Social Brain Functional Maturation in Newborn Infants With and Without a Family History of Autism Spectrum Disorder | Autism Spectrum Disorders | JAMA Network Open | JAMA Network
[Skip to Navigation]
Sign In
Figure 1.  Anatomical Segmentation of Regions of Interest Relevant for Social Function
Anatomical Segmentation of Regions of Interest Relevant for Social Function

The anterior cingulate, the posterior cingulate, the insula, the posterior superior temporal sulcus, the anterior fusiform, and the posterior fusiform are represented inside a 3-dimensional rendering of the neonatal template. Right, left, ventral, and dorsal views were selected to provide full visualization of the anatomical space covered by the regions of interest.

Figure 2.  Higher Levels of Regional Homogeneity in Several Clusters in the Social Brain Among Neonates With Familial Risk for Autism Spectrum Disorder (R+) Compared With Neonates With No Family History (R−)
Higher Levels of Regional Homogeneity in Several Clusters in the Social Brain Among Neonates With Familial Risk for Autism Spectrum Disorder (R+) Compared With Neonates With No Family History (R−)

The 3-dimensional rendered clusters (red) where there was a significant main effect of group (P < .05 after false discovery rate correction) are overlaid on a neonatal template. These clusters were localized to the left parietal lobe (A) and the right posterior fusiform (B). The filled circles in the plots represent the median regional homogeneity value per participant of all the voxels located in the significantly different clusters; boxes represent the quantiles; and the central horizontal line, the median.

Figure 3.  Significantly Different Maturational Trajectory in Cingulate and Insula Among Participants With Familial Risk of Autism Spectrum Disorder
Significantly Different Maturational Trajectory in Cingulate and Insula Among Participants With Familial Risk of Autism Spectrum Disorder

The 3-dimensional rendered clusters overlaid in a neonatal template (red) show the voxels that appeared to have a significantly different age interaction effect between groups. The clusters were located in left anterior insula (A), left anterior cingulate (B), left posterior cingulate (C), right anterior cingulate (D), and right posterior cingulate (E). For each region of interest, we plotted the median regional homogeneity value (lines) with 95% confidence intervals (shaded area) and show how these values increase with postmenstrual age (PMA) for the group with no familial risk, while they remain constant or decrease for the group with familial risk.

Table 1.  Study Sample Descriptive Statistics
Study Sample Descriptive Statistics
Table 2.  Study Sample Participant-Specific Demographic Characteristicsa
Study Sample Participant-Specific Demographic Characteristicsa
1.
Beer  JS, Ochsner  KN.  Social cognition: a multi level analysis.  Brain Res. 2006;1079(1):98-105. doi:10.1016/j.brainres.2006.01.002PubMedGoogle ScholarCrossref
2.
Herrmann  E, Call  J, Hernàndez-Lloreda  MV, Hare  B, Tomasello  M.  Humans have evolved specialized skills of social cognition: the cultural intelligence hypothesis.  Science. 2007;317(5843):1360-1366. doi:10.1126/science.1146282PubMedGoogle ScholarCrossref
3.
Thye  MD, Bednarz  HM, Herringshaw  AJ, Sartin  EB, Kana  RK.  The impact of atypical sensory processing on social impairments in autism spectrum disorder.  Dev Cogn Neurosci. 2018;29:151-167. doi:10.1016/j.dcn.2017.04.010PubMedGoogle ScholarCrossref
4.
Alcalá-López  D, Smallwood  J, Jefferies  E,  et al.  Computing the social brain connectome across systems and states.  Cereb Cortex. 2018;28(7):2207-2232. doi:10.1093/cercor/bhx121PubMedGoogle ScholarCrossref
5.
Lauritsen  MB.  Autism spectrum disorders.  Eur Child Adolesc Psychiatry. 2013;22(suppl 1):S37-S42. doi:10.1007/s00787-012-0359-5PubMedGoogle ScholarCrossref
6.
Lynn  AC, Padmanabhan  A, Simmonds  D,  et al.  Functional connectivity differences in autism during face and car recognition: underconnectivity and atypical age-related changes.  Dev Sci. 2018;21(1):e12508. doi:10.1111/desc.12508PubMedGoogle ScholarCrossref
7.
Fox  SE, Wagner  JB, Shrock  CL, Tager-Flusberg  H, Nelson  CA.  Neural processing of facial identity and emotion in infants at high-risk for autism spectrum disorders.  Front Hum Neurosci. 2013;7(89):89.PubMedGoogle Scholar
8.
Kana  RK, Sartin  EB, Stevens  C  Jr,  et al.  Neural networks underlying language and social cognition during self-other processing in autism spectrum disorders.  Neuropsychologia. 2017;102:116-123. doi:10.1016/j.neuropsychologia.2017.06.008PubMedGoogle ScholarCrossref
9.
Odriozola  P, Uddin  LQ, Lynch  CJ, Kochalka  J, Chen  T, Menon  V.  Insula response and connectivity during social and non-social attention in children with autism.  Soc Cogn Affect Neurosci. 2016;11(3):433-444. doi:10.1093/scan/nsv126PubMedGoogle ScholarCrossref
10.
Daniels  AM, Mandell  DS.  Explaining differences in age at autism spectrum disorder diagnosis: a critical review.  Autism. 2014;18(5):583-597. doi:10.1177/1362361313480277PubMedGoogle ScholarCrossref
11.
Lord  C, Jones  RM.  Annual research review: re-thinking the classification of autism spectrum disorders.  J Child Psychol Psychiatry. 2012;53(5):490-509. doi:10.1111/j.1469-7610.2012.02547.xPubMedGoogle ScholarCrossref
12.
Silbereis  JC, Pochareddy  S, Zhu  Y, Li  M, Sestan  N.  The cellular and molecular landscapes of the developing human central nervous system.  Neuron. 2016;89(2):248-268. doi:10.1016/j.neuron.2015.12.008PubMedGoogle ScholarCrossref
13.
Livingston  LA, Happé  F.  Conceptualising compensation in neurodevelopmental disorders: reflections from autism spectrum disorder.  Neurosci Biobehav Rev. 2017;80:729-742. doi:10.1016/j.neubiorev.2017.06.005PubMedGoogle ScholarCrossref
14.
Happé  F, Ronald  A, Plomin  R.  Time to give up on a single explanation for autism.  Nat Neurosci. 2006;9(10):1218-1220. doi:10.1038/nn1770PubMedGoogle ScholarCrossref
15.
Green  J, Pickles  A, Pasco  G,  et al; British Autism Study of Infant Siblings (BASIS) Team.  Randomised trial of a parent-mediated intervention for infants at high risk for autism: longitudinal outcomes to age 3 years.  J Child Psychol Psychiatry. 2017;58(12):1330-1340. doi:10.1111/jcpp.12728PubMedGoogle ScholarCrossref
16.
Dawson  G, Jones  EJ, Merkle  K,  et al.  Early behavioral intervention is associated with normalized brain activity in young children with autism.  J Am Acad Child Adolesc Psychiatry. 2012;51(11):1150-1159. doi:10.1016/j.jaac.2012.08.018PubMedGoogle ScholarCrossref
17.
Egorov  AV, Draguhn  A.  Development of coherent neuronal activity patterns in mammalian cortical networks: common principles and local hetereogeneity.  Mech Dev. 2013;130(6-8):412-423. doi:10.1016/j.mod.2012.09.006PubMedGoogle ScholarCrossref
18.
Peixoto  RT, Wang  W, Croney  DM, Kozorovitskiy  Y, Sabatini  BL.  Early hyperactivity and precocious maturation of corticostriatal circuits in Shank3B(-/-) mice.  Nat Neurosci. 2016;19(5):716-724. doi:10.1038/nn.4260PubMedGoogle ScholarCrossref
19.
Gonçalves  JT, Anstey  JE, Golshani  P, Portera-Cailliau  C.  Circuit level defects in the developing neocortex of fragile X mice.  Nat Neurosci. 2013;16(7):903-909. doi:10.1038/nn.3415PubMedGoogle ScholarCrossref
20.
Gutierrez  RC, Hung  J, Zhang  Y, Kertesz  AC, Espina  FJ, Colicos  MA.  Altered synchrony and connectivity in neuronal networks expressing an autism-related mutation of neuroligin 3.  Neuroscience. 2009;162(1):208-221. doi:10.1016/j.neuroscience.2009.04.062PubMedGoogle ScholarCrossref
21.
Allène  C, Cattani  A, Ackman  JB,  et al.  Sequential generation of two distinct synapse-driven network patterns in developing neocortex.  J Neurosci. 2008;28(48):12851-12863. doi:10.1523/JNEUROSCI.3733-08.2008PubMedGoogle ScholarCrossref
22.
Hanganu-Opatz  IL.  Between molecules and experience: role of early patterns of coordinated activity for the development of cortical maps and sensory abilities.  Brain Res Rev. 2010;64(1):160-176. doi:10.1016/j.brainresrev.2010.03.005PubMedGoogle ScholarCrossref
23.
Zang  Y, Jiang  T, Lu  Y, He  Y, Tian  L.  Regional homogeneity approach to fMRI data analysis.  Neuroimage. 2004;22(1):394-400. doi:10.1016/j.neuroimage.2003.12.030PubMedGoogle ScholarCrossref
24.
Hughes  EJ, Winchman  T, Padormo  F,  et al.  A dedicated neonatal brain imaging system.  Magn Reson Med. 2017;78(2):794-804. doi:10.1002/mrm.26462PubMedGoogle ScholarCrossref
25.
Satterthwaite  TD, Wolf  DH, Loughead  J,  et al.  Impact of in-scanner head motion on multiple measures of functional connectivity: relevance for studies of neurodevelopment in youth.  Neuroimage. 2012;60(1):623-632. doi:10.1016/j.neuroimage.2011.12.063PubMedGoogle ScholarCrossref
26.
Power  JD, Barnes  KA, Snyder  AZ, Schlaggar  BL, Petersen  SE.  Spurious but systematic correlations in functional connectivity MRI networks arise from subject motion.  Neuroimage. 2012;59(3):2142-2154. doi:10.1016/j.neuroimage.2011.10.018PubMedGoogle ScholarCrossref
27.
Griffanti  L, Salimi-Khorshidi  G, Beckmann  CF,  et al.  ICA-based artefact removal and accelerated fMRI acquisition for improved resting state network imaging.  Neuroimage. 2014;95:232-247. doi:10.1016/j.neuroimage.2014.03.034PubMedGoogle ScholarCrossref
28.
Salimi-Khorshidi  G, Douaud  G, Beckmann  CF, Glasser  MF, Griffanti  L, Smith  SM.  Automatic denoising of functional MRI data: combining independent component analysis and hierarchical fusion of classifiers.  Neuroimage. 2014;90:449-468. doi:10.1016/j.neuroimage.2013.11.046PubMedGoogle ScholarCrossref
29.
Andersson  JLR, Skare  S, Ashburner  J.  How to correct susceptibility distortions in spin-echo echo-planar images: application to diffusion tensor imaging.  Neuroimage. 2003;20(2):870-888. doi:10.1016/S1053-8119(03)00336-7PubMedGoogle ScholarCrossref
30.
Serag  A, Aljabar  P, Ball  G,  et al.  Construction of a consistent high-definition spatio-temporal atlas of the developing brain using adaptive kernel regression.  Neuroimage. 2012;59(3):2255-2265. doi:10.1016/j.neuroimage.2011.09.062PubMedGoogle ScholarCrossref
31.
Andersson  JLR, Jenkinson  M, Smith  S.  Non-linear Registration, aka Spatial Normalisation. Oxford, UK: FMRIB Centre; 2010.
32.
Taylor  PA, Saad  ZS.  FATCAT: (an efficient) functional and tractographic connectivity analysis toolbox.  Brain Connect. 2013;3(5):523-535. doi:10.1089/brain.2013.0154PubMedGoogle ScholarCrossref
33.
Walker  DA.  Converting Kendall’s tau for correlational or meta-analytic analyses.  J Mod Appl Stat Methods. 2003;2(2):525-530. doi:10.22237/jmasm/1067646360Google ScholarCrossref
34.
Gousias  IS, Hammers  A, Counsell  SJ,  et al.  Magnetic resonance imaging of the newborn brain: automatic segmentation of brain images into 50 anatomical regions.  PLoS One. 2013;8(4):e59990. doi:10.1371/journal.pone.0059990PubMedGoogle ScholarCrossref
35.
Gousias  IS, Edwards  AD, Rutherford  MA,  et al.  Magnetic resonance imaging of the newborn brain: manual segmentation of labelled atlases in term-born and preterm infants.  Neuroimage. 2012;62(3):1499-1509. doi:10.1016/j.neuroimage.2012.05.083PubMedGoogle ScholarCrossref
36.
Winkler  AM, Ridgway  GR, Webster  MA, Smith  SM, Nichols  TE.  Permutation inference for the general linear model.  Neuroimage. 2014;92:381-397. doi:10.1016/j.neuroimage.2014.01.060PubMedGoogle ScholarCrossref
37.
Devlin  JT, Jamison  HL, Gonnerman  LM, Matthews  PM.  The role of the posterior fusiform gyrus in reading.  J Cogn Neurosci. 2006;18(6):911-922. doi:10.1162/jocn.2006.18.6.911PubMedGoogle ScholarCrossref
38.
McKeefry  DJ, Zeki  S.  The position and topography of the human colour centre as revealed by functional magnetic resonance imaging.  Brain. 1997;120(Pt 12):2229-2242. doi:10.1093/brain/120.12.2229PubMedGoogle ScholarCrossref
39.
Kanwisher  N, McDermott  J, Chun  MM.  The fusiform face area: a module in human extrastriate cortex specialized for face perception.  J Neurosci. 1997;17(11):4302-4311. doi:10.1523/JNEUROSCI.17-11-04302.1997PubMedGoogle ScholarCrossref
40.
De Haan  M, Johnson  MH, Maurer  D, Perrett  DI.  Recognition of individual faces and average face prototypes by 1- and 3-month-old infants.  Cogn Dev. 2001;16(2):659-678. doi:10.1016/S0885-2014(01)00051-XGoogle ScholarCrossref
41.
Johnson  MH, Griffin  R, Csibra  G,  et al.  The emergence of the social brain network: evidence from typical and atypical development.  Dev Psychopathol. 2005;17(3):599-619. doi:10.1017/S0954579405050297PubMedGoogle ScholarCrossref
42.
Otsuka  Y.  Face recognition in infants: a review of behavioral and near-infrared spectroscopic studies.  Jpn Psychol Res. 2014;56(1):76-90. doi:10.1111/jpr.12024Google ScholarCrossref
43.
von Hofsten  O, von Hofsten  C, Sulutvedt  U, Laeng  B, Brennen  T, Magnussen  S.  Simulating newborn face perception.  J Vis. 2014;14(13):16. doi:10.1167/14.13.16PubMedGoogle ScholarCrossref
44.
Stein  T, Peelen  MV, Sterzer  P.  Adults’ awareness of faces follows newborns’ looking preferences.  PLoS One. 2011;6(12):e29361. doi:10.1371/journal.pone.0029361PubMedGoogle ScholarCrossref
45.
Di Giorgio  E, Frasnelli  E, Salva  OR,  et al.  Difference in visual social predispositions between newborns at low- and high-risk for autism.  Sci Rep. 2016;6:1-9. doi:10.1038/srep26395PubMedGoogle ScholarCrossref
46.
Jonas  J, Rossion  B, Brissart  H,  et al.  Beyond the core face-processing network: intracerebral stimulation of a face-selective area in the right anterior fusiform gyrus elicits transient prosopagnosia.  Cortex. 2015;72:140-155. doi:10.1016/j.cortex.2015.05.026PubMedGoogle ScholarCrossref
47.
Rangarajan  V, Parvizi  J.  Functional asymmetry between the left and right human fusiform gyrus explored through electrical brain stimulation.  Neuropsychologia. 2016;83:29-36. doi:10.1016/j.neuropsychologia.2015.08.003PubMedGoogle ScholarCrossref
48.
Pfeifer  JH, Lieberman  MD, Dapretto  M.  “I know you are but what am I?!”: neural bases of self- and social knowledge retrieval in children and adults.  J Cogn Neurosci. 2007;19(8):1323-1337. doi:10.1162/jocn.2007.19.8.1323PubMedGoogle ScholarCrossref
49.
Doria  V, Beckmann  CF, Arichi  T,  et al.  Emergence of resting state networks in the preterm human brain.  Proc Natl Acad Sci U S A. 2010;107(46):20015-20020. doi:10.1073/pnas.1007921107PubMedGoogle ScholarCrossref
50.
Dawson  G, Webb  SJ, McPartland  J.  Understanding the nature of face processing impairment in autism: insights from behavioral and electrophysiological studies.  Dev Neuropsychol. 2005;27(3):403-424. doi:10.1207/s15326942dn2703_6PubMedGoogle ScholarCrossref
51.
Mundy  P.  A review of joint attention and social-cognitive brain systems in typical development and autism spectrum disorder.  Eur J Neurosci. 2018;47(6):497-514. doi:10.1111/ejn.13720PubMedGoogle ScholarCrossref
52.
Alcauter  S, Lin  W, Keith Smith  J, Gilmore  JH, Gao  W.  Consistent anterior-posterior segregation of the insula during the first 2 years of life.  Cereb Cortex. 2015;25(5):1176-1187. doi:10.1093/cercor/bht312PubMedGoogle ScholarCrossref
53.
Kana  RK, Maximo  JO, Williams  DL,  et al.  Aberrant functioning of the theory-of-mind network in children and adolescents with autism.  Mol Autism. 2015;6(1):59. doi:10.1186/s13229-015-0052-xPubMedGoogle ScholarCrossref
54.
Venkataraman  A, Duncan  JS, Yang  DYJ, Pelphrey  KA.  An unbiased Bayesian approach to functional connectomics implicates social-communication networks in autism.  Neuroimage Clin. 2015;8:356-366. doi:10.1016/j.nicl.2015.04.021PubMedGoogle ScholarCrossref
55.
Wang  Y, Collins  JA, Koski  J, Nugiel  T, Metoki  A, Olson  IR.  Dynamic neural architecture for social knowledge retrieval.  Proc Natl Acad Sci U S A. 2017;114(16):E3305-E3314. doi:10.1073/pnas.1621234114PubMedGoogle ScholarCrossref
56.
Zahn  R, Moll  J, Krueger  F, Huey  ED, Garrido  G, Grafman  J.  Social concepts are represented in the superior anterior temporal cortex.  Proc Natl Acad Sci U S A. 2007;104(15):6430-6435. doi:10.1073/pnas.0607061104PubMedGoogle ScholarCrossref
57.
Olson  IR, McCoy  D, Klobusicky  E, Ross  LA.  Social cognition and the anterior temporal lobes: a review and theoretical framework.  Soc Cogn Affect Neurosci. 2013;8(2):123-133. doi:10.1093/scan/nss119PubMedGoogle ScholarCrossref
58.
Marín  O.  Developmental timing and critical windows for the treatment of psychiatric disorders.  Nat Med. 2016;22(11):1229-1238. doi:10.1038/nm.4225PubMedGoogle ScholarCrossref
59.
Johnson  MH.  Autism as an adaptive common variant pathway for human brain development.  Dev Cogn Neurosci. 2017;25:5-11. doi:10.1016/j.dcn.2017.02.004PubMedGoogle ScholarCrossref
60.
Shukla  DK, Keehn  B, Müller  RA.  Regional homogeneity of fMRI time series in autism spectrum disorders.  Neurosci Lett. 2010;476(1):46-51. doi:10.1016/j.neulet.2010.03.080PubMedGoogle ScholarCrossref
61.
Satterthwaite  TD, Elliott  MA, Gerraty  RT,  et al.  An improved framework for confound regression and filtering for control of motion artifact in the preprocessing of resting-state functional connectivity data.  Neuroimage. 2013;64(1):240-256. doi:10.1016/j.neuroimage.2012.08.052PubMedGoogle ScholarCrossref
62.
Golestani  AM, Kwinta  JB, Khatamian  YB, Chen  JJ.  The effect of low-frequency physiological correction on the reproducibility and specificity of resting-state fMRI metrics: functional connectivity, ALFF, and ReHo.  Front Neurosci. 2017;11:546. doi:10.3389/fnins.2017.00546PubMedGoogle ScholarCrossref
63.
Smith-Collins  APR, Luyt  K, Heep  A, Kauppinen  RA.  High frequency functional brain networks in neonates revealed by rapid acquisition resting state fMRI.  Hum Brain Mapp. 2015;36(7):2483-2494. doi:10.1002/hbm.22786PubMedGoogle ScholarCrossref
64.
André  M, Lamblin  MD, d’Allest  AM,  et al.  Electroencephalography in premature and full-term infants: developmental features and glossary.  Neurophysiol Clin. 2010;40(2):59-124. doi:10.1016/j.neucli.2010.02.002PubMedGoogle ScholarCrossref
65.
Hakamada  S, Watanabe  K, Hara  K, Miyazaki  S, Kumagai  T.  Body movements during sleep in full-term newborn infants.  Brain Dev. 1982;4(1):51-55. doi:10.1016/S0387-7604(82)80101-0PubMedGoogle ScholarCrossref
66.
Rubenstein  E, Chawla  D.  Broader autism phenotype in parents of children with autism: a systematic review of percentage estimates.  J Child Fam Stud. 2018;27(6):1705-1720. doi:10.1007/s10826-018-1026-3PubMedGoogle ScholarCrossref
67.
Lyall  K, Constantino  JN, Weisskopf  MG, Roberts  AL, Ascherio  A, Santangelo  SL.  Parental social responsiveness and risk of autism spectrum disorder in offspring.  JAMA Psychiatry. 2014;71(8):936-942. doi:10.1001/jamapsychiatry.2014.476PubMedGoogle ScholarCrossref
68.
Banks  MS, Salapatek  P.  Acuity and contrast sensitivity in 1-, 2-, and 3-month-old human infants.  Invest Ophthalmol Vis Sci. 1978;17(4):361-365.PubMedGoogle Scholar
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Original Investigation
    Psychiatry
    April 5, 2019

    Social Brain Functional Maturation in Newborn Infants With and Without a Family History of Autism Spectrum Disorder

    Author Affiliations
    • 1Centre for the Developing Brain, School Biomedical Engineering & Imaging Sciences, King’s College London, St Thomas’ Hospital, London, United Kingdom
    • 2Institute of Psychiatry, Psychology & Neuroscience, Department of Forensic and Neurodevelopmental Sciences, King’s College London, Denmark Hill, London, United Kingdom
    • 3Sackler Institute for Translational Neurodevelopment, King’s College London, Denmark Hill, London, United Kingdom
    • 4MRC Centre for Neurodevelopmental Disorders, King’s College London, London, United Kingdom
    • 5Medway Hospital, Gillingham, Kent, United Kingdom
    • 6South London and Maudsley NHS Foundation Trust, London, United Kingdom
    • 7Department of Bioengineering, Imperial College London, London, United Kingdom
    JAMA Netw Open. 2019;2(4):e191868. doi:10.1001/jamanetworkopen.2019.1868
    Key Points español 中文 (chinese)

    Question  Are changes in the maturation of the social brain in infants associated with vulnerability to neurodevelopmental conditions such as autism spectrum disorder?

    Findings  In this cohort study of 36 neonates with and without a family history of autism spectrum disorder, newborns with a family history of autism spectrum disorder had significantly higher neural activity in the right fusiform and left parietal cortex. In addition, the pattern of age-related changes in spontaneous activity in the cingulate and insula was disrupted in infants with a family history of autism spectrum disorder.

    Meaning  Atypical development of functional activity patterns in key regions responsible for social processing may be a vulnerability mechanism for autism.

    Abstract

    Importance  What is inherited or acquired in neurodevelopmental conditions such as autism spectrum disorder (ASD) is not a fixed outcome, but instead is a vulnerability to a spectrum of traits, especially social difficulties. Identifying the biological mechanisms associated with vulnerability requires looking as early in life as possible, before the brain is shaped by postnatal mechanisms and/or the experiences of living with these traits. Animal studies suggest that susceptibility to neurodevelopmental disorders arises when genetic and/or environmental risks for these conditions alter patterns of synchronous brain activity in the perinatal period, but this has never been examined in human neonates.

    Objective  To assess whether alternation of functional maturation of social brain circuits is associated with a family history of ASD in newborns.

    Design, Setting, and Participants  In this cohort study of 36 neonates with and without a family history of ASD, neonates underwent magnetic resonance imaging at St Thomas Hospital in London, England, using a dedicated neonatal brain imaging system between June 23, 2015, and August 1, 2018. Neonates with a first-degree relative with ASD (R+) and therefore vulnerable to autistic traits and neonates without a family history (R−) were recruited for the study. Synchronous neural activity in brain regions linked to social function was compared.

    Main Outcomes and Measures  Regions responsible for social function were selected with reference to a published meta-analysis and the level of synchronous activity within each region was used as a measure of local functional connectivity in a regional homogeneity analysis. Group differences, controlling for sex, age at birth, age at scan, and group × age interactions, were examined.

    Results  The final data set consisted of 18 R+ infants (13 male; median [range] postmenstrual age at scan, 42.93 [40.00-44.86] weeks) and 18 R− infants (13 male; median [range] postmenstrual age at scan, 42.50 [39.29-44.58] weeks). Neonates who were R+ had significantly higher levels of synchronous activity in the right posterior fusiform (t = 2.48; P = .04) and left parietal cortices (t = 3.96; P = .04). In addition, there was a significant group × age interaction within the anterior segment of the left insula (t = 3.03; P = .04) and cingulate cortices (right anterior: t = 3.00; P = .03; left anterior: t = 2.81; P = .03; right posterior: t = 2.77; P = .03; left posterior: t = 2.55; P = .03). In R+ infants, levels of synchronous activity decreased over 39 to 45 weeks’ postmenstrual age, whereas synchronous activity levels increased in R− infants over the same period.

    Conclusions and Relevance  Synchronous activity is required during maturation of functionally connected networks. This study found that in newborn humans, having a first-degree relative with ASD was associated with higher levels of local functional connectivity and dysmaturation of interconnected regions responsible for processing higher-order social information.

    Introduction

    Social cognition refers to “the ability to perceive and process information from others, from one’s self and interpersonal knowledge.”1 The human brain has evolved to become more highly specialized in social interaction and communication than brains of any other species.2 This ability starts maturing early in life and quickly becomes highly complex through building on fundamental processes such as basic sensory integration and emotion recognition.3 Later in development, social function incorporates higher-order processes, such as mental state attribution, involving integrated activity in several regions across the whole brain. These regions make up the “social brain,” which has recently been defined by a meta-analysis of neuroimaging studies in social neuroscience as comprising 36 distinct regions. These can be further subgrouped into 4 main clusters: high-level subnetworks (involving attention, memory, and executive function), intermediate-level networks (responsible for cohesion of information in the social context), and visual-sensory and limbic networks (responsible for perception and integration of relevant cues).4

    Anomalies in social interaction and communication are a core characteristic of autism spectrum disorder (ASD).5 It therefore comes as no surprise that abnormalities in functional connectivity within many of the regions composing the social brain have previously been reported in ASD. For example, ASD deficits in processing facial features have been associated with altered connectivity in the fusiform gyrus,6,7 difficulty with social demands has been associated with poor connectivity in the cingulate cortex,8 and aberrant processing of social stimuli has been associated with atypical activation patterns in the insula.9

    However, the nature and extent of social brain anomalies associated with ASD and related conditions varies considerably. Moreover, the outcomes of those at risk of, or even diagnosed with, this neurodevelopmental condition are not fixed. This is likely a result of the multiple and complex primary and secondary compensatory mechanisms that occur throughout life. For example, well before difficulties become clinically evident at around 3 to 4 years of age,10,11 genetic expression is highly dynamic (with certain genes only expressing in particular developmental stages),12 and exposure to a wide range of potential protective (as well as harmful) environmental factors increases the heterogeneity of individuals with neurodevelopmental conditions.13,14 Also, we now know that parental interventions can modify the outcomes of infants at risk.15,16 Thus, what is inherited or acquired in autistic conditions is unlikely to be a specific disorder, but rather is a vulnerability to a spectrum of traits—and this may be modulated by a variety of factors throughout later development.13 Hence, to truly understand the biological mechanisms that confer vulnerability to ASD, we must look as early in life as possible.

    Early-life studies of vulnerability mechanisms relevant to neurodevelopmental disorders have been largely conducted in animal models. Collectively, these reveal that synchronous activity in the perinatal brain is of fundamental importance as it is associated with local circuit connectivity and the development of functional networks.17 Disruption to these fundamental activity patterns is consistently reported in animals carrying either a gene mutation linked to ASD or exposed to prenatal environmental risk factors for ASD.18-20 Thus, based on the preclinical literature, early atypical synchronous activity may be a signature of vulnerability to neurodevelopmental conditions. However, this hypothesis has not been tested in humans, and specifically human neonates.

    Although in the animal literature there are many types of synchronous activity described in the developing brain (eg, activity dependent on external experience or spontaneous intrinsic activity, characterized by different frequency oscillations and regulated by different receptors that rapidly change in early development21,22), in humans we are much more constrained by limited methods of data collection and analysis. Here we acquired resting state functional magnetic resonance imaging (fMRI) data in neonates during natural sleep. Our measure of synchronous activity was the Kendall correlation between blood oxygenation level dependent (BOLD) signal measured with resting state fMRI (natural sleep) at the individual voxel (volume unit) level and its nearest neighbors, namely, regional homogeneity (ReHo).23 Recognizing that infants with a family history of ASD have a higher likelihood of developing ASD traits than children without a family history, we defined vulnerable infants as those with a first-degree relative with ASD (R+). The R− infants had no family history of these conditions. We used ReHo analysis as a measure of local synchronous activity to compare regional differences in functional connectivity in the social brain. We predicted that synchronous activity patterns would be atypical in the social brain of vulnerable neonates.

    Methods

    A total of 73 neonates were scanned at full term (>37 weeks) as part of the European Autism Interventions (EU-AIMS) Brain Imaging in Babies (BIBs) study approved by the South London National Research Ethics Committee from 2014 to the present. Written informed consent to participate in the study was taken from the parents of all participants prior to data acquisition. The study was conducted in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies. At recruitment, participants were first carefully screened by experienced researchers to determine whether parents reported a family history of ASD. Next, at the time of the scan, a pediatric research nurse with a comprehensive understanding of the UK National Health Service diagnostic provisions for neurodevelopmental conditions interviewed the parent(s). This included taking an extended medical and psychiatric family history and informing each participant’s general practitioner of their participation in the study. A neonate was characterized as at risk if his or her parent provided a clear account that a parent and/or an older sibling had been diagnosed by the appropriate clinical services (namely, the UK Child Development Centres, Child and Adolescent Mental Health Services, or Adult Autism Assessment Services). The preparation of the neonate for scanning and monitoring in the scan took priority on the day and considerable time was spent explaining the protocol and ensuring the parents were well informed about the process. Therefore, if any details of the family risk history were not captured on the scanning day, those parents were also contacted again after the scan.

    Twenty-nine families had either a previous child or a parent diagnosed with ASD, and 44 had no previous clinical conditions in their family history. After applying strict criteria to exclude fMRI data corrupted by head motion (due to high framewise displacement; see Data Processing subsection), results from 11 neonates were discarded from the original 29 in the R+ group and 11 were discarded from the original 44 in the R− group. This resulted in a data set consisting of 51 participants. Of these, 18 infants had a first-degree relative with ASD (and were therefore considered at risk of adverse neurodevelopmental outcomes) and were assigned to the R+ group; the remaining 33 composed the R− group. To control for an imbalance in sex (the R+ group had a higher ratio of male infants) 15 female infants from the R− group were randomly excluded. The final study group, therefore, consisted of 18 R+ and 18 R− participants (the characteristics of the sample are described in detail in Table 1). In the R+ group, 15 infants had a first-degree relative with ASD and 3 had a first-degree relative with both ASD and attention-deficit/hyperactivity disorder. In the R− group, there was no significant family history of neurodevelopmental conditions or major mental illness. Mann-Whitney U tests revealed no significant differences between groups in gestational age at birth, postmenstrual age at scan, weight at birth, and head circumference at scan. The gender distribution between groups was identical (detailed information of each participant is described in Table 2). Infants had no evidence of structural malformation on structural MR images (as reported by a neonatal neuroradiologist).

    Data Acquisition

    Infants were studied during natural sleep following feeding and immobilization in a vacuum evacuated bag (Med-Vac; CFI Medical Solutions). Hearing protection (molded dental putty in the external auditory meatus [President Putty; Coltene Whaledent] and earmuffs [MiniMuffs; Natus Medical Inc]) and physiological monitoring (oxygen saturations, heart rate, and axillary temperature) were applied before data acquisition. Data acquisition sessions were attended by a clinician (pediatric physician or nurse) trained in neonatal resuscitation. While the newborn was in the scanner, the parents were in a waiting room with a video link that allowed them to view their baby at all times. High–temporal resolution resting-state fMRI data were acquired in a 3T Philips Achieva scanner located on the neonatal intensive care unit at St Thomas’ Hospital London with a customized 32-channel neonatal head coil and imaging system.24 We acquired BOLD fMRI data using a multislice gradient-echo echo planar imaging sequence with multiband excitation (multiband factor 9) with the following parameters: repetition time, 392 milliseconds; echo time, 38 milliseconds; spatial resolution, 2.15 mm isotropic; flip angle, 34°; 45 slices; total time, 15 minutes 3 seconds (2300 volumes). High-resolution anatomical T1-weighted images (spatial resolution, 0.8 mm isotropic; field of view, 145 × 122 × 100 mm; and repetition time, 4795 milliseconds) and T2-weighted images (spatial resolution, 0.8 mm isotropic; field of view, 145 × 145 × 108 mm; and parameters of repetition time, 12 000 milliseconds and echo time, 156 milliseconds) were acquired for registration and clinical reporting purposes.24 Further details about the imaging protocol, which takes approximately 1 hour, can be found in the article by Hughes et al.24

    Data Processing

    Data preprocessing was carried out using an in-house pipeline specifically developed and optimized for sources of artifact inherent to neonatal fMRI data, including uncontrolled head motion and developmental differences in brain configuration. As head motion during fMRI data acquisition has been shown to significantly alter analysis and can lead to spurious patterns of connectivity,25,26 the acquired data were then cropped to 1600 continuous volumes with the least amount of motion (estimated by calculating the framewise displacement of each volume) for further analysis. Participants with high framewise displacement (>0.5 mm) in more than 80 volumes of the 1600 total volumes (5% of the cropped data set) were excluded entirely from analysis. Independent component analysis was then used to identify signal artifacts (those related to residual head motion, multiband acquisition, and physiological fluctuations, eg, cardiac pulsation and respiratory motion) and to train a classification algorithm for automatic identification of these sources of artifact, which were then regressed out from the data using the FSL FIX tool.27,28 Next, we used the FSL topup tool to estimate the warp embedding field distortions caused by susceptibility artifacts.29 We use this warp, combined with each participant’s anatomical T2, to nonlinearly transform each participant’s functional data to a 41-week-old neonatal template.30,31

    Statistical Analysis

    Voxelwise regional homogeneity was assessed by calculating the Kendall coefficient of concordance between the BOLD contrast time series of a given voxel with its 26 adjacent neighbors using the Analysis of Functional Neuroimages (AFNI) 3dReHo tool.23,32 Voxelwise ReHo Kendall τ was converted into Pearson r (r = sin[πτ/2]) and then transformed into Fisher z scores prior to statistical comparison.33 A neonatal labeled segmentation atlas34,35 was used to select the regions of interest (ROIs) pertinent to social function. The social brain ROIs were selected based on the recent meta-analysis that reported the social brain to be composed of 36 regions.4 Owing to immature anatomy and unclear boundaries in the neonatal brain, we restricted our study to 16 ROIs that could be clearly delineated on the neonatal atlas. These ROIs encompassed the anterior and posterior fusiform gyri, the anterior and posterior cingulate cortices, the posterior superior temporal sulcus, the insula cortex (Figure 1), the parietal cortex, and the frontal cortex, all bilaterally. The anterior temporal lobe was selected as a control region, as it is thought to rely on the use of abstract concepts in social cognitive processes, which are still very immature at this stage of human development.

    Group differences in ReHo z scores were tested within each ROI using permutation testing as implemented in FSL’s Randomize (v2.1)36 with false discovery rate correction for multiple comparisons (significance threshold set to P < .05 based on 2-tailed analysis). The general linear model was controlled for gestational age at birth, postmenstrual age at scan, and sex. A separate general linear model was used to test the hypothesis that there would be group × age–related differences in ReHo within the social brain network of interest. This model was controlled for sex (false discovery rate–corrected significance threshold, P < .05). Finally, both models were used for a whole-brain exploratory analysis to determine whether any between-group or age-related differences observed within the social ROI analyses could be generalized across the rest of the brain (false discovery rate–corrected significance threshold, P < .05).

    Results
    Main Effect on Synchronous Activity

    The final data set consisted of 18 R+ infants (13 male; median [range] postmenstrual age at scan, 42.93 [40.00-44.86] weeks) and 18 R− infants (13 male; median [range] postmenstrual age at scan, 42.50 [39.29-44.58] weeks). Compared with the R− group, infants in the R+ group had higher ReHo levels in the right posterior fusiform (t = 2.48; P = .04) and the left parietal cortex (t = 3.96; P = .04) (Figure 2). This included the fusiform face area in the right hemisphere (Figure 2). There were no other significant differences in average ReHo values (either higher or lower) between groups in any of the other social ROIs studied. There were also no significant differences in average ReHo in the control regions (anterior temporal lobes) or across the whole brain between groups (eFigure in the Supplement).

    Age Interaction Effect

    A separate voxelwise test of the interaction between groups and the linear effect of postmenstrual age on ReHo values was performed for each of the social brain ROIs. Significant group × age interactions were identified in the left insula (t = 3.03; P = .04), right (t = 3.00; P = .03) and left (t = 2.81; P = .03) anterior cingulate, and right (t = 2.77; P = .03) and left (t = 2.55; P = .03) posterior cingulate cortices (Figure 3), with R+ participants having higher ReHo levels compared with R− neonates at a younger postmenstrual age, while those scanned when older showed the reverse. Within these regions, there was a clear maturational pattern of increasing ReHo values over 39 to 45 weeks’ postmenstrual age in the R− group in comparison with a slight decrease (or minimal change) in R+ infants. This same trend was seen in each group for all other social brain regions, although it did not reach statistical significance. In contrast, there was no group × age interaction in ReHo either across the whole brain or within the preselected anterior temporal lobe control regions (eFigure in the Supplement).

    Discussion

    In this study, we show that by the normal time of birth, regions involved in social cognition had already begun to mature differently in neonates vulnerable to autistic conditions. We report higher ReHo levels localized to the posterior right fusiform gyrus (including the face-specialized region in at-risk neonates) and left parietal cortex. We also report significantly different age-related cross-sectional changes within the bilateral cingulate and the insular cortices. These findings in R+ infants were specific to the social brain; we observed no group differences in ReHo in a predefined control region (anterior temporal lobe), nor any global or maturational differences in local functional connectivity across the whole brain.

    Group Differences in the Social Brain

    Located on the ventral surface of the temporal cortex, the fusiform gyrus is a key structure in high-level visual functions such as face recognition, color processing, and word recognition.37-39 The structure is distinct to hominoids and is likely to be one of the earliest areas involved in social cognitive tasks to mature according to the interactive specialization theory. This theory postulates connectivity and activity patterns in early maturing networks, such as the fusiform,40 will interact with other regions to ensure later maturing networks acquire the intended social cognitive function.41 Among the 4 social clusters, the fusiform is in the visual sensory cluster, which is responsive to biologically relevant cues related to perception-action processes in social cognition and essential for a person to interact with her or his environment.4 Despite their relatively limited visual capacities, neonates have strong preferences toward facelike stimuli. They even demonstrate characteristics of mature face recognition abilities, such as an ability to discriminate between different emotional expressions.42-44 Differences in face processing emerge in infancy in ASD, as evidenced by a recent study reporting different attention levels in R+ newborn infants compared with an R− cohort.45 This is consistent with our finding that there are already differences between the R+ and R− groups that involve the fusiform face area, particularly in the right posterior fusiform, which is known to be associated with face processing.46,47 Thus, increased synchronous activity in the social brain and fusiform face area, specifically, may constitute a vulnerability marker for social traits in ASD, and future studies might benefit from combining this type of analysis with behavioral measures of attention to facelike stimuli.

    In addition to differences in the fusiform gyrus, we also observed significant differences between groups in the left parietal lobe. The left parietal cortex is part of the default mode network, showing strong functional connectivity with the frontal cortex, and has been associated with self and social information retrieval.48 The involvement of this part of the default mode network in our findings fits with evidence that the posterior regions of this network mature prior to frontal cortices.49 Thus, our observations may indicate atypical maturation of a network associated with social priming during rest; this is consistent with recent studies in adolescents with ASD that have found an association between underconnectivity in this area and atypical self-other processing abilities.8

    Group Age-Associated Differences in Early Development

    Face-processing difficulties are thought to be a precursor of the alterations in higher-order processing, such as joint attention, that develop later in infancy.50 Joint attention skills develop in the first 6 months following birth and require the establishment of a common frame of reference between the infant and the adult in order to share information about an object or an event. The primary nodes of the mature joint attention system include the frontal cortex, the insula, the cingulate, the superior temporal cortex, the precuneus, and the amygdala.51 Of these regions, we were constrained by the neonatal atlas used for this study and could only investigate the insula, the cingulate, and part of the superior temporal cortex. These regions are all components of the intermediate level cluster in the social cognition atlas, with the anterior cingulate also being involved in the limbic cluster.4 The interaction effect analysis found that R+ neonates show different age-associated changes in the insula and cingulate. While synchronous activity appears to be increasing with age in the R− group, a change with increasing age is not evident in R+ infants.

    Increasing functional connectivity with age at this early phase of development is in agreement with previously reported results of functional abnormalities in the insula in the neurodevelopmental spectrum. The insula has been found to be an important functional hub region in the infant brain, with significantly increases in mean functional connectivity demonstrated in neonates to 1 year, but not during the second year in typical development.52 Interestingly, studies in older children with ASD have reported both greater activity in the insula when performing an attention task9 and lower activity in the insula when tested for theory of mind.53 This variability also extends to other regions such as the cingulate, within which lower activity has been reported during tasks involving theory of mind,53 while others have reported higher intranetwork cingulate connectivity at rest.54 These discrepancies in results from investigations of older populations may reflect different compensatory activity patterns during task and rest situations, or an aberrant activity pattern that accompanies different maturational profiles in this heterogeneous spectrum. The challenges interpreting data from older cohorts underscore the need to minimize confounding factors and examine vulnerability mechanisms in early development. Until recently, we have had to depend on animal studies to examine early development. While these have successfully provided insight,18,19 they fall short of replicating the complex social brain of humans. Most notably, rodents do not have a specialized face-processing region in the brain, so work in human infants is essential.2

    Absence of Differences in Regions Involved in Higher-Order Processes

    The anterior temporal lobe is considered critical for merging the multicategorical information required to process very abstract social representations of another person.55-57 However, this form of abstract higher-order processing is likely relatively underdeveloped in a newborn infant, which therefore made the region a suitable control area for our purposes. We observed no differences between groups in these regions with any of the models tested, further suggesting that vulnerability to atypical development is time dependent and differs among regions involved in social cognition that mature at different developmental points.

    A Vulnerability Marker for Autistic Traits

    The outcomes of individuals at risk of, or even diagnosed with, ASD are highly diverse.14 Therefore, a vulnerability marker for ASD should ideally reflect a mechanism that allows for heterogeneity of outcomes (ie, it should be modifiable, for example, by postnatal experience). The higher levels of ReHo observed here may meet this requirement. In animal systems, under the influence of the external postnatal environment, spontaneous activity typically becomes sparse once a region is mature but can follow different trajectories depending on genes and environment.58 We suggest that parallels can be drawn with the work carried out in infants here. Thus, we propose that the abnormal prolongation of synchronous neural activity in R+ infants is a potential marker of immaturity, and multifactorial influences in postnatal life will subsequently determine the rate and extent to which a mature pattern of activity is achieved. What causes higher ReHo in the R+ at birth is, however, difficult to know. One possibility is that it is a response to disruption within a prenatal sensitive period59; whether this is a primary pathological or a secondary compensatory mechanism needs to be investigated further in both preclinical and clinical work that not only follows up children at risk but also examines precursors in the prenatal period.

    Limitations

    Our dedicated neonatal MR imaging system,24 optimized high–temporal resolution fMRI, and data analysis pipeline allowed us to mitigate specific challenges associated with neonatal fMRI studies. For example, the high spatial resolution of our fMRI data (2.15 mm isotropic) and the spatial normalization of our data to an age-specific structural atlas30 limited partial volume errors and helped constrain analysis to the cortex.60 Because head motion can lead to signal artifact and affect estimates of functional connectivity,61 we applied strict data exclusion criteria for head motion, analyzing only 1600 continuous volumes of the total 2300 acquired in each data set to ensure that absolute head displacement was less than 0.5 mm for 95% of the data and discarding data sets entirely from infants in whom greater than 5% of the acquired volumes were corrupted by a framewise displacement of more than 0.5 mm. Although this reduced the total duration of the available time series for each infant to 10 minutes 27 seconds, the very high temporal resolution of the data ensured that our ReHo measurements remained robust.62 This high sampling rate also allowed improved removal of physiological noise, which can be problematic in infants with naturally higher breathing and cardiac pulsation rates.63 It is also known that sleep state can alter functional connectivity measures and may represent a potential confounding factor in our study. However, neonatal sleep consists of only 2 stages (active and quiet),64 and it is currently unknown whether connectivity measures differ during these early sleep states in a way similar to how they do in the mature brain. In addition, given that head motion may be a proxy for the level of arousal,65 because we discarded data with large degrees of head motion, our population was likely to be predominately in quiet sleep during study. Nonetheless, we aim to better characterize the effect of physiological parameters and sleep in future studies.

    Another limitation, or rather a novelty, in our study is the inclusion of a parent with ASD as a risk for ASD in the offspring. This is based on the assumption that for a child with ASD (the sibling) there was likely a causal contribution of one or both parental genetic risk. Thus, it is reasonable to consider the parents as conferring risk to their offspring. This is supported by evidence that the parents of children with ASD have high levels of autistic traits.66,67 In addition, many families who go on to have a child with ASD remain singleton families, and these have been relatively neglected in the at-risk design. By including children without a sibling in the at-risk group we hope to be more representative of those families. We also elected to include neonates with a half sibling with ASD and to assign these children to the R+ group. We would expect that if these children do have a lower risk of ASD than the rest of the group, they would lessen the chance of us observing significant differences on the measures analyzed. A further limitation is assigning risk status based on family history, which excludes infants who might have a de novo predisposition to ASD. It is now appreciated that risk for ASD is multifactorial and includes, for example, prenatal exposure to maternal depression and/or antidepressant medication, inflammation, and birth complications. Thus, future studies will be needed to examine whether a dysmaturation within the social brain can be generalized to these other risk groups. The interpretation of our findings may be potentially limited by the small sample size, which may have led to false negative findings as a consequence of low power. To mitigate this, we used nonparametric methods throughout our analyses to avoid bias through making assumptions about the data distribution. However, it will be crucial to replicate the study in a larger data set. We also acknowledge that a truly longitudinal design would be helpful, but for practical purposes having very young infants attend repeatedly even over the first weeks of life is challenging, let alone within the short periods over which the brain changes so dynamically in infancy. We acknowledge the relatively narrow age range of our study. However, we view this as a strength, as during this period of infancy the child develops incredibly rapidly. For example, the sensitivity to contrast in visual stimuli rapidly changes after birth.68

    Conclusions

    We found that newborns with vulnerability to ASD showed significant differences in synchronous activity levels and maturation within core components of the social brain. This implies that a deviation from the typical maturational trajectory of the social brain circuits must be emergent in utero. Future fetal and neonatal imaging studies will help map the development of the social brain and could usefully examine what (diet, stress, or inflammation) might modify activity in the typical infant or even what might normalize aberrant activity in children who are vulnerable to adverse outcomes.

    Back to top
    Article Information

    Accepted for Publication: February 10, 2019.

    Published: April 5, 2019. doi:10.1001/jamanetworkopen.2019.1868

    Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2019 Ciarrusta J et al. JAMA Network Open.

    Corresponding Authors: Grainne McAlonan, MA, MBBS, PhD, Department of Forensic and Neurodevelopmental Science, Institute of Psychiatry, Psychology & Neuroscience, King’s College London, London SE5 8AF, United Kingdom (grainne.mcalonan@kcl.ac.uk); Tomoki Arichi, MBChB, FRCPCH, PhD, Centre for the Developing Brain, Biomedical Engineering and Imaging Sciences, King’s College London, London SE1 7EH, United Kingdom (tomoki.arichi@kcl.ac.uk).

    Author Contributions: Ms Ciarrusta had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Drs Murphy, Edwards, Arichi, and McAlonan contributed equally to the work.

    Concept and design: Ciarrusta, O'Muircheartaigh, Batalle, Murphy, Edwards, Arichi, McAlonan.

    Acquisition, analysis, or interpretation of data: Ciarrusta, O'Muircheartaigh, Dimitrova, Batalle, Cordero-Grande, Price, Hughes, Steinweg, Kangas, Perry, Javed, Stoencheva, Akolekar, Victor, Hajnal, Murphy, Arichi, McAlonan.

    Drafting of the manuscript: Ciarrusta, Arichi, McAlonan.

    Critical revision of the manuscript for important intellectual content: All authors.

    Statistical analysis: Ciarrusta, O'Muircheartaigh, Dimitrova, Batalle, Arichi.

    Obtained funding: Murphy, Edwards, McAlonan.

    Administrative, technical, or material support: Ciarrusta, Dimitrova, Cordero-Grande, Price, Hughes, Kangas, Perry, Javed, Stoencheva, Akolekar, Hajnal, Murphy, Arichi.

    Supervision: Murphy, Edwards, Arichi, McAlonan.

    Conflict of Interest Disclosures: Dr Cordero-Grande reported grants from the European Research Council and the Medical Research Council during the conduct of the study. Dr Price reported grants from the European Research Council, the Medical Research Council, Wellcome EPSRC, and the National Institute for Health Research during the conduct of the study. Dr Hughes reported grants from the European Research Council during the conduct of the study. Dr Steinweg reported grants from the Wellcome Engineering and Physical Sciences Research Council, the Medical Research Council, the European Union's Seventh Framework Programme, and the European Research Council during the conduct of the study. Dr Murphy reported grants from the European Union Innovative Medicines Initiative and the National Institute for Health Research Mental Health Biomedical Research Centre during the conduct of the study; and personal fees from Roche outside the submitted work. Dr Edwards reported grants from the Medical Research Council, the Wellcome Trust, the European Research Council, and the National Institute of Health Research during the conduct of the study. Dr Arichi reported grants from the Medical Research Council UK during the conduct of the study. No other disclosures were reported.

    Funding/Support: Infrastructure support was provided by the National Institute for Health Research Mental Health Biomedical Research Centre at South London, Maudsley NHS Foundation Trust, King's College London, the National Institute for Health Research Mental Health Biomedical Research Centre at Guys, and St Thomas’ Hospitals NHS Foundation Trust. The study was supported in part by the Wellcome Engineering and Physical Sciences Research Council Centre for Medical Engineering at King’s College London (grant WT 203148/Z/16/Z) and the Medical Research Council (United Kingdom) (grants MR/K006355/1 and MR/LO11530/1). Ms Ciarrusta received funding from the BRAINVIEW European Training Network for Early Stage Researchers. Drs O'Muircheartaigh, Murphy, Edwards, and McAlonan received support from the Medical Research Council Centre for Neurodevelopmental Disorders, King’s College London (grant MR/N026063/1). Dr Arichi was supported by a Medical Research Council Clinician Scientist Fellowship (grant MR/P008712/1). Support was also provided by the Sackler Institute for Translational Neurodevelopment at King’s College London, the European Autism Interventions (EU AIMS) trial, and the EU AIMS-2 trial. The EU-AIMS receives support from the Innovative Medicines Initiative Joint Undertaking under grant agreement 115300, resources of which are composed of financial contribution from the European Union’s Seventh Framework Programme (grant FP7/2007-2013), from the European Federation of Pharmaceutical Industries and Associations companies’ in kind contribution, and from Autism Speaks. The AIMS-2-TRIALS received funding from the Innovative Medicines Initiative 2 Joint Undertaking (grant 777394) Developing Human Connectome Project.

    Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    Disclaimer: The views expressed are those of the authors and not necessarily those of the National Health Service, the National Institute for Health Research, or the Department of Health.

    Additional Contributions: We thank the Brain Imaging in Babies team and National Health Service recruitment sites who supported the study, namely, St Georges Hospital London and Medway Maritime Hospital Maidstone. We also thank the families who participated.

    References
    1.
    Beer  JS, Ochsner  KN.  Social cognition: a multi level analysis.  Brain Res. 2006;1079(1):98-105. doi:10.1016/j.brainres.2006.01.002PubMedGoogle ScholarCrossref
    2.
    Herrmann  E, Call  J, Hernàndez-Lloreda  MV, Hare  B, Tomasello  M.  Humans have evolved specialized skills of social cognition: the cultural intelligence hypothesis.  Science. 2007;317(5843):1360-1366. doi:10.1126/science.1146282PubMedGoogle ScholarCrossref
    3.
    Thye  MD, Bednarz  HM, Herringshaw  AJ, Sartin  EB, Kana  RK.  The impact of atypical sensory processing on social impairments in autism spectrum disorder.  Dev Cogn Neurosci. 2018;29:151-167. doi:10.1016/j.dcn.2017.04.010PubMedGoogle ScholarCrossref
    4.
    Alcalá-López  D, Smallwood  J, Jefferies  E,  et al.  Computing the social brain connectome across systems and states.  Cereb Cortex. 2018;28(7):2207-2232. doi:10.1093/cercor/bhx121PubMedGoogle ScholarCrossref
    5.
    Lauritsen  MB.  Autism spectrum disorders.  Eur Child Adolesc Psychiatry. 2013;22(suppl 1):S37-S42. doi:10.1007/s00787-012-0359-5PubMedGoogle ScholarCrossref
    6.
    Lynn  AC, Padmanabhan  A, Simmonds  D,  et al.  Functional connectivity differences in autism during face and car recognition: underconnectivity and atypical age-related changes.  Dev Sci. 2018;21(1):e12508. doi:10.1111/desc.12508PubMedGoogle ScholarCrossref
    7.
    Fox  SE, Wagner  JB, Shrock  CL, Tager-Flusberg  H, Nelson  CA.  Neural processing of facial identity and emotion in infants at high-risk for autism spectrum disorders.  Front Hum Neurosci. 2013;7(89):89.PubMedGoogle Scholar
    8.
    Kana  RK, Sartin  EB, Stevens  C  Jr,  et al.  Neural networks underlying language and social cognition during self-other processing in autism spectrum disorders.  Neuropsychologia. 2017;102:116-123. doi:10.1016/j.neuropsychologia.2017.06.008PubMedGoogle ScholarCrossref
    9.
    Odriozola  P, Uddin  LQ, Lynch  CJ, Kochalka  J, Chen  T, Menon  V.  Insula response and connectivity during social and non-social attention in children with autism.  Soc Cogn Affect Neurosci. 2016;11(3):433-444. doi:10.1093/scan/nsv126PubMedGoogle ScholarCrossref
    10.
    Daniels  AM, Mandell  DS.  Explaining differences in age at autism spectrum disorder diagnosis: a critical review.  Autism. 2014;18(5):583-597. doi:10.1177/1362361313480277PubMedGoogle ScholarCrossref
    11.
    Lord  C, Jones  RM.  Annual research review: re-thinking the classification of autism spectrum disorders.  J Child Psychol Psychiatry. 2012;53(5):490-509. doi:10.1111/j.1469-7610.2012.02547.xPubMedGoogle ScholarCrossref
    12.
    Silbereis  JC, Pochareddy  S, Zhu  Y, Li  M, Sestan  N.  The cellular and molecular landscapes of the developing human central nervous system.  Neuron. 2016;89(2):248-268. doi:10.1016/j.neuron.2015.12.008PubMedGoogle ScholarCrossref
    13.
    Livingston  LA, Happé  F.  Conceptualising compensation in neurodevelopmental disorders: reflections from autism spectrum disorder.  Neurosci Biobehav Rev. 2017;80:729-742. doi:10.1016/j.neubiorev.2017.06.005PubMedGoogle ScholarCrossref
    14.
    Happé  F, Ronald  A, Plomin  R.  Time to give up on a single explanation for autism.  Nat Neurosci. 2006;9(10):1218-1220. doi:10.1038/nn1770PubMedGoogle ScholarCrossref
    15.
    Green  J, Pickles  A, Pasco  G,  et al; British Autism Study of Infant Siblings (BASIS) Team.  Randomised trial of a parent-mediated intervention for infants at high risk for autism: longitudinal outcomes to age 3 years.  J Child Psychol Psychiatry. 2017;58(12):1330-1340. doi:10.1111/jcpp.12728PubMedGoogle ScholarCrossref
    16.
    Dawson  G, Jones  EJ, Merkle  K,  et al.  Early behavioral intervention is associated with normalized brain activity in young children with autism.  J Am Acad Child Adolesc Psychiatry. 2012;51(11):1150-1159. doi:10.1016/j.jaac.2012.08.018PubMedGoogle ScholarCrossref
    17.
    Egorov  AV, Draguhn  A.  Development of coherent neuronal activity patterns in mammalian cortical networks: common principles and local hetereogeneity.  Mech Dev. 2013;130(6-8):412-423. doi:10.1016/j.mod.2012.09.006PubMedGoogle ScholarCrossref
    18.
    Peixoto  RT, Wang  W, Croney  DM, Kozorovitskiy  Y, Sabatini  BL.  Early hyperactivity and precocious maturation of corticostriatal circuits in Shank3B(-/-) mice.  Nat Neurosci. 2016;19(5):716-724. doi:10.1038/nn.4260PubMedGoogle ScholarCrossref
    19.
    Gonçalves  JT, Anstey  JE, Golshani  P, Portera-Cailliau  C.  Circuit level defects in the developing neocortex of fragile X mice.  Nat Neurosci. 2013;16(7):903-909. doi:10.1038/nn.3415PubMedGoogle ScholarCrossref
    20.
    Gutierrez  RC, Hung  J, Zhang  Y, Kertesz  AC, Espina  FJ, Colicos  MA.  Altered synchrony and connectivity in neuronal networks expressing an autism-related mutation of neuroligin 3.  Neuroscience. 2009;162(1):208-221. doi:10.1016/j.neuroscience.2009.04.062PubMedGoogle ScholarCrossref
    21.
    Allène  C, Cattani  A, Ackman  JB,  et al.  Sequential generation of two distinct synapse-driven network patterns in developing neocortex.  J Neurosci. 2008;28(48):12851-12863. doi:10.1523/JNEUROSCI.3733-08.2008PubMedGoogle ScholarCrossref
    22.
    Hanganu-Opatz  IL.  Between molecules and experience: role of early patterns of coordinated activity for the development of cortical maps and sensory abilities.  Brain Res Rev. 2010;64(1):160-176. doi:10.1016/j.brainresrev.2010.03.005PubMedGoogle ScholarCrossref
    23.
    Zang  Y, Jiang  T, Lu  Y, He  Y, Tian  L.  Regional homogeneity approach to fMRI data analysis.  Neuroimage. 2004;22(1):394-400. doi:10.1016/j.neuroimage.2003.12.030PubMedGoogle ScholarCrossref
    24.
    Hughes  EJ, Winchman  T, Padormo  F,  et al.  A dedicated neonatal brain imaging system.  Magn Reson Med. 2017;78(2):794-804. doi:10.1002/mrm.26462PubMedGoogle ScholarCrossref
    25.
    Satterthwaite  TD, Wolf  DH, Loughead  J,  et al.  Impact of in-scanner head motion on multiple measures of functional connectivity: relevance for studies of neurodevelopment in youth.  Neuroimage. 2012;60(1):623-632. doi:10.1016/j.neuroimage.2011.12.063PubMedGoogle ScholarCrossref
    26.
    Power  JD, Barnes  KA, Snyder  AZ, Schlaggar  BL, Petersen  SE.  Spurious but systematic correlations in functional connectivity MRI networks arise from subject motion.  Neuroimage. 2012;59(3):2142-2154. doi:10.1016/j.neuroimage.2011.10.018PubMedGoogle ScholarCrossref
    27.
    Griffanti  L, Salimi-Khorshidi  G, Beckmann  CF,  et al.  ICA-based artefact removal and accelerated fMRI acquisition for improved resting state network imaging.  Neuroimage. 2014;95:232-247. doi:10.1016/j.neuroimage.2014.03.034PubMedGoogle ScholarCrossref
    28.
    Salimi-Khorshidi  G, Douaud  G, Beckmann  CF, Glasser  MF, Griffanti  L, Smith  SM.  Automatic denoising of functional MRI data: combining independent component analysis and hierarchical fusion of classifiers.  Neuroimage. 2014;90:449-468. doi:10.1016/j.neuroimage.2013.11.046PubMedGoogle ScholarCrossref
    29.
    Andersson  JLR, Skare  S, Ashburner  J.  How to correct susceptibility distortions in spin-echo echo-planar images: application to diffusion tensor imaging.  Neuroimage. 2003;20(2):870-888. doi:10.1016/S1053-8119(03)00336-7PubMedGoogle ScholarCrossref
    30.
    Serag  A, Aljabar  P, Ball  G,  et al.  Construction of a consistent high-definition spatio-temporal atlas of the developing brain using adaptive kernel regression.  Neuroimage. 2012;59(3):2255-2265. doi:10.1016/j.neuroimage.2011.09.062PubMedGoogle ScholarCrossref
    31.
    Andersson  JLR, Jenkinson  M, Smith  S.  Non-linear Registration, aka Spatial Normalisation. Oxford, UK: FMRIB Centre; 2010.
    32.
    Taylor  PA, Saad  ZS.  FATCAT: (an efficient) functional and tractographic connectivity analysis toolbox.  Brain Connect. 2013;3(5):523-535. doi:10.1089/brain.2013.0154PubMedGoogle ScholarCrossref
    33.
    Walker  DA.  Converting Kendall’s tau for correlational or meta-analytic analyses.  J Mod Appl Stat Methods. 2003;2(2):525-530. doi:10.22237/jmasm/1067646360Google ScholarCrossref
    34.
    Gousias  IS, Hammers  A, Counsell  SJ,  et al.  Magnetic resonance imaging of the newborn brain: automatic segmentation of brain images into 50 anatomical regions.  PLoS One. 2013;8(4):e59990. doi:10.1371/journal.pone.0059990PubMedGoogle ScholarCrossref
    35.
    Gousias  IS, Edwards  AD, Rutherford  MA,  et al.  Magnetic resonance imaging of the newborn brain: manual segmentation of labelled atlases in term-born and preterm infants.  Neuroimage. 2012;62(3):1499-1509. doi:10.1016/j.neuroimage.2012.05.083PubMedGoogle ScholarCrossref
    36.
    Winkler  AM, Ridgway  GR, Webster  MA, Smith  SM, Nichols  TE.  Permutation inference for the general linear model.  Neuroimage. 2014;92:381-397. doi:10.1016/j.neuroimage.2014.01.060PubMedGoogle ScholarCrossref
    37.
    Devlin  JT, Jamison  HL, Gonnerman  LM, Matthews  PM.  The role of the posterior fusiform gyrus in reading.  J Cogn Neurosci. 2006;18(6):911-922. doi:10.1162/jocn.2006.18.6.911PubMedGoogle ScholarCrossref
    38.
    McKeefry  DJ, Zeki  S.  The position and topography of the human colour centre as revealed by functional magnetic resonance imaging.  Brain. 1997;120(Pt 12):2229-2242. doi:10.1093/brain/120.12.2229PubMedGoogle ScholarCrossref
    39.
    Kanwisher  N, McDermott  J, Chun  MM.  The fusiform face area: a module in human extrastriate cortex specialized for face perception.  J Neurosci. 1997;17(11):4302-4311. doi:10.1523/JNEUROSCI.17-11-04302.1997PubMedGoogle ScholarCrossref
    40.
    De Haan  M, Johnson  MH, Maurer  D, Perrett  DI.  Recognition of individual faces and average face prototypes by 1- and 3-month-old infants.  Cogn Dev. 2001;16(2):659-678. doi:10.1016/S0885-2014(01)00051-XGoogle ScholarCrossref
    41.
    Johnson  MH, Griffin  R, Csibra  G,  et al.  The emergence of the social brain network: evidence from typical and atypical development.  Dev Psychopathol. 2005;17(3):599-619. doi:10.1017/S0954579405050297PubMedGoogle ScholarCrossref
    42.
    Otsuka  Y.  Face recognition in infants: a review of behavioral and near-infrared spectroscopic studies.  Jpn Psychol Res. 2014;56(1):76-90. doi:10.1111/jpr.12024Google ScholarCrossref
    43.
    von Hofsten  O, von Hofsten  C, Sulutvedt  U, Laeng  B, Brennen  T, Magnussen  S.  Simulating newborn face perception.  J Vis. 2014;14(13):16. doi:10.1167/14.13.16PubMedGoogle ScholarCrossref
    44.
    Stein  T, Peelen  MV, Sterzer  P.  Adults’ awareness of faces follows newborns’ looking preferences.  PLoS One. 2011;6(12):e29361. doi:10.1371/journal.pone.0029361PubMedGoogle ScholarCrossref
    45.
    Di Giorgio  E, Frasnelli  E, Salva  OR,  et al.  Difference in visual social predispositions between newborns at low- and high-risk for autism.  Sci Rep. 2016;6:1-9. doi:10.1038/srep26395PubMedGoogle ScholarCrossref
    46.
    Jonas  J, Rossion  B, Brissart  H,  et al.  Beyond the core face-processing network: intracerebral stimulation of a face-selective area in the right anterior fusiform gyrus elicits transient prosopagnosia.  Cortex. 2015;72:140-155. doi:10.1016/j.cortex.2015.05.026PubMedGoogle ScholarCrossref
    47.
    Rangarajan  V, Parvizi  J.  Functional asymmetry between the left and right human fusiform gyrus explored through electrical brain stimulation.  Neuropsychologia. 2016;83:29-36. doi:10.1016/j.neuropsychologia.2015.08.003PubMedGoogle ScholarCrossref
    48.
    Pfeifer  JH, Lieberman  MD, Dapretto  M.  “I know you are but what am I?!”: neural bases of self- and social knowledge retrieval in children and adults.  J Cogn Neurosci. 2007;19(8):1323-1337. doi:10.1162/jocn.2007.19.8.1323PubMedGoogle ScholarCrossref
    49.
    Doria  V, Beckmann  CF, Arichi  T,  et al.  Emergence of resting state networks in the preterm human brain.  Proc Natl Acad Sci U S A. 2010;107(46):20015-20020. doi:10.1073/pnas.1007921107PubMedGoogle ScholarCrossref
    50.
    Dawson  G, Webb  SJ, McPartland  J.  Understanding the nature of face processing impairment in autism: insights from behavioral and electrophysiological studies.  Dev Neuropsychol. 2005;27(3):403-424. doi:10.1207/s15326942dn2703_6PubMedGoogle ScholarCrossref
    51.
    Mundy  P.  A review of joint attention and social-cognitive brain systems in typical development and autism spectrum disorder.  Eur J Neurosci. 2018;47(6):497-514. doi:10.1111/ejn.13720PubMedGoogle ScholarCrossref
    52.
    Alcauter  S, Lin  W, Keith Smith  J, Gilmore  JH, Gao  W.  Consistent anterior-posterior segregation of the insula during the first 2 years of life.  Cereb Cortex. 2015;25(5):1176-1187. doi:10.1093/cercor/bht312PubMedGoogle ScholarCrossref
    53.
    Kana  RK, Maximo  JO, Williams  DL,  et al.  Aberrant functioning of the theory-of-mind network in children and adolescents with autism.  Mol Autism. 2015;6(1):59. doi:10.1186/s13229-015-0052-xPubMedGoogle ScholarCrossref
    54.
    Venkataraman  A, Duncan  JS, Yang  DYJ, Pelphrey  KA.  An unbiased Bayesian approach to functional connectomics implicates social-communication networks in autism.  Neuroimage Clin. 2015;8:356-366. doi:10.1016/j.nicl.2015.04.021PubMedGoogle ScholarCrossref
    55.
    Wang  Y, Collins  JA, Koski  J, Nugiel  T, Metoki  A, Olson  IR.  Dynamic neural architecture for social knowledge retrieval.  Proc Natl Acad Sci U S A. 2017;114(16):E3305-E3314. doi:10.1073/pnas.1621234114PubMedGoogle ScholarCrossref
    56.
    Zahn  R, Moll  J, Krueger  F, Huey  ED, Garrido  G, Grafman  J.  Social concepts are represented in the superior anterior temporal cortex.  Proc Natl Acad Sci U S A. 2007;104(15):6430-6435. doi:10.1073/pnas.0607061104PubMedGoogle ScholarCrossref
    57.
    Olson  IR, McCoy  D, Klobusicky  E, Ross  LA.  Social cognition and the anterior temporal lobes: a review and theoretical framework.  Soc Cogn Affect Neurosci. 2013;8(2):123-133. doi:10.1093/scan/nss119PubMedGoogle ScholarCrossref
    58.
    Marín  O.  Developmental timing and critical windows for the treatment of psychiatric disorders.  Nat Med. 2016;22(11):1229-1238. doi:10.1038/nm.4225PubMedGoogle ScholarCrossref
    59.
    Johnson  MH.  Autism as an adaptive common variant pathway for human brain development.  Dev Cogn Neurosci. 2017;25:5-11. doi:10.1016/j.dcn.2017.02.004PubMedGoogle ScholarCrossref
    60.
    Shukla  DK, Keehn  B, Müller  RA.  Regional homogeneity of fMRI time series in autism spectrum disorders.  Neurosci Lett. 2010;476(1):46-51. doi:10.1016/j.neulet.2010.03.080PubMedGoogle ScholarCrossref
    61.
    Satterthwaite  TD, Elliott  MA, Gerraty  RT,  et al.  An improved framework for confound regression and filtering for control of motion artifact in the preprocessing of resting-state functional connectivity data.  Neuroimage. 2013;64(1):240-256. doi:10.1016/j.neuroimage.2012.08.052PubMedGoogle ScholarCrossref
    62.
    Golestani  AM, Kwinta  JB, Khatamian  YB, Chen  JJ.  The effect of low-frequency physiological correction on the reproducibility and specificity of resting-state fMRI metrics: functional connectivity, ALFF, and ReHo.  Front Neurosci. 2017;11:546. doi:10.3389/fnins.2017.00546PubMedGoogle ScholarCrossref
    63.
    Smith-Collins  APR, Luyt  K, Heep  A, Kauppinen  RA.  High frequency functional brain networks in neonates revealed by rapid acquisition resting state fMRI.  Hum Brain Mapp. 2015;36(7):2483-2494. doi:10.1002/hbm.22786PubMedGoogle ScholarCrossref
    64.
    André  M, Lamblin  MD, d’Allest  AM,  et al.  Electroencephalography in premature and full-term infants: developmental features and glossary.  Neurophysiol Clin. 2010;40(2):59-124. doi:10.1016/j.neucli.2010.02.002PubMedGoogle ScholarCrossref
    65.
    Hakamada  S, Watanabe  K, Hara  K, Miyazaki  S, Kumagai  T.  Body movements during sleep in full-term newborn infants.  Brain Dev. 1982;4(1):51-55. doi:10.1016/S0387-7604(82)80101-0PubMedGoogle ScholarCrossref
    66.
    Rubenstein  E, Chawla  D.  Broader autism phenotype in parents of children with autism: a systematic review of percentage estimates.  J Child Fam Stud. 2018;27(6):1705-1720. doi:10.1007/s10826-018-1026-3PubMedGoogle ScholarCrossref
    67.
    Lyall  K, Constantino  JN, Weisskopf  MG, Roberts  AL, Ascherio  A, Santangelo  SL.  Parental social responsiveness and risk of autism spectrum disorder in offspring.  JAMA Psychiatry. 2014;71(8):936-942. doi:10.1001/jamapsychiatry.2014.476PubMedGoogle ScholarCrossref
    68.
    Banks  MS, Salapatek  P.  Acuity and contrast sensitivity in 1-, 2-, and 3-month-old human infants.  Invest Ophthalmol Vis Sci. 1978;17(4):361-365.PubMedGoogle Scholar
    ×