Identifying Distinct Electrophysiological Endophenotypes in Autism Spectrum Disorder: A Large-Scale Machine Learning Approach Integrating Auditory Brainstem Response and Behavioral Phenotyping

Identifying Distinct Electrophysiological Endophenotypes in Autism Spectrum Disorder: A Large-Scale Machine Learning Approach Integrating Auditory Brainstem Response and Behavioral Phenotyping | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Identifying Distinct Electrophysiological Endophenotypes in Autism Spectrum Disorder: A Large-Scale Machine Learning Approach Integrating Auditory Brainstem Response and Behavioral Phenotyping Qingjie Zhang, Chunmei Ren, Xianrong Liang, Guojun Yun, Kanglong Peng This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9288120/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 5 You are reading this latest preprint version Abstract Background Our study aimed to explore the correlations between Auditory Brainstem Response outcomes and Autism phenotypes. Methods A total of 1883 children with or suspected of being with ASD were enrolled. The related features were acquired by using the Autism Behavior Checklist (ABC), Children Autism Rating Scale (CARS) and ABR outcomes. The cluster analysis was conducted to detect potential subgroups within our samples, and the ANOVA analysis was conducted to reveal the differences among these subgroups using data from clinical assessments. The Pearson Correlation analysis was conducted to depict the associations between ABR results and different autism phenotypes. Results Our results revealed different brainstem vulnerabilities across six ASD clusters. Cluster 6 exhibited robust language-associated temporal alternations, while cluster 4 displayed sensory-related pontine-mesencephalic delays. Besides, compromised olivary function uniquely presented correlations with ABC total scores in cluster 2. Notably, less affected auditory pathways were found in clusters with high-functioning performance. Conclusion Our findings confirmed that compromised ABR components did not display in uniform formats but rather differed across different subgroups. Furthe research is needed to validate these neurophysiological markers, promoting precise personalized interventions tailored to ASD phenotypes. ASD Machine learning ABR Prediction Auditory Sensory Figures Figure 1 Figure 2 Introduction Autism spectrum disorder (ASD) is a highly heritable and heterogeneous neurodevelopmental disorder whose symptoms emerge in early developmental stage and persist along the overall lifespan(Lord et al., 2020 ). ASD is characterized by core disorders in social communication, restricted and repetitive behaviors (RBB), and various co-occurring symptoms(González-Cortés et al., 2019 ; Udhnani & Lee, 2025 ). In addition to these core features, atypical perceptual function is another distinctive characteristic commonly observed in children with ASD(Zhu, Chen, Zhang, Zhang, & Guo, 2025 ). For example, typical developed children often perceive global structures prior to local details, exemplifying the adage of seeing the forest before trees, while children with ASD tend to process the subtle contents first. These results are now integrated into various frameworks for describing the perception function in children with ASD(Mottron, Dawson, Soulières, Hubert, & Burack, 2006 ). The Weak Central Coherence (WCC) proposed that children with ASD cannot integrate subtle contents into concrete concepts(Mottron et al., 2006 ). Controversially, the Enhanced Perceptual Functioning (EPF) hypothesized that children with ASD can outperform their typical developed peers in local perceptions, but the global processing is not compromised(Foster et al., 2016 ). (Zhu et al., 2025 ). Recently, one mate-analysis conducted by Zhu et al. provided reasonable solutions to these debates that global processing is not compromised but rather affected by various tasks parameters(Zhu et al., 2025 ). Recently, electroencephalograms (EEG) and functional magnetic resonate images (fMRI) have emerged as valuable evidence for corroborating these solutions under connectivity perspective. Using electroencephalogram recording, one study found that disrupted connectivity in the right central-right parietal and left central-left frontal connections may interfere with auditory integration(Zhao, Luo, Mei, & Shao, 2025 ). The resting-state fMRI has also shown that children with ASD presented abnormal functional connectivity in primary sensory network involving the auditory cortices(Zhang et al., 2025 ). Recent studies also revealed that age and sex have impacts on these structural connectivity(Sigar et al., 2025 ; Wu et al., 2025 ; Zhang et al., 2025 ). These studies demonstrate how EEG and fMRI can offer valuable perspectives on the links between structures and functions. While previous studies have explored the associations between structures and functions, recent studies have tried to examine the bioelectrical signal using Auditory Brainstem Response (ABR). One study tried to examine the ABR in a small sample of children with ASD and found a notable prolongation in wave Ⅰ and Ⅲ latency as well as Ⅰ/Ⅴ interpeak latency (IPL), but decreased wave Ⅲ/Ⅴ IPL(Liu et al., 2024 ). However, another study found no significance in wave Ⅰ, Ⅲ and Ⅴ, but shorten summating potential (SP) using electrocochleograms before wave Ⅰ(Fujihira, Itoi, Furukawa, Kato, & Kashino, 2021 ). Despite this phenotypic variability in the directionality of temporal shifts in ABR, all observed alternations in auditory response consistently indicated abnormal neural conduction within the auditory pathway. Overall, these study reveals that abnormal auditory processing may originate from the cochlear level rather than high-level cortex, implying that possible activity-induced neural remodeling across the whole auditory pathway, (Blue, Wong, & Dodson, 2024 ; Liu et al., 2024 ; Miron, Beam, & Kohane, 2018 ). These findings in bioelectrical signal demonstrate a paradigm shift from cortico-centric models toward a peripheral-to-central framework. This study aimed to systematically reveal the associations between ABR outcomes and ASD phenotypes. Materia and Methods Participants Participants were recruited from local referral programs of the government service including maternal and childcare service center, education institutions, and community agency. Children with definitive or suspected diagnosis of ASD were referred for comprehensive evaluation through this program. The referred individuals would accept interdisciplinary assessment to achieve definitive diagnosis of ASD and receive tailored intervention. Comprehensive assessment routinely induces the administration of ABR, ABC, CARS, and other standardized tools. Prior to administration, all necessary consents were obtained from all subjects and/or their legal guardian(s). A multidisciplinary team is invited to confirm the diagnosis of ASD. The members included a psychiatrist with ADOS-2 license and two senior neurologists. Prior to administration, all the subjects and/or their legal guardians(s) had signed necessary consents. Subjects were included if they were older than 2 years of age and met the diagnostic criteria of the DSM-5-TR (version 2022), and the clinical presentation consisted of three manifestations of social disorders as well as two manifestations of stereotyped repetitive behaviors, as follows: Socialization disorders Social-emotional interaction disorders Physical Motor Behavioral (nonverbal communication) Social Disorders Social Relationship Development Disorder (Development, Formation, Understanding) Stereotypical Repetitive Behavior (SRB) Repetition of stereotyped motor movements, object manipulation, or verbal expressions Development of repetitive, routine, and patterned stereotyped verbal or nonverbal behaviors Extremely limited, fixed interests, or attention spans Abnormal responses (extreme sensitivity or the opposite) to sensory input, both normal and abnormal (environmental) For specific diagnostic criteria, refer to C.E. Rice's suggested judgment criteria for each entry(Rice et al., 2022 ). Patients with other unrelated conditions were excluded, including peripheral nerve injury, myelitis, spinal embolism syndrome, seizures, and fractures. Clinical measure Autism Behavior Checklist (ABC) The Autism Behavior Checklist (ABC) assessment consists of 57 items that involve possibly all typical autistic behaviors. Items are categorized into five components including relating, sensory, language, body use and object manipulation, and social and self-help. Participants were asked according to the item description given by one researcher and rated the item if their children behaved as the item described. Besides, each item contained its own score ranging from 1 to 4 points according to the item weights. The weighted score of each item is decided by the occurrence frequency in Krug’s study(Krug, Arick, & Almond, 1980 ). For example, item 1 occurred more than item 2, then item 1 is endowed with 4 points, and item 2 is endowed with 2 points. If one item is rated, then participant gets the according score (e.g. item 1 scores 0/4, item 2 scores 0/2). The original cut off score was set at 68, and total score above 67 indicated severe symptoms or higher possibility to get diagnosis as ASD(Wadden, Bryson, & Rodger, 1991 ). The interrater reliability is 0.85, and the intra-rater reliability is 0.82(Abdelmageed, Youssef, Rihan, & Abdelaziz, 2024 ; Krug et al., 1980 ). The Childhood Autism Rating Scale First Edition CARS1 was constructed to collect information from major caregivers’ interviews, direct or indirect observations, and structural interviews. A total of 15 items are involved in CARS1, including relation to people, imitation, emotional response, body use, object use, adaptation to change, visual response, listening response, sensory response, emotion, verbal communication, gesture, activity status, intellectual response, and overall impressions. A four-point rating scale is utilized to quantify the symptoms severity, where one point refers to normal behavior and four points refers to inappropriate behaviors that are different from normal developed children. The total score is the sum of all items, and the higher scores refers to more severe autistic symptoms. CARS1 was delivered by trained/licensed clinicians or researchers with appropriate training for the necessary interview’s techniques with parents and caregivers and judgement criteria. Auditory brainstem response protocol The Neuron Spectrum 5 (Neurosoft LLC, Ivanovo, Russian) was used to record the ABR outcome in a sound-blocked room with comfortable temperature and minimal ambient noise low than 30 db. All the participants received routine examinations to exclude any ear abnormalities. Then, ABR were recorded when they were either sitting in resting-statue or sleeping. Electrodes were placed at the middle of the forehead or Cz in 10–20 system (active), nasal root or Fpz (ground), and mastoids (references) respectively. Electrode impedance was maintained lower than 5 kΩ. Click stimuli (100µs duration, 21.1/s rate) was delivered monaurally via insert headphones at 45dB, 65dB, 85dB. Signals were bandpass filtered (100Hz-3000Hz), and 2000 sweeps were averaged per trial. Recorded outcomes included latencies (Wave Ⅰ, Ⅱ, Ⅲ, Ⅳ, Ⅴ), interpeak latencies (Ⅰ/Ⅲ, Ⅲ/Ⅴ, Ⅰ/Ⅴ). Statistical analysis Model evaluation and selection Table.1 displays six performance indicators that are used to evaluate and select the XGBoost Classifiers using ABR collected from bilateral ears under conditions using different stimulations (45dB, 65dB, and 85dB clicks). These evaluation metrics can present comprehensive perspectives on classification accuracy, discriminative capability, and clinical applicability. Cluster analysis Our study adopted the Ward Linkage method, or the minimum variance method to detect the potential cluster in our sample. To determine the reasonable cluster amount, we utilized the common elbow method based on Bayesian and Akaike information criteria. The ANOVA analysis was conducted to depict the parameters differences among clusters, and significant level was set at p values lower than 0.05 in all tests. The Bonferroni post-hoc analysis was conducted to depict the difference between clusters. Pearson Correlation analysis The Pearson correlation test was conducted to depict the associations between standardized assessment results and ABR outcomes. Result Sample characteristics Our study managed to recruit a sample that consisted of 1883 children and adolescents. Table.2 presents the demographic data collected from this sample. Our sample aged around 64.47 months old (64.47 ± 20.04), and most of them were boys (1600/1883). Our study tried to recruit a sample with balance in terms of age range designed based on Chinese Education System (nursery: 0–3 years, kindergarten: 3–6 years, primary school: 6–12, and junior high school: 12–15 years), but we ultimately obtained a sample with 103 from nursery, 1128 administered in kindergarten, 592 from primary school, and 6 from junior high school. In terms of ASD classification, we managed to obtain a sample with balance using ABC and CARS evaluation criteria as shown in Table.2 . Performance evaluation and selection of XGBoost Classifiers Table.3 and Figure.1 present the performance metrics of XGBoost Classifiers across different feature configurations. Among these models, the XGBoost Classifier trained using data collected from right ear using 65dB clicks demonstrated superior overall performance, achieving the highest F1 score (0.7085) and ranking second in AUROC (0.9208), accuracy (0.9177), and precision (0.7793). Importantly, this configuration exhibited the optimal balance between precision and recall, which is crucial in clinical diagnostic application where both false negatives and false positives carry significant clinical implications. Besides, Models trained on data from conditions using 45dB generally underperformed relative to other scenarios, indicating insufficient feature presentiveness. Conversively, models trained on outcomes from 85dB protocols display competitive AUROC values but compromised precision, suggesting overfitting or redundant features inclusion. Using cluster analysis to identify clusters in samples This study conducted cluster analysis and ANOVA analysis to identify homogeneous subgroups in this sample based on ABC performance. Table.4 shows the phenotypes differences defined by ABC domains in these clusters. Cluster 2 represented the subgroups with severe multidimensional impairment profile. This group exhibited the most pronounced problematic behaviors across sensory, relating and body and object use. Children in this group display significant hyper/hyposensitivity to environmental stimuli, pronounced deficits in inter-person relationship, and prominent SRB. Children with ASD in cluster 4 showed intermediate to high impairment in social related language application, but relatively less affected motor behavior. Cluster 6 consisted of children with ASD accompanying linguistic problems and less affected sensory function. This behavior profile indicates one distinct phenotype characterized by significant communication deficit involving both verbal and non-verbal language disorders. Cluster 1 stood for one moderately sized subgroup with intermediate social communication deficit and less influenced in motor behavior. Cluster 3 demonstrated the subsample with mild to moderate generalized autistic traits. Cluster 5 contained the most participants who display minimal autistic behaviors. In summary, Table.5 depicts one severity gradient from severe multidimensional impairment (cluster 2) through domain specific severe impairments (cluster 4 and 6), intermediate generalized impairment (cluster 1 and 3), to minimal impairment (cluster 5). Pearson correlation outcomes To depict the neurophysiological traits of clinical heterogeneity in children with ASD, correlation analysis was conducted between auditory brainstem response (ABR) metrics and ABC subscale scores across six distinct phenotypic clusters (Figure.2). For correlation coefficients with an absolute magnitude over 0.3, we found that signal transmission from the cochlear nerve to inferior colliculus may be correlated with phenotypical characteristics in children with ASD. Cluster six characterized by predominant language impairment displays the strongest and most complicated correlation metrics. Primary finding was that prolonged wave Ⅲ latency presented robust association with language scores, leading to sequential extended Ⅲ/Ⅴ interval (r equals to 0.51, 0.52 respectively). Besides, these temporal shifts also reflect the disturbed signal conduction from superior olivary to inferior colliculus. Additionally, these temporal delays may involve less responsive reaction originated from cochlear nerves as evident by prolonged wave Ⅰ latency and alternations in Ⅰ/Ⅲ and Ⅰ/Ⅴ intervals. These temporal alternations may explain why latency Ⅴ was compressed by either passive transmission deficits or active central hyperresponsive mechanism. Conversely, compressed wave Ⅰ latency was observed along with shortened Ⅰ/Ⅲ and Ⅰ/Ⅴ intervals, indicating accelerated neural conduction originated from cochlear nerves as relating score get higher. Additionally, wave Ⅲ generated from superior olivary also display negative association with social functioning as evident by shortened Ⅲ/Ⅴ intervals. These bidirectional associations implied distinct electrophysiological mechanism across neural levels within this subgroup. Cluster four mainly displayed moderate social-relational and language impairments. Prolonged wave Ⅲ and Ⅴ latency presented pronounced association with sensory score. These findings reveal that delayed auditory processing at pontine-mesencephalic level may constitute a neurobiological mechanism underlying sensory processing malfunctioning in this subtype. An unexpected finding was that only moderate associations were found between clinical assessments outcomes and ABR components in cluster two that display severe multidimensional impairments. A predominant finding is that, among all the pathway nodes, only cluster 2 displays a robust correlation between superior olivary complex activity and overall ABC scores as evident by prolonged wave Ⅲ and Ⅲ/Ⅴ interval. Our results suggest that compromised olivary may contribute to the broad phenotypic presentation in this group. However, these associations failed to generalize to cluster three characterized by less board impairments. Conversely, we only observed shortened wave Ⅰ/Ⅲ interval caused by hypersensitive cochlear nerves. No robust associations were found in cluster one and five, indicating that preserved auditory pathway integrity in these relatively higher-functioning individuals. Discussion ABR is crucial in predicting clinical outcomes and potential biomarkers in children with ASD, and contributing to understanding auditory process in children ASD and offering implications for targeted intervention and future directions(Crasta, Gavin, & Davies, 2021 ; Williams, Abdelmessih, Key, & Woynaroski, 2021 ; Zhu et al., 2025 ). Previous studies have revealed heterogeneous alternation in ABR outcomes among individuals with ASD compared to typical developed peers(Blue et al., 2024 ; Fujihira et al., 2021 ; Liu et al., 2024 ; Miron et al., 2018 ; Santos et al., 2017 ; Simamora et al., 2024 ). First, our study collected ABR outcomes from bilateral ears using three different stimulations click (e.g.45dB, 65dB, and 85dB), and we established six XGBoost Classifiers using these data. Second, the XGBoost Classifier, with the best performances, were selected using six evaluation metrics. Thirdly, the cluster analysis was used to detect the potential subgroups within our sample, and the ANOVA analysis was used to depict the underlying phenotypes in each cluster. Finally, the Pearson correlation analysis was conducted to reveal the associations between ASD phenotypes and ABR outcomes used to train the best XGBoost Classifier. Notably, our study reveals divergent neurotheological characteristics across distinct phenotypic subgroups. Compromised signal transmission from cochlear nerves to inferior colliculus may associate with language-related impairments, whereas pontine-mesencephalic delays may account for sensory dysfunction in subtypes characterized by social-relational symptoms. Additionally, these neurophysiological findings are absent in relatively high-functioning individuals, revealing the heterogeneous auditory processing across the children with ASD. These findings suggested that ABR outcomes can be potential stratification biomarkers to distinguish among ASD phenotypes, offering neurobiological-proved targets for personalized intervention. Prediction model assembling ABR components for ASD In this study, we used the ABR outcomes collected from bilateral ears using 45dB, 65dB, and 85dB clicks to build predictions models. We found robust performance of models with over 90% prediction accuracy to identify participants with confirmed ASD. Our results reveal that temporal parameter extracted from ABR has been proved clinically features in accurately detecting children with ASD. Compared to other questionnaires or models built based on machine learning methods, models built solely based on ABR outcomes did outperform those incorporated with numerous standardized assessments and electronic database with AUROC reaching over 0.9 (e.g. AUROC: M-CHAT-R/F, 0.907, Social Communication Questionnaire, 0.80, Models built by Shyam et al., 0.895)(Rajagopalan, Zhang, Yahia, & Tammimies, 2024 ). Our developed model has displayed potential for clinical use as ASD screening tools incorporated with bioelectrical examination. In addition to clinical potentials, identifying discriminating predictors, such as ABR outcomes, for ASD detection is crucial for the clinical adaptation of convention screening protocol. Further explainable machine learning models can also inform clinicians about the underlying bioelectrical mechanism contributing to ASD detection. Further, they can promote the individualized intervention and follow-up. ABR outcomes and ASD phenotypes This study reveals distinct patterns of signal transmission within auditory-related pathway across different ASD phenotypes, revealing the neurophysiological heterogeneity in children with ASD. The predominant finding is that auditory pathway dysfunction may manifest in different formats across phenotypes and exhibits selected associations with specific clinical functions, for example those characterized by language-specific impairments or social-relational dysfunctions accompanied by sensory abnormalities. Our results concur with previous findings that prolonged ABR wave latency can be found in children with or suspected with ASD in bilateral ears(Li et al., 2020 ; Miron et al., 2018 ; Noorazar, Jabbari Moghaddam, Kharzaee, & Sohrabpour, 2020). In additions, our findings proved the point of view that reduced or disturbed neural responses may be associated with language-related function in children with ASD(Poulsen et al., 2024 ; Samoylov et al., 2024 ). Besides, aligning with the hypothesis that auditory perception is not always impaired in children with ASD, our findings delineate less affected auditory pathways in phenotypes that characterized by high-functioning performance(Hisaizumi & Tantam, 2024 ). Overall, these complementary observations collectively proposed a hypothesis that auditory perception in children with ASD tend to reflect heterogeneous neural transmission integrity rather than uniform compromise, implicating preserved pathway integrity as a potential neurophysiological marker of preserved function. Still, it is unclear why individuals with ASD show heterogeneity in ABR response found in different phenotypes. This divergent response pattern may be related to a number of factors such as age, gender, and IC division underlying these complicated phenotypes(Cacciato-Salcedo, Lao-Rodríguez, & Malmierca, 2026 ). Collectively, these findings emphasize the necessity of standardized protocols and stratified analysis to disentangle the complex interplay between acoustic variables and neurobiological heterogeneity in ASD. Limitation We acknowledge several limitations in our approach. Our models still need further validation for their generalizability across different populations in multiple clinical settings. Also, clinical history and ABC scores were not promising enough to define the testing target, hence combination with other additional tools are needed, for example eye-tracking results, brain-based biomarker or other standardized assessment results. Conclusion In summary, this study delineates a phenotype-dependent auditory dysfunction in children with ASD, revealing distinct associations among auditory pathway nodes and ABC domains including sensory, language, and social-relational behaviors. Our findings confirmed that compromised ABR components did not display in uniform formats but rather differed across different subgroups. Furthe research is needed to validate these neurophysiological markers, promoting precise personalized interventions tailored to ASD phenotypes. Declarations Ethics approval and consent to participate. All methods conducted in our study were carried out in accordance with relevant guidelines and regulations by qualified clinicians. Our study had achieved ethics approval from Shenzhen Children’s Hospital Ethics Committee. All informed consents were obtained before the administration of the participants from individuals or their legal guardian(s). Consent for publication Not applicable. Availability of data and materials Anonymous Data is available. Declaration of interest statement No potential conflict of interest was reported by the author(s). Funding Not applicable. Authors’ contributions Qingjie Zhang performed the data analysis and wrote the first draft of the manuscript. Ren Chunmei wrote the code scripts, performed the data analysis, and wrote the method sections of the manuscripts. Xianrong Liang collected and categorized the data. Yun Guojun and Peng Kanglong contributed to conception, design of the study, manuscript revision, read, and approved the submitted version. Acknowledgements Give special credit to the parents’ generosity to provide the assessment results of their children. References Abdelmageed RI, Youssef AM, Rihan LS, Abdelaziz AW. Validation of the autism behavior checklist in Egyptian children with autism spectrum disorder. Child Neuropsychol. 2024;1–16. 10.1080/09297049.2024.2309016 . Blue CM, Wong SJ, Dodson K. Auditory brainstem response findings in autism spectrum disorder speech delay population. Am J Otolaryngol. 2024;45(5):104417. 10.1016/j.amjoto.2024.104417 . Cacciato-Salcedo S, Lao-Rodríguez AB, Malmierca MS. Sex- and etiology-specific effects on predictive processing in the inferior colliculus of two rat models of autism. Commun Biol. 2026. 10.1038/s42003-026-09585-z . Crasta JE, Gavin WJ, Davies PL. Expanding our understanding of sensory gating in children with autism spectrum disorders. Clin Neurophysiol. 2021;132(1):180–90. 10.1016/j.clinph.2020.09.020 . Foster NE, Ouimet T, Tryfon A, Doyle-Thomas K, Anagnostou E, Hyde KL. Effects of Age and Attention on Auditory Global-Local Processing in Children with Autism Spectrum Disorder. J Autism Dev Disord. 2016;46(4):1415–28. 10.1007/s10803-015-2684-2 . Fujihira H, Itoi C, Furukawa S, Kato N, Kashino M. Auditory brainstem responses in adults with autism spectrum disorder. Clin Neurophysiol Pract. 2021;6:179–84. 10.1016/j.cnp.2021.04.004 . González-Cortés T, Gutiérrez-Contreras E, Espino-Silva PK, Haro-Santa Cruz J, Álvarez-Cruz D, Rosales-González CC, Ruiz-Flores P. Clinical Profile of Autism Spectrum Disorder in a Pediatric Population from Northern Mexico. J Autism Dev Disord. 2019;49(11):4409–20. 10.1007/s10803-019-04154-2 . Hisaizumi M, Tantam D. Enhanced sensitivity to pitch perception and its possible relation to language acquisition in autism. Autism Dev Lang Impair. 2024;9:23969415241248618. 10.1177/23969415241248618 . Krug DA, Arick J, Almond P. Behavior checklist for identifying severely handicapped individuals with high levels of autistic behavior. J Child Psychol Psychiatry. 1980;21(3):221–9. 10.1111/j.1469-7610.1980.tb01797.x . Li A, Gao G, Fu T, Pang W, Zhang X, Qin Z, Ge R. Continued development of auditory ability in autism spectrum disorder children: A clinical study on click-evoked auditory brainstem response. Int J Pediatr Otorhinolaryngol. 2020;138:110305. 10.1016/j.ijporl.2020.110305 . Liu Y, Liu Z, Diao T, Song P, Hu Q, Jiang N. Exploring the Relationship Between Auditory Brainstem Response Testing and Disease Progression in Pediatric Autism Spectrum Disorder. Altern Ther Health Med. 2024;30(3):51–5. Lord C, Brugha TS, Charman T, Cusack J, Dumas G, Frazier T, Veenstra-VanderWeele J. Autism spectrum disorder. Nat Rev Dis Primers. 2020;6(1):5. 10.1038/s41572-019-0138-4 . Miron O, Beam AL, Kohane IS. Auditory brainstem response in infants and children with autism spectrum disorder: A meta-analysis of wave V. Autism Res. 2018;11(2):355–63. 10.1002/aur.1886 . Mottron L, Dawson M, Soulières I, Hubert B, Burack J. Enhanced perceptual functioning in autism: an update, and eight principles of autistic perception. J Autism Dev Disord. 2006;36(1):27–43. 10.1007/s10803-005-0040-7 . Noorazar SG, Moghaddam J, Kharzaee Y, R., Sohrabpour M. Comparison of Auditory Brain Stem Responses and Otoacoustic Emission of Autism with Healthy Children. Galen Med J. 2020;9:e1937. 10.31661/gmj.v9i0.1937 . Poulsen R, Williams Z, Dwyer P, Pellicano E, Sowman PF, McAlpine D. How auditory processing influences the autistic profile: A review. Autism Res. 2024;17(12):2452–70. 10.1002/aur.3259 . Rajagopalan SS, Zhang Y, Yahia A, Tammimies K. Machine Learning Prediction of Autism Spectrum Disorder From a Minimal Set of Medical and Background Information. JAMA Netw Open. 2024;7(8):e2429229. 10.1001/jamanetworkopen.2024.29229 . Rice CE, Carpenter LA, Morrier MJ, Lord C, DiRienzo M, Boan A, Wiggins LD. Defining in Detail and Evaluating Reliability of DSM-5 Criteria for Autism Spectrum Disorder (ASD) Among Children. J Autism Dev Disord. 2022;52(12):5308–20. 10.1007/s10803-021-05377-y . Samoylov I, Arcara G, Buyanova I, Davydova E, Pereverzeva D, Sorokin A, Arutiunian V. Altered neural synchronization in response to 2 Hz amplitude-modulated tones in the auditory cortex of children with Autism Spectrum Disorder: An MEG study. Int J Psychophysiol. 2024;203:112405. 10.1016/j.ijpsycho.2024.112405 . Santos M, Marques C, Nóbrega Pinto A, Fernandes R, Coutinho MB, Almeida ESC. Autism spectrum disorders and the amplitude of auditory brainstem response wave I. Autism Res. 2017;10(7):1300–5. 10.1002/aur.1771 . Sigar P, Kathrein N, Gragas E, Kupis L, Uddin LQ, Nomi JS. Age-related changes in brain signal variability in autism spectrum disorder. Mol Autism. 2025;16(1):8. 10.1186/s13229-024-00631-3 . Simamora M, Zizlavsky S, Harjoprawito TJA, Wiguna T, Medise BE, Wibawanti R. Correlation between auditory brainstem responses, hyperacusis, and severity of autism spectrum disorder in young children with normal hearing at a tertiary referral center in Indonesia. Med J Malaysia. 2024;79(6):677–82. Udhnani MD, Lee NR. Evaluating Sex Differences in Language Abilities Within Down Syndrome and Autism Spectrum Disorder. Intellect Dev Disabil. 2025;63(1):29–41. 10.1352/1934-9556-63.1.29 . Wadden NP, Bryson SE, Rodger RS. A closer look at the Autism Behavior Checklist: discriminant validity and factor structure. J Autism Dev Disord. 1991;21(4):529–41. 10.1007/bf02206875 . Williams ZJ, Abdelmessih PG, Key AP, Woynaroski TG. Cortical Auditory Processing of Simple Stimuli Is Altered in Autism: A Meta-analysis of Auditory Evoked Responses. Biol Psychiatry Cogn Neurosci Neuroimaging. 2021;6(8):767–81. 10.1016/j.bpsc.2020.09.011 . Wu Y, Lu C, Li M, Li B, Shang X, Jian G, Chen H. Atypical Developmental Patterns of Sensorimotor-Related Networks in Autism Spectrum Disorder: A BrainAGE Study Based on Resting-State fMRI. Autism Res. 2025;18(4):765–73. 10.1002/aur.70008 . Zhang Y, Zhou Q, Gao L, Li J, Li H, Ji GJ, Li D. Abnormal Functional Connectivity of the Primary Sensory Network in Autism Spectrum Disorder: Sex Differences, Early Overdevelopment, and Clinical Significance. Brain Behav. 2025;15(3):e70363. 10.1002/brb3.70363 . Zhao Q, Luo Y, Mei X, Shao Z. Functional connectivity alterations in high-functioning preschool boys with autism spectrum disorder. Appl Neuropsychol Child. 2025;1–11. 10.1080/21622965.2025.2581091 . Zhu M, Chen F, Zhang Y, Zhang Z, Guo C. Auditory Global-Local Processing in Autism Spectrum Disorder: A Systematic Review and Meta-Analysis. J Autism Dev Disord. 2025. 10.1007/s10803-025-06901-0 . Tables Tables 1 to 5 are available in the Supplementary Files section. Additional Declarations No competing interests reported. Supplementary Files table1.xlsx table2.xlsx table3.xlsx table4.xlsx table5.xlsx Cite Share Download PDF Status: Under Review Version 1 posted Reviewers agreed at journal 12 May, 2026 Reviewers invited by journal 15 Apr, 2026 Editor assigned by journal 09 Apr, 2026 Submission checks completed at journal 09 Apr, 2026 First submitted to journal 01 Apr, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9288120","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":626819580,"identity":"5eee03bf-ce5b-4c13-8986-722675a5e136","order_by":0,"name":"Qingjie Zhang","email":"","orcid":"","institution":"Shenzhen Children's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Qingjie","middleName":"","lastName":"Zhang","suffix":""},{"id":626819581,"identity":"91e3ee38-cafb-452f-af53-e31a562afe7f","order_by":1,"name":"Chunmei Ren","email":"","orcid":"","institution":"Shenzhen Luohu Hospital of Traditional Chinese Medicine","correspondingAuthor":false,"prefix":"","firstName":"Chunmei","middleName":"","lastName":"Ren","suffix":""},{"id":626819582,"identity":"7f66e32f-57a2-45c3-8b04-05b2fcd2208d","order_by":2,"name":"Xianrong Liang","email":"","orcid":"","institution":"Shenzhen Children's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Xianrong","middleName":"","lastName":"Liang","suffix":""},{"id":626819583,"identity":"1c7b0443-e796-40d5-a8b6-62797e96a5cd","order_by":3,"name":"Guojun Yun","email":"","orcid":"","institution":"Shenzhen Children's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Guojun","middleName":"","lastName":"Yun","suffix":""},{"id":626819584,"identity":"106e32b5-de1e-4563-9f61-60065e225a34","order_by":4,"name":"Kanglong Peng","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABIElEQVRIie2RsU7DMBCGHUVyGC6ZLUVqXsFWpCLUKjwHmyGSWTrAwuwuyZLujRh4BjbGVF4NPECXVHkB2BElBhUikUBHJPxJXk732ff7ELJY/iDYm8/r59dpgjzZKdftodCvBKAUW2KRIqg+Ku+N/AdlRIQID7ByJOF7Khj0OETgutFJs2ku76bRMfJX9Wmm0OHu3W9ZinhyQXDA1iKOSy1YgYKUGuVoIfsV0Gy9pOCy6xkO/UxxQDAmRqGPA4ORGQ2BE+e21EbZ7qOIsxAq6twQMErVUe6HBms/uZQ8pdBm8bOUFarNwh/Oger++FFuVim3SZSrTeNnSeTli1X9dDUZUc37J9tBP6902304GA1spfvc1+Cm+eVXwWKxWP4PbzfuWOC82dnsAAAAAElFTkSuQmCC","orcid":"","institution":"Shenzhen Children's Hospital","correspondingAuthor":true,"prefix":"","firstName":"Kanglong","middleName":"","lastName":"Peng","suffix":""}],"badges":[],"createdAt":"2026-04-01 07:26:30","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9288120/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9288120/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":107639732,"identity":"f34d1d75-3de0-4564-ab74-d29119eb0006","added_by":"auto","created_at":"2026-04-23 13:11:04","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":389799,"visible":true,"origin":"","legend":"\u003cp\u003eLegend not included with this version.\u003c/p\u003e","description":"","filename":"figure1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9288120/v1/57a167df9228cc75080f964b.jpg"},{"id":107639752,"identity":"d4bd00d7-cf66-4828-8792-dd6aec567cd9","added_by":"auto","created_at":"2026-04-23 13:11:07","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":4924176,"visible":true,"origin":"","legend":"\u003cp\u003eLegend not included with this version.\u003c/p\u003e","description":"","filename":"figure2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9288120/v1/82804cca42ad6610413d2cd9.jpg"},{"id":107707599,"identity":"60360bed-e9df-4ba9-9680-edf2ed5e63ed","added_by":"auto","created_at":"2026-04-24 09:20:43","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":5524478,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9288120/v1/05cb6a93-8df6-49b7-8db1-4d9a1b914f75.pdf"},{"id":107639935,"identity":"9c1b5d2d-4688-4b1c-8b20-8e9be23b23b0","added_by":"auto","created_at":"2026-04-23 13:11:23","extension":"xlsx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":9708,"visible":true,"origin":"","legend":"","description":"","filename":"table1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9288120/v1/6a12e221ceeb7a3c905d73a9.xlsx"},{"id":107639751,"identity":"06f63a65-b480-4b6e-acaf-6aaddaf1c368","added_by":"auto","created_at":"2026-04-23 13:11:07","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":13044,"visible":true,"origin":"","legend":"","description":"","filename":"table2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9288120/v1/80dae95c86c1f099da70ef5e.xlsx"},{"id":107639753,"identity":"97390e60-8b26-44cf-bbf4-36ee08c8193a","added_by":"auto","created_at":"2026-04-23 13:11:07","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":9696,"visible":true,"origin":"","legend":"","description":"","filename":"table3.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9288120/v1/b3394d1fefcf5c13d3c239b9.xlsx"},{"id":107639734,"identity":"111a546a-21d2-4a94-b24b-41b991d07ec4","added_by":"auto","created_at":"2026-04-23 13:11:04","extension":"xlsx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":10970,"visible":true,"origin":"","legend":"","description":"","filename":"table4.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9288120/v1/4d41a67400446c817e4218cf.xlsx"},{"id":107639699,"identity":"2d13328d-7387-4e60-92cb-638191e9db15","added_by":"auto","created_at":"2026-04-23 13:11:02","extension":"xlsx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":9580,"visible":true,"origin":"","legend":"","description":"","filename":"table5.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9288120/v1/371a7985dc57560139e13519.xlsx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Identifying Distinct Electrophysiological Endophenotypes in Autism Spectrum Disorder: A Large-Scale Machine Learning Approach Integrating Auditory Brainstem Response and Behavioral Phenotyping","fulltext":[{"header":"Introduction","content":"\u003cp\u003eAutism spectrum disorder (ASD) is a highly heritable and heterogeneous neurodevelopmental disorder whose symptoms emerge in early developmental stage and persist along the overall lifespan(Lord et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). ASD is characterized by core disorders in social communication, restricted and repetitive behaviors (RBB), and various co-occurring symptoms(Gonz\u0026aacute;lez-Cort\u0026eacute;s et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Udhnani \u0026amp; Lee, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). In addition to these core features, atypical perceptual function is another distinctive characteristic commonly observed in children with ASD(Zhu, Chen, Zhang, Zhang, \u0026amp; Guo, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). For example, typical developed children often perceive global structures prior to local details, exemplifying the adage of seeing the forest before trees, while children with ASD tend to process the subtle contents first. These results are now integrated into various frameworks for describing the perception function in children with ASD(Mottron, Dawson, Souli\u0026egrave;res, Hubert, \u0026amp; Burack, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2006\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe Weak Central Coherence (WCC) proposed that children with ASD cannot integrate subtle contents into concrete concepts(Mottron et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2006\u003c/span\u003e). Controversially, the Enhanced Perceptual Functioning (EPF) hypothesized that children with ASD can outperform their typical developed peers in local perceptions, but the global processing is not compromised(Foster et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). (Zhu et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Recently, one mate-analysis conducted by Zhu et al. provided reasonable solutions to these debates that global processing is not compromised but rather affected by various tasks parameters(Zhu et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eRecently, electroencephalograms (EEG) and functional magnetic resonate images (fMRI) have emerged as valuable evidence for corroborating these solutions under connectivity perspective. Using electroencephalogram recording, one study found that disrupted connectivity in the right central-right parietal and left central-left frontal connections may interfere with auditory integration(Zhao, Luo, Mei, \u0026amp; Shao, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). The resting-state fMRI has also shown that children with ASD presented abnormal functional connectivity in primary sensory network involving the auditory cortices(Zhang et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Recent studies also revealed that age and sex have impacts on these structural connectivity(Sigar et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Wu et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Zhang et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). These studies demonstrate how EEG and fMRI can offer valuable perspectives on the links between structures and functions.\u003c/p\u003e \u003cp\u003eWhile previous studies have explored the associations between structures and functions, recent studies have tried to examine the bioelectrical signal using Auditory Brainstem Response (ABR). One study tried to examine the ABR in a small sample of children with ASD and found a notable prolongation in wave Ⅰ and Ⅲ latency as well as Ⅰ/Ⅴ interpeak latency (IPL), but decreased wave Ⅲ/Ⅴ IPL(Liu et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). However, another study found no significance in wave Ⅰ, Ⅲ and Ⅴ, but shorten summating potential (SP) using electrocochleograms before wave Ⅰ(Fujihira, Itoi, Furukawa, Kato, \u0026amp; Kashino, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Despite this phenotypic variability in the directionality of temporal shifts in ABR, all observed alternations in auditory response consistently indicated abnormal neural conduction within the auditory pathway. Overall, these study reveals that abnormal auditory processing may originate from the cochlear level rather than high-level cortex, implying that possible activity-induced neural remodeling across the whole auditory pathway, (Blue, Wong, \u0026amp; Dodson, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Liu et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Miron, Beam, \u0026amp; Kohane, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). These findings in bioelectrical signal demonstrate a paradigm shift from cortico-centric models toward a peripheral-to-central framework.\u003c/p\u003e \u003cp\u003eThis study aimed to systematically reveal the associations between ABR outcomes and ASD phenotypes.\u003c/p\u003e"},{"header":"Materia and Methods","content":"\u003cp\u003eParticipants\u003c/p\u003e \u003cp\u003eParticipants were recruited from local referral programs of the government service including maternal and childcare service center, education institutions, and community agency. Children with definitive or suspected diagnosis of ASD were referred for comprehensive evaluation through this program. The referred individuals would accept interdisciplinary assessment to achieve definitive diagnosis of ASD and receive tailored intervention. Comprehensive assessment routinely induces the administration of ABR, ABC, CARS, and other standardized tools. Prior to administration, all necessary consents were obtained from all subjects and/or their legal guardian(s).\u003c/p\u003e \u003cp\u003eA multidisciplinary team is invited to confirm the diagnosis of ASD. The members included a psychiatrist with ADOS-2 license and two senior neurologists.\u003c/p\u003e \u003cp\u003ePrior to administration, all the subjects and/or their legal guardians(s) had signed necessary consents. Subjects were included if they were older than 2 years of age and met the diagnostic criteria of the DSM-5-TR (version 2022), and the clinical presentation consisted of three manifestations of social disorders as well as two manifestations of stereotyped repetitive behaviors, as follows:\u003c/p\u003e \u003cp\u003eSocialization disorders\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\u003col\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eSocial-emotional interaction disorders\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003ePhysical Motor Behavioral (nonverbal communication) Social Disorders\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003eSocial Relationship Development Disorder (Development, Formation, Understanding)\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003c/ol\u003e \u003cp\u003e\u003c/p\u003e \u003cp\u003eStereotypical Repetitive Behavior (SRB)\u003c/p\u003e \u003cp\u003e\u003c/p\u003e\u003col\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003e Repetition of stereotyped motor movements, object manipulation, or verbal expressions\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eDevelopment of repetitive, routine, and patterned stereotyped verbal or nonverbal behaviors\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eExtremely limited, fixed interests, or attention spans\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eAbnormal responses (extreme sensitivity or the opposite) to sensory input, both normal and abnormal (environmental)\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003c/ol\u003e\u003cp\u003e\u003c/p\u003e \u003cp\u003eFor specific diagnostic criteria, refer to C.E. Rice's suggested judgment criteria for each entry(Rice et al., \u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e). Patients with other unrelated conditions were excluded, including peripheral nerve injury, myelitis, spinal embolism syndrome, seizures, and fractures.\u003c/p\u003e \u003cp\u003eClinical measure\u003c/p\u003e \u003cp\u003eAutism Behavior Checklist (ABC)\u003c/p\u003e \u003cp\u003eThe Autism Behavior Checklist (ABC) assessment consists of 57 items that involve possibly all typical autistic behaviors. Items are categorized into five components including relating, sensory, language, body use and object manipulation, and social and self-help. Participants were asked according to the item description given by one researcher and rated the item if their children behaved as the item described. Besides, each item contained its own score ranging from 1 to 4 points according to the item weights. The weighted score of each item is decided by the occurrence frequency in Krug’s study(Krug, Arick, \u0026amp; Almond, \u003cspan class=\"CitationRef\"\u003e1980\u003c/span\u003e). For example, item 1 occurred more than item 2, then item 1 is endowed with 4 points, and item 2 is endowed with 2 points. If one item is rated, then participant gets the according score (e.g. item 1 scores 0/4, item 2 scores 0/2). The original cut off score was set at 68, and total score above 67 indicated severe symptoms or higher possibility to get diagnosis as ASD(Wadden, Bryson, \u0026amp; Rodger, \u003cspan class=\"CitationRef\"\u003e1991\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe interrater reliability is 0.85, and the intra-rater reliability is 0.82(Abdelmageed, Youssef, Rihan, \u0026amp; Abdelaziz, \u003cspan class=\"CitationRef\"\u003e2024\u003c/span\u003e; Krug et al., \u003cspan class=\"CitationRef\"\u003e1980\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe Childhood Autism Rating Scale First Edition\u003c/p\u003e \u003cp\u003eCARS1 was constructed to collect information from major caregivers’ interviews, direct or indirect observations, and structural interviews. A total of 15 items are involved in CARS1, including relation to people, imitation, emotional response, body use, object use, adaptation to change, visual response, listening response, sensory response, emotion, verbal communication, gesture, activity status, intellectual response, and overall impressions. A four-point rating scale is utilized to quantify the symptoms severity, where one point refers to normal behavior and four points refers to inappropriate behaviors that are different from normal developed children. The total score is the sum of all items, and the higher scores refers to more severe autistic symptoms. CARS1 was delivered by trained/licensed clinicians or researchers with appropriate training for the necessary interview’s techniques with parents and caregivers and judgement criteria.\u003c/p\u003e \u003cp\u003eAuditory brainstem response protocol\u003c/p\u003e \u003cp\u003eThe Neuron Spectrum 5 (Neurosoft LLC, Ivanovo, Russian) was used to record the ABR outcome in a sound-blocked room with comfortable temperature and minimal ambient noise low than 30 db. All the participants received routine examinations to exclude any ear abnormalities. Then, ABR were recorded when they were either sitting in resting-statue or sleeping. Electrodes were placed at the middle of the forehead or Cz in 10–20 system (active), nasal root or Fpz (ground), and mastoids (references) respectively. Electrode impedance was maintained lower than 5 kΩ.\u003c/p\u003e \u003cp\u003eClick stimuli (100µs duration, 21.1/s rate) was delivered monaurally via insert headphones at 45dB, 65dB, 85dB. Signals were bandpass filtered (100Hz-3000Hz), and 2000 sweeps were averaged per trial. Recorded outcomes included latencies (Wave Ⅰ, Ⅱ, Ⅲ, Ⅳ, Ⅴ), interpeak latencies (Ⅰ/Ⅲ, Ⅲ/Ⅴ, Ⅰ/Ⅴ).\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eModel evaluation and selection\u003c/p\u003e \u003cp\u003e \u003cb\u003eTable.1\u003c/b\u003e displays six performance indicators that are used to evaluate and select the XGBoost Classifiers using ABR collected from bilateral ears under conditions using different stimulations (45dB, 65dB, and 85dB clicks). These evaluation metrics can present comprehensive perspectives on classification accuracy, discriminative capability, and clinical applicability.\u003c/p\u003e \u003cp\u003eCluster analysis\u003c/p\u003e \u003cp\u003eOur study adopted the Ward Linkage method, or the minimum variance method to detect the potential cluster in our sample. To determine the reasonable cluster amount, we utilized the common elbow method based on Bayesian and Akaike information criteria. The ANOVA analysis was conducted to depict the parameters differences among clusters, and significant level was set at p values lower than 0.05 in all tests. The Bonferroni post-hoc analysis was conducted to depict the difference between clusters.\u003c/p\u003e \u003cp\u003ePearson Correlation analysis\u003c/p\u003e \u003cp\u003eThe Pearson correlation test was conducted to depict the associations between standardized assessment results and ABR outcomes.\u003c/p\u003e \u003c/div\u003e"},{"header":"Result","content":"\u003cp\u003eSample characteristics\u003c/p\u003e\u003cp\u003eOur study managed to recruit a sample that consisted of 1883 children and adolescents. Table.2 presents the demographic data collected from this sample. Our sample aged around 64.47 months old (64.47 ± 20.04), and most of them were boys (1600/1883). Our study tried to recruit a sample with balance in terms of age range designed based on Chinese Education System (nursery: 0–3 years, kindergarten: 3–6 years, primary school: 6–12, and junior high school: 12–15 years), but we ultimately obtained a sample with 103 from nursery, 1128 administered in kindergarten, 592 from primary school, and 6 from junior high school. In terms of ASD classification, we managed to obtain a sample with balance using ABC and CARS evaluation criteria as shown in \u003cb\u003eTable.2\u003c/b\u003e.\u003c/p\u003e\u003cp\u003ePerformance evaluation and selection of XGBoost Classifiers\u003c/p\u003e\u003cp\u003e \u003cb\u003eTable.3\u003c/b\u003e and \u003cb\u003eFigure.1\u003c/b\u003e present the performance metrics of XGBoost Classifiers across different feature configurations. Among these models, the XGBoost Classifier trained using data collected from right ear using 65dB clicks demonstrated superior overall performance, achieving the highest F1 score (0.7085) and ranking second in AUROC (0.9208), accuracy (0.9177), and precision (0.7793). Importantly, this configuration exhibited the optimal balance between precision and recall, which is crucial in clinical diagnostic application where both false negatives and false positives carry significant clinical implications.\u003c/p\u003e\u003cp\u003eBesides, Models trained on data from conditions using 45dB generally underperformed relative to other scenarios, indicating insufficient feature presentiveness. Conversively, models trained on outcomes from 85dB protocols display competitive AUROC values but compromised precision, suggesting overfitting or redundant features inclusion.\u003c/p\u003e\u003cp\u003eUsing cluster analysis to identify clusters in samples\u003c/p\u003e\u003cp\u003eThis study conducted cluster analysis and ANOVA analysis to identify homogeneous subgroups in this sample based on ABC performance. \u003cb\u003eTable.4\u003c/b\u003e shows the phenotypes differences defined by ABC domains in these clusters.\u003c/p\u003e\u003cp\u003eCluster 2 represented the subgroups with severe multidimensional impairment profile. This group exhibited the most pronounced problematic behaviors across sensory, relating and body and object use. Children in this group display significant hyper/hyposensitivity to environmental stimuli, pronounced deficits in inter-person relationship, and prominent SRB.\u003c/p\u003e\u003cp\u003eChildren with ASD in cluster 4 showed intermediate to high impairment in social related language application, but relatively less affected motor behavior.\u003c/p\u003e\u003cp\u003eCluster 6 consisted of children with ASD accompanying linguistic problems and less affected sensory function. This behavior profile indicates one distinct phenotype characterized by significant communication deficit involving both verbal and non-verbal language disorders.\u003c/p\u003e\u003cp\u003eCluster 1 stood for one moderately sized subgroup with intermediate social communication deficit and less influenced in motor behavior. Cluster 3 demonstrated the subsample with mild to moderate generalized autistic traits. Cluster 5 contained the most participants who display minimal autistic behaviors.\u003c/p\u003e\u003cp\u003eIn summary, \u003cb\u003eTable.5\u003c/b\u003e depicts one severity gradient from severe multidimensional impairment (cluster 2) through domain specific severe impairments (cluster 4 and 6), intermediate generalized impairment (cluster 1 and 3), to minimal impairment (cluster 5).\u003c/p\u003e\u003cp\u003ePearson correlation outcomes\u003c/p\u003e\u003cp\u003eTo depict the neurophysiological traits of clinical heterogeneity in children with ASD, correlation analysis was conducted between auditory brainstem response (ABR) metrics and ABC subscale scores across six distinct phenotypic clusters \u003cb\u003e(Figure.2).\u003c/b\u003e For correlation coefficients with an absolute magnitude over 0.3, we found that signal transmission from the cochlear nerve to inferior colliculus may be correlated with phenotypical characteristics in children with ASD.\u003c/p\u003e\u003cp\u003eCluster six characterized by predominant language impairment displays the strongest and most complicated correlation metrics. Primary finding was that prolonged wave Ⅲ latency presented robust association with language scores, leading to sequential extended Ⅲ/Ⅴ interval (r equals to 0.51, 0.52 respectively). Besides, these temporal shifts also reflect the disturbed signal conduction from superior olivary to inferior colliculus. Additionally, these temporal delays may involve less responsive reaction originated from cochlear nerves as evident by prolonged wave Ⅰ latency and alternations in Ⅰ/Ⅲ and Ⅰ/Ⅴ intervals. These temporal alternations may explain why latency Ⅴ was compressed by either passive transmission deficits or active central hyperresponsive mechanism. Conversely, compressed wave Ⅰ latency was observed along with shortened Ⅰ/Ⅲ and Ⅰ/Ⅴ intervals, indicating accelerated neural conduction originated from cochlear nerves as relating score get higher. Additionally, wave Ⅲ generated from superior olivary also display negative association with social functioning as evident by shortened Ⅲ/Ⅴ intervals. These bidirectional associations implied distinct electrophysiological mechanism across neural levels within this subgroup.\u003c/p\u003e\u003cp\u003eCluster four mainly displayed moderate social-relational and language impairments. Prolonged wave Ⅲ and Ⅴ latency presented pronounced association with sensory score. These findings reveal that delayed auditory processing at pontine-mesencephalic level may constitute a neurobiological mechanism underlying sensory processing malfunctioning in this subtype.\u003c/p\u003e\u003cp\u003eAn unexpected finding was that only moderate associations were found between clinical assessments outcomes and ABR components in cluster two that display severe multidimensional impairments. A predominant finding is that, among all the pathway nodes, only cluster 2 displays a robust correlation between superior olivary complex activity and overall ABC scores as evident by prolonged wave Ⅲ and Ⅲ/Ⅴ interval. Our results suggest that compromised olivary may contribute to the broad phenotypic presentation in this group. However, these associations failed to generalize to cluster three characterized by less board impairments. Conversely, we only observed shortened wave Ⅰ/Ⅲ interval caused by hypersensitive cochlear nerves. No robust associations were found in cluster one and five, indicating that preserved auditory pathway integrity in these relatively higher-functioning individuals.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eABR is crucial in predicting clinical outcomes and potential biomarkers in children with ASD, and contributing to understanding auditory process in children ASD and offering implications for targeted intervention and future directions(Crasta, Gavin, \u0026amp; Davies, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Williams, Abdelmessih, Key, \u0026amp; Woynaroski, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Zhu et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Previous studies have revealed heterogeneous alternation in ABR outcomes among individuals with ASD compared to typical developed peers(Blue et al., \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Fujihira et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Liu et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Miron et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Santos et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Simamora et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). First, our study collected ABR outcomes from bilateral ears using three different stimulations click (e.g.45dB, 65dB, and 85dB), and we established six XGBoost Classifiers using these data. Second, the XGBoost Classifier, with the best performances, were selected using six evaluation metrics. Thirdly, the cluster analysis was used to detect the potential subgroups within our sample, and the ANOVA analysis was used to depict the underlying phenotypes in each cluster. Finally, the Pearson correlation analysis was conducted to reveal the associations between ASD phenotypes and ABR outcomes used to train the best XGBoost Classifier. Notably, our study reveals divergent neurotheological characteristics across distinct phenotypic subgroups. Compromised signal transmission from cochlear nerves to inferior colliculus may associate with language-related impairments, whereas pontine-mesencephalic delays may account for sensory dysfunction in subtypes characterized by social-relational symptoms. Additionally, these neurophysiological findings are absent in relatively high-functioning individuals, revealing the heterogeneous auditory processing across the children with ASD. These findings suggested that ABR outcomes can be potential stratification biomarkers to distinguish among ASD phenotypes, offering neurobiological-proved targets for personalized intervention.\u003c/p\u003e \u003cp\u003ePrediction model assembling ABR components for ASD\u003c/p\u003e \u003cp\u003eIn this study, we used the ABR outcomes collected from bilateral ears using 45dB, 65dB, and 85dB clicks to build predictions models. We found robust performance of models with over 90% prediction accuracy to identify participants with confirmed ASD. Our results reveal that temporal parameter extracted from ABR has been proved clinically features in accurately detecting children with ASD. Compared to other questionnaires or models built based on machine learning methods, models built solely based on ABR outcomes did outperform those incorporated with numerous standardized assessments and electronic database with AUROC reaching over 0.9 (e.g. AUROC: M-CHAT-R/F, 0.907, Social Communication Questionnaire, 0.80, Models built by Shyam et al., 0.895)(Rajagopalan, Zhang, Yahia, \u0026amp; Tammimies, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Our developed model has displayed potential for clinical use as ASD screening tools incorporated with bioelectrical examination. In addition to clinical potentials, identifying discriminating predictors, such as ABR outcomes, for ASD detection is crucial for the clinical adaptation of convention screening protocol. Further explainable machine learning models can also inform clinicians about the underlying bioelectrical mechanism contributing to ASD detection. Further, they can promote the individualized intervention and follow-up.\u003c/p\u003e \u003cp\u003eABR outcomes and ASD phenotypes\u003c/p\u003e \u003cp\u003eThis study reveals distinct patterns of signal transmission within auditory-related pathway across different ASD phenotypes, revealing the neurophysiological heterogeneity in children with ASD. The predominant finding is that auditory pathway dysfunction may manifest in different formats across phenotypes and exhibits selected associations with specific clinical functions, for example those characterized by language-specific impairments or social-relational dysfunctions accompanied by sensory abnormalities. Our results concur with previous findings that prolonged ABR wave latency can be found in children with or suspected with ASD in bilateral ears(Li et al., \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Miron et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Noorazar, Jabbari Moghaddam, Kharzaee, \u0026amp; Sohrabpour, 2020). In additions, our findings proved the point of view that reduced or disturbed neural responses may be associated with language-related function in children with ASD(Poulsen et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Samoylov et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Besides, aligning with the hypothesis that auditory perception is not always impaired in children with ASD, our findings delineate less affected auditory pathways in phenotypes that characterized by high-functioning performance(Hisaizumi \u0026amp; Tantam, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Overall, these complementary observations collectively proposed a hypothesis that auditory perception in children with ASD tend to reflect heterogeneous neural transmission integrity rather than uniform compromise, implicating preserved pathway integrity as a potential neurophysiological marker of preserved function. Still, it is unclear why individuals with ASD show heterogeneity in ABR response found in different phenotypes. This divergent response pattern may be related to a number of factors such as age, gender, and IC division underlying these complicated phenotypes(Cacciato-Salcedo, Lao-Rodr\u0026iacute;guez, \u0026amp; Malmierca, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2026\u003c/span\u003e). Collectively, these findings emphasize the necessity of standardized protocols and stratified analysis to disentangle the complex interplay between acoustic variables and neurobiological heterogeneity in ASD.\u003c/p\u003e \u003cp\u003eLimitation\u003c/p\u003e \u003cp\u003eWe acknowledge several limitations in our approach. Our models still need further validation for their generalizability across different populations in multiple clinical settings. Also, clinical history and ABC scores were not promising enough to define the testing target, hence combination with other additional tools are needed, for example eye-tracking results, brain-based biomarker or other standardized assessment results.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIn summary, this study delineates a phenotype-dependent auditory dysfunction in children with ASD, revealing distinct associations among auditory pathway nodes and ABC domains including sensory, language, and social-relational behaviors. Our findings confirmed that compromised ABR components did not display in uniform formats but rather differed across different subgroups. Furthe research is needed to validate these neurophysiological markers, promoting precise personalized interventions tailored to ASD phenotypes.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eEthics approval and consent to participate.\u003c/h2\u003e\n\u003cp\u003eAll methods conducted in our study were carried out in accordance with relevant guidelines and regulations by qualified clinicians. Our study had achieved ethics approval from Shenzhen Children\u0026rsquo;s Hospital Ethics Committee. All informed consents were obtained before the administration of the participants from individuals or their legal guardian(s).\u003c/p\u003e\n\u003cp\u003eConsent for publication\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003eAvailability of data and materials\u003c/p\u003e\n\u003cp\u003eAnonymous Data is available.\u003c/p\u003e\n\u003ch2\u003eDeclaration of interest statement\u003c/h2\u003e\n\u003cp\u003eNo potential conflict of interest was reported by the author(s).\u003c/p\u003e\n\u003ch2\u003eFunding\u003c/h2\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003eAuthors\u0026rsquo; contributions\u003c/p\u003e\n\u003cp\u003eQingjie Zhang performed the data analysis and wrote the first draft of the manuscript. Ren Chunmei wrote the code scripts, performed the data analysis, and wrote the method sections of the manuscripts. Xianrong Liang collected and categorized the data. Yun Guojun and Peng Kanglong contributed to conception, design of the study, manuscript revision, read, and approved the submitted version.\u003c/p\u003e\n\u003ch2\u003eAcknowledgements\u003c/h2\u003e\n\u003cp\u003eGive special credit to the parents\u0026rsquo; generosity to provide the assessment results of their children.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAbdelmageed RI, Youssef AM, Rihan LS, Abdelaziz AW. Validation of the autism behavior checklist in Egyptian children with autism spectrum disorder. Child Neuropsychol. 2024;1\u0026ndash;16. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1080/09297049.2024.2309016\u003c/span\u003e\u003cspan address=\"10.1080/09297049.2024.2309016\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBlue CM, Wong SJ, Dodson K. Auditory brainstem response findings in autism spectrum disorder speech delay population. Am J Otolaryngol. 2024;45(5):104417. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.amjoto.2024.104417\u003c/span\u003e\u003cspan address=\"10.1016/j.amjoto.2024.104417\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCacciato-Salcedo S, Lao-Rodr\u0026iacute;guez AB, Malmierca MS. Sex- and etiology-specific effects on predictive processing in the inferior colliculus of two rat models of autism. Commun Biol. 2026. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s42003-026-09585-z\u003c/span\u003e\u003cspan address=\"10.1038/s42003-026-09585-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCrasta JE, Gavin WJ, Davies PL. Expanding our understanding of sensory gating in children with autism spectrum disorders. Clin Neurophysiol. 2021;132(1):180\u0026ndash;90. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.clinph.2020.09.020\u003c/span\u003e\u003cspan address=\"10.1016/j.clinph.2020.09.020\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFoster NE, Ouimet T, Tryfon A, Doyle-Thomas K, Anagnostou E, Hyde KL. Effects of Age and Attention on Auditory Global-Local Processing in Children with Autism Spectrum Disorder. J Autism Dev Disord. 2016;46(4):1415\u0026ndash;28. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10803-015-2684-2\u003c/span\u003e\u003cspan address=\"10.1007/s10803-015-2684-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFujihira H, Itoi C, Furukawa S, Kato N, Kashino M. Auditory brainstem responses in adults with autism spectrum disorder. Clin Neurophysiol Pract. 2021;6:179\u0026ndash;84. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cnp.2021.04.004\u003c/span\u003e\u003cspan address=\"10.1016/j.cnp.2021.04.004\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGonz\u0026aacute;lez-Cort\u0026eacute;s T, Guti\u0026eacute;rrez-Contreras E, Espino-Silva PK, Haro-Santa Cruz J, \u0026Aacute;lvarez-Cruz D, Rosales-Gonz\u0026aacute;lez CC, Ruiz-Flores P. Clinical Profile of Autism Spectrum Disorder in a Pediatric Population from Northern Mexico. J Autism Dev Disord. 2019;49(11):4409\u0026ndash;20. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10803-019-04154-2\u003c/span\u003e\u003cspan address=\"10.1007/s10803-019-04154-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHisaizumi M, Tantam D. Enhanced sensitivity to pitch perception and its possible relation to language acquisition in autism. Autism Dev Lang Impair. 2024;9:23969415241248618. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1177/23969415241248618\u003c/span\u003e\u003cspan address=\"10.1177/23969415241248618\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKrug DA, Arick J, Almond P. Behavior checklist for identifying severely handicapped individuals with high levels of autistic behavior. J Child Psychol Psychiatry. 1980;21(3):221\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/j.1469-7610.1980.tb01797.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1469-7610.1980.tb01797.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi A, Gao G, Fu T, Pang W, Zhang X, Qin Z, Ge R. Continued development of auditory ability in autism spectrum disorder children: A clinical study on click-evoked auditory brainstem response. Int J Pediatr Otorhinolaryngol. 2020;138:110305. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.ijporl.2020.110305\u003c/span\u003e\u003cspan address=\"10.1016/j.ijporl.2020.110305\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu Y, Liu Z, Diao T, Song P, Hu Q, Jiang N. Exploring the Relationship Between Auditory Brainstem Response Testing and Disease Progression in Pediatric Autism Spectrum Disorder. Altern Ther Health Med. 2024;30(3):51\u0026ndash;5.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLord C, Brugha TS, Charman T, Cusack J, Dumas G, Frazier T, Veenstra-VanderWeele J. Autism spectrum disorder. Nat Rev Dis Primers. 2020;6(1):5. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41572-019-0138-4\u003c/span\u003e\u003cspan address=\"10.1038/s41572-019-0138-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMiron O, Beam AL, Kohane IS. Auditory brainstem response in infants and children with autism spectrum disorder: A meta-analysis of wave V. Autism Res. 2018;11(2):355\u0026ndash;63. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/aur.1886\u003c/span\u003e\u003cspan address=\"10.1002/aur.1886\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMottron L, Dawson M, Souli\u0026egrave;res I, Hubert B, Burack J. Enhanced perceptual functioning in autism: an update, and eight principles of autistic perception. J Autism Dev Disord. 2006;36(1):27\u0026ndash;43. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10803-005-0040-7\u003c/span\u003e\u003cspan address=\"10.1007/s10803-005-0040-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNoorazar SG, Moghaddam J, Kharzaee Y, R., Sohrabpour M. Comparison of Auditory Brain Stem Responses and Otoacoustic Emission of Autism with Healthy Children. Galen Med J. 2020;9:e1937. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.31661/gmj.v9i0.1937\u003c/span\u003e\u003cspan address=\"10.31661/gmj.v9i0.1937\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePoulsen R, Williams Z, Dwyer P, Pellicano E, Sowman PF, McAlpine D. How auditory processing influences the autistic profile: A review. Autism Res. 2024;17(12):2452\u0026ndash;70. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/aur.3259\u003c/span\u003e\u003cspan address=\"10.1002/aur.3259\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRajagopalan SS, Zhang Y, Yahia A, Tammimies K. Machine Learning Prediction of Autism Spectrum Disorder From a Minimal Set of Medical and Background Information. JAMA Netw Open. 2024;7(8):e2429229. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1001/jamanetworkopen.2024.29229\u003c/span\u003e\u003cspan address=\"10.1001/jamanetworkopen.2024.29229\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRice CE, Carpenter LA, Morrier MJ, Lord C, DiRienzo M, Boan A, Wiggins LD. Defining in Detail and Evaluating Reliability of DSM-5 Criteria for Autism Spectrum Disorder (ASD) Among Children. J Autism Dev Disord. 2022;52(12):5308\u0026ndash;20. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10803-021-05377-y\u003c/span\u003e\u003cspan address=\"10.1007/s10803-021-05377-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSamoylov I, Arcara G, Buyanova I, Davydova E, Pereverzeva D, Sorokin A, Arutiunian V. Altered neural synchronization in response to 2 Hz amplitude-modulated tones in the auditory cortex of children with Autism Spectrum Disorder: An MEG study. Int J Psychophysiol. 2024;203:112405. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.ijpsycho.2024.112405\u003c/span\u003e\u003cspan address=\"10.1016/j.ijpsycho.2024.112405\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSantos M, Marques C, N\u0026oacute;brega Pinto A, Fernandes R, Coutinho MB, Almeida ESC. Autism spectrum disorders and the amplitude of auditory brainstem response wave I. Autism Res. 2017;10(7):1300\u0026ndash;5. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/aur.1771\u003c/span\u003e\u003cspan address=\"10.1002/aur.1771\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSigar P, Kathrein N, Gragas E, Kupis L, Uddin LQ, Nomi JS. Age-related changes in brain signal variability in autism spectrum disorder. Mol Autism. 2025;16(1):8. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s13229-024-00631-3\u003c/span\u003e\u003cspan address=\"10.1186/s13229-024-00631-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSimamora M, Zizlavsky S, Harjoprawito TJA, Wiguna T, Medise BE, Wibawanti R. Correlation between auditory brainstem responses, hyperacusis, and severity of autism spectrum disorder in young children with normal hearing at a tertiary referral center in Indonesia. Med J Malaysia. 2024;79(6):677\u0026ndash;82.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUdhnani MD, Lee NR. Evaluating Sex Differences in Language Abilities Within Down Syndrome and Autism Spectrum Disorder. Intellect Dev Disabil. 2025;63(1):29\u0026ndash;41. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1352/1934-9556-63.1.29\u003c/span\u003e\u003cspan address=\"10.1352/1934-9556-63.1.29\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWadden NP, Bryson SE, Rodger RS. A closer look at the Autism Behavior Checklist: discriminant validity and factor structure. J Autism Dev Disord. 1991;21(4):529\u0026ndash;41. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/bf02206875\u003c/span\u003e\u003cspan address=\"10.1007/bf02206875\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWilliams ZJ, Abdelmessih PG, Key AP, Woynaroski TG. Cortical Auditory Processing of Simple Stimuli Is Altered in Autism: A Meta-analysis of Auditory Evoked Responses. Biol Psychiatry Cogn Neurosci Neuroimaging. 2021;6(8):767\u0026ndash;81. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.bpsc.2020.09.011\u003c/span\u003e\u003cspan address=\"10.1016/j.bpsc.2020.09.011\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu Y, Lu C, Li M, Li B, Shang X, Jian G, Chen H. Atypical Developmental Patterns of Sensorimotor-Related Networks in Autism Spectrum Disorder: A BrainAGE Study Based on Resting-State fMRI. Autism Res. 2025;18(4):765\u0026ndash;73. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/aur.70008\u003c/span\u003e\u003cspan address=\"10.1002/aur.70008\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang Y, Zhou Q, Gao L, Li J, Li H, Ji GJ, Li D. Abnormal Functional Connectivity of the Primary Sensory Network in Autism Spectrum Disorder: Sex Differences, Early Overdevelopment, and Clinical Significance. Brain Behav. 2025;15(3):e70363. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/brb3.70363\u003c/span\u003e\u003cspan address=\"10.1002/brb3.70363\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhao Q, Luo Y, Mei X, Shao Z. Functional connectivity alterations in high-functioning preschool boys with autism spectrum disorder. Appl Neuropsychol Child. 2025;1\u0026ndash;11. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1080/21622965.2025.2581091\u003c/span\u003e\u003cspan address=\"10.1080/21622965.2025.2581091\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhu M, Chen F, Zhang Y, Zhang Z, Guo C. Auditory Global-Local Processing in Autism Spectrum Disorder: A Systematic Review and Meta-Analysis. J Autism Dev Disord. 2025. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10803-025-06901-0\u003c/span\u003e\u003cspan address=\"10.1007/s10803-025-06901-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003eTables 1 to 5 are available in the Supplementary Files section.\u003c/p\u003e\n"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"journal-of-neurodevelopmental-disorders","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jndd","sideBox":"Learn more about [Journal of Neurodevelopmental Disorders](http://jneurodevdisorders.biomedcentral.com/)","snPcode":"11689","submissionUrl":"https://submission.nature.com/new-submission/11689/3","title":"Journal of Neurodevelopmental Disorders","twitterHandle":"@BioMedCentral","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"ASD, Machine learning, ABR, Prediction, Auditory, Sensory","lastPublishedDoi":"10.21203/rs.3.rs-9288120/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9288120/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eOur study aimed to explore the correlations between Auditory Brainstem Response outcomes and Autism phenotypes.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eA total of 1883 children with or suspected of being with ASD were enrolled. The related features were acquired by using the Autism Behavior Checklist (ABC), Children Autism Rating Scale (CARS) and ABR outcomes. The cluster analysis was conducted to detect potential subgroups within our samples, and the ANOVA analysis was conducted to reveal the differences among these subgroups using data from clinical assessments. The Pearson Correlation analysis was conducted to depict the associations between ABR results and different autism phenotypes.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eOur results revealed different brainstem vulnerabilities across six ASD clusters. Cluster 6 exhibited robust language-associated temporal alternations, while cluster 4 displayed sensory-related pontine-mesencephalic delays. Besides, compromised olivary function uniquely presented correlations with ABC total scores in cluster 2. Notably, less affected auditory pathways were found in clusters with high-functioning performance.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eOur findings confirmed that compromised ABR components did not display in uniform formats but rather differed across different subgroups. Furthe research is needed to validate these neurophysiological markers, promoting precise personalized interventions tailored to ASD phenotypes.\u003c/p\u003e","manuscriptTitle":"Identifying Distinct Electrophysiological Endophenotypes in Autism Spectrum Disorder: A Large-Scale Machine Learning Approach Integrating Auditory Brainstem Response and Behavioral Phenotyping","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-23 13:10:26","doi":"10.21203/rs.3.rs-9288120/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"114795142518953441681282983740095197195","date":"2026-05-12T19:44:58+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-04-15T14:54:03+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-09T16:49:33+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-04-09T16:48:49+00:00","index":"","fulltext":""},{"type":"submitted","content":"Journal of Neurodevelopmental Disorders","date":"2026-04-01T07:14:36+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"journal-of-neurodevelopmental-disorders","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jndd","sideBox":"Learn more about [Journal of Neurodevelopmental Disorders](http://jneurodevdisorders.biomedcentral.com/)","snPcode":"11689","submissionUrl":"https://submission.nature.com/new-submission/11689/3","title":"Journal of Neurodevelopmental Disorders","twitterHandle":"@BioMedCentral","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"f32fbda9-4821-402c-8fe5-da5aa06f3f20","owner":[],"postedDate":"April 23rd, 2026","published":true,"recentEditorialEvents":[{"type":"reviewerAgreed","content":"114795142518953441681282983740095197195","date":"2026-05-12T19:44:58+00:00","index":40,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-23T13:10:27+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-23 13:10:26","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9288120","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9288120","identity":"rs-9288120","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}