Share:
Share this content in WeChat
X
• Clinical Article •
Research on autism subtypes based on individualized extreme functional connectivity
DU Xiangyi  ZHOU Yuan  ZHU Yuanqiang  LI Jiani  LI Leilei  ZHENG Jianmin 

DOI:10.12015/issn.1674-8034.2026.11.004.


[Abstract] Objective Autism spectrum disorder (ASD) is characterized by marked neurobiological heterogeneity. This study aimed to identify individual-level extreme functional connectivity features in patients with ASD using the Tolerance Interval (TI) method and to explore neurophysiological subtypes and their value for clinical diagnosis through cluster analysis.Materials and Methods This study is a retrospective study based on publicly available databases, which included 907 participants from the Autism Brain Imaging Data Exchange (ABIDE) database [451 in the ASD group and 456 in the healthy controls (HC) group]. Individual-level extreme functional connections were identified using the TI method. Multiple clustering algorithms were applied to evaluate the subtype structure within the ASD group, and the associations between subtype-specific features and Autism Diagnostic Observation Schedule (ADOS) scores were further examined. Finally, a support vector machine (SVM) classification model based on the fusion of multiple kernel functions was constructed to evaluate the clinical discriminative utility of the identified subtypes.Results ASD-related extreme functional connections identified by the TI method showed a dispersed distribution pattern, with hub nodes mainly located in the right inferior frontal gyrus, right middle frontal gyrus, and left superior temporal gyrus. Cluster analysis revealed two well-defined neurophysiological subtypes in ASD. Subtype 1 was characterized by widespread functional dysregulation involving multiple systems, and some extreme functional connections within this subtype were significantly associated with ADOS communication and stereotyped behavior scores (P < 0.001). Subtype 2 showed highly concentrated network impairments, with abnormal features focused on core nodes of the salience network, including the basal ganglia, thalamus, and amygdala, as well as reduced connectivity within the default mode network. The classification accuracy for the overall ASD versus HC model was 56.4%, indicating limited discriminative ability of the unstratified model. In contrast, subtype-based stratification significantly improved classification performance: the accuracy was 61.8% for Subtype 1 and 83.7% for Subtype 2.Conclusions Subtyping based on extreme functional connectivity not only helps clarify the pathological mechanisms of ASD but also significantly improves the performance of machine learning models for computer-aided diagnosis of autism.
[Keywords] autism spectrum disorder;functional magnetic resonance imaging;magnetic resonance imaging;extreme functional connectivity;tolerance interval;machine learning

DU Xiangyi   ZHOU Yuan   ZHU Yuanqiang   LI Jiani   LI Leilei   ZHENG Jianmin*  

Department of Diagnostic Radiology, the First Affiliated Hospital of Air Force Military Medical University, Xi'an 710032, China

Corresponding author: ZHENG J M, E-mail: jmzheng1986@126.com

Conflicts of interest   None.

Received  2026-01-22
Accepted  2026-09-10
DOI: 10.12015/issn.1674-8034.2026.11.004
DOI:10.12015/issn.1674-8034.2026.11.004.

[1]
CHEN J, ZHANG C, WANG R, et al. Molecular basis underlying functional connectivity of fusiform gyrus subregions: A transcriptome-neuroimaging spatial correlation study[J]. Cortex, 2022, 152: 59-73. DOI: 10.1016/j.cortex.2022.03.016.
[2]
BARAN B, NGUYEN Q T H, MYLONAS D, et al. Increased resting-state thalamocortical functional connectivity in children and young adults with autism spectrum disorder[J]. Autism Res, 2023, 16(2): 271-279. DOI: 10.1002/aur.2875.
[3]
WANG J, KAWATA N Y S, CAO X, et al. White-Matter fiber tract and resting-state functional connectivity abnormalities in young children with autism spectrum disorder[J/OL]. Neuroimage, 2025, 310: 121109 [2026-01-22]. https://doi.org/10.1016/j.neuroimage.2025.121109. DOI: 10.1016/j.neuroimage.2025.121109.
[4]
KHANDAN KHADEM-REZA Z, SHAHRAM M A, ZARE H. Altered resting-state functional connectivity of the brain in children with autism spectrum disorder[J]. Radiol Phys Technol, 2023, 16(2): 284-291. DOI: 10.1007/s12194-023-00717-2.
[5]
FELDMAN D, PRIGGE M, ALEXANDER A, et al. Flexible nonlinear modeling reveals age-related differences in resting-state functional brain connectivity in autistic males from childhood to mid-adulthood[J/OL]. Mol Autism, 2025, 16(1): 24 [2026-01-22]. https://doi.org/10.1186/s13229-025-00657-1. DOI: 10.1186/s13229-025-00657-1.
[6]
XU Y Y, ZHANG G M, WU G R, et al. Research progress on multimodal structural-functional MRI studies of hierarchical large-scalebrain network abnormalities in autism spectrum disorder[J]. Chin J Magn Reson Imaging, 2026, 17(1): 128-133. DOI: 10.12015/issn.1674-8034.2026.01.020.
[7]
LIU Q, LAI H, LE J, et al. Identifying brain functional subtypes and corresponding task performance profiles in autism spectrum disorder[J]. Mol Psychiatry, 2025, 30(11): 5034-5044. DOI: 10.1038/s41380-025-03086-x.
[8]
TADDEI M, CUESTA P, ANNUNZIATA S, et al. Correlation between autistic traits and brain functional connectivity in preschoolers with autism spectrum disorder: a resting state MEG study[J]. Neurol Sci, 2024, 45(9): 4549-4561. DOI: 10.1007/s10072-024-07528-2.
[9]
JIN J, WANG X Q, YANG X, et al. Abnormal individualized peak functional connectivity toward potential repetitive transcranial magnetic stimulation treatment of autism spectrum disorder[J]. Hum Brain Mapp, 2023, 44(16): 5450-5459. DOI: 10.1002/hbm.26455.
[10]
WANG J, WANG X, WANG R, et al. Atypical Resting-State Functional Connectivity of Intra/Inter-Sensory Networks Is Related to Symptom Severity in Young Boys With Autism Spectrum Disorder[J/OL]. Front Physiol, 2021, 12: 626338 [2026-01-22]. https://doi.org/10.3389/fphys.2021.626338. DOI: 10.3389/fphys.2021.626338.
[11]
KORNISCH M, GONZALEZ C, IKUTA T. Functional connectivity of the posterior cingulate cortex in autism spectrum disorder[J/OL]. Psychiatry Res Neuroimaging, 2024, 342: 111848 [2026-01-22]. https://doi.org/10.1016/j.pscychresns.2024.111848. DOI: 10.1016/j.pscychresns.2024.111848.
[12]
LAWRENCE K E, HERNANDEZ L M, BOWMAN H C, et al. Sex Differences in Functional Connectivity of the Salience, Default Mode, and Central Executive Networks in Youth with ASD[J]. Cereb Cortex, 2020, 30(9): 5107-5120. DOI: 10.1093/cercor/bhaa105.
[13]
WAN L, LI Y, ZHU G, et al. Multimodal investigation of dynamic brain network alterations in autism spectrum disorder: Linking connectivity dynamics to symptoms and developmental trajectories[J/OL]. Neuroimage, 2024, 302: 120895 [2026-01-22]. https://doi.org/10.1016/j.neuroimage.2024.120895. DOI: 10.1016/j.neuroimage.2024.120895.
[14]
ZHANG J, ZHANG Z, SUN H, et al. Personalized functional network mapping for autism spectrum disorder and attention-deficit/hyperactivity disorder[J/OL]. Transl Psychiatry, 2024, 14(1): 92 [2026-01-22]. https://doi.org/10.1038/s41398-024-02797-z. DOI: 10.1038/s41398-024-02797-z.
[15]
LU H, DONG Q, GAO L, et al. Sex heterogeneity of dynamic brain activity and functional connectivity in autism spectrum disorder[J]. Autism Res, 2024, 17(9): 1796-1809. DOI: 10.1002/aur.3227.
[16]
WANG N, YAO D, MA L, et al. Multi-site clustering and nested feature extraction for identifying autism spectrum disorder with resting-state fMRI[J/OL]. Med Image Anal, 2022, 75: 102279 [2026-01-22]. https://doi.org/10.1016/j.media.2021.102279. DOI: 10.1016/j.media.2021.102279.
[17]
REN P, BI Q, PANG W, et al. Stratifying ASD and characterizing the functional connectivity of subtypes in resting-state fMRI[J/OL]. Behav Brain Res, 2023, 449: 114458 [2026-01-22]. https://doi.org/10.1016/j.bbr.2023.114458. DOI: 10.1016/j.bbr.2023.114458.
[18]
FENG W, LIU G, ZENG K, et al. A review of methods for classification and recognition of ASD using fMRI data[J/OL]. J Neurosci Methods, 2022, 368: 109456 [2026-01-22]. https://doi.org/10.1016/j.jneumeth.2021.109456. DOI: 10.1016/j.jneumeth.2021.109456.
[19]
ZHANG J, LI W, ROSKOS L K, et al. Immunogenicity assay cut point determination using nonparametric tolerance limit[J]. J Immunol Methods, 2017, 442: 29-34. DOI: 10.1016/j.jim.2017.01.001.
[20]
FANG K, NIU L, WEN B, et al. Individualized resting-state functional connectivity abnormalities unveil two major depressive disorder subtypes with contrasting abnormal patterns of abnormality[J/OL]. Transl Psychiatry, 2025, 15(1): 45 [2026-01-22]. https://doi.org/10.1038/s41398-025-03268-9. DOI: 10.1038/s41398-025-03268-9.
[21]
MARON-KATZ A, ZHANG Y, NARAYAN M, et al. Individual Patterns of Abnormality in Resting-State Functional Connectivity Reveal Two Data-Driven PTSD Subgroups[J]. Am J Psychiatry, 2020, 177(3): 244-253. DOI: 10.1176/appi.ajp.2019.19010060.
[22]
DI MARTINO A, YAN C G, LI Q, et al. The autism brain imaging data exchange: towards a large-scale evaluation of the intrinsic brain architecture in autism[J]. Mol Psychiatry, 2014, 19(6): 659-667. DOI: 10.1038/mp.2013.78.
[23]
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[J/OL]. Sci Data, 2017, 4: 170010 [2026-01-22]. https://doi.org/10.1038/sdata.2017.10. DOI: 10.1038/sdata.2017.10.
[24]
HONG S J, DE WAEL R VOS, BETHLEHEM R A I, et al. Atypical functional connectome hierarchy in autism[J/OL]. Nat Commun, 2019, 10(1): 1022 [2026-01-22]. https://doi.org/10.1038/s41467-019-08944-1. DOI: 10.1038/s41467-019-08944-1.
[25]
YAN C G, WANG X D, ZUO X N, et al. DPABI: Data Processing & Analysis for (Resting-State) Brain Imaging[J]. Neuroinformatics, 2016, 14(3): 339-351. DOI: 10.1007/s12021-016-9299-4.
[26]
ZHANG D, LIU J, FAN L, et al. Quantitative description of the relationship between the enhancement of distraction-suppression and brain local state alteration after transcranial direct current stimulation[J/OL]. Front Neurosci, 2022, 16: 984893 [2026-01-22]. https://doi.org/10.3389/fnins.2022.984893. DOI: 10.3389/fnins.2022.984893.
[27]
ZHU P Y, CHEN X, ZOU L L, et al. Default mode network functional connectivity changes in adolescent bipolar disorder patientsafter short-term treatment[J]. Chin J Magn Reson Imaging, 2025, 16(12):14-21. DOI: 10.12015/issn.1674-8034.2025.12.003.
[28]
ZHANG Y, XUE Y, WU X, et al. Selecting Multiple Node Statistics Jointly from Functional Connectivity Networks for Brain Disorders Identification[J]. Brain Topogr, 2022, 35(5-6): 559-571. DOI: 10.1007/s10548-022-00914-z.
[29]
FAN L, LI H, ZHUO J, et al. The Human Brainnetome Atlas: A New Brain Atlas Based on Connectional Architecture[J]. Cereb Cortex, 2016, 26(8): 3508-3526. DOI: 10.1093/cercor/bhw157.
[30]
WANG Y, MA H, ZHANG D F, et al. Normalizing radiomics features from multiscale structural MRI of the adolescent brain[J]. Chin J Magn Reson Imaging, 2023, 14(8):100-107. DOI: 10.12015/issn.1674-8034.2023.08.016.
[31]
ZERAATI M, DAVOODI A. ASD-GraphNet: A novel graph learning approach for Autism Spectrum Disorder diagnosis using fMRI data[J/OL]. Comput Biol Med, 2025, 196(Pt B): 110723 [2026-01-22]. https://doi.org/10.1016/j.compbiomed.2025.110723. DOI: 10.1016/j.compbiomed.2025.110723.
[32]
LIAO W, WU G R, XU Q, et al. DynamicBC: a MATLAB toolbox for dynamic brain connectome analysis[J]. Brain Connect, 2014, 4(10): 780-790. DOI: 10.1089/brain.2014.0253.
[33]
BENKARIM O, PAQUOLA C, PARK B Y, et al. Connectivity alterations in autism reflect functional idiosyncrasy[J/OL]. Commun Biol, 2021, 4(1): 1078 [2026-01-22]. https://doi.org/10.1038/s42003-021-02572-6. DOI: 10.1038/s42003-021-02572-6.
[34]
XU S, LI M, YANG C, et al. Altered Functional Connectivity in Children With Low-Function Autism Spectrum Disorders[J/OL]. Front Neurosci, 2019, 13: 806 [2026-01-22]. https://doi.org/10.3389/fnins.2019.00806. DOI: 10.3389/fnins.2019.00806.
[35]
DENG J, LEI T, DU X. Effects of sensory integration training on balance function and executive function in children with autism spectrum disorder: evidence from Footscan and fNIRS[J/OL]. Front Psychol, 2023, 14: 1269462 [2026-01-22]. https://doi.org/10.3389/fpsyg.2023.1269462. DOI: 10.3389/fpsyg.2023.1269462.
[36]
TRYFON A, FOSTER N E V, SHARDA M, et al. Speech perception in autism spectrum disorder: An activation likelihood estimation meta-analysis[J]. Behav Brain Res, 2018, 338: 118-127. DOI: 10.1016/j.bbr.2017.10.025.
[37]
KITAMURA S, MATSUOKA K, TAKAHASHI M, et al. Association of adverse childhood experience-related increase in neurite density with sensory over-responsivity in autism spectrum disorder: A neurite orientation dispersion and density imaging study[J]. J Psychiatr Res, 2023, 161: 316-323. DOI: 10.1016/j.jpsychires.2023.03.029.
[38]
SHEN L, ZHANG J, FAN S, et al. Cortical thickness abnormalities in autism spectrum disorder[J]. Eur Child Adolesc Psychiatry, 2024, 33(1): 65-77. DOI: 10.1007/s00787-022-02133-0.
[39]
LEE J E, KIM S, PARK S, et al. Atypical maturation of the functional connectome hierarchy in autism[J/OL]. Mol Autism, 2025, 16(1): 21 [2026-01-22]. https://doi.org/10.1186/s13229-025-00641-9. DOI: 10.1186/s13229-025-00641-9.
[40]
JIANG A, MA X, LI S, et al. Age-atypical brain functional networks in autism spectrum disorder: a normative modeling approach[J]. Psychol Med, 2024, 54(9): 2042-2053. DOI: 10.1017/s0033291724000138.
[41]
AHN Y A, MOFFITT J M, TAO Y, et al. Objective Measurement of Social Gaze and Smile Behaviors in Children with Suspected Autism Spectrum Disorder During Administration of the Autism Diagnostic Observation Schedule, 2nd Edition[J]. J Autism Dev Disord, 2024, 54(6): 2124-2137. DOI: 10.1007/s10803-023-05990-z.
[42]
WAGNER L, BANCHIK M, OKADA N J, et al. Associations between thalamocortical functional connectivity and sensory over-responsivity in infants at high likelihood for ASD[J]. Cereb Cortex, 2023, 33(12): 8075-8086. DOI: 10.1093/cercor/bhad100.
[43]
SHANG J, SHEN E, YU Y, et al. Relationship between abnormal intrinsic functional connectivity of subcortices and autism symptoms in high-functioning adults with autism spectrum disorder[J/OL]. Psychiatry Res Neuroimaging, 2024, 337: 111762 [2026-01-22]. https://doi.org/10.1016/j.pscychresns.2023.111762. DOI: 10.1016/j.pscychresns.2023.111762.
[44]
LEE Y, PARK B Y, JAMES O, et al. Autism Spectrum Disorder Related Functional Connectivity Changes in the Language Network in Children, Adolescents and Adults[J/OL]. Front Hum Neurosci, 2017, 11: 418 [2026-01-22]. https://doi.org/10.3389/fnhum.2017.00418. DOI: 10.3389/fnhum.2017.00418.
[45]
WANG K, LI K, NIU X. Altered Functional Connectivity in a Triple-Network Model in Autism With Co-occurring Attention Deficit Hyperactivity Disorder[J/OL]. Front Psychiatry, 2021, 12 [2026-01-22]. https://doi.org/10.3389/fpsyt.2021.736755. DOI: 10.3389/fpsyt.2021.736755.
[46]
ATTANASIO M, MAZZA M, LE DONNE I, et al. Salience Network in Autism: preliminary results on functional connectivity analysis in resting state[J]. Eur Arch Psychiatry Clin Neurosci, 2025, 275(8): 2479-2492. DOI: 10.1007/s00406-024-01949-y.
[47]
HAGHIGHAT H, MIRZAREZAEE M, ARAABI B N, et al. Functional Networks Abnormalities in Autism Spectrum Disorder: Age-Related Hypo and Hyper Connectivity[J]. Brain Topogr, 2021, 34(3): 306-322. DOI: 10.1007/s10548-021-00831-7.
[48]
GIRAULT J B, NISHINO T, TALOVIĆ M, et al. Functional connectivity between the visual and salience networks and autistic social features at school-age[J/OL]. J Neurodev Disord, 2025, 17(1): 23 [2026-01-22]. https://doi.org/10.1186/s11689-025-09613-9. DOI: 10.1186/s11689-025-09613-9.
[49]
WANG L, QIN Y, YANG S, et al. Posterior default mode network is associated with the social performance in male children with autism spectrum disorder: A dynamic causal modeling analysis based on triple-network model[J/OL]. Hum Brain Mapp, 2024, 45(8): e26750 [2026-01-22]. https://doi.org/10.1002/hbm.26750. DOI: 10.1002/hbm.26750.
[50]
ZHOU R, SUN C, SUN M, et al. Altered intra- and inter-network connectivity in autism spectrum disorder[J]. Aging (Albany NY), 2024, 16(11): 10004-10015. DOI: 10.18632/aging.205913.
[51]
ILIOSKA I, OLDEHINKEL M, LLERA A, et al. Connectome-wide Mega-analysis Reveals Robust Patterns of Atypical Functional Connectivity in Autism[J]. Biol Psychiatry, 2023, 94(1): 29-39. DOI: 10.1016/j.biopsych.2022.12.018.
[52]
NAIR A, JALAL R, LAWRENCE K E, et al. Triple-Network Functional Connectivity in Adolescents With Autism Spectrum Disorder Compared to Early-Onset Psychosis[J/OL]. Autism Res, 2026, 19(4): e70163 [2026-01-22]. https://doi.org/10.1002/aur.70163. DOI: 10.1002/aur.70163.
[53]
DUBEY S, SARKHEL S, SENGUPTA S, et al. Broader autism phenotype, the default mode network, and brain entropy: A network-level hypothesis[J/OL]. Res Neurodivers, 2026, 2 [2026-01-22]. https://doi.org/10.1016/j.rin.2026.100044. DOI: 10.1016/j.rin.2026.100044.

PREV Alterations in white matter functional networks and functional covariance networks in children with self-limited epilepsy with centrotemporal spikes
NEXT An interpretable machine-learning model based on structural and metabolic features of the cortical-striatal-cerebellar circuit for predicting recurrence risk after recovery in adolescents with first-episode major depressive disorder
  



Tel & Fax: +8610-67113815    E-mail: editor@cjmri.cn