Skip to main content
Download PDF

Protocol for the development of joint attention-based subclassification of autism spectrum disorder and validation using multi-modal data

BMC Psychiatry volume 23, Article number: 589 (2023) Cite this article

Abstract

Background

Heterogeneity in clinical manifestation and underlying neuro-biological mechanisms are major obstacles to providing personalized interventions for individuals with autism spectrum disorder (ASD). Despite various efforts to unify disparate data modalities and machine learning techniques for subclassification, replicable ASD clusters remain elusive. Our study aims to introduce a novel method, utilizing the objective behavioral biomarker of gaze patterns during joint attention, to subclassify ASD. We will assess whether behavior-based subgrouping yields clinically, genetically, and neurologically distinct ASD groups.

Methods

We propose a study involving 60 individuals with ASD recruited from a specialized psychiatric clinic to perform joint attention tasks. Through the examination of gaze patterns in social contexts, we will conduct a semi-supervised clustering analysis, yielding two primary clusters: good gaze response group and poor gaze response group. Subsequent comparison will occur across these clusters, scrutinizing neuroanatomical structure and connectivity using structural as well as functional brain imaging studies, genetic predisposition through single nucleotide polymorphism data, and assorted socio-demographic and clinical information.

Conclusions

The aim of the study is to investigate the discriminative properties and the validity of the joint attention-based subclassification of ASD using multi-modality data.

Trial registration

Clinical trial, KCT0008530, Registered 16 June 2023, https://cris.nih.go.kr/cris/index/index.do.

Peer Review reports

Background

The early onset [1], diverse clinical and neurobiological features [2,3,4], and rising prevalence of autism spectrum disorder (ASD) [5, 6] highlight the urgent need and associated challenges for its subclassification. Recent advancements in genomics [7,8,9] and neurobiology [10,11,12,13] underscore the disorder’s complexity, revealing various etiological pathways. This wide spectrum of ASD symptoms poses both a hurdle and incentive for subclassification. Subclassification methodologies are distinguished by the type of data analyzed and the algorithms used for grouping [12]. The data can be behavior-based or biology-based, each indirectly indicating variations in the other [14, 15]. Moreover, subclassification algorithms range from supervised and unsupervised to hybrid models, each using univariate or multivariate statistics, with their respective pros and cons [16].

Genomic subclassification of ASD

Genetic factors are the primary risk contributors to ASD, encompassing a range of both rare and common genetic variants. This includes rare de novo mutations, copy number variants, protein-truncating single nucleotide polymorphisms (SNPs), indels, and even non-coding de novo mutations in chromatin interactions [17, 18]. Previous studies have linked these genetic elements to the wide-ranging clinical manifestations of ASD. The genetic origins of ASD are complex and multifactorial; however, our grasp of how these genetic elements functionally influence ASD is still limited. Additionally, the way these genetic factors interact with each other and non-genetic factors to shape risk remains uncertain. However, considering the enormous number of variants implicated in ASD, coupling these genetic data with other types of data obtained through different modalities may help us discern specific genetic vulnerabilities and their corresponding functional consequences in ASD [19].

Neuro-subclassification of ASD

A few studies have explored neuroanatomical differences between individuals with ASD and neurotypical individuals, as well as within the ASD group itself [11,12,13, 20]. A significant portion of this research has been conducted using the Autism Brain Imaging Data Exchange (ABIDE) dataset, an open-source compilation of brain imaging data from ASD and neurotypical individuals [21, 22]. Using this resource, Hong et al. identified three distinct ASD subclasses with varying clinical outcomes, based on cortical thickness, intensity contrast, surface area, and geodesic distance [20]. Choi et al., applying an unsupervised clustering approach to ABIDE’s resting state fMRI (rsfMRI) data and utilizing ‘connectome-based gradient’ and ‘functional random forest’ methodologies, were able to differentiate various forms of ASD [12]. However, their results lacked complete replicability as the clinical scores of the subgroups identified through cluster analysis in a follow-up dataset deviated from those in the original dataset. Additionally, these ASD clusters failed to show significant differences in neural connectivity [12, 20]. These outcomes imply that purely unsupervised clustering using intricate data like fMRI might not yield clinically meaningful ASD subtypes.

Clinical score-driven subclassification of ASD

A research team adopted a top-down approach in their pursuit of developing a new ASD subclassification method, using a clinical-score based clustering method and multimodality data [23]. While this approach allowed for the identification of distinct ASD subgroups based on balance between social-communicative and restricted repetitive behaviors – the two main symptom domains of ASD), these ASD subgroups showed no evident differences in neural circuitry and genetic components [23].

Limitations of previous studies

Previous attempts at cluster analysis using unsupervised learning on intricate data such as fMRI did not yield consistent or replicable subclassifications; even with identical experimental parameters, the replication dataset did not reproduce the same ASD subgroups as the original discovery dataset [11,12,13]. Without any labels to guide the clustering, individuals were grouped into distinct subclusters based on non-linear data patterns that lacked clinical significance. Meanwhile, studies using supervised learning methods for cluster analysis also fell short, failing to identify subgroups with unique clinical phenotypes and neurobiological mechanisms [23]. It is worth noting that these studies used human-rated autism-related scale sub-scores as labels for supervised learning. Although these scores are clinically valid and frequently used, they may not adequately capture objective and qualitative differences in ASD phenotypes.

Possible objective behavioral biomarker for subclassification of ASD

Joint attention, as indicated by prior research, may serve as an objective behavioral biomarker for ASD subclassification [24, 25]. This term refers to the sharing of attentional focus with others on objects or events and is believed to aid social learning [26, 27]. We have established a digitized method to assess joint attention, implementing a novel protocol to prompt three types of joint attention - initiation of joint attention (IJA), low-level and high-level response to joint attention (RJAlow and RJAhigh) - as defined in the Early Social Communication Scales (ESCS) [28]. This method uses video-recorded task-related behaviors [26, 27], which are then utilized to train a deep learning model to identify ASD and evaluate its symptom severity [25].

Identifying the most affected level—biological, neural, or behavioral—in an individual with Autism Spectrum Disorder (ASD) to determine the suitable targeted intervention is a significant challenge. To address this, we have initiated the development of the Yonsei-Seoul Multi-modal Subclassification (YSMS) protocol, which was designed to streamline data collection from newly diagnosed ASD patients. Specifically, the focus is on the patterns of joint attention gaze demonstrated during behavioral tasks, with the aim to discern whether these patterns can differentiate various ASD subtypes, each unique in their biological and neurological attributes. We intend to implement a pilot version of this protocol in an exploratory study involving 60 individuals diagnosed with ASD from June 2023 to May 2024 to compare objective behavioral biomarker-based ASD subgroups using multimodality data. Such endeavor is expected to drive the creation of a novel ASD subclassification system, considering each individual’s biological, neural, and behavioral characteristics.

Methods

Recruitment

We have designed a single center, prospective, non-randomized experimental study, where individuals with ASD diagnosis will be recruited from June 2023 – May 2024. A group of 60 ASD individuals will be recruited from outpatient clinic at the Child and Adolescent Psychiatry Division of Seoul National University Hospital (SNUH). Enrollment criteria are: (1) age of 30 ~ 71 months, (2) children diagnosed with ASD by a psychiatrist and confirmed by Autism Diagnostic Observation Schedule II (ADOS2). Exclusion criteria are: (1) history of organic brain injury, (2) having visual or auditory deficit, (3) having neurological (motor/musculoskeletal) conditions, (4) history of adverse reaction to sedation anesthesia. A schematic showing the flow of patients is depicted in Fig. 1.

Fig. 1
figure 1

Recruitment and Overall Study Schema

Full size image

Multi-modal Data Collection – different types of data

Socio-demographics and family history

We gather information about each participant’s socio-demographics and family history to consider for environmental risk factors that may have contributed to different social gaze patterns during joint attention tasks. Patient information such as sex and age at time of assessment will be gathered. Family information such as parental age, parental education level as well as socio-economic status will be gathered from parent interviews. Moreover, number of siblings as well as neuropsychiatric history of siblings will also be obtained.

Clinical assessments - neuropsychiatric tests

The diagnosis of ASD is confirmed using ADOS-2, the gold standard diagnostic tool for ASD diagnosis [29]. Autistic tendencies are measured using the Korean versions of the Social Responsiveness Scale (K-SRS) as well as Social Communication Questionnaire (K-SCQ) [30]. The Social Responsiveness Scale (SRS) is a 65-item questionnaire, scored from zero to three based on the frequency of the described behavior, that evaluates the social interactions of children over the past six months, with higher scores indicating lower social capability [30]. The SCQ is a 40-item screening instrument that is based on Autism Diagnostic Interview-Revised (ADI-R), a tool for more in-depth assessment of ASD symptoms, and selects key items that deviate from normal development [30, 31].

Child behavioral problems will be checked with Child Behavior Checklist (CBCL) as well as Vineland Adaptive Behavior Scale (VABS) [32]. The child’s motor functions will be assessed through Developmental Coordination Disorder Questionnaire (DCDQ) [32]. To assess the cognitive levels of participants, the Korean Wechsler Preschool and Primary Scale of Intelligence–Fourth Edition (K-WPPSI-IV) will be used [33].

Joint attention tasks and Video Data

We will conduct three types of joint attention tasks in the order of initiation of joint attention, low-level response to joint attention, and high-level response to joint attention according to the Yonsei Seoul Multi-modal Subclassification (YSMS) Joint Attention Task Standard Operating Procedure. Development and validation of AI model trained on video data collected using this protocol has been published elsewhere [24, 25]. Supplementary Fig. 1 shows setup for joint attention experiments based on our proposed protocol. Initiation of joint attention task requires use of only toy 1, while low level response to joint attention task requiring use of toy 1 and toy 2 (distraction) and high-level response to joint attention task requiring use of four picture prompts.

Video (RGB or RGB-D)-based Gaze Estimation during Joint attention situations

Following the YSMS Social Gaze-Based Clustering Standard Operating Procedure, we will gather joint attention videoclips for purpose of determining gaze patterns of each participant during joint attention tasks. Joint attention tasks were slightly modified from previous study [24, 25], whereby picture prompts are placed on the same plane as the examiner and toy objects so that all gaze points will fall on the same imaginary vertical hyperplane parallel to the participant’s head/eyes, as shown in Supplementary Fig. 8. With such videorecording set-up, RGB-D information (if using RGB-D device) as well as head’s yaw, pitch, and roll may be computed via computer vision techniques such as algorithm for estimating 3D position of the head and face-tracker [34]. Head position and orientation is tracked through the frame sequence and at every time point this information is stored for later analyzing different patterns of social (joint attention situations) gaze among participants.

Magnetic resonance imaging Acquisition and Processing

We use magnetic resonance imaging (MRI) to study differences in brain structure [35], function, and connectivity [36,37,38]. These methods are important for determining whether different clusters or subgroups of ASD based on social gaze patterns differ at systems-level, providing the underlying neurological mechanisms of different types of ASD. We will be gathering all MRI data from a single center—SNUH.

Structural MRI (sMRI) and Diffusion Tensor Imaging (DTI)

MRI scans are acquired on 3T scanners from Siemens (https://www.siemens-healthineers.com/). We will carry out several procedures to optimize structural and functional sequences for the best Siemens-specific options and to address challenges related to standardization and quality assurance of image-acquisition.

Global descriptors of brain anatomy can be elucidated through measures of total grey and white matter volume [35]. Alongside these metrics, we will probe into distinctions in cortical thickness, intensity contrast, geodesic distance, and cortical surface areas. These anatomical indices possess unique genetic determinants, phylogeny, and developmental trajectories [35, 36]. Furthermore, we will calculate structural connectivity indices utilizing both structural MRI scans and Diffusion Tensor Imaging (DTI). Variations in intrinsic grey matter connectivity can be inferred by examining the disparities in both local and global wiring costs [37]. Concurrently, distinctions in short and long-range white matter tracts can be evaluated using tractography analysis of particular pathways [38].

Resting state functional magnetic resonance imaging (rsfMRI)

Standard gradient-echo echo planar imaging paradigm will be utilized for the acquisition of resting state fMRI scans [39]. We will carry out several procedures to find optimal parameters such as field of view, repetition time, echo time, flip angle, etc. During the resting state scan, participants will be sedated with their eyes closed. The Configurable Pipeline for the Analysis of Connectomes (CPAC, http://fcp-indi.github.com) preprocessing pipeline will be employed for fMRI data processing to remove non-neural noise, such as motion artifacts, from the BOLD-signal [12, 20]. Multiple functional networks have been identified that are characterized by coherent patterns of intrinsic activity between ‘nodes’ that resemble patterns of activity that are engaged during specific cognitive functions [12]. We aim to identify whether social gaze-based subgroups differ in (hyper/hypo-) connectivity within and across these networks [10, 11, 13].

Whole genome sequencing (WGS)

We acquire blood samples from the participant for genomic analyses—we specifically wish to identify all variants associated with ASD and calculate polygenic risk scores for each participant. These methods are important for determining whether different clusters or subgroups of ASD based on social gaze patterns differ at the genetic, blueprint-level; ASD individuals may show varying social gaze patterns based on their entire mutation profile.

Collection and Processing of Peripheral Whole Blood Samples: Each participant will contribute a 3ml sample of peripheral whole blood. These samples will be subjected to genomic DNA extraction and whole genome sequencing, performed with the NextSeq 550Dx System (Illumina, San Diego, CA, USA). The procedure for whole genome sequencing will adhere to an industry standard quality-controlled sequence analysis pipeline and utilize a current reference sequence for the purposes of mapping and variant calling [40].

Identification of single nucleotide polymorphisms (SNPs) and insertions and deletions (indels) will be conducted with HaplotypeCaller and MuTect2 from the GATK package (3.8-0, https://github.com/broadinstitute/gatk/releases), as well as with VarScan2 (2.4.0, https://github.com/dkoboldt/varscan/releases). A variety of databases will be used for the analysis and annotation of variants, including but not limited to, the Online Mendelian Inheritance in Man (OMIM), the Human Gene Mutation Database (HGMD), Clinvar, dbSNP, 1000 Genomes, the Exome Aggregation Consortium (ExAC), the Exome Sequencing Project (ESP), the Korean Reference Genome Database (KRGDB), Autism Speaks MSSNG resource, and the Simons Simplex Collection (SSC). Variants will be classified following the standards and guidelines set forth by the American College of Medical Genetics (ACMG) [41]. Following annotation, the sequencing data will undergo additional processing to compute polygenic risk scores for ASD of the top SNPs after thresholding the SNP component [42]. Such computation aims to aid in differentiating between individuals with Autism Spectrum Disorder (ASD) based on their distinct genetic susceptibility or polygenic risk variability.

The different types of data for comparison between subgroups of ASD are listed in Table 1.

Table 1 Different types of data acquired in proposed study
Full size table

Considerations related to the statistical analysis

Statistical models and analyses

Social Gaze-based semi-supervised clustering analysis

From our previous study, we explored certain relationships between individuals with ASD and clinical scores as well as their performance on joint attention tasks, namely initiation of joint attention and response to joint attention [24, 25]. Such analyses revealed that variables that explain for the variance within group of ASD individuals is their performance in joint attention tasks. Hence forth, we will use join attention success rate for the three different joint attention tasks and perform principal component analysis followed by semi-supervised k-means clustering analysis [43] - using good and poor gaze response head-pose values as labels - to obtain two ASD subgroups, one with decreased joint attention and the other with intact joint attention. An alternative method would be to use a range of threshold value k to divide the group of ASD individuals in two sub-groups where one group shows relatively better joint attention abilities.

Comparison between good gaze response group and poor gaze response group

Concerning group comparison between two subgroups of ASD based on gaze response during joint attention tasks, several elements will be investigated. These include socio-demographic and clinical characteristics, social gaze patterns (gaze frequency towards or away from foci of attention). Furthermore, structural MRI profiles comprising of cortical thickness and surface area, intensity contrast, and geodesic distance, as well as rsfMRI profiles detailing stepwise functional connectivity difference, will be assessed.

Continuous variables will be represented using means, standard deviations, medians, and ranges. To compare categorical variables, the chi-squared test will be utilized. All statistical analyses and computations of the validation measures will be conducted using Python 3.6.8, in conjunction with SciPy version 1.4.1 (https://docs.scipy.org/doc/scipy-1.4.1/) and Statsmodels 0.11.1 (https://www.statsmodels.org/dev/release/version0.11.1.html). A p-value < 0.05 will be used as the threshold for statistical significance.

Power calculation – sample size estimation

For this exploratory study, the lack of preceding data precludes the establishment of a definitive sample size. Nevertheless, referencing guidelines for comparable studies, we advocate a minimum sample size of 12, considering the study’s feasibility, precision of means and variances, and compliance with regulatory requirements [44]. Reflecting on a previous investigation that effectively classified three distinct severity subclasses of ASD with a cohort of 45 individuals [25], our study plans to recruit 60 participants. This number takes into consideration potential participant dropouts or missing data essential for statistical analysis.

Ethical considerations

This study, adhering to the Helsinki Declaration and its subsequent amendments, mandates informed consent from participants and their legal guardians based on verbal and written details provided. Participants and their legal guardians reserve the right to withdraw at any point, with no obligation to justify their decision and no impact on future treatment. Approval for the study has been granted by the Seoul National University Hospital IRB Review Board (IRB No. H-2210-137-1374). The investigators anticipate no discomfort for the subjects from the tests and tasks involved, with no short- or long-term risks identified in relation to this study.

Outcome

We posit that ASD individuals with decreased joint attention, characterized by atypical social gaze patterns, will exhibit unique genetic, structural, and functional neural patterns distinct from ASD individuals with mostly intact joint attention, who demonstrate appropriate social gaze patterns. The differentiation is likely rooted in the connectivity within the visual and social brain pathways. It is anticipated that ASD individuals with decreased joint attention will possess SNPs associated with the visual pathway. Furthermore, individuals with greater RRB than SC impairment are expected to exhibit less impairment in social gaze ability. Conversely, participants with a higher SC impairment relative to RRB are predicted to demonstrate a pronounced deficiency in social gaze ability. A layout of expected outcome table is depicted in Table 2.

Table 2 Expected outcome table layout
Full size table

Discussion

Prompted by a clinical demand for ASD subclassification based on distinct neurobiological and behavioral traits, this study is the first to prospectively gather multi-modal data of ASD individuals, utilizing objective behavioral biomarkers for group clustering. Greater comprehension of ASD types, characterized by their unique behavioral anomalies and underlying mechanisms, is essential for devising patient-specific treatments. Ignoring the interplay of patient-specific needs, overt impairments (social-communication or attentiveness to social cues), genetic susceptibilities, and biological-neurological conditions can result in inefficiencies and parental frustration. Our methodology seeks to mitigate these issues by facilitating successful ASD subclassification based on unique behavioral-neurological-genetic patterns.

Data Availability

The dataset used and/or analyzed during the current are available from the corresponding author on reasonable request.

Abbreviations

ADOS2:

Autism Diagnostic Observation Schedule-2

ASD:

Autism Spectrum Disorder

CBCL:

Child Behavior Checklist

DCDQ:

Developmental Coordination Disorder Questionnaire

DTI:

Diffusion tensor imaging

FSIQ:

Full Scale Intelligence Quotient

GATK:

Genome Analysis Toolkit

IJA:

Initiation of Joint Attention

JA:

Joint Attention

PCA:

Principal Component Analysis

RJAlow :

Low Level Response to Joint Attention

RJAhigh :

High Level Response to Joint Attention

rsfMRI:

Resting State Functional Magnetic Resonance Imaging

SCQ:

Social Communication Questionnaire

SNP:

Single Nucleotide Polymorphism

SRS:

Social Responsiveness Scale

sMRI:

Structural Magnetic Resonance Imaging

VABS:

Vineland Adaptive Behavior Scales

VIQ:

Verbal Intelligence Quotient

WPPSI:

Wechsler Preschool and Primary Scale of Intelligence

YSMS:

Yonsei Seoul Multi-modal Subclassification

References

  1. Ozonoff S, Iosif AM, Baguio F et al. A prospective study of the emergence of early behavioral signs of Autism. Vol 49.; 2010.

  2. Wolfers T, Floris DL, Dinga R, et al. From pattern classification to stratification: towards conceptualizing the heterogeneity of Autism Spectrum Disorder. Neurosci Biobehav Rev. 2019;104:240–54. https://doi.org/10.1016/j.neubiorev.2019.07.010.

    Article  PubMed  Google Scholar 

  3. Lombardo MV, Lai MC, Baron-Cohen S. Big data approaches to decomposing heterogeneity across the autism spectrum. Mol Psychiatry. 2019;24(10):1435–50. https://doi.org/10.1038/s41380-018-0321-0.

    Article  PubMed  PubMed Central  Google Scholar 

  4. Warrier V, Zhang X, Reed P, et al. Genetic correlates of phenotypic heterogeneity in autism. Nat Genet. 2022;54(9):1293–304. https://doi.org/10.1038/s41588-022-01072-5.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  5. Zablotsky B, Black LI, Maenner MJ, et al. Prevalence and trends of developmental disabilities among children in the United States: 2009–2017. Pediatrics. 2019;144(4). https://doi.org/10.1542/peds.2019-0811.

  6. Hodges H, Fealko C, Soares N. Autism spectrum disorder: definition, epidemiology, causes, and clinical evaluation. Transl Pediatr. 2020;9:55–S65.

    Article  Google Scholar 

  7. Warrier V, Zhang X, Reed P, et al. Genetic correlates of phenotypic heterogeneity in autism. Nat Genet. 2022;54(9):1293–304. https://doi.org/10.1038/s41588-022-01072-5.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  8. Kim I, Bin, Lee T, Lee J, et al. Non-coding de novo mutations in chromatin interactions are implicated in autism spectrum disorder. Mol Psychiatry. 2022;27(11):4680–94. https://doi.org/10.1038/s41380-022-01697-2.

    Article  CAS  PubMed  Google Scholar 

  9. Antaki D, Guevara J, Maihofer AX et al. A phenotypic spectrum of autism is attributable to the combined effects of rare variants, polygenic risk and sex. Nat Genet.

  10. Hong SJ, de Wael RV, Bethlehem RAI, et al. Atypical functional connectome hierarchy in autism. Nat Commun. 2019;10(1). https://doi.org/10.1038/s41467-019-08944-1.

  11. Choi H, Byeon K, Park B, yong, et al. Diagnosis-informed connectivity subtyping discovers subgroups of autism with reproducible symptom profiles. NeuroImage. 2022;256. https://doi.org/10.1016/j.neuroimage.2022.119212.

  12. Hong SJ, Vogelstein JT, Gozzi A, et al. Toward Neurosubtypes in Autism. Biol Psychiatry. 2020;88(1):111–28. https://doi.org/10.1016/j.biopsych.2020.03.022.

    Article  PubMed  Google Scholar 

  13. Reardon AM, Li K, Langley J, Hu XP. Subtyping Autism Spectrum Disorder Via Joint modeling of clinical and connectomic profiles. Brain Connect. 2022;12(2):193–205. https://doi.org/10.1089/brain.2020.0997.

    Article  PubMed  Google Scholar 

  14. Bilder RM, Sabb FW, Cannon TD, et al. Phenomics: the systematic study of phenotypes on a genome-wide scale. Neuroscience. 2009;164(1):30–42.

    Article  CAS  PubMed  Google Scholar 

  15. Van Dam NT, O’Connor D, Marcelle ET, et al. Data-Driven phenotypic categorization for neurobiological analyses: beyond DSM-5 labels. Biol Psychiatry. 2017;81(6):484–94. https://doi.org/10.1016/j.biopsych.2016.06.027.

    Article  PubMed  Google Scholar 

  16. Varoquaux G, Thirion B. How machine learning is shaping cognitive neuroimaging. Vol 3.; 2014. http://www.gigasciencejournal.com/content/3/1/28.

  17. Warrier V, Zhang X, Reed P, et al. Genetic correlates of phenotypic heterogeneity in autism. Nat Genet. 2022;54(9):1293–304. https://doi.org/10.1038/s41588-022-01072-5.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  18. Fu JM, Satterstrom FK, Peng M, et al. Rare coding variation provides insight into the genetic architecture and phenotypic context of autism. Nat Genet Published online. 2022. https://doi.org/10.1038/s41588-022-01104-0.

    Article  Google Scholar 

  19. Choi L, An JY. Genetic architecture of autism spectrum disorder: Lessons from large-scale genomic studies. Neurosci Biobehav Rev. 2021;128:244–57. https://doi.org/10.1016/j.neubiorev.2021.06.028.

    Article  CAS  PubMed  Google Scholar 

  20. Hong SJ, Valk SL, Di Martino A, Milham MP, Bernhardt BC. Multidimensional neuroanatomical subtyping of autism spectrum disorder. Cereb Cortex. 2018;28(10):3578–88. https://doi.org/10.1093/cercor/bhx229.

    Article  PubMed  Google Scholar 

  21. Di Martino A, Yan C, Li Q, et al. The autism brain imaging data exchange: towards a large-scale evaluation of the intrinsic brain architecture in autism. Mol Psychiatry. 2014;6:659–67. https://doi.org/10.1038/mp.2013.78.

    Article  Google Scholar 

  22. Di Martino A, O’Connor D, Chen B, et al. Enhancing studies of the connectome in autism using the autism brain imaging data exchange II. Sci Data. 2017;4(1):170010. https://doi.org/10.1038/sdata.2017.10.

    Article  PubMed  PubMed Central  Google Scholar 

  23. Bertelsen N, Landi I, Bethlehem RAI, et al. Imbalanced social-communicative and restricted repetitive behavior subtypes of autism spectrum disorder exhibit different neural circuitry. Commun Biol. 2021;4(1). https://doi.org/10.1038/s42003-021-02015-2.

  24. Ko C, Kang S, Hong SB et al. AI-assisted Initiation to Joint Attention Evaluation for Autism Spectrum Disorder Detection. In: 2022 IEEE 3rd International Conference on Human-Machine Systems (ICHMS). IEEE; 2022:1–6. https://doi.org/10.1109/ICHMS56717.2022.9980778.

  25. Ko C, Lim JH, Hong SB, Park YR. Development and validation of a joint attention-based deep learning system for detection and symptom severity assessment of autism spectrum disorder. JAMA Netw Open. 2023;6(5):e2315174. https://doi.org/10.1001/jamanetworkopen.2023.15174.

    Article  PubMed  PubMed Central  Google Scholar 

  26. Nyström P, Thorup E, Bölte S, Falck-Ytter T. Joint attention in infancy and the emergence of Autism. Biol Psychiatry. 2019;86(8):631–8. https://doi.org/10.1016/j.biopsych.2019.05.006.

    Article  PubMed  Google Scholar 

  27. Mundy P, Block J, Delgado C, Pomares Y, Van Hecke AV, Parlade MV. Individual differences and the development of joint attention in infancy. Child Dev. 2007;78(3):938–54. https://doi.org/10.1111/j.1467-8624.2007.01042.x.

    Article  PubMed  PubMed Central  Google Scholar 

  28. Mundy P, Delgado C, Block J et al. DRAFT A Manual for the EARLY SOCIAL COMMUNICATION SCALES (ESCS).; 2003. http://edscholars.ucdavis.edu/vrlab/home.

  29. Hus V, Gotham K, Lord C. Standardizing ADOS domain scores: separating severity of social affect and restricted and repetitive behaviors. J Autism Dev Disord.

  30. Schanding GT, Nowell KP, Goin-Kochel RP. Utility of the Social Communication Questionnaire-Current and Social Responsiveness Scale as teacher-report screening tools for autism spectrum disorders. J Autism Dev Disord. 2012;42(8):1705–16. https://doi.org/10.1007/s10803-011-1412-9.

    Article  PubMed  Google Scholar 

  31. De Bildt A, Systema S, Zander E, et al. Autism diagnostic interview-revised (ADI-R) algorithms for toddlers and young preschoolers: application in a non-US sample of 1,104 children. J Autism Dev Diord. 2015;45(7):2076–91.

    Article  Google Scholar 

  32. Frazier TW, Georgiades S, Bishop SL, Hardan AY. Behavioral and cognitive characteristics of females and males with autism in the simons simplex collection. J Am Acad Child Adolesc Psychiatry. 2014;53(3). https://doi.org/10.1016/j.jaac.2013.12.004.

  33. Park H, Seo Y, Lee JA, Study of Concurrent Validities of K-WPPSI-IV. Korean J Child Stud. 2015;36(1):65–83. https://doi.org/10.5723/kjcs.2015.36.1.65.

    Article  Google Scholar 

  34. Zhang D, Li J, Shan Z et al. Implementations of Dlib Deep Learning Face Recognition Technology. International Conference on Robots & Intelligent System (ICRIS). Date of conference: 7–8 Nov. 2020. IEEE. https://doi.org/10.1109/ICRIS52159.2020.00030.

  35. Courchesne E, Campbell K, Solso S. Brain growth across the life span in autism: age-specific changes in anatomical pathology. Brain Res. 2011;1380:138–45.

    Article  CAS  PubMed  Google Scholar 

  36. Belmonte MK, Allen G, Beckel-Mitchener A. Boulanger LM < Carper RA, Webb SJ. Autism and abnormal development of brain connectivity. J Neurosci. 2004;24:9228–31.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  37. Ecker C, Ronan L, Feng Y, et al. Intrinsic gray-matter connectivity of the brain in adults with autism spectrum disorder. Proc Natl Acad Sci USA. 2013;110:13222–7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  38. Catani M, Dell’Acqua F, Budisavljevic S, et al. Frontal networks in adults with autism spectrum disorder. Brain. 2016;139:616–30.

    Article  PubMed  PubMed Central  Google Scholar 

  39. Subah FZ, Deb K, Dhar PK, Koshiba T. A Deep Learning Approach to Predict Autism Spectrum Disorder using multisite resting-state fMRI. Appl Sci. 2021;11(8):3636. https://doi.org/10.3390/app11083636.

    Article  CAS  Google Scholar 

  40. Ahn YJ, Markkandan K, Baek I, et al. An efficient and tunable parameter to improve variant calling for whole genome and exome sequencing data. Genes & Genomics. 2018;40:39–47.

    Article  CAS  Google Scholar 

  41. Richards S, Aziz N, Bale S, et al. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Sci. 2015;17(5):405–24. https://doi.org/10.1038/gim.2015.30.

    Article  Google Scholar 

  42. Chen SF, Dias R, Evans D et al. Genotype imputation and variability in polygenic risk score estimation. 2020;12(100). https://doi.org/10.1186/s13073-020-00801-x.

  43. Hanmin Y, Hao L, Qianting S. An improved semi-supervised K-means clustering algorithm. 2016 IEEE Information Technology, Networking, Electronic, Automation Control Conference. Date of conference: 20–22 May 2016. IEEE. https://doi.org/10.1109/ITNEC.2016.7560315.

  44. Hajian-Tilaki K. Sample size estimation in diagnostic test studies of biomedical informatics. J Biomed Inform. 2014;48:193–204. https://doi.org/10.1016/j.jbi.2014.02.013.

    Article  PubMed  Google Scholar 

Download references

Acknowledgements

We would like to acknowledge and show our appreciation to all the nurses, research assistants, radiologists and MRI technicians of Seoul National University Hospital Child and Adolescent Psychiatry Clinic for supporting and embracing the extra clinical workload created by this research project. We would like to extend our sincerest gratitude to all children and parents who have shown interest to participate in our project.

Funding

Funding was provided by a grant from the MD-PhD Physician-Scientist Training Program from the Korean Health Industry Development Institute (KHIDI), Ministry of Health and Welfare of the Republic of Korea.

Author information

Authors and Affiliations

  1. Department of Biomedical Systems Informatics, Yonsei University College of Medicine, Seoul, South Korea

    Chanyoung Ko & Yu Rang Park

  2. Division of Child and Adolescent Psychiatry, Department of Psychiatry, Seoul National University Hospital, Seoul, South Korea

    Soyeon Kang & Soon-Beom Hong

  3. Department of Psychiatry and Institute of Human Behavioral Medicine, Seoul National University College of Medicine, Seoul, South Korea

    Soon-Beom Hong

Authors
  1. Chanyoung Ko
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Soyeon Kang
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Soon-Beom Hong
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Yu Rang Park
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

CK initiated the study, procured the funding, conceptualized, and constructed the original manuscript, to which all authors contributed significant intellectual input. SK contributed significantly to paper conceptualization. SH contributed to conceptualization of the study design and editing the paper. YP procured the funding, contributed to paper conceptualization and the final editing of the paper. All authors read and approved the final manuscript.

Corresponding authors

Correspondence to Soon-Beom Hong or Yu Rang Park.

Ethics declarations

Competing interests

The authors declare no competing interests.

Ethics approval and consent to participate

The study is approved by the Seoul National University Hospital IRB Review Board (IRB No. H-2210-137-1374). The study will be conducted in accordance with the Helsinki Declaration 1964, including subsequent revisions. Informed consent will be obtained from all the participants and their legal guardians involved in the study. The participants and their legal guardians may, at any time, choose to withdraw from the study without being required to explain and without affecting the person’s future treatment. The study investigators are under the impression that the questionnaires, interviews, videorecording tasks, and other data acquisition methods will not lead to any lasting discomfort for the participants. There are no known expected short- or long-term risks associated with the present study.

Consent for publication

Not applicable.

Funding

Funding was provided by a grant from the MD-PhD Physician-Scientist Training Program from the Korean Health Industry Development Institute (KHIDI), Ministry of Health and Welfare of the Republic of Korea. This work was also supported by the Korean Ministry of Science and Information and Communication Technology Institute via Institute of Information & communications Technology Planning & Evaluation (IITP) grant (No. 2022-0-00234, ‘Development of digital therapeutics for improvement of social interaction skills in patients with autism spectrum disorder’).

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Ko, C., Kang, S., Hong, SB. et al. Protocol for the development of joint attention-based subclassification of autism spectrum disorder and validation using multi-modal data. BMC Psychiatry 23, 589 (2023). https://doi.org/10.1186/s12888-023-04978-4

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12888-023-04978-4

Keywords

BMC Psychiatry

ISSN: 1471-244X

Contact us