Prediction of Autism at 3 Years from Behavioural and Developmental Measures in High-Risk Infants: A Longitudinal Cross-Domain Classifier Analysis

Patterns of spontaneous neural activity associated with social communication abilities among infants and toddlers showing signs of autism

Early disruptions to social communication development, including delays in joint attention and language, are among the earliest markers of autism spectrum disorder (autism, henceforth). Although social communication differences are a core feature of autism, there is marked heterogeneity in social communication-related development among infants and toddlers exhibiting autism symptoms. Neural markers of individual differences in joint attention and language abilities may provide important insight into heterogeneity in autism symptom expression during infancy and toddlerhood. This study examined patterns of spontaneous electroencephalography (EEG) activity associated with joint attention and language skills in 70 community-referred 12- to 23-month-olds with autism symptoms and elevated scores on an autism diagnostic instrument. Data-driven cluster-based permutation analyses revealed significant positive associations between relative alpha power (6-9 Hz) and concurrent response to joint attention skills, receptive language, and expressive language abilities. Exploratory analyses also revealed significant negative associations between relative alpha power and measures of core autism features (i.e., social communication difficulties and restricted/repetitive behaviors). These findings shed light on the neural mechanisms underlying typical and atypical social communication development in emerging autism and provide a foundation for future work examining neural predictors of social communication growth and markers of intervention response.

Using the Griffiths Mental Development Scales to Evaluate a Developmental Profile of Children with Autism Spectrum Disorder and Their Symptomatologic Severity

Early diagnosis is crucial for Autism spectrum disorder (ASD) and is achieved through a screening of developmental indicators to recognise children who are at risk of autism. One of the most widely used instruments in clinical practice for assessing child development is the Griffiths Mental Development Scale (GMDS). We sought (a) to assess longitudinally whether children diagnosed with ASD, with a mean age of 33.50 months (SD 7.69 months), show a developmental delay of abilities measured by the GMDS over time and (b) to analyse which skills of the GMDS could be associate to the symptomatologic severity of ASD. Our results showed lower scores of General Quotient and all sub-quotients of GMDS from first (T0) to second assessment (T1), except for the Performance sub-quotient. Three sub-quotients (Personal-Social, Hearing and Language and Practical Reasoning) also associate symptom severity at the time when the diagnosis of ASD is made.

Parent-Child Interactions May Help to Explain Relations Between Parent Characteristics and Clinically Observed Child Autistic Behaviours

The importance of supporting parent-child interactions has been noted in the context of prodromal autism, but little consideration has been given to the possible contributing role of parental characteristics, such as psychological distress. This cross-sectional study tested models in which parent-child interaction variables mediated relations between parent characteristics and child autistic behaviour in a sample of families whose infant demonstrated early signs of autism (N = 103). The findings suggest that associations between parent characteristics (psychological distress; aloofness) and child autistic behaviours may be mediated by the child's inattentiveness or negative affect during interactions. These findings have important implications in developing and implementing interventions in infancy which target the synchrony of parent-child interaction with the goal to support children's social communication development.

Energy-Efficient EEG-Based Scheme for Autism Spectrum Disorder Detection Using Wearable Sensors

The deployment of wearable wireless systems that collect physiological indicators to aid in diagnosing neurological disorders represents a potential solution for the new generation of e-health systems. Electroencephalography (EEG), a recording of the brain's electrical activity, is a promising physiological test for the diagnosis of autism spectrum disorders. It can identify the abnormalities of the neural system that are associated with autism spectrum disorders. However, streaming EEG samples remotely for classification can reduce the wireless sensor's lifespan and creates doubt regarding the application's feasibility. Therefore, decreasing data transmission may conserve sensor energy and extend the lifespan of wireless sensor networks. This paper suggests the development of a sensor-based scheme for early age autism detection. The proposed scheme implements an energy-efficient method for signal transformation allowing relevant feature extraction for accurate classification using machine learning algorithms. The experimental results indicate an accuracy of 96%, a sensitivity of 100%, and around 95% of F1 score for all used machine learning models. The results also show that our scheme energy consumption is 97% lower than streaming the raw EEG samples.

Advancing research on early autism through an integrated risk and resilience perspective

To date, a deficit-oriented approach dominates autism spectrum disorder (ASD) research, including studies of infant siblings of children with ASD at high risk (HR) for the disabilities associated with this disorder. Despite scientific advances regarding early ASD-related risk, there remains little systematic investigation of positive development, limiting the scope of research and quite possibly a deeper understanding of pathways toward and away from ASD-related impairments. In this paper, we argue that integrating a resilience framework into early ASD research has the potential to enhance knowledge on prodromal course, phenotypic heterogeneity, and developmental processes of risk and adaptation. We delineate a developmental systems resilience framework with particular reference to HR infants. To illustrate the utility of a resilience perspective, we consider the "female protective effect" and other evidence of adaptation in the face of ASD-related risk. We suggest that a resilience framework invites focal questions about the nature, timing, levels, interactions, and mechanisms by which positive adaptation occurs in relation to risk and developmental pathways toward and away from ASD-related difficulties. We conclude with recommendations for future research, including more focus on adaptive development and multisystem processes, pathways away from disorder, and reconsideration of extant evidence within an integrated risk-and-resilience framework.

Factors related to retention in a longitudinal study of infants at familial risk for autism

Background

Reporting retention data is critical to determining the soundness of a study's conclusions (internal validity) and broader generalizability (external validity). Although selective attrition can lead to overestimates of effects, biased conclusions, or overly expansive generalizations, retention rates are not reported in many longitudinal studies.

Methods

We examined multiple child- and family-level factors potentially associated with retention in a longitudinal study of younger siblings of children with autism spectrum disorder (ASD; n = 304) or typical development (n = 163). The sample was followed from the first year of life to 36 months of age, for up to 7 visits.

Results

Of the 467 infant siblings who were consented and participated in at least one research visit, 397 (85.0%) were retained to study completion at 36 months. Retention rates did not differ by familial risk group (ASD-risk vs. Low-risk), sex, race, ethnicity, age at enrollment, number of children in the family, maternal employment, marital status, or parent concerns about the child at enrollment. A stepwise regression model identified 4 variables that, together, provided the most parsimonious predictive model of study retention: maternal education, maternal age at child's birth, travel distance to the study site, and diagnostic outcome classification at the final study visit.

Conclusions

The retained and not-retained groups did not differ on most demographic and clinical variables, suggesting few threats to internal and external validity. The significantly higher rate of retention of children diagnosed with ASD (95%) than typically developing children (83%) may, however, present biases when studying recurrence risk. We conclude by describing engagement and tracking methods that can be used to maximize retention in longitudinal studies of children at risk of ASD.

Modulation of EEG theta by naturalistic social content is not altered in infants with family history of autism

Theta oscillations (spectral power and connectivity) are sensitive to the social content of an experience in typically developing infants, providing a possible marker of early social brain development. Autism is a neurodevelopmental condition affecting early social behaviour, but links to underlying social brain function remain unclear. We explored whether modulations of theta spectral power and connectivity by naturalistic social content in infancy are related to family history for autism. Fourteen-month-old infants with (family history; FH; N = 75) and without (no family history; NFH; N = 26) a first-degree relative with autism watched social and non-social videos during EEG recording. We calculated theta (4-5 Hz) spectral power and connectivity modulations (social-non-social) and associated them with outcomes at 36 months. We replicated previous findings of increased theta power and connectivity during social compared to non-social videos. Theta modulations with social content were similar between groups, for both power and connectivity. Together, these findings suggest that neural responses to naturalistic social stimuli may not be strongly altered in 14-month-old infants with family history of autism.

Retained Primitive Reflexes and Potential for Intervention in Autistic Spectrum Disorders

We provide evidence to support the contention that many aspects of Autistic Spectrum Disorder (ASD) are related to interregional brain functional disconnectivity associated with maturational delays in the development of brain networks. We think a delay in brain maturation in some networks may result in an increase in cortical maturation and development in other networks, leading to a developmental asynchrony and an unevenness of functional skills and symptoms. The paper supports the close relationship between retained primitive reflexes and cognitive and motor function in general and in ASD in particular provided to indicate that the inhibition of RPRs can effect positive change in ASD.

Design, Development, and a Pilot Study of a Low-Cost Robot for Child–Robot Interaction in Autism Interventions

Socially assistive robots are widely deployed in interventions with children on the autism spectrum, exploiting the benefits of this technology in social behavior intervention plans, while reducing their autistic behavior. Furthermore, innovations in modern technologies such as machine learning enhance these robots with great capabilities. Since the results of this implementation are promising, their total cost makes them unaffordable for some organizations while the needs are growing progressively. In this paper, a low-cost robot for autism interventions is proposed, benefiting from the advantages of machine learning and low-cost hardware. The mechanical design of the robot and the development of machine learning models are presented. The robot was evaluated by a small group of educators for children with ASD. The results of various model implementations, together with the design evaluation of the robot, are encouraging and indicate that this technology would be advantageous for deployment in child–robot interaction scenarios.

Super responders: Predicting language gains from JASPER among limited language children with autism spectrum disorder

Early intervention can provide a great benefit for children with autism spectrum disorder (ASD). However, no single intervention is effective for all children. Even when an intervention is effective overall, individual child response varies. Some children make incredible progress, and others make slow or no progress. Therefore, it is important that the field move towards developing methods to personalize intervention. Operationalizing meaningful change and predicting intervention response are critical steps in designing systematic and personalized early intervention. The present research used improvement in expressive language to group children that received a targeted social communication early intervention, Joint Attention, Symbolic Play, Engagement, and Regulation (JASPER), into super responders and slow responders. Using baseline data from traditional standardized assessments of cognition and behavioral data from validated experimental measures of play and social communication, we used conditional inference tree models to predict responder status. From a sample of 99 preschool age, limited language children with ASD, play diversity was the most significant predictor of responder status. Children that played functionally with a wider variety of toys had increased odds of being a super responder to JASPER. A combination of lower play diversity and impairments in fine motor abilities increased the odds of children being slow responders to JASPER. Results from the present study can inform future efforts to individualize intervention and systematic approaches to augmenting treatment in real time. LAY SUMMARY: To help us answer the question of for whom an intervention works best, we examined 99 children, age three to five, who qualified as being limited spoken language communicators, and received a targeted intervention for social communication and language. We used child characteristics before intervention to predict which children would improve their language the most and found that the ability to play appropriately with a wider variety of toys predicted the best improvements in expressive language. These findings will help better inform future work to individualize intervention based on the unique needs of each child.

You may point, but do not touch: Impact of gesture-types and cognition on language in typical and atypical development

BACKGROUND

Evidence shows that the relation with the referent (object manipulation, contact/no contact pointing) and the different hand features (index finger/open palm) when pointing indicate different levels of cognitive and linguistic attainment in typical development (TD). This evidences the close link between pointing, cognition and language in TD, but this relation is understudied in autism. Moreover, the longitudinal pathway these abilities follow remains unexplored and it is unclear what specific role (predictor or mediator) pointing and cognition have in both typical and atypical language development.

AIMS

The first aim was to investigate whether pointing hand features (index finger/open palm) and relation with the referent (manipulation, contact and no contact pointing) similarly predict language in children with and without autism. The second aim was to explore whether cognition mediates the longitudinal relationship between pointing and language development.

METHODS & PROCEDURES

Sixteen children with autism, 13 children at high risk (HR) for autism and 18 TD children participated in an interactive gesture-elicitation task and were tested on standardised cognitive and expressive language batteries in a longitudinal design. A two-step analysis consisted of a stepwise linear regression and mediation analyses. First, the linear regression identified which hand features and types of relation with the referent predicted expressive language in all groups. Second, three mediation analyses (one per group) assessed the predictor/mediator role of the variables that met significance in the regression analysis.

OUTCOMES & RESULTS

Both cognition and index finger pointing were direct longitudinal predictors of further expressive language skills in the autism group. In TD and HR groups this relation was mediated by age.

CONCLUSIONS & IMPLICATIONS

Findings highlight the role of age in communicative development, but suggest a key role of cognition and index finger use in the longitudinal relationship between pointing gestures and expressive language development in children with autism. This has important clinical implications and supports the view that index finger pointing production might be a useful tool in the intervention for communicative and language abilities in autism.

WHAT THIS PAPER ADDS

What is already known on the subject There is evidence that no contact pointing is associated with complex socio-cognitive abilities that underpin communication in TD. Similarly, studies in TD show that index finger pointing is closely linked with language acquisition. However, it is unclear whether these associations are present in autism. In addition, the mediating (or predictive) role of cognition in the pointing-language relation has not yet been explored neither in typical nor in atypical development. What this paper adds to existing knowledge This paper shows that index finger pointing and cognition are direct longitudinal predictors of expressive language in the autism group. In the other groups this relation is mediated by age. This suggests that there is a window of opportunity for pointing to predict expressive language whereas the predictive value of cognition expands in development. Based on this, children with autism would share the same language predictors as TD children, but with delays. What are the potential or actual clinical implications of this work? This study reveals that index finger, age and cognition reliably predict spoken language in autism, which may indicate that early prelinguistic intervention based on pointing production and the improvement of cognitive skills might have a positive impact on spoken language in this population.

The Vocabulary of Infants with an Elevated Likelihood and Diagnosis of Autism Spectrum Disorder: A Systematic Review and Meta-Analysis of Infant Language Studies Using the CDI and MSEL

Diagnoses of autism spectrum disorder (ASD) are typically accompanied by atypical language development, which can be noticeable even before diagnosis. The siblings of children diagnosed with ASD are at elevated likelihood for ASD diagnosis and have been shown to have higher prevalence rates than the general population. In this paper, we systematically reviewed studies looking at the vocabulary size and development of infants with autism. One inclusion criterion was that infants were grouped either pre-diagnostically as elevated or typical likelihood or post-diagnostically as ASD or without ASD. This review focused on studies that tested infants up to 24 months of age and that assessed vocabulary either via the parent-completed MacArthur–Bates Communicative Developmental Inventory (CDI) or the clinician-administered Mullen Scales of Early Learning (MSEL). Our systematic search yielded 76 studies. A meta-analysis was performed on these studies that compared the vocabulary scores of EL and TL infants pre-diagnostically and the scores of ASD and non-ASD infants post-diagnostically. Both pre- and post-diagnostically, it was found that the EL and ASD infants had smaller vocabularies than their TL and non-ASD peers, respectively. The effect sizes across studies were heterogenous, prompting additional moderator analyses of age and sub-group analyses of the language measure used (CDI or MSEL) as potential moderators of the effect size. Age was found to be a moderator both in the pre- and post-diagnostical groups, however, language measure was not a moderator in either diagnostic group. Interpretations and future research directions are discussed based on these findings.

Trajectories of Vocabulary Development in Children with Autism Spectrum Disorder Across Multiple Measures

This longitudinal study examined how receptive and expressive vocabulary assessments capture vocabulary development in children with Autism Spectrum Disorder (ASD) and typically developing (TD) children. Using mixed regression modelling, we explored when children with ASD significantly different from TD children. We also examined the variability of individual trajectories of vocabulary development in children with ASD. Children with ASD showed slowed trajectories and significantly differed from TD children by 24 months on all assessments except for picture-based assessments. Children with ASD also showed high heterogeneity in trajectories, with some showing inconsistent patterns of growth, stagnation, and regression across assessments. This suggests that conclusions based on individual assessments of vocabulary can vary and assessment characteristics must be considered when monitoring vocabulary development.

Food for Thought: Machine Learning in Autism Spectrum Disorder Screening of Infants

Diagnoses of autism spectrum disorders (ASD) are typically made after toddlerhood by examining behavioural patterns. Earlier identification of ASD enables earlier intervention and better outcomes. Machine learning provides a data-driven approach of diagnosing autism at an earlier age. This review aims to summarize recent studies and technologies utilizing machine learning based strategies to screen infants and children under the age of 18 months for ASD, and identify gaps that can be addressed in the future. We reviewed nine studies based on our search criteria, which includes primary studies and technologies conducted within the last 10 years that examine children with ASD or at high risk of ASD with a mean age of less than 18 months old. The studies must use machine learning analysis of behavioural features of ASD as major methodology. A total of nine studies were reviewed, of which the sensitivity ranges from 60.7% to 95.6%, the specificity ranges from 50% to 100%, and the accuracy ranges from 60.9% to 97.7%. Factors that contribute to the inconsistent findings include the varied presentation of ASD among patients and study design differences. Previous studies have shown moderate accuracy, sensitivity and specificity in the differentiation of ASD and non-ASD individuals under the age of 18 months. The application of machine learning and artificial intelligence in the screening of ASD in infants is still in its infancy, as observed by the granularity of data available for review. As such, much work needs to be done before the aforementioned technologies can be applied into clinical practice to facilitate early screening of ASD.

Motor Skill Differences in Autism Spectrum Disorder: a Clinically Focused Review

PURPOSE OF REVIEW

This review synthesizes recent, clinically relevant findings on the scope, significance, and centrality of motor skill differences in autism spectrum disorder (ASD).

RECENT FINDINGS

Motor challenges in ASD are pervasive, clinically meaningful, and highly underrecognized, with up to 87% of the autistic population affected but only a small percentage receiving motor-focused clinical care. Across development, motor differences are associated with both core autism symptoms and broader functioning, though the precise nature of those associations and the specificity of motor profiles to ASD remain unestablished. Findings suggest that motor difficulties in ASD are quantifiable and treatable, and that detection and intervention efforts targeting motor function may also positively influence social communication. Recent evidence supports a need for explicit recognition of motor impairment within the diagnostic framework of ASD as a clinical specifier. Motor differences in ASD warrant greater clinical attention and routine incorporation into screening, evaluation, and treatment planning.

Annual Research Review: Anterior Modifiers in the Emergence of Neurodevelopmental Disorders (AMEND)-a systems neuroscience approach to common developmental disorders

We present the Anterior Modifiers in the Emergence of Neurodevelopmental Disorders (AMEND) framework, designed to reframe the field of prospective studies of neurodevelopmental disorders. In AMEND we propose conceptual, statistical and methodological approaches to separating markers of early-stage perturbations from later developmental modifiers. We describe the evidence for, and features of, these interacting components before outlining analytical approaches to studying how different profiles of early perturbations and later modifiers interact to produce phenotypic outcomes. We suggest this approach could both advance our theoretical understanding and clinical approach to the emergence of developmental psychopathology in early childhood.

Use of Empirical Mode Decomposition in ERP Analysis to Classify Familial Risk and Diagnostic Outcomes for Autism Spectrum Disorder

Event-related potentials (ERPs) activated by faces and gaze processing are found in individuals with autism spectrum disorder (ASD) in the early stages of their development and may serve as a putative biomarker to supplement behavioral diagnosis. We present a novel approach to the classification of visual ERPs collected from 6-month-old infants using intrinsic mode functions (IMFs) derived from empirical mode decomposition (EMD). Selected features were used as inputs to two machine learning methods (support vector machines and k-nearest neighbors (k-NN)) using nested cross validation. Different runs were executed for the modelling and classification of the participants in the control and high-risk (HR) groups and the classification of diagnosis outcome within the high-risk group: HR-ASD and HR-noASD. The highest accuracy in the classification of familial risk was 88.44%, achieved using a support vector machine (SVM). A maximum accuracy of 74.00% for classifying infants at risk who go on to develop ASD vs. those who do not was achieved through k-NN. IMF-based extracted features were highly effective in classifying infants by risk status, but less effective by diagnostic outcome. Advanced signal analysis of ERPs integrated with machine learning may be considered a first step toward the development of an early biomarker for ASD.

Uncovering neurodevelopmental paths to autism spectrum disorder through an integrated analysis of developmental measures and neural sensitivity to faces

BACKGROUND

Autism spectrum disorder (ASD) is highly heterogeneous in its etiology and manifestation. The neurobiological processes underlying ASD development are reflected in multiple features, from behaviour and cognition to brain functioning. An integrated analysis of these features may optimize the identification of these processes.

METHODS

We examined cognitive and adaptive functioning and ASD symptoms between 8 and 36 months in 161 infants at familial high risk for ASD and 71 low-risk controls; we also examined neural sensitivity to eye gaze at 8 months in a subsample of 140 high-risk and 61 low-risk infants. We used linked independent component analysis to extract patterns of variation across domains and development, and we selected the patterns significantly associated with clinical classification at 36 months.

RESULTS

An early process at 8 months, indicating high levels of functioning and low levels of symptoms linked to higher attention to gaze shifts, was reduced in infants who developed ASD. A longitudinal process of increasing functioning and low levels of symptoms was reduced in infants who developed ASD, and another process suggesting a stagnation in cognitive functioning at 24 months was increased in infants who developed ASD.

LIMITATIONS

Although the results showed a clear significant trend relating to clinical classification, we found substantial overlap between groups.

CONCLUSION

We uncovered underlying processes that acted together early in development and were associated with clinical outcomes. Our results highlighted the complexity of emerging ASD, which goes beyond the borders of clinical categories. Future work should integrate genetic data to investigate the specific genetic risks linked to these processes.

Eye Movements and Behavioural Responses to Gaze-Contingent Expressive Faces in Typically Developing Infants and Infant Siblings

Studies with infant siblings of children with Autism Spectrum Disorder have attempted to identify early markers for the disorder and suggest that autistic symptoms emerge between 12 and 24 months of age. Yet, a reliable first-year marker remains elusive. We propose that in order to establish first-year manifestations of this inherently social disorder, we need to develop research methods that are sufficiently socially demanding and realistically interactive. Building on Keemink et al. [2019, Developmental Psychology, 55, 1362-1371], we employed a gaze-contingent eye-tracking paradigm in which infants could interact with face stimuli. Infants could elicit emotional expressions (happiness, sadness, surprise, fear, disgust, anger) from on-screen faces by engaging in eye contact. We collected eye-tracking data and video-recorded behavioural response data from 122 (64 male, 58 female) typically developing infants and 31 infant siblings (17 male, 14 female) aged 6-, 9- and 12-months old. All infants demonstrated a significant Expression by AOI interaction (F(10, 1470) = 10.003, P < 0.001, ŋ_p ²  = 0.064). Infants' eye movements were "expression-specific" with infants distributing their fixations to AOIs differently per expression. Whereas eye movements provide no evidence of deviancies, behavioural response data show significant aberrancies in reciprocity for infant siblings. Infant siblings show reduced social responsiveness at the group level (F(1, 147) = 4.10, P = 0.042, ŋ_p ²  = 0.028) and individual level (Fischer's Exact, P = 0.032). We conclude that the gaze-contingency paradigm provides a realistically interactive experience capable of detecting deviancies in social responsiveness early, and we discuss our results in relation to subsequent infant sibling development. LAY SUMMARY: We investigated how infant siblings of children with autism spectrum disorder respond to interactive faces presented on a computer screen. Our study demonstrates that infant siblings are less responsive when interacting with faces on a computer screen (e.g., they smile and imitate less) in comparison to infants without an older sibling with autism. Reduced responsiveness within social interaction could potentially have implications for how parents and carers interact with these infants. Autism Res 2021, 14: 973-983. © 2020 International Society for Autism Research and Wiley Periodicals LLC.

Leveraging epigenetics to examine differences in developmental trajectories of social attention: A proof-of-principle study of DNA methylation in infants with older siblings with autism

Preliminary evidence suggests that changes in DNA methylation, a widely studied epigenetic mechanism, contribute to the etiology of Autism Spectrum Disorder (ASD). However, data is primarily derived from post-mortem brain samples or peripheral tissue from adults. Deep-phenotyped longitudinal infant cohorts are essential to understand how epigenetic modifications relate to early developmental trajectories and emergence of ASD symptoms. We present a proof-of-principle study designed to evaluate the potential of prospective epigenetic studies of infant siblings of children with ASD. Illumina genome-wide 450 K DNA methylation data from buccal swabs was generated for 63 male infants at multiple time-points from 8 months to 2 years of age (total N = 107 samples). 11 of those infants received a diagnosis of ASD at 3 years. We conducted a series of analyses to characterize DNA methylation signatures associated with categorical outcome and neurocognitive measures from parent-report questionnaire, eye-tracking and electro-encephalography. Effects observed across the entire genome (epigenome-wide association analyses) suggest that collecting DNA methylation samples within infant-sibling designs allows for the detection of meaningful signals with smaller sample sizes than previously estimated. Mapping networks of co-methylated probes associated with neural correlates of social attention implicated enrichment of pathways involved in brain development. Longitudinal modelling found covariation between phenotypic traits and DNA methylation levels in the proximity of genes previously associated with cognitive development, although larger samples and more complete datasets are needed to obtain generalizable results. In conclusion, assessment of DNA methylation profiles at multiple time-points in infant-sibling designs is a promising avenue to comprehend developmental origins and mechanisms of ASD.

Data-Driven Diagnostics and the Potential of Mobile Artificial Intelligence for Digital Therapeutic Phenotyping in Computational Psychiatry

Data science and digital technologies have the potential to transform diagnostic classification. Digital technologies enable the collection of big data, and advances in machine learning and artificial intelligence enable scalable, rapid, and automated classification of medical conditions. In this review, we summarize and categorize various data-driven methods for diagnostic classification. In particular, we focus on autism as an example of a challenging disorder due to its highly heterogeneous nature. We begin by describing the frontier of data science methods for the neuropsychiatry of autism. We discuss early signs of autism as defined by existing pen-and-paper-based diagnostic instruments and describe data-driven feature selection techniques for determining the behaviors that are most salient for distinguishing children with autism from neurologically typical children. We then describe data-driven detection techniques, particularly computer vision and eye tracking, that provide a means of quantifying behavioral differences between cases and controls. We also describe methods of preserving the privacy of collected videos and prior efforts of incorporating humans in the diagnostic loop. Finally, we summarize existing digital therapeutic interventions that allow for data capture and longitudinal outcome tracking as the diagnosis moves along a positive trajectory. Digital phenotyping of autism is paving the way for quantitative psychiatry more broadly and will set the stage for more scalable, accessible, and precise diagnostic techniques in the field.

Towards a Multivariate Biomarker-Based Diagnosis of Autism Spectrum Disorder: Review and Discussion of Recent Advancements

An ever-evolving understanding of autism spectrum disorder (ASD) pathophysiology necessitates that diagnostic standards also evolve from being observation-based to include quantifiable clinical measurements. The multisystem nature of ASD motivates the use of multivariate methods of statistical analysis over common univariate approaches for discovering clinical biomarkers relevant to this goal. In addition to characterization of important behavioral patterns for improving current diagnostic instruments, multivariate analyses to date have allowed for thorough investigation of neuroimaging-based, genetic, and metabolic abnormalities in individuals with ASD. This review highlights current research using multivariate statistical analyses to quantify the value of these behavioral and physiological markers for ASD diagnosis. A detailed discussion of a blood-based diagnostic test for ASD using specific metabolite concentrations is also provided. The advancement of ASD biomarker research promises to provide earlier and more accurate diagnoses of the disorder.

Social attention: What is it, how can we measure it, and what can it tell us about autism and ADHD?

Neurodevelopmental disorders like autism spectrum disorder (ASD) and attention deficit hyperactivity disorder (ADHD) affect 2-10% of children worldwide but are still poorly understood. Prospective studies of infants with an elevated familial likelihood of ASD or ADHD can provide insight into early mechanisms that canalize development down a typical or atypical course. Such work holds potential for earlier identification and intervention to support optimal outcomes in individuals with neurodevelopmental disorders. Disrupted attention may be involved in developmental trajectories to ASD and ADHD. Specifically, altered attention to social stimuli has been suggested as a possible endophenotype of ASD, lying between genetic factors impacting brain development and later symptoms. Similarly, changes in domain-general aspects of attention are commonly seen in ADHD and emerging evidence suggests these may begin in infancy. Could these patterns point to a common risk factor for both disorders? Or does social attention reflect the activity of a particular network of brain systems that is distinct to those underpinning general attention skills? One challenge to addressing such questions is our lack of understanding of the relation between social and general attention. In this chapter we review evidence from infants with later ASD and ADHD that illuminates this question.

Autism spectrum disorder risk prediction: A systematic review of behavioral and neural investigations

A reliable diagnosis of autism spectrum disorder (ASD) is difficult to make until after toddlerhood. Detection in an earlier age enables early intervention, which is typically more effective. Recent studies of the development of brain and behavior in infants and toddlers have provided important insights in the diagnosis of autism. This extensive review focuses on published studies of predicting the diagnosis of autism during infancy and toddlerhood younger than 3 years using behavioral and neuroimaging approaches. After screening a total of 782 papers, 17 neuroimaging and 43 behavioral studies were reviewed. The features for prediction consist of behavioral measures using screening tools, observational and experimental methods, brain volumetric measures, and neural functional activation and connectivity patterns. The classification approaches include logistic regression, linear discriminant function, decision trees, support vector machine, and deep learning based methods. Prediction performance has large variance across different studies. For behavioral studies, the sensitivity varies from 20% to 100%, and specificity ranges from 48% to 100%. The accuracy rates range from 61% to 94% in neuroimaging studies. Possible factors contributing to this inconsistency may be partially due to the heterogeneity of ASD, different targeted populations (i.e., high-risk group for ASD and general population), age when the features were collected, and validation procedures. The translation to clinical practice requires extensive further research including external validation with large sample size and optimized feature selection. The use of multi-modal features, e.g., combination of neuroimaging and behavior, is worth further investigation to improve the prediction accuracy.

ERP markers are associated with neurodevelopmental outcomes in 1-5 month old infants in rural Africa and the UK

INTRODUCTION

Infants and children in low- and middle-income countries are frequently exposed to a range of poverty-related risk factors, increasing their likelihood of poor neurodevelopmental outcomes. There is a need for culturally objective markers, which can be used to study infants from birth, thereby enabling early identification and ultimately intervention during a critical time of neurodevelopment.

METHOD

In this paper, we investigate developmental changes in auditory event related potentials (ERP) associated with habituation and novelty detection in infants between 1 and 5 months living in the United Kingdom and The Gambia, West Africa. Previous research reports that whereas newborns' ERP responses are increased when presented with stimuli of higher intensity, this sensory driven response decreases over the first few months of life, giving rise to a cognitively driven, novelty-based response. Anthropometric measures were obtained concurrently with the ERP measures at 1 and 5 months of age. Neurodevelopmental outcome was measured using the Mullen Scales of Early Learning (MSEL) at 5 months of age.

RESULTS

The described developmental change was observed in the UK cohort, who exhibited an intensity-based response at 1 month and a novelty-based response at 5 months of age. This change was accompanied by greater habituation to stimulus intensity at 5 compared to 1 month. In the Gambian cohort we did not see a change from an intensity-to a novelty-based response, and no change in habituation to stimulus intensity across the two age points. The degree of change from an intensity towards a novelty-based response was further found to be associated with MSEL scores at 5 months of infant age, whereas infants' growth between 1 and 5 months was not.

DISCUSSION

Our study highlights the utility of ERP-based markers to study young infants in rural Africa. By implementing a well-established paradigm in a previously understudied population we have demonstrated its use as a culturally objective tool to better understand early learning in diverse settings world-wide. Results offer insight into the neurodevelopmental processes underpinning early neurocognitive development, which may in the future contribute to early identification of infants at heightened risk of adverse neurodevelopmental outcome.

Association of Genetic Risks With Autism Spectrum Disorder and Early Neurodevelopmental Delays Among Children Without Intellectual Disability

IMPORTANCE

Autism spectrum disorder (ASD) is highly heritable, and modest contributions of common genetic variants to ASD have been reported. However, the association of genetic risks derived from common risk variants with ASD traits in children from the general population is not clear, and the association of these genetic risks with neurodevelopment in infants has not been well understood.

OBJECTIVE

To test whether a polygenic risk score (PRS) for ASD is associated with neurodevelopmental progress at age 18 months and ASD traits at age 6 years among children from the general population.

DESIGN, SETTING, AND PARTICIPANTS

In this cohort study, 876 children in the Hamamatsu Birth Cohort for Mothers and Children in Hamamatsu, Japan, underwent testing for the association of an ASD PRS with neurodevelopmental progress and ASD traits. Data collection began in December 2007 and is ongoing. Data analysis was conducted from April to December 2019.

MAIN OUTCOMES AND MEASURES

Summary data from the largest genome-wide association study were used to generate ASD PRSs, and significance of thresholds was calculated for each outcome. The Autism Diagnostic Observation Schedule 2 was used to measure ASD traits at age 6 years, and the Mullen Scales of Early Learning was used to measure neurodevelopmental progress at age 18 months.

RESULTS

Of 876 participants (mean [SD] gestational age at birth, 38.9 [1.6] weeks; 438 [50.0%] boys; 868 [99.1%] Japanese), 734 were analyzed. The ASD PRS was associated with ASD traits (R2 = 0.024; β, 0.71; SE, 0.24; P = .03). The association of ASD PRS with infant neurodevelopment was most pronounced in gross motor (R2 = 0.015; β, -1.25; SE, 0.39; P = .01) and receptive language (R2 = 0.014; β, -1.19; SE, 0.39; P = .02) scores on the Mullen Scales of Early Learning. Gene set enrichment analyses found that several pathways, such as cell maturation (R2 = 0.057; β, -5.28; SE, 1.40; P < .001) and adenylyl cyclase activity and cyclic adenosine monophosphate concentration (R2 = 0.064; β, -5.30; SE 1.30; P < .001), were associated with ASD traits. Gene sets associated with inflammation were commonly enriched with ASD traits and gross motor skills (eg, chemokine motif ligand 2 production: R2 = 0.051; β, -6.04; SE, 1.75; P = .001; regulation of monocyte differentiation: R2 = 0.052; β, -6.63; SE, 1.90; P = .001; and B-cell differentiation: R2 = 0.051; β, 7.37; SE, 2.15; P = .001); glutamatergic signaling-associated gene sets were commonly enriched with ASD traits and receptive language skills (eg, regulation of glutamate secretion: R2 = 0.052; β, -5.82; SE, 1.68; P = .001; ionotropic glutamate receptor signaling pathway: R2 = 0.047; β, 3.54; SE, 1.09; P = .001; and negative regulation of glutamate secretion: R2 = 0.045; β, -5.38; SE, 1.74; P = .002).

CONCLUSIONS AND RELEVANCE

In this study, the ASD PRS was associated with ASD traits among children from the general population. Genetic risks for ASD might be associated with delays in some neurodevelopmental domains, such as gross motor and receptive language skills.

The Use of Artificial Intelligence in Screening and Diagnosis of Autism Spectrum Disorder: A Literature Review

Objectives:

The detection of autism spectrum disorder (ASD) is based on behavioral observations. To build a more objective datadriven method for screening and diagnosing ASD, many studies have attempted to incorporate artificial intelligence (AI) technologies. Therefore, the purpose of this literature review is to summarize the studies that used AI in the assessment process and examine whether other behavioral data could potentially be used to distinguish ASD characteristics.

Methods:

Based on our search and exclusion criteria, we reviewed 13 studies.

Results:

To improve the accuracy of outcomes, AI algorithms have been used to identify items in assessment instruments that are most predictive of ASD. Creating a smaller subset and therefore reducing the lengthy evaluation process, studies have tested the efficiency of identifying individuals with ASD from those without. Other studies have examined the feasibility of using other behavioral observational features as potential supportive data.

Conclusion:

While previous studies have shown high accuracy, sensitivity, and specificity in classifying ASD and non-ASD individuals, there remain many challenges regarding feasibility in the real-world that need to be resolved before AI methods can be fully integrated into the healthcare system as clinical decision support systems.

Pre-emptive intervention versus treatment as usual for infants showing early behavioural risk signs of autism spectrum disorder: a single-blind, randomised controlled trial

BACKGROUND

Great interest exists in the potential efficacy of prediagnostic interventions within the autism spectrum disorder prodrome, but available evidence relates to children at high familial risk. We aimed to test the efficacy of a pre-emptive intervention designed for infants showing early behavioural signs of autism spectrum disorder.

METHODS

In this single-blind, randomised controlled trial done at two specialist centres in Australia, infants aged 9-14 months were enrolled if they were showing at least three early behavioural signs of autism spectrum disorder on the Social Attention and Communication Surveillance-Revised (SACS-R) 12-month checklist. Infants were randomly assigned (1:1) to receive a parent-mediated video-aided intervention (iBASIS-VIPP) or treatment as usual. Group allocation was done by minimisation, stratified by site, sex, age, and the number of SACS-R risk behaviours. Assessments were done at baseline (before treatment allocation) and at the 6 month endpoint. The primary outcome was Autism Observation Scale for Infants (AOSI), which measures early behavioural signs associated with autism spectrum disorder. Secondary outcomes were a range of infant and caregiver outcomes measured by Manchester Assessment of Caregiver-Infant interaction (MACI), Mullen Scales of Early Learning (MSEL), Vineland Adaptive Behaviour Scales, 2nd edition (VABS-2), MacArthur-Bates Communicative Development Inventory (MCDI), and Parenting Sense of Competence (PSOC) scale. This trial is registered with Australian New Zealand Clinical Trials Registry, number ANZCTR12616000819426.

FINDINGS

Between June 9, 2016, and March 30, 2018, 103 infants were randomly assigned, 50 to the iBASIS-VIPP group and 53 to the treatment-as-usual group. After the intervention, we observed no significant differences between groups on early autism spectrum disorder behavioural signs measured by the AOSI (difference estimate -0·74, 95% CI -2·47 to 0·98). We also observed no significant differences on secondary outcomes measuring caregiver non-directiveness (0·16, -0·33 to 0·65), caregiver sensitive responding (0·24, -0·15 to 0·63), and infant attentiveness (-0·19, -0·63 to 0·25) during parent-child interactions (MACI), as well as on researcher-administered measures of receptive (1·30, -0·48 to 3·08) and expressive language (0·54, -0·73 to 1·80), visual reception (0·31, -0·77 to 1·40), and fine motor skills (0·55, -0·32 to 1·41) using the MSEL. Compared with the treatment-as-usual group, the iBASIS-VIPP group had lower infant positive affect (-0·69, -1·27 to -0·10) on the MACI, but higher caregiver-reported receptive (37·17, 95% CI 10·59 to 63·75) and expressive vocabulary count (incidence rate ratio 2·31, 95% CI 1·22 to 4·33) on MCDI, and functional language use (difference estimate 6·43, 95% CI 1·06 to 11·81) on VABS. There were no significant group differences on caregiver-reported measures of MCDI infant gesture use (3·22, -0·60 to 7·04) and VABS social behaviour (3·28, -1·43 to 7·99). We observed no significant differences between groups on self-reported levels of parenting satisfaction (difference estimate 0·21, 95% CI -0·09 to 0·52), interest (-0·23, -0·62 to 0·16) and efficacy (-0·08, -0·38 to 0·22) on PSOC.

INTERPRETATION

A pre-emptive intervention for the autism spectrum disorder prodrome had no immediate treatment effect on early autism spectrum disorder symptoms, the quality of parent-child interactions, or researcher-administered measures of developmental skills. However, we found a positive effect on parent-rated infant communication skills. Ongoing follow-up of this infant cohort will assess longer-term developmental effects.

FUNDING

Western Australia Children's Research Fund, Autism Cooperative Research Centre, La Trobe University, and Angela Wright Bennett Foundation.

Identifying Atypical Development: A Role of Day-Care Workers?

Identifying the early signs of developmental disability is important for ensuring timely diagnosis and early intervention. Day-care workers may be in a prime position to notice potential developmental deviations, but it is unclear if they can accurately recognize subtle early signs of atypical development. Sixty day-care workers examined home-videos of very young children with fragile X syndrome and typically developing children. Results indicated that most day-care workers can distinguish typical and atypical development in general and might therefore have an important role in early identification. Special work experience and advanced pedagogical training appeared to boost day-care workers' sensitivity to detect atypical features in early development and to provide effective daily surveillance.

Preschool Staff Spot Social Communication Difficulties, But Not Restricted and Repetitive Behaviors in Young Autistic Children

To fulfill the criteria for autism spectrum disorder (ASD), symptoms must be present across domains and contexts. We assessed preschool staff’s ratings of social communication and interaction (SCI) and restricted and repetitive behaviors (RRBs) in 3-year-old siblings of children with ASD, either diagnosed (n = 12) or not diagnosed (n = 36) with ASD, and typically developing siblings with no family history of ASD (n = 16). Ratings of SCI were more accurate than RRBs in differentiating the ASD group from the two other groups, and only the SCI ratings correlated with clinical assessment of social behavior. We conclude that while preschool staff ratings of SCI behaviors are adequate, ratings of RRBs should be treated with more caution.

Gesture and Language Trajectories in Early Development: An Overview From the Autism Spectrum Disorder Perspective

The well-documented gesture-language relation in typical communicative development (TD) remains understudied in autism spectrum disorder (ASD). Research on early communication skills shows that gesture production is a strong predictor of language in TD, but little is known about the association between gestures and language in ASD. This review focuses on exploring this relation by addressing two topics: the reliability of gestures as predictor of language competences in ASD and the types of potential differences (quantitative, qualitative, or both) in the gesture-language trajectory in children on the autism spectrum compared to typically developing children. We find evidence that gesture production is indeed a reliable predictor of early communicative skills and that both quantitative and qualitative differences have been established in research in the development of verbal and non-verbal communication skills in ASD, with lower gesture rates at the quantitative level, and a trajectory that starts deviating from the TD trajectory only at some point after the first year of life.

Eurosibs: Towards robust measurement of infant neurocognitive predictors of autism across Europe

•

The Eurosibs consortium is a nine-site European neurocognitive study of infants with an older sibling with ASD.

•

Data quality assessments show that that neurocognitive measures hold promise for cross-site consistency in diverse populations.

•

We present robust data analysis pipelines and highlight challenges and opportunities for future multisite research efforts.

Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder that affects social communication skills and flexible behaviour. Developing new treatment approaches for ASD requires early identification of the factors that influence later behavioural outcomes. One fruitful research paradigm has been the prospective study of infants with a first degree relative with ASD, who have around a 20% likelihood of developing ASD themselves. Early findings have identified a range of candidate neurocognitive markers for later ASD such as delayed attention shifting or neural responses to faces, but given the early stage of the field most sample sizes are small and replication attempts remain rare. The Eurosibs consortium is a European multisite neurocognitive study of infants with an older sibling with ASD conducted across nine sites in five European countries. In this manuscript, we describe the selection and standardization of our common neurocognitive testing protocol. We report data quality assessments across sites, showing that neurocognitive measures hold great promise for cross-site consistency in diverse populations. We discuss our approach to ensuring robust data analysis pipelines and boosting future reproducibility. Finally, we summarise challenges and opportunities for future multi-site research efforts.

Spotting Signs of Autism in 3-Year-Olds: Comparing Information from Parents and Preschool Staff

Preschool informants may provide valuable information about symptoms of autism spectrum disorder (ASD) in young children. We compared the diagnostic accuracy of ratings by preschool staff with those by parents of 3-year-old children using the Achenbach System of Empirically Based Assessment Preschool Forms. The sample consisted of 32 children at familial risk for ASD without diagnosis, 10 children at risk for ASD with diagnosis, and 14 low-risk typically developing controls. Preschool staff ratings were more accurate than parent ratings at differentiating children with and without ASD, and more closely associated with clinician-rated symptoms. These results point to the value of information from preschool informants in early detection and diagnostic assessments.

Latent trajectories of adaptive behaviour in infants at high and low familial risk for autism spectrum disorder

Background

Autism spectrum disorder (ASD) is characterised by persisting difficulties in everyday functioning. Adaptive behaviour is heterogeneous across individuals with ASD, and it is not clear to what extent early development of adaptive behaviour relates to ASD outcome in toddlerhood. This study aims to identify subgroups of infants based on early development of adaptive skills and investigate their association with later ASD outcome.

Methods

Adaptive behaviour was assessed on infants at high (n = 166) and low (n = 74) familial risk for ASD between 8 and 36 months using the Vineland Adaptive Behavior Scales (VABS-II). The four domains of VABS-II were modelled in parallel using growth mixture modelling to identify distinct classes of infants based on adaptive behaviour. Then, we associated class membership with clinical outcome and ASD symptoms at 36 months and longitudinal measures of cognitive development.

Results

We observed three classes characterised by decreasing trajectories below age-appropriate norms (8.3%), stable trajectories around age-appropriate norms (73.8%), and increasing trajectories reaching average scores by age 2 (17.9%). Infants with declining adaptive behaviour had a higher risk (odds ratio (OR) = 4.40; confidence interval (CI) 1.90; 12.98) for ASD and higher parent-reported symptoms in the social, communication, and repetitive behaviour domains at 36 months. Furthermore, there was a discrepancy between adaptive and cognitive functioning as the class with improving adaptive skills showed stable cognitive development around average scores.

Conclusions

Findings confirm the heterogeneity of trajectories of adaptive functioning in infancy, with a higher risk for ASD in toddlerhood linked to a plateau in the development of adaptive functioning after the first year of life.

Early behavioral indices of inherited liability to autism

OBJECTIVE

The observed heterogeneity of autism spectrum disorder (ASD)-and the diversity of rare germline mutations with which it has been associated-has been difficult to reconcile with knowledge of its pronounced heritability in the population.

METHODS

This article reviews and synthesizes recent family and developmental studies incorporating behavioral indices of inherited risk for ASD.

RESULTS

Autism may arise from critical combinations of early inherited neurobehavioral susceptibilities-some specific to autism, some not-each of which may be traceable to partially-independent sets of genetic variation. These susceptibilities and their respective genetic origins may not relate to the characterizing symptoms of autism (after it develops) in a straightforward way, and may account for "missing heritability" in molecular genetic studies.

CONCLUSIONS

Within-individual aggregations of a finite set of early inherited neurobehavioral susceptibilities-each individually common in the population-may account for a significant share of the heritability of ASD. Comprehensive identification of these underlying traits my help elucidate specific early intervention targets in individual patients, especially if autism represents a developmental consequence of earlier-interacting susceptibilities. Scientific understanding of the early ontogeny of autism will benefit from epidemiologically-rigorous, genetically-informative studies of robust endophenotypic candidates whose inter-relationships in infancy are mapped and normed.

Cerebrospinal fluid and the early brain development of autism

BACKGROUND

There is currently a renaissance of interest in the many functions of cerebrospinal fluid (CSF). Altered flow of CSF, for example, has been shown to impair the clearance of pathogenic inflammatory proteins involved in neurodegenerative diseases, such as amyloid-β. In addition, the role of CSF in the newly discovered lymphatic system of the brain has become a prominently researched area in clinical neuroscience, as CSF serves as a conduit between the central nervous system and immune system.

MAIN BODY

This article will review the importance of CSF in regulating normal brain development and function, from the prenatal period throughout the lifespan, and highlight recent research that CSF abnormalities in autism spectrum disorder (ASD) are present in infancy, are detectable by conventional structural MRI, and could serve as an early indicator of altered neurodevelopment.

CONCLUSION

The identification of early CSF abnormalities in children with ASD, along with emerging knowledge of the underlying pathogenic mechanisms, has the potential to serve as early stratification biomarkers that separate children with ASD into biological subtypes that share a common pathophysiology. Such subtypes could help parse the phenotypic heterogeneity of ASD and map on to targeted, biologically based treatments.

Mobile detection of autism through machine learning on home video: A development and prospective validation study

BACKGROUND

The standard approaches to diagnosing autism spectrum disorder (ASD) evaluate between 20 and 100 behaviors and take several hours to complete. This has in part contributed to long wait times for a diagnosis and subsequent delays in access to therapy. We hypothesize that the use of machine learning analysis on home video can speed the diagnosis without compromising accuracy. We have analyzed item-level records from 2 standard diagnostic instruments to construct machine learning classifiers optimized for sparsity, interpretability, and accuracy. In the present study, we prospectively test whether the features from these optimized models can be extracted by blinded nonexpert raters from 3-minute home videos of children with and without ASD to arrive at a rapid and accurate machine learning autism classification.

METHODS AND FINDINGS

We created a mobile web portal for video raters to assess 30 behavioral features (e.g., eye contact, social smile) that are used by 8 independent machine learning models for identifying ASD, each with >94% accuracy in cross-validation testing and subsequent independent validation from previous work. We then collected 116 short home videos of children with autism (mean age = 4 years 10 months, SD = 2 years 3 months) and 46 videos of typically developing children (mean age = 2 years 11 months, SD = 1 year 2 months). Three raters blind to the diagnosis independently measured each of the 30 features from the 8 models, with a median time to completion of 4 minutes. Although several models (consisting of alternating decision trees, support vector machine [SVM], logistic regression (LR), radial kernel, and linear SVM) performed well, a sparse 5-feature LR classifier (LR5) yielded the highest accuracy (area under the curve [AUC]: 92% [95% CI 88%-97%]) across all ages tested. We used a prospectively collected independent validation set of 66 videos (33 ASD and 33 non-ASD) and 3 independent rater measurements to validate the outcome, achieving lower but comparable accuracy (AUC: 89% [95% CI 81%-95%]). Finally, we applied LR to the 162-video-feature matrix to construct an 8-feature model, which achieved 0.93 AUC (95% CI 0.90-0.97) on the held-out test set and 0.86 on the validation set of 66 videos. Validation on children with an existing diagnosis limited the ability to generalize the performance to undiagnosed populations.

CONCLUSIONS

These results support the hypothesis that feature tagging of home videos for machine learning classification of autism can yield accurate outcomes in short time frames, using mobile devices. Further work will be needed to confirm that this approach can accelerate autism diagnosis at scale.