <.1 r = -.717 (-.413, -.877), p =.3 CCC = -.436 (-.39, -.714), p =.334 CCC = -.7 (-.7, -.836), p <.1 CCC = -.94 (-.18, -.816), p =.4 Mullen Non-Verbal: r =.794 (.683,.87), p <.1 r =.34 (-.16,.6), p =.613 Assessment of Cognitive Functioning CCC =.733 (.96,.89), p <.1 3 CCC =.39 (.41, 1 ADOS Social Affect: r = -.686 (-.3, -.796), p <.1.64), p =.3 r = -.61 (-.3, -.79), p <.1 Mullen Receptive: r =.76 (.69, 1.849), p <.1 CCC = -.61 (-.446, -.7), p <.1 3 3 Mullen Non-Verbal: r =.794 (.683,.87), p <.1 r =.34 (-.16,.6), p =.613 r =.1 (.197,.74), p =.3 CCC = -.91 (-.93, -.784), p =.6 1 CCC =.733 (.96,.89), p <.1 8 CCC =.68 (.19,.797), p <.1 Severity of ASD: ADOS Total: r = -.673 (-.14, -.787), p <.1 CCC =.39 (.41,.64), p =.3 CCC =.447 (.13,.69), p =.133 1 1 6 r = -.6 (-.7, -.769), p =.11 1 Mullen Receptive: r =.76 (.69,.849), p <.1 1 Mullen Expressive: r =.744 (.6, 4.838), p <.1 CCC = -.619 (-.444, -.749), p <.1 1 r =.1 (.197,.74), p =.3 1 r =.73 (.68,.773), p =.9 CCC = -.9 (-.9, -.747), p =.6 1 Prognostic Indicators CCC =.68 (.19,.797), p <.1 CCC =.78 (.84,.88), p <.1 ADOS Severity: r = -.13 (-.77, -.691), p =.1 CCC =.447 (.13,.69), p =.133 CCC =.9 (.8,.747), p =.6 1 r = -. (-.11, -.7), p =.18 Mullen Expressive: r =.744 (.6,.838), p <.1 CCC = -.47 (-.319, -.71), p <.1 Social Ability r =.73 Index (.68,.773), p =.9 CCC = -.436 (-.39, -.714), p =.334 3 4 Mullen Non-Verbal: r =.79 (.4,.74), p <.1 CCC =.78 (.84,.88), p <.1 4 r =.9 (-.18,.), p =.643 Non Verbal Ability Index Receptive Language Ability Index Expressive Language Ability Index Assessment of CCC =.9 (.8,.747), p =.6 Assessment of Verbal Ability: 3 3 3 CCC =.39 (.371,.674), p <.1 Assessment Social Assessment of Social Disability Functioning of Cognitive Functioning Receptive CCC & = Expressive.81 (-.9,.4), Language p =.71 Mullen Receptive: r =.633 (.496,.739), p <.1 3 4 Mullen Non-Verbal: r =.79 (.4,.74), p <.1 4 ADOS Social Affect: r = -.819 (-.71, -.893), p <.1 r =.9 (-.18, 3.), 3 p =.643 r =.487 (.6, 3 Mullen Non-Verbal: r =.794 (.683,.87), p <.1.683), p =.7 r = -.717 (-.413, -.877), p =.3 1 1 CCC = 1 r =.34 (-.16,.6), p =.613 3.611 (.469,.73), p <.1 3 1 3 CCC =.39 (.371,.674), p <.1 CCC = -.784 (-.649, -.871), p <.1 CCC =.81 (-.9,.4), p =.71 CCC =.733 (.96,.89), p <.1 CCC =.468 (.,.67), p =.1 1 1 CCC = -.636 (-.81, -.838), p =. CCC =.39 (.41,.64), p =.3 Mullen Receptive: r =.633 (.496,.739), p <.1 Mullen Expressive: r =.683 (.6,.777), p <.1 ADOS Total: r = -.8 (-.68, -.884), p <.1 1 r =.487 (.6,.683), p =.7 r =.1 (.43, 1 Mullen Receptive: r =.76 (.69,.849), p <.1.69), p =. 1 1 r = -.73 (-.444, -.886), p =.1 1 1 CCC =.611 (.469,.73), 1 p <.1 r =.1 (.197,.74), p =.3 1 CCC =.66 (.33, 1.761), p <.1 CCC = -.776 (-.637, -.866), p <.1 Non Verbal Ability Index Receptive Language Ability Index CCC Expressive =.468 Language (.,.67), Ability Index p =.1 CCC =.68 (.19,.797), p <.1 CCC =.48 (.17,.678), p =.8 CCC = -.63 (-.6, CCC =.447 -.831), (.13, p =.6.69), p =.133 1 Mullen 1 Expressive: r =.683 (.6,.777), p <.1 ADOS Severity: Mullen r = Expressive: -.79 (-.67, r = -.87),.744 (.6, p <.1.838), p <.1 r =.1 (.43,.69), p =. Positive Controls CCC =.66 (.33,.761), p <.1 r = -.717 (-.413, r =.73 -.877), (.68, p =.3.773), p =.9 Non Verbal Ability Index Receptive Language Ability Index Expressive Language Ability Index CCC =.48 (.17,.678), p =.8 CCC = -.7 (-.7, CCC =.78 -.836), (.84, p <.1.88), p <.1 Ability Non Verbal Index Ability Index Ability Receptive Index Language Ability Index Ability Expressive IndexLanguage Ability Index CCC = -.94 (-.18, CCC =.9 -.816), (.8, p =.4.747), p =.6 Social Disability Index Social Disability Index Social Disability Age: r =.97 (-.11,.96), p =.388 Positive Controls r =.83 (-.19,.371), p =.917 3 CCC =.117 (-.9,.314), p =.67 3 1 4 ADOS Social Affect: Mullen r = Non-Verbal: -.686 (-.3, r = -.796),.79 (.4, p <.1.74), p <.1 4 CCC =.43 (-.7,.33), p =.787 r = -.61 (-.3, r =.9 -.79), (-.18, p <.1.), p =.643 3 1 CCC = -.61 (-.446, -.7), p <.1 3 3 CCC =.39 (.371,.674), p <.1 Age: r =.97 (-.11,.96), p =.388 Quality: r =.3 (-.,.391), p =. CCC = -.91 (-.93, CCC =.81 -.784), (-.9, p =.6.4), p =.71 1 r =.83 (-.19,.371), p =.917 r =.81 8 (-.1,.369), p =.997 4 CCC =.17 (-., CCC =.117 (-.9,.314), p =.67.3), p =.136 ADOS Total: Mullen r = Receptive: -.673 (-.14, r = -.787),.633 (.496, p <.1.739), p <.1 1 CCC =.43 (-.7, 1.33), p =.787 CCC = -.38 6 (-.33,.61), p =.84 r = -.6 (-.7, r =.487 -.769), (.6, p =.11.683), p =.7 1 1 1 1 4 CCC = -.619 (-.444, CCC =.611 -.749), (.469, p <.1.73), p <.1 18 3 Quality: r =.3 (-.,.391), p =. CCC = -.9 (-.9, CCC =.468 -.747), (., p =.6.67), p =.1 r =.81 (-.1,.369), p =.997 1 16 CCC =.17 (-.,.3), p =.136 ADOS Severity: Mullen r = Expressive: -.13 (-.77, r = -.691),.683 (.6, p =.1.777), p <.1 CCC = -.38 (-.33,.61), p =.84 r = -. (-.11, r =.1 -.7), (.43, p =.18.69), p =. 14 CCC = -.47 (-.319, CCC =.66 -.71), (.33, p <.1.761), p <.1 Social Disability Ability Non Verbal Index Ability Index Index Social Disability Ability Receptive Age: IndexLanguage Ability Index r = -.63 Social (-.37, Disability Ability Expressive.114), IndexLanguage Ability Index p =.4843 CCC = -.436 (-.39, CCC =.48 -.714), (.17, p =.334.678), p =.8 Social 3 Ability Index r = -.99 (-.34,.17), p =.4487 Training Sample Positive Controls CCC = -.96 (-.68,.8), p =.88 4 Assessment of Cognitive Functioning Testing Sample CCC = -.73 (-.319,.18), p =.7 Age: r = -.63 (-.37,.114), p =.4843 Quality: r =.1 (-.7,.318), p =.96 3 3 r = -.99 (-.34,.17), p =.4487 r =.179 (-.77,.41), p =.1681 Training 3 3 Mullen Non-Verbal: r =.794 (.683,.87), p <.1 Training Sample Regression testing set: mean rados Total = -.6, Age: p <.1 r =.97 (-.11,.96), p =.388 CCC = -.96 (-.68,.8), p =.88 CCC =.169 (-.8,.33), p =.61 r =.34 (-.16,.6), p =.613 r =.83 (-.19,.371), p =.917 4 Testing Sample CCC = -.73 Testing (-.319, Regression.18), p 4 =.7 CCC =.173 (-.83,.47), p =.183 CCC.733 (.96,.89), p <.1 testing set: mean rmullen Verbal =., CCC =.117 (-.9,.314), p =.67 CCC p =.39.1 (.41,.64), p =.3 Quality: r =.1 (-.7,.318), p =.96 CCC =.43 (-.7,.33), p =.787 3 18 r =.179 (-.77,.41), p =.1681 1 Mullen Receptive: r =.76 (.69,.849), p <.1 Training Regression 3 Quality: r =.3 (-.,.391), p =. 1 CCC =.169 (-.8, r =.1 (.197,.74), p =.3 1.33), p =.61 1 r =.81 (-.1,.369), p =.997 Testing Regression CCC =.68 (.19,.797), p <.1 16 CCC =.173 (-.83,.47), p =.183 CCC =.17 (-.,.3), p =.136 CCC =.447 (.13,.69), p =.133 1 CCC = -.38 (-.33,.61), p =.84 14 Mullen Expressive: r =.744 (.6,.838), p <.1 Historical Data Figure XXXX. Measures of correlation between eye-tracking indices and standardized assessments of autistic social disability (using the Autism Diagnostic Observation Schedule, r =.73 (.68,.773), p =.9 ADOS) and Social of verbal Ability and Index non-verbal cognitive skills (using the Mullen Scales of Early Learning, Mullen), as well as with chronological age and eye-tracking data quality rating (with which no CCC =.78 (.84,.88), p <.1 associations are expected). Higher scores on the ADOS denote greater social disability. Non-verbal Social cognitive Ability Index skills on the Mullen are given Social as age Ability equivalence Index scores (in months). Verbal/language Non Verbal Ability Index Receptive Language Ability Index Expressive Language Ability Index CCC =.9 (.8,.747), p =.6 skills on the Mullen are given separately as receptive and expressive language skills, again as age equivalence scores (in months). Plots are separated by age group and measure. In all plots and Historical Data Figure XXXX. Measures of correlation in between reporting eye-tracking of correlation indices coefficients, and standardized light gray denotes assessments results of for autistic the training social sample, dark gray denotes results for the testing sample. Correlation coefficients are given with 9% confidence intervals in parentheses. Abbreviations: (Mullen) AE = age equivalence (in months); INFANT disability (using the Autism Diagnostic Observation Schedule, ADOS) and of verbal and non-verbal cognitive skills (using the Mullen Scales of Early Learning, Mullen), as well as with chronological r age = Spearman s and eye-tracking rho; CCC data = Lin s quality Concordance rating (with Correlation Coefficient. 3 which no 4 Mullen Non-Verbal: Age: r =.79 (.4, r = -.63.74), (-.37, p <.1.114), p =.4843 associations are expected). Higher scores on the ADOS denote greater social disability. Non-verbal cognitive skills on the Mullen are given as 3age equivalence scores (in 4 months). Verbal/language r =.9 (-.18, r = -.99.), (-.34, p =.643.17), p =.4487 Training Sample skills on the Mullen are given separately as receptive and expressive language skills, again as age equivalence scores (in months). SIBLING 3 Plots are separated by STUDY age group and measure. In all plots and 3 3 CCC =.39 CCC (.371, = -.96.674), (-.68, p <.1.8), p =.88 in reporting of correlation coefficients, light gray denotes results for the training sample, dark gray denotes results for the testing sample. Correlation coefficients are given with 9% confidence 4 Testing Sample CCC =.81 CCC (-.9, = -.73.4), (-.319, p =.71.18), p =.7 intervals in parentheses. Abbreviations: (Mullen) AE = age equivalence (in months); r = Spearman s rho; CCC = Lin s Concordance Correlation Coefficient. Public Health Opportunities Mullen Receptive: Quality: r =.633 (.496, r =.1.739), (-.7, p <.1.318), p =.96 A National Institutes of Health 3 r =.487 (.6, r =.179.683), (-.77, p =.7.41), p =.1681 Training Regression Autism 1 1 CCC =.611 CCC (.469, =.169.73), (-.8, p <.1.33), p =.61 1 Center of Excellence Testing Regression CCC =.468 CCC (., =.173.67), (-.83, p =.1.47), p =.183 1 Mullen Expressive: r =.683 (.6,.777), p <.1 r =.1 (.43,.69), p =. CCC =.66 (.33,.761), p <.1 Non Verbal Ability Index Receptive Language Ability Index Expressive Language Ability Index CCC =.48 (.17,.678), p =.8 ADOS Socia Mullen Non Verbal AE Mullen Non Verbal AE Age (months) Age (months) ADOS To Mullen Receptive 18-Month-Olds Language AE Mullen Receptive 4-Month-Olds Language AE Quality 18-Month-Olds Quality 4-Month-Olds A Mullen Non Verbal AE Age (months) ADOS Sev Mullen Expressive Language AE Mullen Expressive Language AE Mullen Receptive Language AE Quality 18-Month-Olds 4-Month-Olds 18-Month-Olds 4-Month-Olds ADOS Social Affect Score ADOS Social Affect Score Mullen Non Verbal AE Mullen Non Verbal AE 18-Month 4-Month-Olds 18-Month-Olds 4-Month-Olds 18-Month-Olds 4-Month-Olds ADOS Social A ADOS Social Affect Score Mullen Non Verbal AE Mullen Non Verbal AE Age (months) Age (months) Mullen Expressive Language AE ADOS Total Score ADOS Total Score Mullen Receptive Language AE Mullen Receptive Language AE ADOS Tot ADOS Total Score Mullen Receptive Language AE Mullen Receptive Language AE Quality Quality ADOS Severity Score ADOS Severity Score Mullen Expressive Language AE Mullen Expressive Language AE ADOS Seve ADOS Severity Score Mullen Expressive Language AE Mullen Expressive Language AE High-throughput, low-cost, deployment of universal screening in the community Early detection, early intervention, optimal outcome Prevention or attenuation of intellectual disability in ASD 18-Month-Olds 4-Month-Olds Positive Historical Controls Data Figure XXXX. Measures of correlation between eye-tracking indices and standardized assessments of autistic social disability (using the Autism Diagnostic Observation Schedule, ADOS) and of verbal and non-verbal cognitive skills (using the Mullen Scales of Early Learning, Mullen), as well as with chronological age and eye-tracking data quality rating (with which no associations are expected). Higher scores on the ADOS denote greater social disability. Non-verbal cognitive skills on the Mullen are given as age equivalence scores (in months). Verbal/language skills on the Mullen are given separately as receptive and expressive language skills, again as age equivalence scores (in months). Plots are separated by age group and measure. In all plots and in reporting of correlation coefficients, light gray denotes results for the training sample, dark gray denotes results for the Age: testing sample. Correlation r =.97 coefficients (-.11,.96), are given p =.388 with 9% confidence Support a system that does not intervals in parentheses. Abbreviations: (Mullen) r =.83 (-.19,.371), p =.917 4 AE = age equivalence (in months); r = Spearman s rho; CCC = Lin s Concordance Correlation Coefficient. CCC =.117 (-.9,.314), p =.67 have sufficient expert clinicians CCC =.43 (-.7,.33), p =.787 Age (months) Age (months) 18 16 14 3 Quality Quality 3 4 3 A new, promising view of autism, with universal design implications Genetic influence informs modality of early treatment Training Sample Testing Sample Reduce the child, family, health, Training Regression education, and societal costs of Testing Regression autism Quality: r =.3 (-.,.391), p =. r =.81 (-.1,.369), p =.997 CCC =.17 (-.,.3), p =.136 CCC = -.38 (-.33,.61), p =.84 Age: r = -.63 (-.37,.114), p =.4843 r = -.99 (-.34,.17), p =.4487 CCC = -.96 (-.68,.8), p =.88 CCC = -.73 (-.319,.18), p =.7 Quality: r =.1 (-.7,.318), p =.96 r =.179 (-.77,.41), p =.1681 CCC =.169 (-.8,.33), p =.61 CCC =.173 (-.83,.47), p =.183 Historical Data Figure XXXX. Measures of correlation between eye-tracking indices and standardized assessments of autistic social disability (using the Autism Diagnostic Observation Schedule, ADOS) and of verbal and non-verbal cognitive skills (using the Mullen Scales of Early Learning, Mullen), as well as with chronological age and eye-tracking data quality rating (with which no associations are expected). Higher scores on the ADOS denote greater social disability. Non-verbal cognitive skills on the Mullen are given as age equivalence scores (in months). Verbal/language skills on the Mullen are given separately as receptive and expressive language skills, again as age equivalence scores (in months). Plots are separated by age group and measure. In all plots and in reporting of correlation coefficients, light gray denotes results for the training sample, dark gray denotes results for the testing sample. Correlation coefficients are given with 9% confidence intervals in parentheses. Abbreviations: (Mullen) AE = age equivalence (in months); r = Spearman s rho; CCC = Lin s Concordance Correlation Coefficient.
Objectifying and quantifying diagnosis in toddlers: community-viable proxies Quantitative Measures of Social Disability High-throughput, low-cost, deployment of universal screening in the community Objective, quantitative measures of risk Early detection, early intervention, optimal outcome Prevention or attenuation of intellectual disability in ASD 61 From High-Impact Publication to FDAregistered Clinical Trial of Diagnostic Intended Indication for Use Site selection 7 sites confirmed Marcus Seattle Children s Cincinnati Children s Southwest Regional Autism Center (Phoenix) UC-San Francisco University of Vermont TREAT Center Toronto 63 The Social Developmental Testing Device is a medical device designed to measure visual attention to social information in the environment relative to normative, age-specific benchmarks. These measurements assess presence (primary efficacy outcome) & severity (secondary efficacy outcome) of Autism Spectrum Disorder (ASD) in 16- to 3-month-old children.
A Bioethical Imperative: Access to Early Treatment - Promoting Social Engagement New science informing modality of treatment - Genetic influences over social visual engagement Reciprocal Social Interaction Typically Developing -month old baby -month-old baby later diagnosed with ASD 6 66 How to link these quantifications of behavior to the genetic bases of autism? Measuring the genetic structure of social visual engagement Eye tracking assay is under tight genetic control N= TD toddlers, including 166 epidemiologically-ascertained twins, 8 MZ & 84 DZ Age: 1.3 months (4.6) toddlers: 8 monozygotic twins (41 MZ pairs) Warren Jones, PhD John Constantino, MD 84 dizygotic twins (4 DZ pairs) 84 non-sibling comparison children (4 non-sib control pairs) age 1.3(4.3) months non-sibs matched <1 day Twin-Twin Concordance MZ ICC:.91 (.8-.9) DZ ICC:.3 (.7-.9) Non-sibling pair:.16 (.-.44)
Concordance in social visual engagement as a function of zygosity. 69 7 Moment-by-Moment: Funnels of Attention MZ concordance of looking behavior at timescales of milliseconds When watching complex social scenes MZ twins are more likely than DZ twin to shift their eyes at the same moment in the same direction onto the same targets They are creating their social world of experiences
73 74 The markers of social visual engagement that are most highly heritable 3 3 The markers of social visual engagement 9 that th are most highly heritable Mouth (%) 4 1 TD Cohort: MZ, DZ, non-siblings (N =, ) Mouth (%) 4 1 TD Cohort: Density Distribution Percentiles th th th 7 th Mouth (%) 4 3 1 1 3 4 Eyes (%) 1 3 4 Eyes (%).8 are also those that most clearly distinguish.6 typically-developing children from those with autism. Mouth (%) TD Cohort: MZ, DZ, non-siblings (N =, ) ASD Comparison Cohort (N = 43, ) Eyes (%) ASD Replication Cohort Mouth (%) 4 3 1 True Positive Rate 1.4. TD Cohort: Density Distribution Percentiles th th th 7 th 9 th ASD Comparison Cohort AUC =.88 (.84-.9)..4.6.8 1 False Positive Rate ASD Replication Cohort Empirical Fitted 9% Marcus CI Autism Center Mouth (%) 1 3 4 Eyes (%) 4 3 1 1 3 4 Eyes (%) 1 3 4 Eyes (%) Classification of ASD Comparison Cohort 1 AUC =.88 (.84-.9)..4.6.8 1 False Positive Rate.8 are 3 also those that most clearly distinguish.6 typically-developing children from those with autism. Mouth (%) 4 1 ASD Comparison Cohort (N = 43, ) ASD Replication Cohort (N = 4, ) 1 3 4 Eyes (%) True Positive Rate True Positive Rate.8.6.4. 1.4. Empirical Fitted 9% CI ASD Replication Cohort AUC =.86 (.8-.91)..4.6.8 1 False Positive Rate Empirical Fitted 9% CI
ASD Comparison Cohort (N = 43, ) 1 ASD Comparison Cohort.6 The markers of social visual engagement that.4 are most highly heritable Mouth (%) Mouth (%) 4 3 1 Eyes (%) ASD Replication Cohort (N = 4, ) 1 3 4 Eyes (%) AUC =.88 (.84-.9)..4.6.8 1 False Positive Rate are also those that most clearly distinguish typically-developing children from those with autism. True Positive Rate True Positive Rate.8. 1.8.6.4. Empirical Fitted 9% CI Classification of high Heritability (eye- & mouth-looking) ASD Replication Cohort Empirical Fitted 9% CI AUC =.86 (.8-.91)..4.6.8 1 False Positive Rate Scarr & McCartney, 1983. + high Probability (shifting eyes at same moments, in same directions, towards same content) = profound influence on human biological niche construction From reducing age of diagnosis to improving access to early intervention Family Engagement ALLEVIATE AUTISM (National Research Council, 1) so how do we achieve hours per week in which the child is engaged actively and productively in meaningful activities? Family Engagement Less than % of children who will need special services in school in the US are identified before the age of 3 years 79 8
Augmenting Access to Early Treatment Family Primary Care Physician Amy Wetherby, PhD Early Intervention Provider Jennifer Stapel- Wax, PsyD the Community: Families, Pediatricians, Early Intervention Providers Parent-Delivered Early Social Interaction Wetherby et al., 14 84
Treating deviations from normative social engagement: Parent- Delivered Social Interaction Amy Wetherby, PhD 8 9 Goals for Early Treatment: Every waking hour in the home and in the community Teaching Strategies & Supports to Promote Active Engagement Child Behaviors ACTIVE ENGAGEMENT 1. Emotional Regulation. Productivity 3. Social Connectedness 4. Gaze to Face. Response to Verbal 6. Directed 7. Flexibility 8. Generative Ideas Parent Behaviors TRANSACTIONAL SUPPORTS 1. Participation & Role. Make Activity 3. Follow Child s Attention 4. Promote Initiations. Balance of Turns 6. Support Comprehension 7. Modeling 8. Expectations & Supports for better skills!model and expand language and play skills!extend activity, child s roles, & transitions!balance demands and supports Supports for social reciprocity!natural reinforcers!waiting for initiation and balance of turns!clear message to ensure comprehension Supports for a common agenda!positioning!follow child s attentional focus!motivating activity with clear roles & turns 87
Social Interaction is the Platform for Brain Development Our brains become who we are. (J LeDoux) Brain structure and function are physical instantiations of lived experience. 9
Neonates preferentially orient towards stimuli that More Preferred Less Preferred More Preferred Less Preferred mother s voice stranger s voice complex non-speech pure tone, structured noise silence mother s voice stranger s voice complex non-speech pure tone, structured noise silence sound like caregivers. DeCasper & Fifer, 198. Vouloumanos & Werker, 7. Butterfield & Siperstein, 197 Eisenberg, 1976. Neonates preferentially orient towards stimuli that Neonates preferentially orient towards stimuli that More Preferred Less Preferred More Preferred Less Preferred mother s scent stranger s scent biological motion inverted biological motion scrambled biological motion sound like caregivers. smell like caregivers. sound like caregivers. smell like caregivers. move like caregivers. Macfarlane, 197. Porter & Winberg, 1999. Simion, Regolin, & Bulf, 8.
Neonates preferentially orient towards stimuli that Neonates preferentially orient towards stimuli that More Preferred Less Preferred More Preferred Less Preferred face-like face-like configural face-like, scrambled face-like, inverted mother, engaging stranger, eyes open stranger, eyes averted stranger, eyes closed Goren, Sarty, & Wu, 197. Johnson, Dziurawiec, Ellis, & Morton, 1991. Simion, Valenza, Umiltà, & Barba, 1998. Valenza, Simion, Cassia, & Umiltà, 1996. Farroni et al,. sound like caregivers. smell like caregivers. move like caregivers. look like caregivers. Bushnell, Sai, & Mullin, 1989. Simion, Valenza, Umiltà, & Barba, 1998. Farroni, Csibra, Simion, & Johnson,. Batki, Baron-Cohen, et al,. Sai, 199. Sai,. Walton, Bower, & Bower, 199. sound like caregivers. smell like caregivers. move like caregivers. look like caregivers. interact like caregivers. Typically-Developing -Month-Old
Normative Growth Charts of Social Visual Engagement 7 6 Fixation Time (%) 4 3 1 mean 9% CI TD eyes TD, N=63 Change in Fixation (% per month) 4 - -4 3 4 6 9 1 1 18 4 Increasing Decreasing 3 4 6 9 1 1 18 4 Growth Charts Fixation Time (%) Change in Fixation (% per month) % of Children Who % of Children Who 7 6 4 3 1 4 - -4 1 9 8 7 6 4 3 3 4 6 9 1 1 18 4 3 4 6 9 1 1 18 4 Regard Caregiver s Face Smile Laugh Say Mama / Dada Play (Feed Doll) Interpersonal Pretend Increasing Decreasing Growth Charts Change in Fixation % of Children Who % of Children Who (% per month) Fixation Time (%) Fixation Time (%) 7 6 4 3 1 4 - -4 >98 9 8 7 6 4 3 3 4 6 9 1 1 18 4 3 4 6 9 1 1 18 4 Babble Imitate Speech Sounds Produce 1 Word Produce >6 Words Increasing Decreasing 1 1 3 4 6 9 1 1 18 4 < 3 4 6 9 1 1 18 4
-Month-Old with Autism Eye-Looking in Typically-Developing Infants and Infants Later Diagnosed with ASD 7 6 Fixation Time (%) 4 3 1 mean 9% CI TD eyes TD, N=63 Change in Fixation (% per month) 4 - -4 3 4 6 9 1 1 18 4 Dt TD eyes Increasing Decreasing 3 4 6 9 1 1 18 4 Eye-Looking Attention to in eyes Typically-Developing is present but in decline in Infants 6-month-old infants later diagnosed with autism and Infants Later Diagnosed with ASD Warren Jones 1,,3 & Ami Klin 1,,3 7 6 Attention to eyes is present but in decline in 6-month-old infants later diagnosed with autism First Replication Cohort Warren Jones 1,,3 & Ami Klin 1,,3 7 6 Fixation Time (%) 4 3 1 mean 9% CI TD eyes ASD1 eyes Fixation Time (%) 4 3 1 mean 9% CI TD eyes ASD1 eyes TD, N=63 ASD1, N=11 Change in Fixation (% per month) 4 - -4 Jones & Klin, Nature, 13. 3 4 6 9 1 1 18 4 3 4 6 9 1 1 18 4 Dt TD eyes Dt ASD1 eyes TD, N=63 ASD1, N=11 Change in Fixation (% per month) 4 - -4 Jones & Klin, Nature, 13. 3 4 6 9 1 1 18 4 3 4 6 9 1 1 18 4 Dt TD eyes Dt ASD1 eyes
First Replication Cohort First Replication Cohort 7 6 7 6 Fixation Time (%) 4 3 1 mean 9% CI TD eyes ASD1 eyes ASD eyes Fixation Time (%) 4 3 1 TD eyes ASD1 eyes ASD eyes TD, N=63 ASD1, N=11 ASD, N=13 Change in Fixation (% per month) 4 - -4 3 4 6 9 1 1 18 4 Dt TD eyes Dt ASD1 eyes Dt ASD eyes 3 4 6 9 1 1 18 4 TD, N=63 ASD1, N=11 ASD, N=13 Change in Fixation (% per month) 4 - -4 3 4 6 9 1 1 18 4 3 4 6 9 1 1 18 4 Dt TD eyes Dt ASD1 eyes Dt ASD eyes Growth Charts Fixation Time (%) 7 6 4 3 Our ultimate goal 1 Change in Fixation (% per month) % of Children Who % of Children Who 4 - -4 1 9 8 7 6 1 3 4 6 9 1 1 18 4 Regard Caregiver s Face Quantifying Disruption of Early-Emerging, Highly-Conserved Smile 3 4 6 9 1 1 18 4 Laugh 4 3 Mechanisms of Social Adaptive Action Say Mama / Dada Play (Feed Doll) Interpersonal Pretend 3 4 6 9 1 1 18 4 To make autism an issue of diversity, not of disability 11