Association of Polygenic Liability for Autism With Face-Sensitive Cortical Responses From Infancy

Autism is a heritable condition affecting 1% of people worldwide. Despite a pressing need for early intervention, the developmental paths through which genetic variants are associated with emerging behavioral symptoms in infancy remain opaque. The latency of the N170 event-related potential response to faces is replicably altered in individuals with autism¹ and has potential as a stratification biomarker for prognostic social functioning.² The N170 precursor (N290) to faces vs nonfaces is also altered prior to symptom emergence in infants subsequently diagnosed with autism.³ These early differences in brain processing represent a plausible developmental mechanism linking genetic liability and behavioral autism symptoms. We investigated whether N290 latency to faces vs nonfaces is associated with autism polygenic scores and cross-disorder polygenic scores in infants with and without a family history of autism.

In this cohort study, 104 infants with and without a family history of autism provided DNA and participated in an electroencephalography (EEG) task³ presenting face and nonface images as part of a longitudinal prospective study (the British Autism Study of Infant Siblings [BASIS]). Diagnostic assessments at age 3 years determined whether infants with a family history of autism were diagnosed with autism, showed typical development, or showed other signs of atypical development (Table). Ethical approval was obtained from the Health Research Authority of the English National Health Service. Parents gave written informed consent.

Infants viewed face or nonface (scrambled pixels of the face) images while brain electrical activity was measured continuously with a 128-channel Hydrocel Sensor Net System (Electrical Geodesics Inc). N290 latency was extracted for each condition (220 to 319 milliseconds; more than 10 good-quality EEG trials; mean of 19 occipitotemporal electrodes), and the difference in N290 latency between face and nonface stimuli was computed (face-nonface [F-N] N290 latency).

Genome-wide genotype data were obtained from saliva and buccal cheek-swab DNA.⁴ Standardized polygenic scores were calculated using PRSice-2 software in R version 3.6.3 (The R Foundation) for 234 unrelated infants of European ancestry, assigned by the investigators based on principal component analysis on a combined sample of infants’ and Hapmap3 genotypes. Autism polygenic scores and cross-disorder polygenic scores were generated using the Autism⁵ and Cross-Disorder⁶ European-based genome-wide association studies (GWAS) at a range of P value thresholds (.001 < threshold P ≤ 1). Linkage disequilibrium estimation for clumping (r²< 0.1; 250-kilobase distance from index variant) was based on the 1000 Genomes Project reference panel. Five ancestry principal components were included as covariates.

Regression analyses tested the association between F-N N290 latency and autism polygenic scores and cross-disorder polygenic scores at the GWAS P value thresholds that explained the highest variance (Nagelkerke R²) in infants with autism and a family history of autism and those with atypical development (whether autism or other). Model fit improvement was tested using χ² when adding autism polygenic scores to the logistic model that tested the association of F-N N290 latency with autism. Tests were 2-tailed and significance was set at P < .05. Details on diagnostic assessment, EEG, and genetic data preprocessing are available in the eMethods in the Supplement and online at https://github.com/annagui/PGS_EEG.

Of 104 infants included in this study, 53 were female (51.0%). The mean (SD) age was 8.3 (1.2) months. As previously reported,² infants later diagnosed with autism showed diminished differentiation between N290 latency to face and nonface stimuli relative to infants without a family history of autism (Table; Figure, A). Higher autism polygenic scores (threshold P = .01; number of single nucleotide variants = 4806; Nagelkerke R² = 0.054; P = .01) (Figure, B) was associated with shorter N290 latency to face vs nonface stimuli (β = −3.89; SE = 1.94; P = .047) (Figure, C). Cross-disorder polygenic score (threshold P = .50; number of single nucleotide variants = 59 669; Nagelkerke R² = 0.015; P = .15) was even more strongly associated with F-N N290 latency (β = −5.05; SE = 1.81; P = .006) (Figure, D). Testing the association between these precursors and autism (dependent variable), the model fit significantly improved when adding polygenic scores to F-N N290 latency as an independent variable (McFadden R² = 0.121; P = .008).

Altered cortical responses to social vs nonsocial stimuli in infancy may be one brain processing pathway through which genetic liability leads to behavioral autism symptoms and may suggest a suitable target for early identification. This study has limitations. The relatively small size and composition of the sample somewhat limit the generalizability of findings. Future studies should leverage larger GWAS and population infant samples, including those of non-European ancestry.

Accepted for Publication: March 30, 2021.

Published Online: June 7, 2021. doi:10.1001/jamapediatrics.2021.1338

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2021 Gui A et al. JAMA Pediatrics.

Corresponding Author: Anna Gui, PhD, Centre for Brain and Cognitive Development, Birkbeck College, University of London, Malet Street, London WC1E 7HX, United Kingdom (agui01@mail.bbk.ac.uk).

Author Contributions: Dr Gui 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.

Concept and design: Gui, Meaburn, Charman, Johnson, Jones.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Gui, Jones.

Critical revision of the manuscript for important intellectual content: Meaburn, Tye, Charman, Johnson, Jones.

Statistical analysis: Gui.

Obtained funding: Meaburn, Charman, Johnson, Jones.

Administrative, technical, or material support: Gui, Tye, Jones.

Supervision: Meaburn, Tye, Charman, Johnson, Jones.

Conflict of Interest Disclosures: Dr Gui reports grants from Marie Skłodowska-Curie and Economic and Social Research Council during the conduct of the study. Dr Meaburn received the Simons Foundation Autism Research Initiative (SFARI) Pilot Award during the conduct of the study. Dr Charman reports grants from Medical Research Council, European Union Horizon 2020, and Innovative Medicines Initiative during the conduct of the study; as well as personal fees from F. Hoffmann-La Roche and Servier and royalties from Sage Publications and Guilford Publications outside the submitted work. Dr Jones reports grants from Medical Research Council, Simons Foundation, Economic and Social Research Council, European Union Horizon 2020, and Innovative Medicines Initiative during the conduct of the study. No other disclosures were reported.

Funding/Support: The genetic data collection and generation and manuscript preparation were funded by grant 642996 from the European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie actions, grant ES/R009368/1 from the Economic and Social Research Council to Drs Gui and Jones, and by grant 511504 from the Simons Foundation Autism Research Initiative (SFARI) Pilot Award to Drs Jones and Meaburn. The experimental and behavioral data collection for all phases of the British Autism Study of Infant Siblings (BASIS) was funded by grants G0701484 and MR/K021389/1 from the Medical Research Council Programme, the BASIS funding consortium led by Autistica (http://www.basisnetwork.org). The electroencephalography and behavioral data collection used in this publication received partial funding from the Innovative Medicines Initiative Joint Undertaking under grant agreement 115300 for the European Autism Interventions—A Multicentre Study for Developing New Medications (EU-AIMS) project, which receives resources from the European Union’s Seventh Framework Programme (FP7/2007-2013) and the European Federation of Pharmaceutical Industries and Associations companies (EFPIA), and by the Innovative Medicines Initiative 2 Joint Undertaking under grant agreement 777394 for the project AIMS-2-TRIALS, which receives support from the European Union’s Horizon 2020 research and innovation program, the European Federation of Pharmaceutical Industries and Associations companies’ in-kind contributions, Autism Speaks, Autistica, and the Simons Foundation for Autism Research Initiative.

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: Any views expressed are those of the authors and not necessarily those of the funders.

Additional Contributions: We thank Jennifer K. Lowe, PhD (Department of Neurology, Psychiatry, and Human Genetics and Center for Autism Research and Treatment, Semel Institute, David Geffen School of Medicine, University of California, Los Angeles), for training on genotype data analysis; she was not compensated for her contribution to this work. We are very grateful to all the BASIS families for their participation in our study.