Age-related individual-specific subspace of autism spectrum disorder based on common orthogonal basis extraction algorithm improves the accuracy of clinical symptoms prediction

Age-related individual-specific subspace of autism spectrum disorder based on common orthogonal basis extraction algorithm improves the accuracy of clinical symptoms prediction | 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 Age-related individual-specific subspace of autism spectrum disorder based on common orthogonal basis extraction algorithm improves the accuracy of clinical symptoms prediction Tingting Luo, Jie Zhang, Manxue Zhang, Lei Li, Shengnan Zhao, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8832166/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 11 You are reading this latest preprint version Abstract Background Autism spectrum disorder (ASD) is a group of highly heterogeneous neurodevelopmental disorders with onset in early childhood. Functional magnetic resonance imaging (fMRI) studies have revealed that ASD is related to altered functional connectivity (AFC), and that individual-specific change isolated from AFC with data-driven method of individuals with ASD has significantly enhanced the predictive ability for behavioral symptoms. Although few studies have incorporated age as a factor, it is critically important for ASD, a neurodevelopmental disorder, as age significantly influences the disorder's onset and progression. Methods In this study, we analyzed 437 participants (208 ASD, 229 typical development) from the Autism Brain Imaging Data Exchange, employed the common orthogonal basis extraction (COBE) algorithm to isolate age-related individual-specific features and examine their predictive abilities for clinical behaviors. A validation analysis was performed in an independent sample. Results We found that the age-related, individual-specific feature set significantly improved behavioral prediction. The most substantial improvement was observed in predicting social behavior among adolescents with ASD, which showed a peak increase of 133% and an average increase of 41% compared to AFC. These findings were replicated in an independent validation dataset. Conclusion The age-related individual-specific subspace demonstrates superior predictive power for clinical symptoms. This underscores the critical importance of incorporating both inter-individual variability and the developmental perspective into ASD biomarker exploration and targeted intervention research. Autism spectrum disorder Age Individual-specific Functional connectivity Brain network Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Autism spectrum disorder (ASD) is a group of neurodevelopmental disorders with onset in early childhood, characterized as impaired social communication and restricted and repetitive behaviors (RRB)[ 1 – 3 ]. Research has revealed that alterations in brain connectivity patterns emerge prior to the onset of behavioral symptoms in ASD[ 4 ]. Aberrant structural and functional brain development likely contributes significantly to the neural mechanisms underlying ASD[ 5 ]. Functional magnetic resonance imaging (fMRI) studies have revealed that ASD is related to brain functional connectivity (FC)[ 6 ]. Furthermore, these FC patterns exhibiting strong age-dependent correlations encompassing numerous reported patterns of both hypoconnectivity and hyperconnectivity across large-scale brain networks, like within networks, between networks, and at the whole-brain level[ 7 – 11 ]. For example, a study found increased FC in the auditory network (AN) and the sensorimotor network in children with ASD aged 3–7 years, and notably enhanced positive connectivity between the primary visual network and higher visual network[ 12 ]. Research on older children (8–13 years) found that children with ASD showed decreased FC within the default mode network (DMN), increased FC between the DMN and the salient network (SN), and decreased FC in the DMN midline core (medial prefrontal cortex-posterior cingulate cortex; mPFC-PCC) that correlated negatively with ASD symptom severity[ 13 ]. Haghighat et al. reported that FC abnormalities in ASD primarily manifest as hyperconnectivity in childhood, whereas both hyperconnectivity and hypoconnectivity were observed in adolescence and adulthood, and they proposed that age is a significant factor influencing FC disturbances[ 9 ]. Therefore, it is essential to map the dynamic neurodevelopmental trajectories in ASD and highlights the importance of age-stratified research in elucidating ASD heterogeneity. Traditional case-control brain imaging analyses focus on group-level comparisons by describing the concept of the "average patient", ignoring the underlying biological heterogeneity among individual patients[ 14 ]. The traditional case-control paradigm fails to account for these sources of heterogeneity of individual patients and consequently neglects the heterogeneity inherent in ASD-related neuroimaging features[ 15 ]. Individual differences in brain structure and function provide rich information related to neuropathology[ 16 – 18 ]. For instance, a study exploring the heterogeneity of brain structure in ASD based on individual structural covariance network found that ASD showed significantly altered structural covariance edges mainly involved in the frontal and subcortical regions, which highlights the crucial role of frontal and subcortical regions in the heterogeneity of ASD[ 16 ]. Additionally, studies based on FC found individual-specific fMRI-Subspaces improve FC prediction of behavior[ 17 , 18 ]. Therefore, it is important to detect individual-specific difference, especially for disorders with high heterogeneity like ASD, to explore the neurological pathological basis. The common orthogonal basis extraction (COBE) algorithm could identify and separate individual-shared and individual-specific subspaces from multiblock data [ 19 , 20 ]. These individual-shared components help reveal connections among members of the dataset and can characterize the dataset itself, while individual-specific part aids in identifying each unique member within the dataset. Previous studies have applied COBE to resting-state functional magnetic resonance imaging (rs-fMRI) data for feature separation[ 18 , 21 ], and found that individual-specific connectivity improved the prediction of clinical symptoms, highlighting the importance for dissecting heterogeneity of ASD[ 18 ]. However, previous study failed to take age into account, which is highly important to ASD. This study will focus on altered functional connectivity (AFC) in ASD from a developmental perspective. Reference to previous studies[ 22 ], we divided participants into four age groups. We employ the COBE algorithm to dissociate individual-shared features across different age groups (age-invariant) and individual-specific features (age-related) from AFC of individuals with ASD. We aim to (1) examine whether age-related individual-specific AFC enhances the predictive power for clinical symptoms; (2) identify relationships between these age-related, individual-specific AFC patterns and core behavioral symptoms across different developmental stages; and (3) validate the findings in an independent dataset. Methods Participants MRI data and phenotypic data from 3 sites with the largest participant numbers were downloaded from the autism brain imaging data exchange (ABIDE) datasets ( https://fcon_1000.projects.nitrc.org/indi/abide/ )[ 10 , 23 ]. We selected three identical sites with a large sample size and high data quality from ABIDE I and ABIDE II respectively, where participant inclusion/exclusion criteria and MRI scanning parameters remained consistent. Finally, 437 participants (age range: 6–39 years; ASD = 208, typical development (TD) = 229) were included in this study. Detail of sites included is provided in Supplementary Materials TABLE S1 . In 2000, Jeffrey[ 24 ] put forward the theory of emerging adulthood, believing that 18–25 years old is an independent developmental stage with unique psychosocial characteristics. Based on this, participants were divided into 4 age groups: 6–11 years, 12–17 years, 18–25 years, and 26–39 years. Validation dataset: External validation was performed in an independent sample collected at Laboratory of Child and Adolescent Psychiatry, Mental Health Center and Psychiatric Laboratory, West China Hospital, Sichuan University. The ASD subjects were mainly recruited from the outpatient department of Mental Health Center of West China Hospital of Sichuan University and West China Second University Hospital of Sichuan University. They were initially diagnosed by experienced child psychiatrists according to the Diagnostic and Statistical Manual of Mental Disorders − 5 (DSM-5)[ 1 ] or the International Statistical Classification of Diseases and Related Health Problems 10th Revision (ICD-10)[ 25 ]. Subsequently, all participants underwent the Autism Diagnostic Observation Schedule (ADOS)[ 26 ], a gold-standard diagnostic tool. Participants were included in the study only if they scored above the established diagnostic thresholds for ASD. For children aged 5 and below, we used the Gesell Development Schedules (GDS) to assess their developmental levels[ 27 ], which are expressed in developmental quotient (DQ). For children aged 6 and above, we used the third edition of the Chinese-Wechsler Intelligence Scale for Children (C-WISC)[ 28 ] to evaluate the cognitive levels of the subjects, which are represented by intelligence quotient (IQ). Subjects were excluded based on the following criteria: (1) neuropsychiatric disorders, such as epilepsy, encephalitis; (2) history of craniocerebral injury; (3) monogenetic diseases, such as fragile X syndrome or Angelman syndrome; (4) taking psychiatric medications during assessment; or (5) refusal to undergo MRI or the presence of contraindications for MRI, including fixed metal dentures/orthodontic appliances, pacemakers, or cochlear implants. Finally, 245 participants (age range: 2–17 years; ASD = 145, TD = 99) were included in this study. We divided the subjects into 3 groups based on age: 2–5 years (ASD = 66, TD = 30), 6–11 years (ASD = 61, TD = 55), and 12–17 years (ASD = 19, TD = 14). MRI Data Preprocessing The parameters of MRI data collection see Supplementary Materials. Preprocessing of rs-fMRI data was performed on the MATLAB 2023a platform through RESTplus—an enhanced toolkit for rs-fMRI data processing[ 29 ]. The preprocessing pipeline included: (1) Removal of the first 10 time points; (2) Slice timing correction; (3) Head motion correction; (4) Spatial normalization, using Diffeomorphic Anatomical Registration Through Exponentiated Lie algebra (DARTEL)[ 30 ] for registration to Montreal Neurological Institute(MNI) standard space with 3.0×3.0×3.0 mm³ resampling; (5) Spatial smoothing (6 mm full-width at half-maximum kernel); (6) Linear trend removal; (7) Band-pass filtering (0.01–0.08 Hz); and (8) Nuisance covariate regression, including cerebrospinal fluid signals, white matter signals, and Friston-24 head motion parameters. Quality control measures: Excluded participants with poor registration quality; Excluded participants exhibiting head framewise displacement (FD) > 3 mm translation or > 3° rotation. Information about number of excluded participants due to excessive head motion and missing time points see Supplementary Materials Table S2. Compute Altered Functional Connectivity In this study, the Dosenbach’ atlas[ 31 ] was utilized to parcellate the brain into 160 regions of interest (ROIs), which were subsequently categorized into six functionally distinct subnetworks: DMN, frontoparietal network (FPN), cingulo-opercular network (CON), sensorimotor network (SMN), occipital network (ON), and cerebellum (CB). Pairwise Pearson correlations between ROI time courses yielded a symmetric 160 × 160 FC matrix for each subject, with correlation coefficients converted to z-scores through Fisher’s r-to-z transformation to approximate normality. Sex and sites were included as covariates in the regression. In line with previous work[ 18 ], AFC was computed for each ASD participant as follows: $$\:AFC=\frac{{FC}_{ASD}-mean\left({FC}_{TD}\right)}{SD\left({FC}_{TD}\right)}$$ 1 AFC represents the altered level of FC strength between one individual with ASD and mean FC strength of TD. For each ASD participant, a 160 × 160 AFC matrix was obtained. Common Orthogonal Basis Extraction AFC matrices computed above were stratified into four age groups: 6–11, 12–17, 18–25, and 26–39 years. We applied the COBE [ 19 , 20 ] algorithm to these age-stratified AFC matrices. Originally designed for multiblock data (i.e., collections of matrices) decomposition, COBE separates shared and block-specific subspaces[ 20 ]. In this study, we demarcated age groups as experimental "blocks." This allowed the COBE algorithm to extract both a shared subspace across all groups and an age-related subspace that captures the connectivity changes specific to each cohort. The illustration of COBE sees Supplementary Materials Fig. S1 . Code for COBE is publicly available: https://github.com/ClinicalBrainLab/OCD_Cerebellar-Visual-Community . Predicting clinical symptoms using the individual-specific subspace The AFC matrix encompasses both AFC in the individual-shared subspace which is age-invariant and AFC in the individual-specific subspace which is age-related. To investigate the predictive efficacy of age-related individual-specific subspaces for clinical symptoms, we used elastic net regression to optimize model performance[ 32 ]. This algorithm was applied separately to original AFC matrices and AFC in the individual-specific subspace across different age groups to construct predictive models for clinical symptoms. In this study, the core autism symptoms were assessed using the autism diagnostic observation schedule (ADOS)[ 26 ]. This instrument's three factors—communication, social, and restricted and repetitive behaviors (RRB)—served as clinical symptom labels. AFC matrices and AFC in the individual-specific subspaces from the four age groups were used as input features for training separate elastic net regression models. Model hyperparameters (regularization coefficients λ₁ and λ₂) were determined via grid search. Five-fold cross-validation was implemented to enhance model generalizability: Data were partitioned into five equal subsets; four subsets were used for training and the remaining subset for prediction, with this process repeated five times. Model accuracy was defined by Pearson correlation coefficient between the predicted score and clinical symptom score within the test fold[ 33 ]. Finally, the average accuracy was calculated over five runs. Correlation analysis between the individual-specific subspace and clinical symptoms To investigate the relationship between AFC in the individual-specific subspace of different age groups and clinical symptoms, we computed the norms of brain networks of AFC in the individual-specific subspaces. Each element in the AFC matrix represents the changed FC value between two ROIs for an individual with ASD. For all elements corresponding to brain network i and brain network j, we concatenated them row-wise into a one-dimensional vector, denoted as X = (x₁, x₂, ..., xₙ). We then calculated their L2-norm: $$\:\left|\right|X|{|}_{2}=\sqrt{{x}_{1}^{2}+{x}_{2}^{2}+...+{x}_{n}^{2}}$$ 2 The L2-norm, representing the Euclidean distance from the coordinate system origin to vector X, is calculated as the square root of the sum of its squared components. This metric quantifies a vector’s magnitude (length) and reflects its signal "energy" or "intensity". Within the distinctive subspaces of each age group, we computed the L2-norm for every brain network across all ASD participant. Subsequently, Pearson correlation analyses were performed between these L2-norm values (derived from individual-specific subspaces) and clinical behavioral measures. In this study, core autism symptoms were assessed using the Autism Diagnostic Observation Schedule (ADOS)[ 26 ], which comprising three subscales: Communication, Social Interaction, and Restricted Repetitive Behaviors (RRB). Validation analysis Validation analysis was performed to estimate the robustness of our findings in an independent cohort. We decomposed AFC into the individual-shared and individual-specific subspaces. We evaluated the validity of our findings in the validation dataset. Applying the COBE algorithm similarly extracted age-independent individual-shared subspaces and age-related individual-specific subspaces. Similarly, we conducted predictive models to exam the predictive power of age-related individual-specific subspace for clinical symptoms through elastic net regression. Results Demographic information After excluding participants with excessive head motion, 437 participants (age range: 6-39 years; ASD=208, TD=229) were included in this study. For details on participant demographics from ABIDE, see Table 1. For details on participant demographics from validation dataset, see TABLE 2. Extract three individual-shared components We found that the estimations were robust when common component (C) is three, with the highest similarity 0.8 (Fig. S2). Thus, we focused on three common components in subsequent analyses. Across all four age groups, we consistently identified three ASD-shared components from the AFC matrix (see Fig. 1). In Component 1, functional connectivity was predominantly positive across both intra- and inter-network connections, except within SMN (Fig. 1(a), (b)). The top 30 strongest positive connections were primarily distributed between DMN and SMN, and between DMN and FPN (Fig. 1(c)). In Component 2, CB exhibited predominantly negative functional connectivity with other brain networks, while connections among other networks were mainly positive (Fig. 1(a), (b)). The top 30 positive connections were concentrated within the SMN, between SMN and CON, and between SMN and ON; conversely, the top 30 negative connections were predominantly distributed between CB and CON, and between CB and SMN (Fig. 1(c)). In Component 3, functional connectivity between the CB and other networks was primarily positive. The top 30 positive connections were mainly located between CB and CON, and between CB and SMN, whereas the top 30 negative connections were predominantly distributed between DMN and ON. AFC in the individual-specific subspace enhance the predictive ability of clinical symptoms Compared to raw AFC, the individual-specific subspace derived from AFC demonstrated significantly higher accuracy in predicting clinical symptoms, with a trend toward greater improvement in older age groups (Fig. 2). Specifically, the individual-specific subspace outperformed raw AFC in predicting specific domains across different age groups: in the subgroup of childhood (6-11 years), it showed higher predictive accuracy for communication (0.27 vs. 0.23; a 17% increase), social (0.37 vs. 0.26; a 42% increase), and RRB (0.33 vs. 0.32; a 3% increase), but not for the ADOS total score (0.35 vs. 0.29; a 17% decrease), with an average increase of 8.6% in predictive accuracy . In the subgroup of adolescence (12-17 years), it demonstrated superior accuracy for communication (0.31 vs. 0.18; a 72% increase), social (0.35 vs. 0.15; a 133% increase), and the ADOS total score (0.4 vs. 0.32; a 25% increase), but not for RRB (0.29 VS. 0.31; a 6% decrease), with an average increase of 41%. In the subgroup of emerging adulthood (18-25 years), it achieved higher accuracy for communication (0.35 VS. 0.29; a 21% increase), social (0.43 VS. 0.41; a 5% increase), RRB (0.32 VS. 0.22; a 45% increase), and the ADOS total score (0.51 VS. 0.37; a 38% increase), with an average increase of 25%. In early adulthood (26-39 years), it better predicted communication (0.38 VS. 0.35; a 8.6% increase), social (0.43 VS. 0.39; a 10% increase), and RRB (0.47 VS. 0.46; a 2% increase), but not for the ADOS total score (0.47 VS. 0.54; a 13% decrease), with an average increase of 0.6%. Notably, the individual-specific subspace in adolescents with ASD predicts social behavior, showing the highest level of improvement at 133%, as well as an average increase of 41%. The correlation between clinical symptoms and the L2-norm of brain networks in individual-specific subspace We computed the L2-norm of brain networks within the individual-specific subspace, where the L2-norm represents the "strength" of connectivity of brain network. Correlations between the L2-norm of brain networks in individual-specific subspace and behaviors were calculated separately for each age group. As shown in Fig. 3: In the “6-11 years” group, significant positive correlations were observed between CON-CB connectivity strength and both communication (r = 0.27, p = 0.020) and social skills (r = 0.28, p = 0.018). In the “12-17 years” group, negative correlations with communication were found across most inter- and intra-network connections, with the exception of intra-FPN, intra-ON, intra-CB, and inter-ON-CB connections. Significant negative correlations with social skills were identified for DMN-FPN (r = -0.25, p = 0.037), DMN-ON (r = -0.37, p = 0.002), DMN-CB (r = -0.30, p = 0.01), FPN-CB (r = -0.28, p = 0.019), CON-SMN (r = -0.30, p = 0.012), CON-ON (r = -0.35, p = 0.003), CON-CB (r = -0.33, p = 0.006), and SMN-CB (r = -0.26, p = 0.033). Significant negative correlations with RRB were observed for FPN-CB (r = -0.24, p = 0.049) and ON-CB (r = -0.26, p = 0.032). In the “18-25 years” group, significant negative correlations were found between FPN-CB connectivity strength and both communication (r = -0.52, p = 0.004) and social skills (r = -0.39, p = 0.041). In the “26-39 years” group, significant positive correlations with communication were identified for intra-DMN (r = 0.57, p = 0.011), intra-SMN (r = 0.49, p = 0.032), inter-DMN-FPN (r = 0.47, p = 0.040), and inter-DMN-SMN (r = 0.48, p = 0.038). Validation analysis In an independent dataset, we extracted the individual-shared subspace from AFC matrix among different age groups, and obtained three common components in the individual-shared subspace in the same way. These three common components have similar connection patterns among the three common components with that obtained from ABIDE (Fig. 4). Furthermore, the individual-specific subspace significantly enhanced predictive accuracy for clinical symptoms compared to raw AFC (Fig. 5). Specifically, the individual-specific subspace outperformed raw AFC in predicting specific domains across different age groups: in early childhood (2-5 years), it achieved higher accuracy for communication (0.39 vs. 0.26; a 50% increase), social (0.41 vs. 0.32; a 28% increase), RRB (0.31 vs. 0.2; a 55% increase), and the ADOS total score (0.22 vs. 0.21; a 4.8% increase), with an average increase of 34%. In childhood (6-11 years), it showed higher predictive accuracy for communication (0.46 vs. 0.42; a 9.5% increase), and RRB (0.32 vs. 0.14; a 129% increase), but not for social (0.3 vs. 0.32; a 6.3% decrease) and the ADOS total score (0.45 vs. 0.47; a 4.3% decrease), with an average increase of 13%. In adolescence (12-17 years), it demonstrated superior accuracy for communication (0.72 vs. 0.46; a 57% increase), social (0.52 vs. 0.46; a 13% increase), RRB (0.48 VS. 0.38; a 26% increase), and the ADOS total score (0.61 vs. 0.4; a 53% increase), with an average increase of 37%. Notably, the individual-specific subspace in adolescents with ASD predicts clinical behaviors showing the highest level of average improvement at 37%. The correction analysis results of validation dataset were shown in Fig. S3. Discussion Using COBE, AFC matrix of individual with ASD is projected onto two subspaces: one is an individual-shared subspace, representing the shared pattern of connectivity alterations in ASD across ages; the other is an age-related, individual-specific subspace, representing the unique connectivity features of each ASD individual after removing the shared components. We found that the age-related individual-specific AFC demonstrated greater predictive power for clinical symptoms compared to raw AFC. Furthermore, the age-related individual-specific subspace is associated with communication deficits, social interaction, and RRB. Overall, our findings demonstrate that it is necessary to take age into account to understand the heterogeneous FC of ASD. Capturing and utilizing individual-specific brain connectivity features at a given developmental stage is central to dissecting the clinical heterogeneity of ASD. Individual-shared common components across ages The three common components of the cross-age, ASD-shared subspace exhibit distinct patterns of alterations. Notably, the connectivity patterns between the DMN-CON, FPN-CON, and SMN-FPN networks were similar across the three shared components, characterized by predominantly positive connectivity. The brain networks of DMN, CON, FPN, and SMN have been extensively studied in ASD, and abnormalities within these networks have been consistently reported in numerous resting-state fMRI studies. DMN is a core network for human introspection, social interaction, and consciousness. Dysfunction within the DMN is closely implicated in various neurological and psychiatric disorders, such as ASD, depression, and schizophrenia[ 34 – 37 ]. Multiple brain regions comprising the DMN are strongly linked to key ASD theories, particularly the Theory of Mind (ToM) [ 38 , 39 ] and the theory of neural connectivity deficits[ 40 – 42 ]. The DMN is active during resting states and demonstrates increased engagement during social cognition and self-referential processing[ 43 – 45 ]. In contrast, the FPN is recruited during active information processing and is critically implicated in executive functions, cognitive flexibility, working memory, goal-directed behavior, and attentional control, serves as the core circuitry for executing higher-order cognitive tasks. Higher-order cognitive impairments, particularly deficits in executive functions[ 46 – 48 ], are characteristic of ASD. Dysfunctional connectivity of FPN is closely linked to core symptoms of ASD[ 49 ], suggesting the FPN may underlie fundamental aspects of ASD symptomatology. The CON, comprising anterior cingulate cortex (ACC), insula, and thalamus, is involved in conflict monitoring, cognitive control, and efficient attentional resource allocation. It plays a critical role in detecting the behavioral relevance of internal or external stimuli, orchestrating the dynamic switching between the DMN and FPN, and optimally allocating attentional resources[ 50 , 51 ]. Aberrant FC within this network is associated with impaired emotional and social information processing in ASD, highlighting its essential role in social cognition and complex cognitive processes[ 52 ]. The SMN supports motor control, sensory information processing, and motor learning. Abnormal SMN connectivity in ASD manifests as sensorimotor impairments[ 53 , 54 ], including repetitive/stereotyped movements[ 1 ], aberrant oculomotor control[ 55 ], postural sway[ 56 ], and impaired motor coordination[ 57 ]. A study[ 58 ] found that children and adolescents with ASD had abnormally increased FC between the PCC, a hub region of the DMN, and bilateral occipital regions (e.g., middle occipital gyrus, lingual gyrus), reflecting excessive integration of visual information. Additionally, the hyperconnectivity between the PCC and language-related regions (e.g., bilateral inferior frontal gyrus, inferior parietal lobule, anterior/posterior cingulate cortex) reflected aberrant "cross-network communication" between the DMN and other brain networks, potentially leading to an imbalance between self-referential processing and external task switching during language processing. A recent resting-state fMRI meta-analysis[ 59 ] revealed that children with ASD exhibited decreased functional activity in the left insula (a hub of CON), bilateral ACC/ mPFC, left angular gyrus, and right inferior temporal gyrus, alongside increased functional activity in the right supplementary motor area (a hub of SMN) and precuneus. This pattern suggests functional abnormalities within the DMN, CON, FPN and SMN networks in ASD. Research has further demonstrated that reduced DMN FC is significantly associated with diminished social motivation and impaired mentalizing abilities in individuals with ASD[ 60 ]. For instance, decreased FC of posterior DMN nodes (e.g., the temporoparietal junction) predicts deficits in understanding social intentions in ASD patients[ 61 ]. Our findings provide evidence for the social and sensorimotor impairments in ASD, underscoring the converging role of the DMN, CON, FPN and SMN in the neuropathological mechanisms of the disorder. AFC in the ASD-specific subspace enhance the predictive ability of clinical symptoms ASD exhibits substantial heterogeneity[ 2 , 62 ], and characterizing brain-behavior relationships is essential for understanding and treating psychiatric conditions[ 63 ]. Brain-behavior mapping in ASD constitutes a critical step for localizing specific behavioral circuits[ 61 ], guiding clinical research and practice through the identification of targets for individualized diagnosis and intervention. Currently, it lacks biologically validated diagnostic or therapeutic approaches for ASD. A crucial aspect of biomarker development involves demonstrating that candidate biomarkers predict relevant behavioral outcomes and disease trajectories[ 64 ]. Predictive modeling provides a statistically rigorous framework for characterizing individual differences, particularly in neurodevelopmental conditions[ 65 , 66 ]. Successful predictive models require that features derived from neuroimaging data exhibit inter-individual variability, preserving key individual characteristics[ 67 ]. For instance, prediction models inevitably yield poor performance if input features are identical across participants, precluding effective behavioral forecasting. Previous research has established that individualized FC profiles demonstrate strong predictive power within models, compelling a paradigm shift from group-level analyses toward leveraging individual differences to decipher high clinical heterogeneity, taking developmental factors into account as well[ 17 , 68 ]. Consequently, we incorporated individual-specific features when mapping neuroimaging data to ASD-related behavioral symptoms. Furthermore, accounting for developmental stage is paramount in predictive modeling for neurodevelopmental disorders. Our study stratified participants into four age groups to examine the predictive capacity of individual-specific subspaces for behavioral symptoms across distinct developmental periods. This approach aligns with Kazeminejad et al.[ 69 ], who constructed separate predictive models for different age cohorts and demonstrated distinct functional brain network architectures in ASD classification models for 5–15 year-olds versus 15–30 year-olds, supporting age-stratified modeling to enhance predictive accuracy. In our study, the strongest predictive power of the individual-specific subspace was observed in adolescence. This variation in predictive improvement across age groups suggests that the relationship between brain connectivity and behavioral symptoms in ASD is not static but undergoes dynamic evolution throughout development. The neural substrates supporting behavioral performance thus differ across childhood, adolescence, and adulthood, consistent with extensive neurodevelopmental literature[ 9 , 70 ]. Our finding that the largest gains in predictive accuracy occurred during adolescence is consistent with neurodevelopmental models of ASD, which posit that this period is marked by a pronounced and critical reorganization of brain networks[ 70 , 71 ]. The intimate link between these network changes and behavioral symptoms, as captured by our model, is further supported by studies showing age-specific alterations in functional connectivity that correlate with symptom severity[ 72 , 73 ]. The correlation between the individual-specific subspace brain network and clinical symptoms Brain network abnormalities in ASD exhibit marked individual specificity. Large-scale cohort studies reveal that approximately 30% of individuals with ASD demonstrate significant brain network age dysynchrony, characterized by developmentally lagged connectivity patterns in DMN and SN relative to chronological age[ 60 ]. Furthermore, the developmental trajectory of the DMN correlates with the severity of RRB[ 62 ]. Neuroplasticity-driven dynamic changes may manifest as childhood hyperconnectivity transitioning to hypoconnectivity during adolescence[ 62 ]. Recent research demonstrated that while neurotypical controls show significant prefrontal and insular activation in individuals under 25 years, the ASD group exhibits absent activation in these regions during this developmental period. Beyond age 25, only left insular activation emerges in ASD. This pattern suggests delayed maturation of executive function-associated regions (FPN) in ASD, supporting the delayed neurodevelopment hypothesis[ 74 ]. This study demonstrates that within the individual-specific subspace of each age group, the connectivity between CB and other networks consistently correlate with communication abilities. CB was initially considered primarily responsible for motor function. However, emerging evidence reveals its critical role in multimodal sensory integration[ 75 – 77 ]. Beyond motor dysregulation, CB dysfunction contributes to deficits in executive function, visuospatial processing, and emotional regulation[ 77 ]. Impaired CB integration may underlie multimodal deficits affecting motor coordination, language, and even social behavior[ 78 ]. Critically, cerebello-cortical connections (e.g., with DMN, CON, and FPN) effectively predict social and cognitive functioning[ 61 ]. In this study, the norm of brain networks of AFC within the individual-specific subspace exhibited primarily positive correlations with clinical symptoms for ASD individuals aged 6–11 and 26–39 years, whereas an inverse relationship was observed for those aged 12–17 and 18–25 years (Fig. 3). However, a discrepancy was observed in the validation analysis (Fig. S3). This could be attributed to the sampling strategy, although the results were also strongly age-dependent. In summary, this study reveals distinct neurodevelopmental patterns of brain networks across different stages in ASD, laying the groundwork for identifying developmentally-informed neuroimaging biomarkers. Limitations Our results have several limitations. First, although ABIDE provides a multi-site neuroimaging data across a wide age range, the sample size for several age groups—particularly those aged 18–25 and 26–39—is relatively small. Future studies would benefit from larger datasets with more adult participants. Furthermore, longitudinal data are needed to clarify whether AFC in the individual-specific subspace reflects ASD heterogeneity, neurodevelopmental stage effects, or a complex interplay between the two. Conclusions The superior predictive power of the individual-specific subspace for clinical symptoms, which was most pronounced in the adolescent cohort, underscores a pivotal finding: the brain-behavior relationship in ASD is not static but is fundamentally shaped by both individual neurobiological variability and developmental stage. This strongly suggests that the quest for reliable biomarkers in ASD must move beyond group-level averages to embrace a more nuanced, personalized approach that accounts for the unique developmental trajectory of each individual. Consequently, our results highlight the critical importance of integrating developmental neuroscience principles into clinical translation. Targeting future interventions, particularly during pivotal windows of neurodevelopment such as adolescence, may yield the greatest therapeutic benefits by aligning with the brain's inherent plastic potential during this dynamic reorganizational period. Abbreviations ABIDE Autism Brain Imaging Data Exchange ACC Anterior cingulate corte ADOS Autism Diagnostic Observation Schedule AFC Altered functional connectivity ASD Autism spectrum disorder AN Auditory network CB Cerebellum COBE Common orthogonal basis extraction CON Cingulo-opercular network C-WISC Chinese-Wechsler Intelligence Scale for Children DQ Developmental quotient DMN Default mode network DSM-5 Diagnostic and Statistical Manual of Mental Disorders - 5 FC Functional connectivity FD Framewise displacement FPN Frontoparietal network fMRI Functional magnetic resonance imaging GDS Gesell Development Schedules ICD-10 International Statistical Classification of Diseases and Related Health Problems 10th Revision IQ Intelligence quotient mPFC Medial prefrontal cortex ON Occipital network PCC Posterior cingulate cortex ROIs Regions of interest RRB Restricted and repetitive behaviors rs-fMRI Resting-state functional magnetic resonance imaging SMN Sensorimotor network SN Salient network TD Typical development ToM Theory of Mind Declarations Authors' contributions TL download data from ABIDE, preprocessed the MRI data, and drafted the manuscript. JZ contributed to data analysis, scientific editing and language editing of the manuscript. MZ assisted in study conception and design. LL, SZ, ZY and YJ contributed to the data collection. BL conceived and designed the analyses. YH contributed to funding, conception, and design. All authors listed have made a substantial, direct, or intellectual contribution to the work, and approved it for publication. Funding This work was supported by the Medical and Industrial Integration Project of Chengdu City (Grant No. ZYGX2022YGRH020); the Sichuan Provincial Department of Science and Technology's Special Project for Central Support of Local Science and Technology Development (Grant No. 2023ZYD0123); the Program of Chengdu Science and Technology (Grant No. 2022-YF09-00010-SN). Data Availability Statement The data that support the findings of this study are available in the ABIDE dataset (https://fcon_1000.projects.nitrc.org/indi/abide/). Ethics approval and consent to participate We obtained informed consent from the parents or caregivers of the children who participated. This study was carried out in accordance with the Declaration of Helsinki and approved by the Medical Ethics Committee of the West China Hospital, Sichuan University. Clinical trial number Not applicable. Consent for publication Not applicable. Competing interests The authors declare no competing interests. References American Psychiatric, A.; American Psychiatric Association, D. S. M. T. F., Diagnostic and statistical manual of mental disorders : DSM-5 . Fifth ed.; American Psychiatric Publishing: 2013. Lord, C.; Elsabbagh, M.; Baird, G., et al., Autism spectrum disorder [J]. Lancet (London, England), 2018, 392 (10146): 508-520. Hirota, T.; King, B. H., Autism Spectrum Disorder: A Review [J]. JAMA, 2023, 329 (2): 157-168. Emerson, R. W.; Adams, C.; Nishino, T., et al., Functional neuroimaging of high-risk 6-month-old infants predicts a diagnosis of autism at 24 months of age [J]. Science translational medicine, 2017, 9 (393). Williams, J. A.; Burgess, S.; Suckling, J., et al., Inflammation and Brain Structure in Schizophrenia and Other Neuropsychiatric Disorders: A Mendelian Randomization Study [J]. JAMA psychiatry, 2022, 79 (5): 498-507. Rasero, J.; Jimenez-Marin, A.; Diez, I., et al., The Neurogenetics of Functional Connectivity Alterations in Autism: Insights From Subtyping in 657 Individuals [J]. Biol Psychiatry, 2023, 94 (10): 804-813. O'Hearn, K.; Lynn, A., Age differences and brain maturation provide insight into heterogeneous results in autism spectrum disorder [J]. Frontiers in human neuroscience, 2022, 16: 957375. Sato, W.; Uono, S., The atypical social brain network in autism: advances in structural and functional MRI studies [J]. Curr Opin Neurol, 2019, 32 (4): 617-621. Haghighat, H.; Mirzarezaee, M.; Araabi, B. N., et al., Functional Networks Abnormalities in Autism Spectrum Disorder: Age-Related Hypo and Hyper Connectivity [J]. Brain topography, 2021, 34 (3): 306-322. 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]. Molecular psychiatry, 2014, 19 (6): 659-67. Nair, A.; Jolliffe, M.; Lograsso, Y. S. S., et al., A Review of Default Mode Network Connectivity and Its Association With Social Cognition in Adolescents With Autism Spectrum Disorder and Early-Onset Psychosis [J]. Frontiers in psychiatry, 2020, 11: 614. 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]. Frontiers in physiology, 2021, 12: 626338. Yerys, B. E.; Gordon, E. M.; Abrams, D. N., et al., Default mode network segregation and social deficits in autism spectrum disorder: Evidence from non-medicated children [J]. Neuroimage Clin, 2015, 9: 223-32. Wolfers, T.; Beckmann, C. F.; Hoogman, M., et al., Individual differences v. the average patient: mapping the heterogeneity in ADHD using normative models [J]. Psychol Med, 2020, 50 (2): 314-323. Tang, S.; Sun, N.; Floris, D. L., et al., Reconciling Dimensional and Categorical Models of Autism Heterogeneity: A Brain Connectomics and Behavioral Study [J]. Biological psychiatry, 2020, 87 (12): 1071-1082. Guo, X.; Zhang, X.; Chen, H., et al., Exploring the heterogeneity of brain structure in autism spectrum disorder based on individual structural covariance network [J]. Cereb Cortex, 2023, 33 (12): 7311-7321. Kashyap, R.; Kong, R.; Bhattacharjee, S., et al., Individual-specific fMRI-Subspaces improve functional connectivity prediction of behavior [J]. NeuroImage, 2019, 189: 804-812. Shan, X.; Uddin, L. Q.; Ma, R., et al., Disentangling the Individual-Shared and Individual-Specific Subspace of Altered Brain Functional Connectivity in Autism Spectrum Disorder [J]. Biological psychiatry, 2024, 95 (9): 870-880. Zhou, G.; Zhao, Q.; Zhang, Y., et al., Linked Component Analysis From Matrices to High-Order Tensors: Applications to Biomedical Data [J]. Proceedings of the IEEE, 2016, 104 (2): 310-331. Zhou, G.; Cichocki, A.; Zhang, Y., et al., Group Component Analysis for Multiblock Data: Common and Individual Feature Extraction [J]. IEEE transactions on neural networks and learning systems, 2016, 27 (11): 2426-2439. Kashyap, R.; Bhattacharjee, S.; Yeo, B. T. T., et al., Maximizing dissimilarity in resting state detects heterogeneous subtypes in healthy population associated with high substance use and problems in antisocial personality [J]. Human brain mapping, 2020, 41 (5): 1261-1273. Zhang, A.; Liu, L.; Chang, S., et al., Connectivity-Based Brain Network Supports Restricted and Repetitive Behaviors in Autism Spectrum Disorder Across Development [J]. Frontiers in psychiatry, 2022, 13: 874090. 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]. Scientific data, 2017, 4: 170010. Arnett, J. J., Emerging adulthood. A theory of development from the late teens through the twenties [J]. The American psychologist, 2000, 55 (5): 469-80. Organization, W. H., ICD-10 : international statistical classification of diseases and related health problems : tenth revision [J]. Acta Chirurgica Iugoslavica, 2010, 56 (3): 65-9. Lord, C.; Risi, S.; Lambrecht, L., et al., The autism diagnostic observation schedule-generic: a standard measure of social and communication deficits associated with the spectrum of autism [J]. Journal of autism and developmental disorders, 2000, 30 (3): 205-23. Rachel; S.; Ball, The Gesell developmental schedules: Arnold Gesell (1880–1961) [J]. Journal of Abnormal Child Psychology, 1977, 5 (3): 233-239. Woolger, C., Wechsler Intelligence Scale for Children-Third Edition (wisc-iii) [J]. Springer US. Jia, X. Z.; Wang, J.; Sun, H. Y., et al., RESTplus: an improved toolkit for resting-state functional magnetic resonance imaging data processing [J]. Science bulletin, 2019, 64 (14): 953-954. Ashburner, J., A fast diffeomorphic image registration algorithm [J]. NeuroImage, 2007, 38 (1): 95-113. Dosenbach, N. U.; Nardos, B.; Cohen, A. L., et al., Prediction of individual brain maturity using fMRI [J]. Science (New York, N.Y.), 2010, 329 (5997): 1358-61. Zou, H.; Hastie, T., Regularization and variable selection via the elastic net [J]. Journal of the Royal Statistical Society, 2005, 67 (5): 768-768. Finn, E. S.; Shen, X.; Scheinost, D., et al., Functional connectome fingerprinting: identifying individuals using patterns of brain connectivity [J]. Nature neuroscience, 2015. Yang, B.; Wang, M.; Zhou, W., et al., Disrupted network integration and segregation involving the default mode network in autism spectrum disorder [J]. Journal of affective disorders, 2023, 323: 309-319. Zhou, H. X.; Chen, X.; Shen, Y. Q., et al., Rumination and the default mode network: Meta-analysis of brain imaging studies and implications for depression [J]. NeuroImage, 2020, 206: 116287. King, S.; Mothersill, D.; Holleran, L., et al., Early life stress, low-grade systemic inflammation and weaker suppression of the default mode network (DMN) during face processing in Schizophrenia [J]. Translational psychiatry, 2023, 13 (1): 213. Gattuso, J. J.; Perkins, D.; Ruffell, S., et al., Default Mode Network Modulation by Psychedelics: A Systematic Review [J]. The international journal of neuropsychopharmacology, 2023, 26 (3): 155-188. Wing, L.; Gould, J., Severe impairments of social interaction and associated abnormalities in children: epidemiology and classification [J]. J Autism Dev Disord, 1979, 9 (1): 11-29. Simon; Baron-Cohen; Alan, et al., Mechanical, behavioural and Intentional understanding of picture stories in autistic children [J]. British Journal of Developmental Psychology, 1986. Frith, C., Is autism a disconnection disorder? [J]. The Lancet. Neurology, 2004, 3 (10): 577. Geschwind, D. H.; Levitt, P., Autism spectrum disorders: developmental disconnection syndromes [J]. Current opinion in neurobiology, 2007, 17 (1): 103-11. Kana, R. K.; Uddin, L. Q.; Kenet, T., et al., Brain connectivity in autism [J]. Frontiers in human neuroscience, 2014, 8: 349. Iacoboni, M.; Lieberman, M. D.; Knowlton, B. J., et al., Watching social interactions produces dorsomedial prefrontal and medial parietal BOLD fMRI signal increases compared to a resting baseline [J]. NeuroImage, 2004, 21 (3): 1167-73. Spreng, R. N.; Grady, C. L., Patterns of brain activity supporting autobiographical memory, prospection, and theory of mind, and their relationship to the default mode network [J]. Journal of cognitive neuroscience, 2010, 22 (6): 1112-23. Spreng, R. N., The fallacy of a "task-negative" network [J]. Frontiers in psychology, 2012, 3: 145. Demetriou, E. A.; Lampit, A.; Quintana, D. S., et al., Autism spectrum disorders: a meta-analysis of executive function [J]. Mol Psychiatry, 2018, 23 (5): 1198-1204. Xie, R.; Sun, X.; Yang, L., et al., Characteristic Executive Dysfunction for High-Functioning Autism Sustained to Adulthood [J]. Autism research : official journal of the International Society for Autism Research, 2020, 13 (12): 2102-2121. Sadozai, A. K.; Sun, C.; Demetriou, E. A., et al., Executive function in children with neurodevelopmental conditions: a systematic review and meta-analysis [J]. Nature human behaviour, 2024, 8 (12): 2357-2366. Lin, H. Y.; Perry, A.; Cocchi, L., et al., Development of frontoparietal connectivity predicts longitudinal symptom changes in young people with autism spectrum disorder [J]. Transl Psychiatry, 2019, 9 (1): 86. Seeley, W. W.; Menon, V.; Schatzberg, A. F., et al., Dissociable intrinsic connectivity networks for salience processing and executive control [J]. The Journal of neuroscience : the official journal of the Society for Neuroscience, 2007, 27 (9): 2349-56. Eckert, M. A.; Menon, V.; Walczak, A., et al., At the heart of the ventral attention system: the right anterior insula [J]. Human brain mapping, 2009, 30 (8): 2530-41. Attanasio, M.; Mazza, M.; Le Donne, I., et al., Salience Network in Autism: preliminary results on functional connectivity analysis in resting state [J]. European archives of psychiatry and clinical neuroscience, 2024. Lim, Y. H.; Partridge, K.; Girdler, S., et al., Standing Postural Control in Individuals with Autism Spectrum Disorder: Systematic Review and Meta-analysis [J]. J Autism Dev Disord, 2017, 47 (7): 2238-2253. Cook, J., From movement kinematics to social cognition: the case of autism [J]. Philosophical transactions of the Royal Society of London. Series B, Biological sciences, 2016, 371 (1693). Schmitt, L. M.; Cook, E. H.; Sweeney, J. A., et al., Saccadic eye movement abnormalities in autism spectrum disorder indicate dysfunctions in cerebellum and brainstem [J]. Mol Autism, 2014, 5 (1): 47. Bojanek, E. K.; Wang, Z.; White, S. P., et al., Postural control processes during standing and step initiation in autism spectrum disorder [J]. J Neurodev Disord, 2020, 12 (1): 1. Dziuk, M. A.; Gidley Larson, J. C.; Apostu, A., et al., Dyspraxia in autism: association with motor, social, and communicative deficits [J]. Developmental medicine and child neurology, 2007, 49 (10): 734-9. Gao, Y.; Linke, A.; Jao Keehn, R. J., et al., The language network in autism: Atypical functional connectivity with default mode and visual regions [J]. Autism research : official journal of the International Society for Autism Research, 2019, 12 (9): 1344-1355. Guo, Z.; Tang, X.; Xiao, S., et al., Systematic review and meta-analysis: multimodal functional and anatomical neural alterations in autism spectrum disorder [J]. Mol Autism, 2024, 15 (1): 16. 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. Horien, C.; Floris, D. L.; Greene, A. S., et al., Functional Connectome-Based Predictive Modeling in Autism [J]. Biol Psychiatry, 2022, 92 (8): 626-642. Guo, X.; Zhai, G.; Liu, J., et al., Inter-individual heterogeneity of functional brain networks in children with autism spectrum disorder [J]. Mol Autism, 2022, 13 (1): 52. Vieira, S.; Bolton, T. A. W.; Schöttner, M., et al., Multivariate brain-behaviour associations in psychiatric disorders [J]. Transl Psychiatry, 2024, 14 (1): 231. Koutsouleris, N.; Meisenzahl, E. M.; Davatzikos, C., et al., Use of neuroanatomical pattern classification to identify subjects in at-risk mental states of psychosis and predict disease transition [J]. Archives of general psychiatry, 2009, 66 (7): 700-12. Scheinost, D.; Noble, S.; Horien, C., et al., Ten simple rules for predictive modeling of individual differences in neuroimaging [J]. NeuroImage, 2019, 193: 35-45. Rosenberg, M. D.; Casey, B. J.; Holmes, A. J., Prediction complements explanation in understanding the developing brain [J]. Nature communications, 2018, 9 (1): 589. Finn, E. S.; Todd Constable, R., Individual variation in functional brain connectivity: implications for personalized approaches to psychiatric disease [J]. Dialogues in clinical neuroscience, 2016, 18 (3): 277-287. Cui, W.; Ma, Y.; Ren, J., et al., Personalized Functional Connectivity Based Spatio-Temporal Aggregated Attention Network for MCI Identification [J]. IEEE transactions on neural systems and rehabilitation engineering : a publication of the IEEE Engineering in Medicine and Biology Society, 2023, 31: 2257-2267. Kazeminejad, A.; Sotero, R. C., Topological Properties of Resting-State fMRI Functional Networks Improve Machine Learning-Based Autism Classification [J]. Frontiers in neuroscience, 2018, 12: 1018. Uddin, L. Q.; Supekar, K.; Menon, V., Reconceptualizing functional brain connectivity in autism from a developmental perspective [J]. Frontiers in human neuroscience, 2013, 7: 458. Paus, T.; Keshavan, M.; Giedd, J. N., Why do many psychiatric disorders emerge during adolescence? [J]. Nature reviews. Neuroscience, 2008, 9 (12): 947-57. Nomi, J. S.; Uddin, L. Q., Developmental changes in large-scale network connectivity in autism [J]. Neuroimage Clin, 2015, 7: 732-41. Padmanabhan, A.; Lynch, C. J.; Schaer, M., et al., The Default Mode Network in Autism [J]. Biol Psychiatry Cogn Neurosci Neuroimaging, 2017, 2 (6): 476-486. May, K. E.; Kana, R. K., Frontoparietal Network in Executive Functioning in Autism Spectrum Disorder [J]. Autism research : official journal of the International Society for Autism Research, 2020, 13 (10): 1762-1777. Ishikawa, T.; Shimuta, M.; Häusser, M., Multimodal sensory integration in single cerebellar granule cells in vivo [J]. eLife, 2015, 4. Ronconi, L.; Casartelli, L.; Carna, S., et al., When one is Enough: Impaired Multisensory Integration in Cerebellar Agenesis [J]. Cerebral cortex (New York, N.Y. : 1991), 2017, 27 (3): 2041-2051. Xiao, L.; Scheiffele, P., Local and long-range circuit elements for cerebellar function [J]. Current opinion in neurobiology, 2018, 48: 146-152. Modi, M. E.; Sahin, M., Translational use of event-related potentials to assess circuit integrity in ASD [J]. Nature reviews. Neurology, 2017, 13 (3): 160-170. Tables Table 1 Participant Demographics of ABIDE ASD (M ± SD) TD (M ± SD) t/χ² p Cohen’ d All age （ 6-39 years ） N=208 N=229 - - - Age 15.29±6.99 15.87±7.14 -0.868 0.386 -0.083 Sex（M/F） 192/16 191/38 7.975 0.005** - IQ 102 ± 17 113±14 -7.249 < 0.001** -0.698 6-11 years N= 85 N= 83 - - - Age 9.60±1.61 9.59±1.54 0.039 0.969 0.006 Sex（M/F） 78/7 69/14 2.861 0.091 - IQ 104 ±18 115±14 -4.056 <0.001** -.630 12-17 years N=72 N=76 - - - Age 14.77±1.69 14.35±1.59 1.562 0.120 0.257 Sex（M/F） 69/3 61/15 8.390 0.004** - IQ 99±16 110±13 -4.721 <0.001** -1.703 18-25 years N=31 N=42 - - - Age 21.75±2.15 21.35±2.24 0.757 0.451 0.179 Sex（M/F） 27/4 37/5 0.016 0.898 - IQ 106±16 114±11 -2.3 0.024* -0.550 26-39 years N=20 N=28 - - - Age 31.35 ± 4.25 30.47 ± 3.67 0.774 0.443 0.227 Sex（M/F） 18/2 24/4 0.196 0.658 - IQ 103 ± 15 117 ± 15 -3.218 0.002** -0.942 * p <0.05, ** p <0.01. Abbreviations: IQ, intelligence quotient; M, male; F, female; M: mean; SD: standard deviation Table 2 Participant demographics of validation database ASD (M ± SD) TD (M ± SD) t/χ² p Cohen’ d All age (2-17years) N=146 N=99 - - - Age 7.16 ± 3.69 7.63 ± 3.31 -1.035 0.302 -0.135 Sex（M/F） 127/19 70/29 9.925 0.002* - 2 - 5 years N= 66 N= 30 - - - Age 4.06 ± 1.04 4.26 ± 0.91 -0.925 0.358 -0.204 Sex（M/F） 54/12 17/13 6.774 0.009* - DQ 67.40 ± 14.05 102.87±11.09 -12.132 <0.001** -2.691 6 – 11 years N=61 N=55 - - - Age 8.26 ± 1.60 7.91±1.57 1.197 0.234 0.223 Sex（M/F） 55/6 43/12 3.168 0.075 - IQ 82.58±24.54 116.34±13.28 -9.009 <0.001** -1.703 12 – 17 years N=19 N=14 - - - Age 14.37±1.80 13.79±1.76 0.937 0.356 0.430 Sex（M/F） 17/2 12/4 1.281 0.258 - IQ 90.53 ± 22.48 110.71 ± 9.47 -3.150 0.004** -1.110 Abbreviations: ASD: autism spectrum disorder; TD: typical development; IQ: intelligence quotient; DQ: development quotient; M, male; F, female; M: mean; SD: standard deviation. Additional Declarations No competing interests reported. Supplementary Files SupplementaryMaterials.docx Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 30 Mar, 2026 Reviews received at journal 25 Mar, 2026 Reviews received at journal 24 Mar, 2026 Reviewers agreed at journal 15 Mar, 2026 Reviewers agreed at journal 11 Mar, 2026 Reviewers agreed at journal 18 Feb, 2026 Reviewers invited by journal 17 Feb, 2026 Editor invited by journal 12 Feb, 2026 Editor assigned by journal 10 Feb, 2026 Submission checks completed at journal 10 Feb, 2026 First submitted to journal 09 Feb, 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-8832166","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":593234325,"identity":"bed004d6-6aa0-4944-a73a-e491798ab20a","order_by":0,"name":"Tingting Luo","email":"","orcid":"","institution":"West China Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Tingting","middleName":"","lastName":"Luo","suffix":""},{"id":593234326,"identity":"e0d11aec-9f92-4dad-810e-b7df8fd9e5eb","order_by":1,"name":"Jie Zhang","email":"","orcid":"","institution":"Southwestern University of Finance and Economics","correspondingAuthor":false,"prefix":"","firstName":"Jie","middleName":"","lastName":"Zhang","suffix":""},{"id":593234327,"identity":"c156bed7-b368-4fef-94a4-c5ab8be93638","order_by":2,"name":"Manxue Zhang","email":"","orcid":"","institution":"West China Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Manxue","middleName":"","lastName":"Zhang","suffix":""},{"id":593234328,"identity":"f21249b6-6059-45ca-ba20-f9137eecff87","order_by":3,"name":"Lei Li","email":"","orcid":"","institution":"West China Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Lei","middleName":"","lastName":"Li","suffix":""},{"id":593234329,"identity":"0caef466-bea6-4b24-9958-f2bcdc37973e","order_by":4,"name":"Shengnan Zhao","email":"","orcid":"","institution":"West China Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Shengnan","middleName":"","lastName":"Zhao","suffix":""},{"id":593234330,"identity":"37653550-19d6-4801-b544-03911471c87f","order_by":5,"name":"Zhaozhi Yang","email":"","orcid":"","institution":"West China Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Zhaozhi","middleName":"","lastName":"Yang","suffix":""},{"id":593234331,"identity":"819c284c-88eb-46c3-ae10-5528d93d804d","order_by":6,"name":"Yuchu Jiang","email":"","orcid":"","institution":"West China Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Yuchu","middleName":"","lastName":"Jiang","suffix":""},{"id":593234332,"identity":"c7f2674a-e88a-4692-a011-148b1dd9bcce","order_by":7,"name":"Bin Liu","email":"","orcid":"","institution":"Southwestern University of Finance and Economics","correspondingAuthor":false,"prefix":"","firstName":"Bin","middleName":"","lastName":"Liu","suffix":""},{"id":593234333,"identity":"c9655eef-6168-490c-80c4-2899722ccde8","order_by":8,"name":"Yi Huang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAx0lEQVRIiWNgGAWjYBACAygtB6HYSNBiTLqWxAaitZhLJD97+LXtcPr89jMGDB/KDjPwz27Ar8VyRpq5sWzb4dwNZ3IMGGecO8wgcecAAYfdSDCTlgRpkeAxYOZtO8xgIJFASEv6N5CWdPkZQC1/idOSYyb5se1wAsMNoBZGorSceVMmzXAu3XDDmbSCgz3n0nkkbhDScjx9m+SPMmt5+fbDGx8AGXL8MwhoAQFmXmh0HABiHsLqgYDxxx+i1I2CUTAKRsFIBQBvkUKVmHzD7QAAAABJRU5ErkJggg==","orcid":"","institution":"West China Hospital of Sichuan University","correspondingAuthor":true,"prefix":"","firstName":"Yi","middleName":"","lastName":"Huang","suffix":""}],"badges":[],"createdAt":"2026-02-09 15:38:42","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8832166/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8832166/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":103505968,"identity":"32210f7c-0ff2-43bc-974f-870ea1d11875","added_by":"auto","created_at":"2026-02-26 13:33:40","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":710772,"visible":true,"origin":"","legend":"\u003cp\u003eCommon components spanning the individual-shared subspace. \u003cstrong\u003e(a)\u003c/strong\u003e: Patterns of positive and negative connectivity associated with 3 common components of altered functional connectivity (AFC) in the individual-shared subspace. Hot colors indicate positive connectivity in this component and cold colors indicate negative connectivity in this component; \u003cstrong\u003e(b)\u003c/strong\u003e: Patterns of positive and negative connectivity associated with common components of AFC in the individual-shared subspace averaged within and between subnetworks; \u003cstrong\u003e(c)\u003c/strong\u003e: The top 30 positive and negative edges of 3 common components of AFC in the individual-shared subspace. Abbreviations: CON, cingulo-opercular network; DMN, default mode network; FPN, frontoparietal network; ON, occipital network; SMN, sensorimotor network; CB, cerebellum.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8832166/v1/5ab864cc553dc7ea0eccf99b.png"},{"id":103301844,"identity":"24180bfb-2012-4d4a-b233-6a5793a064c7","added_by":"auto","created_at":"2026-02-24 08:19:19","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":139966,"visible":true,"origin":"","legend":"\u003cp\u003eCross-validated prediction accuracies for clinical symptoms. It shows the predictive capabilities of the original AFC and individual-specific subspaces for autistic behaviors across different age groups.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8832166/v1/fd865879441bc871e901f23e.png"},{"id":103301839,"identity":"dabf5b09-b612-42b8-a69d-1e8777d80c31","added_by":"auto","created_at":"2026-02-24 08:19:19","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":280107,"visible":true,"origin":"","legend":"\u003cp\u003eRelationships between the L2-norm of brain networks in the individual-specific subspace and the behaviors.\u003cstrong\u003e (a)\u003c/strong\u003e: In the group of aged 6-11 years, the correlations between the L2-norm of brain networks in the individual-specific subspace and communication, social and RRB; \u003cstrong\u003e(b)\u003c/strong\u003e: In the group of aged 12-17 years, the correlations between the L2-norm of brain networks in the individual-specific subspace and communication, social and RRB; \u003cstrong\u003e(c)\u003c/strong\u003e: In the group of aged 18-25 years, the correlations between the L2-norm of brain networks in the individual-specific subspace and communication, social and RRB; \u003cstrong\u003e(d)\u003c/strong\u003e: In the group of aged 26-39 years, the correlations between the L2-norm of brain networks in the individual-specific subspace and communication, social and RRB. *represents corrected \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.05. Abbreviations: ADOS, Autism Diagnostic Observation Schedule; AFC, altered functional connectivity; RRB, restricted and repetitive behavior; CON, cingulo-opercular network; DMN, default mode network; FPN, frontoparietal network; ON, occipital network; RRB, restricted, repetitive behavior; SMN, sensorimotor network; CB, cerebellum.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8832166/v1/2189c437986292e2dd7741d8.png"},{"id":103301842,"identity":"8902ca33-31c7-41e5-9fb8-3d5ca901b2de","added_by":"auto","created_at":"2026-02-24 08:19:19","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":776278,"visible":true,"origin":"","legend":"\u003cp\u003eCommon components spanning the individual-shared subspace in validation dataset. \u003cstrong\u003e(a)\u003c/strong\u003e: Patterns of positive and negative connectivity associated with 3 common components of AFC in the individual-shared subspace. Hot colors indicate positive connectivity in this component and cold colors indicate negative connectivity in this component;\u003cstrong\u003e (b)\u003c/strong\u003e: Patterns of positive and negative connectivity associated with common components of AFC in the individual-shared subspace averaged within and between subnetworks; \u003cstrong\u003e(c)\u003c/strong\u003e: The top 30 positive and negative edges of 3 common components of AFC in the individual-shared subspace. Abbreviations: CON, cingulo-opercular network; DMN, default mode network; FPN, frontoparietal network; ON, occipital network; SMN, sensorimotor network.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8832166/v1/21e6ecc558f1d3081e21bf03.png"},{"id":103301843,"identity":"e3d5b85d-9e3a-4341-b03b-097dab0fbf63","added_by":"auto","created_at":"2026-02-24 08:19:19","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":115095,"visible":true,"origin":"","legend":"\u003cp\u003eCross-validated prediction accuracies for clinical symptoms in validation dataset. It shows the predictive capabilities of the original AFC and individual-specific subspaces for autistic behaviors across different age groups. Abbreviations: ADOS, Autism Diagnostic Observation Schedule; AFC, altered functional connectivity; RRB, restricted and repetitive behavior. SFE, individual-specific subspace separated from AFC; AFC_2_5, the original AFC matrix of individuals with ASD aged 2-5 years; SPE_2_5, the individual-specific subspace separated from AFC of individuals with ASD aged 2-5 years; AFC_6_11, the original AFC matrix of individuals with ASD aged 6-11 years; SPE_6_11, the individual-specific subspace separated from AFC of individuals with ASD aged 6-11 years; AFC_12_17, the original AFC matrix of individuals with ASD aged 12-17 years; SPE_12_17, the individual-specific subspace separated from AFC of individuals with ASD aged 12-17 years.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8832166/v1/21e0dee834bbdfbb60072ed8.png"},{"id":103509487,"identity":"9c7a45b8-88e9-48f7-b4ef-3ba3a63d64c7","added_by":"auto","created_at":"2026-02-26 13:59:14","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2818307,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8832166/v1/15e433e9-959c-4fa3-94dc-6c7b2f7d09cc.pdf"},{"id":103506331,"identity":"a3483d39-ce0a-44b5-b77d-fca6b1953c15","added_by":"auto","created_at":"2026-02-26 13:35:21","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":746842,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterials.docx","url":"https://assets-eu.researchsquare.com/files/rs-8832166/v1/a7ea26ba0867e18a1d08b1f7.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Age-related individual-specific subspace of autism spectrum disorder based on common orthogonal basis extraction algorithm improves the accuracy of clinical symptoms prediction","fulltext":[{"header":"Introduction","content":"\u003cp\u003eAutism spectrum disorder (ASD) is a group of neurodevelopmental disorders with onset in early childhood, characterized as impaired social communication and restricted and repetitive behaviors (RRB)[\u003cspan additionalcitationids=\"CR2\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Research has revealed that alterations in brain connectivity patterns emerge prior to the onset of behavioral symptoms in ASD[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Aberrant structural and functional brain development likely contributes significantly to the neural mechanisms underlying ASD[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Functional magnetic resonance imaging (fMRI) studies have revealed that ASD is related to brain functional connectivity (FC)[\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Furthermore, these FC patterns exhibiting strong age-dependent correlations encompassing numerous reported patterns of both hypoconnectivity and hyperconnectivity across large-scale brain networks, like within networks, between networks, and at the whole-brain level[\u003cspan additionalcitationids=\"CR8 CR9 CR10\" citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. For example, a study found increased FC in the auditory network (AN) and the sensorimotor network in children with ASD aged 3\u0026ndash;7 years, and notably enhanced positive connectivity between the primary visual network and higher visual network[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Research on older children (8\u0026ndash;13 years) found that children with ASD showed decreased FC within the default mode network (DMN), increased FC between the DMN and the salient network (SN), and decreased FC in the DMN midline core (medial prefrontal cortex-posterior cingulate cortex; mPFC-PCC) that correlated negatively with ASD symptom severity[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. Haghighat et al. reported that FC abnormalities in ASD primarily manifest as hyperconnectivity in childhood, whereas both hyperconnectivity and hypoconnectivity were observed in adolescence and adulthood, and they proposed that age is a significant factor influencing FC disturbances[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. Therefore, it is essential to map the dynamic neurodevelopmental trajectories in ASD and highlights the importance of age-stratified research in elucidating ASD heterogeneity.\u003c/p\u003e \u003cp\u003eTraditional case-control brain imaging analyses focus on group-level comparisons by describing the concept of the \"average patient\", ignoring the underlying biological heterogeneity among individual patients[\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. The traditional case-control paradigm fails to account for these sources of heterogeneity of individual patients and consequently neglects the heterogeneity inherent in ASD-related neuroimaging features[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Individual differences in brain structure and function provide rich information related to neuropathology[\u003cspan additionalcitationids=\"CR17\" citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. For instance, a study exploring the heterogeneity of brain structure in ASD based on individual structural covariance network found that ASD showed significantly altered structural covariance edges mainly involved in the frontal and subcortical regions, which highlights the crucial role of frontal and subcortical regions in the heterogeneity of ASD[\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. Additionally, studies based on FC found individual-specific fMRI-Subspaces improve FC prediction of behavior[\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. Therefore, it is important to detect individual-specific difference, especially for disorders with high heterogeneity like ASD, to explore the neurological pathological basis. The common orthogonal basis extraction (COBE) algorithm could identify and separate individual-shared and individual-specific subspaces from multiblock data [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. These individual-shared components help reveal connections among members of the dataset and can characterize the dataset itself, while individual-specific part aids in identifying each unique member within the dataset. Previous studies have applied COBE to resting-state functional magnetic resonance imaging (rs-fMRI) data for feature separation[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e], and found that individual-specific connectivity improved the prediction of clinical symptoms, highlighting the importance for dissecting heterogeneity of ASD[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. However, previous study failed to take age into account, which is highly important to ASD.\u003c/p\u003e \u003cp\u003eThis study will focus on altered functional connectivity (AFC) in ASD from a developmental perspective. Reference to previous studies[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e], we divided participants into four age groups. We employ the COBE algorithm to dissociate individual-shared features across different age groups (age-invariant) and individual-specific features (age-related) from AFC of individuals with ASD. We aim to (1) examine whether age-related individual-specific AFC enhances the predictive power for clinical symptoms; (2) identify relationships between these age-related, individual-specific AFC patterns and core behavioral symptoms across different developmental stages; and (3) validate the findings in an independent dataset.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003eParticipants\u003c/p\u003e \u003cp\u003eMRI data and phenotypic data from 3 sites with the largest participant numbers were downloaded from the autism brain imaging data exchange (ABIDE) datasets (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://fcon_1000.projects.nitrc.org/indi/abide/\u003c/span\u003e\u003cspan address=\"https://fcon_1000.projects.nitrc.org/indi/abide/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e)[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. We selected three identical sites with a large sample size and high data quality from ABIDE I and ABIDE II respectively, where participant inclusion/exclusion criteria and MRI scanning parameters remained consistent. Finally, 437 participants (age range: 6\u0026ndash;39 years; ASD\u0026thinsp;=\u0026thinsp;208, typical development (TD)\u0026thinsp;=\u0026thinsp;229) were included in this study. Detail of sites included is provided in Supplementary Materials TABLE \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e. In 2000, Jeffrey[\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e] put forward the theory of emerging adulthood, believing that 18\u0026ndash;25 years old is an independent developmental stage with unique psychosocial characteristics. Based on this, participants were divided into 4 age groups: 6\u0026ndash;11 years, 12\u0026ndash;17 years, 18\u0026ndash;25 years, and 26\u0026ndash;39 years.\u003c/p\u003e \u003cp\u003eValidation dataset: External validation was performed in an independent sample collected at Laboratory of Child and Adolescent Psychiatry, Mental Health Center and Psychiatric Laboratory, West China Hospital, Sichuan University. The ASD subjects were mainly recruited from the outpatient department of Mental Health Center of West China Hospital of Sichuan University and West China Second University Hospital of Sichuan University. They were initially diagnosed by experienced child psychiatrists according to the Diagnostic and Statistical Manual of Mental Disorders\u0026thinsp;\u0026minus;\u0026thinsp;5 (DSM-5)[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e] or the International Statistical Classification of Diseases and Related Health Problems 10th Revision (ICD-10)[\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. Subsequently, all participants underwent the Autism Diagnostic Observation Schedule (ADOS)[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e], a gold-standard diagnostic tool. Participants were included in the study only if they scored above the established diagnostic thresholds for ASD. For children aged 5 and below, we used the Gesell Development Schedules (GDS) to assess their developmental levels[\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e], which are expressed in developmental quotient (DQ). For children aged 6 and above, we used the third edition of the Chinese-Wechsler Intelligence Scale for Children (C-WISC)[\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e] to evaluate the cognitive levels of the subjects, which are represented by intelligence quotient (IQ). Subjects were excluded based on the following criteria: (1) neuropsychiatric disorders, such as epilepsy, encephalitis; (2) history of craniocerebral injury; (3) monogenetic diseases, such as fragile X syndrome or Angelman syndrome; (4) taking psychiatric medications during assessment; or (5) refusal to undergo MRI or the presence of contraindications for MRI, including fixed metal dentures/orthodontic appliances, pacemakers, or cochlear implants. Finally, 245 participants (age range: 2\u0026ndash;17 years; ASD\u0026thinsp;=\u0026thinsp;145, TD\u0026thinsp;=\u0026thinsp;99) were included in this study. We divided the subjects into 3 groups based on age: 2\u0026ndash;5 years (ASD\u0026thinsp;=\u0026thinsp;66, TD\u0026thinsp;=\u0026thinsp;30), 6\u0026ndash;11 years (ASD\u0026thinsp;=\u0026thinsp;61, TD\u0026thinsp;=\u0026thinsp;55), and 12\u0026ndash;17 years (ASD\u0026thinsp;=\u0026thinsp;19, TD\u0026thinsp;=\u0026thinsp;14).\u003c/p\u003e \u003cp\u003eMRI Data Preprocessing\u003c/p\u003e \u003cp\u003eThe parameters of MRI data collection see Supplementary Materials. Preprocessing of rs-fMRI data was performed on the MATLAB 2023a platform through RESTplus\u0026mdash;an enhanced toolkit for rs-fMRI data processing[\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. The preprocessing pipeline included: (1) Removal of the first 10 time points; (2) Slice timing correction; (3) Head motion correction; (4) Spatial normalization, using Diffeomorphic Anatomical Registration Through Exponentiated Lie algebra (DARTEL)[\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e] for registration to Montreal Neurological Institute(MNI) standard space with 3.0\u0026times;3.0\u0026times;3.0 mm\u0026sup3; resampling; (5) Spatial smoothing (6 mm full-width at half-maximum kernel); (6) Linear trend removal; (7) Band-pass filtering (0.01\u0026ndash;0.08 Hz); and (8) Nuisance covariate regression, including cerebrospinal fluid signals, white matter signals, and Friston-24 head motion parameters. Quality control measures: Excluded participants with poor registration quality; Excluded participants exhibiting head framewise displacement (FD)\u0026thinsp;\u0026gt;\u0026thinsp;3 mm translation or \u0026gt;\u0026thinsp;3\u0026deg; rotation. Information about number of excluded participants due to excessive head motion and missing time points see Supplementary Materials Table S2.\u003c/p\u003e \u003cp\u003eCompute Altered Functional Connectivity\u003c/p\u003e \u003cp\u003eIn this study, the Dosenbach\u0026rsquo; atlas[\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e] was utilized to parcellate the brain into 160 regions of interest (ROIs), which were subsequently categorized into six functionally distinct subnetworks: DMN, frontoparietal network (FPN), cingulo-opercular network (CON), sensorimotor network (SMN), occipital network (ON), and cerebellum (CB). Pairwise Pearson correlations between ROI time courses yielded a symmetric 160 \u0026times; 160 FC matrix for each subject, with correlation coefficients converted to z-scores through Fisher\u0026rsquo;s r-to-z transformation to approximate normality. Sex and sites were included as covariates in the regression. In line with previous work[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e], AFC was computed for each ASD participant as follows:\u003cdiv id=\"Equ1\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ1\" name=\"EquationSource\"\u003e\n$$\\:AFC=\\frac{{FC}_{ASD}-mean\\left({FC}_{TD}\\right)}{SD\\left({FC}_{TD}\\right)}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e1\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eAFC represents the altered level of FC strength between one individual with ASD and mean FC strength of TD. For each ASD participant, a 160 \u0026times; 160 AFC matrix was obtained.\u003c/p\u003e \u003cp\u003eCommon Orthogonal Basis Extraction\u003c/p\u003e \u003cp\u003eAFC matrices computed above were stratified into four age groups: 6\u0026ndash;11, 12\u0026ndash;17, 18\u0026ndash;25, and 26\u0026ndash;39 years. We applied the COBE [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e] algorithm to these age-stratified AFC matrices. Originally designed for multiblock data (i.e., collections of matrices) decomposition, COBE separates shared and block-specific subspaces[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. In this study, we demarcated age groups as experimental \"blocks.\" This allowed the COBE algorithm to extract both a shared subspace across all groups and an age-related subspace that captures the connectivity changes specific to each cohort. The illustration of COBE sees Supplementary Materials Fig. \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e. Code for COBE is publicly available: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/ClinicalBrainLab/OCD_Cerebellar-Visual-Community\u003c/span\u003e\u003cspan address=\"https://github.com/ClinicalBrainLab/OCD_Cerebellar-Visual-Community\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e \u003cp\u003ePredicting clinical symptoms using the individual-specific subspace\u003c/p\u003e \u003cp\u003eThe AFC matrix encompasses both AFC in the individual-shared subspace which is age-invariant and AFC in the individual-specific subspace which is age-related. To investigate the predictive efficacy of age-related individual-specific subspaces for clinical symptoms, we used elastic net regression to optimize model performance[\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]. This algorithm was applied separately to original AFC matrices and AFC in the individual-specific subspace across different age groups to construct predictive models for clinical symptoms. In this study, the core autism symptoms were assessed using the autism diagnostic observation schedule (ADOS)[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]. This instrument's three factors\u0026mdash;communication, social, and restricted and repetitive behaviors (RRB)\u0026mdash;served as clinical symptom labels. AFC matrices and AFC in the individual-specific subspaces from the four age groups were used as input features for training separate elastic net regression models. Model hyperparameters (regularization coefficients λ₁ and λ₂) were determined via grid search. Five-fold cross-validation was implemented to enhance model generalizability: Data were partitioned into five equal subsets; four subsets were used for training and the remaining subset for prediction, with this process repeated five times. Model accuracy was defined by Pearson correlation coefficient between the predicted score and clinical symptom score within the test fold[\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]. Finally, the average accuracy was calculated over five runs.\u003c/p\u003e \u003cp\u003eCorrelation analysis between the individual-specific subspace and clinical symptoms\u003c/p\u003e \u003cp\u003eTo investigate the relationship between AFC in the individual-specific subspace of different age groups and clinical symptoms, we computed the norms of brain networks of AFC in the individual-specific subspaces. Each element in the AFC matrix represents the changed FC value between two ROIs for an individual with ASD. For all elements corresponding to brain network i and brain network j, we concatenated them row-wise into a one-dimensional vector, denoted as X = (x₁, x₂, ..., xₙ). We then calculated their L2-norm:\u003cdiv id=\"Equ2\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ2\" name=\"EquationSource\"\u003e\n$$\\:\\left|\\right|X|{|}_{2}=\\sqrt{{x}_{1}^{2}+{x}_{2}^{2}+...+{x}_{n}^{2}}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e2\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eThe L2-norm, representing the Euclidean distance from the coordinate system origin to vector X, is calculated as the square root of the sum of its squared components. This metric quantifies a vector\u0026rsquo;s magnitude (length) and reflects its signal \"energy\" or \"intensity\". Within the distinctive subspaces of each age group, we computed the L2-norm for every brain network across all ASD participant. Subsequently, Pearson correlation analyses were performed between these L2-norm values (derived from individual-specific subspaces) and clinical behavioral measures. In this study, core autism symptoms were assessed using the Autism Diagnostic Observation Schedule (ADOS)[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e], which comprising three subscales: Communication, Social Interaction, and Restricted Repetitive Behaviors (RRB).\u003c/p\u003e \u003cp\u003eValidation analysis\u003c/p\u003e \u003cp\u003eValidation analysis was performed to estimate the robustness of our findings in an independent cohort. We decomposed AFC into the individual-shared and individual-specific subspaces. We evaluated the validity of our findings in the validation dataset. Applying the COBE algorithm similarly extracted age-independent individual-shared subspaces and age-related individual-specific subspaces. Similarly, we conducted predictive models to exam the predictive power of age-related individual-specific subspace for clinical symptoms through elastic net regression.\u003c/p\u003e"},{"header":"Results","content":"\u003ch2\u003eDemographic information \u0026nbsp;\u003c/h2\u003e\n\u003cp\u003eAfter excluding participants with excessive head motion, 437 participants (age range: 6-39 years; ASD=208, TD=229) were included in this study. For details on participant demographics from ABIDE, see Table 1.\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003eFor details on participant demographics from validation dataset, see TABLE 2.\u003c/p\u003e\n\u003ch2\u003eExtract three individual-shared components\u003c/h2\u003e\n\u003cp\u003eWe found that the estimations were robust when common component (C) is three, with the highest similarity 0.8 (Fig. S2). Thus, we focused on three common components in subsequent analyses. Across all four age groups, we consistently identified three ASD-shared components from the AFC matrix (see Fig. 1). In Component 1, functional connectivity was predominantly positive across both intra- and inter-network connections, except within SMN (Fig. 1(a), (b)). The top 30 strongest positive connections were primarily distributed between DMN and SMN, and between DMN and FPN (Fig. 1(c)). In Component 2, CB exhibited predominantly negative functional connectivity with other brain networks, while connections among other networks were mainly positive (Fig. 1(a), (b)). The top 30 positive connections were concentrated within the SMN, between SMN and CON, and between SMN and ON; conversely, the top 30 negative connections were predominantly distributed between CB and CON, and between CB and SMN (Fig. 1(c)). In Component 3, functional connectivity between the CB and other networks was primarily positive. The top 30 positive connections were mainly located between CB and CON, and between CB and SMN, whereas the top 30 negative connections were predominantly distributed between DMN and ON.\u003c/p\u003e\n\u003ch2\u003eAFC in the individual-specific subspace enhance the predictive ability of clinical symptoms\u003c/h2\u003e\n\u003cp\u003eCompared to raw AFC, the individual-specific subspace derived from AFC demonstrated significantly higher accuracy in predicting clinical symptoms, with a trend toward greater improvement in older age groups (Fig. 2). Specifically, the individual-specific subspace outperformed raw AFC in predicting specific domains across different age groups: in the subgroup of childhood (6-11 years), it showed higher predictive accuracy for communication (0.27 vs. 0.23; a 17% increase), social (0.37 vs. 0.26; a 42% increase), and RRB (0.33 vs. 0.32; a 3% increase), but not for the ADOS total score (0.35 vs. 0.29; \u0026nbsp;a 17%\u0026nbsp;decrease), with an average increase of 8.6% in predictive accuracy . In the subgroup of adolescence (12-17 years), it demonstrated superior accuracy for communication (0.31 vs. 0.18; a 72% increase), social (0.35 vs. 0.15; a 133% increase), and the ADOS total score (0.4 vs. 0.32; a 25% increase), but not for RRB (0.29 VS. 0.31; a 6% decrease), with an average increase of 41%. In the subgroup of emerging adulthood (18-25 years), it achieved higher accuracy for communication (0.35 VS. 0.29; a 21% increase), social (0.43 VS. 0.41; a 5% increase), RRB (0.32 VS. 0.22; a 45% increase), and the ADOS total score (0.51 VS. 0.37; a 38% increase), with an average increase of 25%. In early adulthood (26-39 years), it better predicted communication (0.38 VS. 0.35; a 8.6% increase), social (0.43 VS. 0.39; a 10% increase), and RRB (0.47 VS. 0.46; a 2% increase), but not for the ADOS total score (0.47 VS. 0.54; a 13% decrease), with an average increase of 0.6%. Notably, the individual-specific subspace in adolescents with ASD predicts social behavior, showing the highest level of improvement at 133%, as well as an average increase of 41%.\u003c/p\u003e\n\u003ch2\u003eThe correlation between clinical symptoms and the L2-norm of brain networks in individual-specific subspace\u003c/h2\u003e\n\u003cp\u003eWe computed the L2-norm of brain networks within the individual-specific subspace, where the L2-norm represents the \u0026quot;strength\u0026quot; of connectivity of brain network. Correlations between the L2-norm of brain networks in individual-specific subspace and behaviors were calculated separately for each age group. As shown in Fig. 3: In the \u0026ldquo;6-11 years\u0026rdquo; group, significant positive correlations were observed between CON-CB connectivity strength and both communication (r = 0.27, p = 0.020) and social skills (r = 0.28, p = 0.018). In the \u0026ldquo;12-17 years\u0026rdquo; group, negative correlations with communication were found across most inter- and intra-network connections, with the exception of intra-FPN, intra-ON, intra-CB, and inter-ON-CB connections. Significant negative correlations with social skills were identified for DMN-FPN (r = -0.25,\u003cem\u003e\u0026nbsp;p\u003c/em\u003e = 0.037), DMN-ON (r = -0.37,\u003cem\u003e\u0026nbsp;p\u003c/em\u003e = 0.002), DMN-CB (r = -0.30, \u003cem\u003ep\u003c/em\u003e = 0.01), FPN-CB (r = -0.28, \u003cem\u003ep\u003c/em\u003e = 0.019), CON-SMN (r = -0.30, \u003cem\u003ep\u003c/em\u003e = 0.012), CON-ON (r = -0.35, \u003cem\u003ep\u003c/em\u003e = 0.003), CON-CB (r = -0.33, \u003cem\u003ep\u003c/em\u003e = 0.006), and SMN-CB (r = -0.26, \u003cem\u003ep\u003c/em\u003e = 0.033). Significant negative correlations with RRB were observed for FPN-CB (r = -0.24, \u003cem\u003ep\u003c/em\u003e = 0.049) and ON-CB (r = -0.26, \u003cem\u003ep\u003c/em\u003e = 0.032). In the \u0026ldquo;18-25 years\u0026rdquo; group, significant negative correlations were found between FPN-CB connectivity strength and both communication (r = -0.52, \u003cem\u003ep\u003c/em\u003e = 0.004) and social skills (r = -0.39, \u003cem\u003ep\u003c/em\u003e = 0.041). In the \u0026ldquo;26-39 years\u0026rdquo; group, significant positive correlations with communication were identified for intra-DMN (r = 0.57, \u003cem\u003ep\u003c/em\u003e = 0.011), intra-SMN (r = 0.49, \u003cem\u003ep\u003c/em\u003e = 0.032), inter-DMN-FPN (r = 0.47, \u003cem\u003ep\u003c/em\u003e = 0.040), and inter-DMN-SMN (r = 0.48, \u003cem\u003ep\u003c/em\u003e = 0.038).\u003c/p\u003e\n\u003ch2\u003eValidation analysis\u003c/h2\u003e\n\u003cp\u003eIn an independent dataset, we extracted the individual-shared subspace from AFC matrix among different age groups, and obtained three common components in the individual-shared subspace in the same way. These three common components have similar connection patterns among the three common components with that obtained from ABIDE (Fig. 4). Furthermore, the individual-specific subspace significantly enhanced predictive accuracy for clinical symptoms compared to raw AFC (Fig. 5). Specifically, the individual-specific subspace outperformed raw AFC in predicting specific domains across different age groups: in early childhood (2-5 years), it achieved higher accuracy for communication (0.39 vs. 0.26; a 50% increase), social (0.41 vs. 0.32; a 28% increase), RRB (0.31 vs. 0.2; a 55% increase), and the ADOS total score (0.22 vs. 0.21; a 4.8% increase), with an average increase of 34%. In childhood (6-11 years), it showed higher predictive accuracy for communication (0.46 vs. 0.42; a 9.5% increase), and RRB (0.32 vs. 0.14; a 129% increase), but not for social (0.3 vs. 0.32; a 6.3% decrease) and the ADOS total score (0.45 vs. 0.47; a 4.3% decrease), with an average increase of 13%. In adolescence (12-17 years), it demonstrated superior accuracy for communication (0.72 vs. 0.46; a 57% increase), social (0.52 vs. 0.46; a 13% increase), RRB (0.48 VS. 0.38; a 26% increase), and the ADOS total score (0.61 vs. 0.4; a 53% increase), with an average increase of 37%. Notably, the individual-specific subspace in adolescents with ASD predicts clinical behaviors showing the highest level of average improvement at 37%. The correction analysis results of validation dataset were shown in Fig.\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003eS3.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eUsing COBE, AFC matrix of individual with ASD is projected onto two subspaces: one is an individual-shared subspace, representing the shared pattern of connectivity alterations in ASD across ages; the other is an age-related, individual-specific subspace, representing the unique connectivity features of each ASD individual after removing the shared components. We found that the age-related individual-specific AFC demonstrated greater predictive power for clinical symptoms compared to raw AFC. Furthermore, the age-related individual-specific subspace is associated with communication deficits, social interaction, and RRB. Overall, our findings demonstrate that it is necessary to take age into account to understand the heterogeneous FC of ASD. Capturing and utilizing individual-specific brain connectivity features at a given developmental stage is central to dissecting the clinical heterogeneity of ASD.\u003c/p\u003e \u003cp\u003eIndividual-shared common components across ages\u003c/p\u003e \u003cp\u003eThe three common components of the cross-age, ASD-shared subspace exhibit distinct patterns of alterations. Notably, the connectivity patterns between the DMN-CON, FPN-CON, and SMN-FPN networks were similar across the three shared components, characterized by predominantly positive connectivity. The brain networks of DMN, CON, FPN, and SMN have been extensively studied in ASD, and abnormalities within these networks have been consistently reported in numerous resting-state fMRI studies.\u003c/p\u003e \u003cp\u003eDMN is a core network for human introspection, social interaction, and consciousness. Dysfunction within the DMN is closely implicated in various neurological and psychiatric disorders, such as ASD, depression, and schizophrenia[\u003cspan additionalcitationids=\"CR35 CR36\" citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. Multiple brain regions comprising the DMN are strongly linked to key ASD theories, particularly the Theory of Mind (ToM) [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e] and the theory of neural connectivity deficits[\u003cspan additionalcitationids=\"CR41\" citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e]. The DMN is active during resting states and demonstrates increased engagement during social cognition and self-referential processing[\u003cspan additionalcitationids=\"CR44\" citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e]. In contrast, the FPN is recruited during active information processing and is critically implicated in executive functions, cognitive flexibility, working memory, goal-directed behavior, and attentional control, serves as the core circuitry for executing higher-order cognitive tasks. Higher-order cognitive impairments, particularly deficits in executive functions[\u003cspan additionalcitationids=\"CR47\" citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e], are characteristic of ASD. Dysfunctional connectivity of FPN is closely linked to core symptoms of ASD[\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e], suggesting the FPN may underlie fundamental aspects of ASD symptomatology. The CON, comprising anterior cingulate cortex (ACC), insula, and thalamus, is involved in conflict monitoring, cognitive control, and efficient attentional resource allocation. It plays a critical role in detecting the behavioral relevance of internal or external stimuli, orchestrating the dynamic switching between the DMN and FPN, and optimally allocating attentional resources[\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e, \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e]. Aberrant FC within this network is associated with impaired emotional and social information processing in ASD, highlighting its essential role in social cognition and complex cognitive processes[\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e]. The SMN supports motor control, sensory information processing, and motor learning. Abnormal SMN connectivity in ASD manifests as sensorimotor impairments[\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e, \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e], including repetitive/stereotyped movements[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e], aberrant oculomotor control[\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e], postural sway[\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e], and impaired motor coordination[\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eA study[\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e] found that children and adolescents with ASD had abnormally increased FC between the PCC, a hub region of the DMN, and bilateral occipital regions (e.g., middle occipital gyrus, lingual gyrus), reflecting excessive integration of visual information. Additionally, the hyperconnectivity between the PCC and language-related regions (e.g., bilateral inferior frontal gyrus, inferior parietal lobule, anterior/posterior cingulate cortex) reflected aberrant \"cross-network communication\" between the DMN and other brain networks, potentially leading to an imbalance between self-referential processing and external task switching during language processing. A recent resting-state fMRI meta-analysis[\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e] revealed that children with ASD exhibited decreased functional activity in the left insula (a hub of CON), bilateral ACC/ mPFC, left angular gyrus, and right inferior temporal gyrus, alongside increased functional activity in the right supplementary motor area (a hub of SMN) and precuneus. This pattern suggests functional abnormalities within the DMN, CON, FPN and SMN networks in ASD. Research has further demonstrated that reduced DMN FC is significantly associated with diminished social motivation and impaired mentalizing abilities in individuals with ASD[\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e]. For instance, decreased FC of posterior DMN nodes (e.g., the temporoparietal junction) predicts deficits in understanding social intentions in ASD patients[\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e]. Our findings provide evidence for the social and sensorimotor impairments in ASD, underscoring the converging role of the DMN, CON, FPN and SMN in the neuropathological mechanisms of the disorder.\u003c/p\u003e \u003cp\u003eAFC in the ASD-specific subspace enhance the predictive ability of clinical symptoms\u003c/p\u003e \u003cp\u003eASD exhibits substantial heterogeneity[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e], and characterizing brain-behavior relationships is essential for understanding and treating psychiatric conditions[\u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e]. Brain-behavior mapping in ASD constitutes a critical step for localizing specific behavioral circuits[\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e], guiding clinical research and practice through the identification of targets for individualized diagnosis and intervention. Currently, it lacks biologically validated diagnostic or therapeutic approaches for ASD. A crucial aspect of biomarker development involves demonstrating that candidate biomarkers predict relevant behavioral outcomes and disease trajectories[\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e]. Predictive modeling provides a statistically rigorous framework for characterizing individual differences, particularly in neurodevelopmental conditions[\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e, \u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e66\u003c/span\u003e]. Successful predictive models require that features derived from neuroimaging data exhibit inter-individual variability, preserving key individual characteristics[\u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e67\u003c/span\u003e]. For instance, prediction models inevitably yield poor performance if input features are identical across participants, precluding effective behavioral forecasting.\u003c/p\u003e \u003cp\u003ePrevious research has established that individualized FC profiles demonstrate strong predictive power within models, compelling a paradigm shift from group-level analyses toward leveraging individual differences to decipher high clinical heterogeneity, taking developmental factors into account as well[\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e]. Consequently, we incorporated individual-specific features when mapping neuroimaging data to ASD-related behavioral symptoms. Furthermore, accounting for developmental stage is paramount in predictive modeling for neurodevelopmental disorders. Our study stratified participants into four age groups to examine the predictive capacity of individual-specific subspaces for behavioral symptoms across distinct developmental periods. This approach aligns with Kazeminejad et al.[\u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e69\u003c/span\u003e], who constructed separate predictive models for different age cohorts and demonstrated distinct functional brain network architectures in ASD classification models for 5\u0026ndash;15 year-olds versus 15\u0026ndash;30 year-olds, supporting age-stratified modeling to enhance predictive accuracy.\u003c/p\u003e \u003cp\u003eIn our study, the strongest predictive power of the individual-specific subspace was observed in adolescence. This variation in predictive improvement across age groups suggests that the relationship between brain connectivity and behavioral symptoms in ASD is not static but undergoes dynamic evolution throughout development. The neural substrates supporting behavioral performance thus differ across childhood, adolescence, and adulthood, consistent with extensive neurodevelopmental literature[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e70\u003c/span\u003e]. Our finding that the largest gains in predictive accuracy occurred during adolescence is consistent with neurodevelopmental models of ASD, which posit that this period is marked by a pronounced and critical reorganization of brain networks[\u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e70\u003c/span\u003e, \u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e71\u003c/span\u003e]. The intimate link between these network changes and behavioral symptoms, as captured by our model, is further supported by studies showing age-specific alterations in functional connectivity that correlate with symptom severity[\u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e72\u003c/span\u003e, \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e73\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe correlation between the individual-specific subspace brain network and clinical symptoms\u003c/p\u003e \u003cp\u003eBrain network abnormalities in ASD exhibit marked individual specificity. Large-scale cohort studies reveal that approximately 30% of individuals with ASD demonstrate significant brain network age dysynchrony, characterized by developmentally lagged connectivity patterns in DMN and SN relative to chronological age[\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e]. Furthermore, the developmental trajectory of the DMN correlates with the severity of RRB[\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e]. Neuroplasticity-driven dynamic changes may manifest as childhood hyperconnectivity transitioning to hypoconnectivity during adolescence[\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e]. Recent research demonstrated that while neurotypical controls show significant prefrontal and insular activation in individuals under 25 years, the ASD group exhibits absent activation in these regions during this developmental period. Beyond age 25, only left insular activation emerges in ASD. This pattern suggests delayed maturation of executive function-associated regions (FPN) in ASD, supporting the delayed neurodevelopment hypothesis[\u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e74\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThis study demonstrates that within the individual-specific subspace of each age group, the connectivity between CB and other networks consistently correlate with communication abilities. CB was initially considered primarily responsible for motor function. However, emerging evidence reveals its critical role in multimodal sensory integration[\u003cspan additionalcitationids=\"CR76\" citationid=\"CR75\" class=\"CitationRef\"\u003e75\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR77\" class=\"CitationRef\"\u003e77\u003c/span\u003e]. Beyond motor dysregulation, CB dysfunction contributes to deficits in executive function, visuospatial processing, and emotional regulation[\u003cspan citationid=\"CR77\" class=\"CitationRef\"\u003e77\u003c/span\u003e]. Impaired CB integration may underlie multimodal deficits affecting motor coordination, language, and even social behavior[\u003cspan citationid=\"CR78\" class=\"CitationRef\"\u003e78\u003c/span\u003e]. Critically, cerebello-cortical connections (e.g., with DMN, CON, and FPN) effectively predict social and cognitive functioning[\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eIn this study, the norm of brain networks of AFC within the individual-specific subspace exhibited primarily positive correlations with clinical symptoms for ASD individuals aged 6\u0026ndash;11 and 26\u0026ndash;39 years, whereas an inverse relationship was observed for those aged 12\u0026ndash;17 and 18\u0026ndash;25 years (Fig.\u0026nbsp;3). However, a discrepancy was observed in the validation analysis (Fig. S3). This could be attributed to the sampling strategy, although the results were also strongly age-dependent. In summary, this study reveals distinct neurodevelopmental patterns of brain networks across different stages in ASD, laying the groundwork for identifying developmentally-informed neuroimaging biomarkers.\u003c/p\u003e\n\u003ch3\u003eLimitations\u003c/h3\u003e\n\u003cp\u003eOur results have several limitations. First, although ABIDE provides a multi-site neuroimaging data across a wide age range, the sample size for several age groups\u0026mdash;particularly those aged 18\u0026ndash;25 and 26\u0026ndash;39\u0026mdash;is relatively small. Future studies would benefit from larger datasets with more adult participants. Furthermore, longitudinal data are needed to clarify whether AFC in the individual-specific subspace reflects ASD heterogeneity, neurodevelopmental stage effects, or a complex interplay between the two.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eThe superior predictive power of the individual-specific subspace for clinical symptoms, which was most pronounced in the adolescent cohort, underscores a pivotal finding: the brain-behavior relationship in ASD is not static but is fundamentally shaped by both individual neurobiological variability and developmental stage. This strongly suggests that the quest for reliable biomarkers in ASD must move beyond group-level averages to embrace a more nuanced, personalized approach that accounts for the unique developmental trajectory of each individual. Consequently, our results highlight the critical importance of integrating developmental neuroscience principles into clinical translation. Targeting future interventions, particularly during pivotal windows of neurodevelopment such as adolescence, may yield the greatest therapeutic benefits by aligning with the brain's inherent plastic potential during this dynamic reorganizational period.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eABIDE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eAutism Brain Imaging Data Exchange\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eACC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eAnterior cingulate corte\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eADOS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eAutism Diagnostic Observation Schedule\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eAFC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eAltered functional connectivity\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eASD\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eAutism spectrum disorder\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eAN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eAuditory network\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eCB\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eCerebellum\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eCOBE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eCommon orthogonal basis extraction\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eCON\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eCingulo-opercular network\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eC-WISC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eChinese-Wechsler Intelligence Scale for Children\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eDQ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eDevelopmental quotient\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eDMN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eDefault mode network\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eDSM-5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eDiagnostic and Statistical Manual of Mental Disorders - 5\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eFC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eFunctional connectivity\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eFD\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eFramewise displacement\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eFPN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eFrontoparietal network\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003efMRI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eFunctional magnetic resonance imaging\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eGDS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eGesell Development Schedules\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eICD-10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eInternational Statistical Classification of Diseases and Related Health Problems 10th Revision\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eIQ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eIntelligence quotient\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003emPFC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eMedial prefrontal cortex\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eON\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eOccipital network\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003ePCC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003ePosterior cingulate cortex\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eROIs\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eRegions of interest\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eRRB\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eRestricted and repetitive behaviors\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003ers-fMRI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eResting-state functional magnetic resonance imaging\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eSMN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eSensorimotor network\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eSN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eSalient network\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eTD\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eTypical development\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eToM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 435px;\"\u003e\n \u003cp\u003eTheory of Mind\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAuthors\u0026apos; contributions\u003c/h2\u003e\n\u003cp\u003eTL download data from ABIDE, preprocessed the MRI data, and drafted the manuscript. JZ contributed to data analysis, scientific editing and language editing of the manuscript. MZ assisted in study conception and design. LL,\u0026nbsp;SZ, ZY and YJ contributed to the data collection. BL conceived and designed the analyses. YH contributed to funding, conception, and design. All authors listed have made a substantial, direct, or intellectual contribution to the work, and approved it for publication.\u003c/p\u003e\n\u003ch2\u003eFunding\u003c/h2\u003e\n\u003cp\u003eThis work was supported by\u0026nbsp;the Medical and Industrial Integration Project of Chengdu City (Grant\u0026nbsp;No. ZYGX2022YGRH020); the Sichuan Provincial Department of Science and Technology\u0026apos;s Special Project for Central Support of Local Science and Technology Development (Grant No. 2023ZYD0123); the Program of Chengdu Science and Technology (Grant\u0026nbsp;No.\u0026nbsp;2022-YF09-00010-SN).\u003c/p\u003e\n\u003ch2\u003eData Availability Statement\u003c/h2\u003e\n\u003cp\u003eThe data that support the findings of this study are available in the ABIDE dataset (https://fcon_1000.projects.nitrc.org/indi/abide/).\u003c/p\u003e\n\u003ch2\u003eEthics approval and consent to participate\u003c/h2\u003e\n\u003cp\u003eWe obtained informed consent from the parents or caregivers of the children who participated. This study was carried out in accordance with the Declaration of Helsinki and approved by the Medical Ethics Committee of the West China Hospital, Sichuan University.\u003c/p\u003e\n\u003ch2\u003eClinical trial number\u003c/h2\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003ch2\u003eConsent for publication\u003c/h2\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003ch2\u003eCompeting interests\u003c/h2\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAmerican Psychiatric, A.; American Psychiatric Association, D. S. M. T. F., Diagnostic and statistical manual of mental disorders : \u003cem\u003eDSM-5\u003c/em\u003e. Fifth ed.; American Psychiatric Publishing: 2013.\u003c/li\u003e\n\u003cli\u003eLord, C.; Elsabbagh, M.; Baird, G., et al., Autism spectrum disorder [J]. Lancet (London, England), 2018, 392 (10146): 508-520.\u003c/li\u003e\n\u003cli\u003eHirota, T.; King, B. H., Autism Spectrum Disorder: A Review [J]. JAMA, 2023, 329 (2): 157-168.\u003c/li\u003e\n\u003cli\u003eEmerson, R. W.; Adams, C.; Nishino, T., et al., Functional neuroimaging of high-risk 6-month-old infants predicts a diagnosis of autism at 24 months of age [J]. Science translational medicine, 2017, 9 (393).\u003c/li\u003e\n\u003cli\u003eWilliams, J. A.; Burgess, S.; Suckling, J., et al., Inflammation and Brain Structure in Schizophrenia and Other Neuropsychiatric Disorders: A Mendelian Randomization Study [J]. JAMA psychiatry, 2022, 79 (5): 498-507.\u003c/li\u003e\n\u003cli\u003eRasero, J.; Jimenez-Marin, A.; Diez, I., et al., The Neurogenetics of Functional Connectivity Alterations in Autism: Insights From Subtyping in 657 Individuals [J]. Biol Psychiatry, 2023, 94 (10): 804-813.\u003c/li\u003e\n\u003cli\u003eO\u0026apos;Hearn, K.; Lynn, A., Age differences and brain maturation provide insight into heterogeneous results in autism spectrum disorder [J]. Frontiers in human neuroscience, 2022, 16: 957375.\u003c/li\u003e\n\u003cli\u003eSato, W.; Uono, S., The atypical social brain network in autism: advances in structural and functional MRI studies [J]. Curr Opin Neurol, 2019, 32 (4): 617-621.\u003c/li\u003e\n\u003cli\u003eHaghighat, H.; Mirzarezaee, M.; Araabi, B. N., et al., Functional Networks Abnormalities in Autism Spectrum Disorder: Age-Related Hypo and Hyper Connectivity [J]. Brain topography, 2021, 34 (3): 306-322.\u003c/li\u003e\n\u003cli\u003eDi 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]. Molecular psychiatry, 2014, 19 (6): 659-67.\u003c/li\u003e\n\u003cli\u003eNair, A.; Jolliffe, M.; Lograsso, Y. S. S., et al., A Review of Default Mode Network Connectivity and Its Association With Social Cognition in Adolescents With Autism Spectrum Disorder and Early-Onset Psychosis [J]. Frontiers in psychiatry, 2020, 11: 614.\u003c/li\u003e\n\u003cli\u003eWang, 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]. Frontiers in physiology, 2021, 12: 626338.\u003c/li\u003e\n\u003cli\u003eYerys, B. E.; Gordon, E. M.; Abrams, D. N., et al., Default mode network segregation and social deficits in autism spectrum disorder: Evidence from non-medicated children [J]. Neuroimage Clin, 2015, 9: 223-32.\u003c/li\u003e\n\u003cli\u003eWolfers, T.; Beckmann, C. F.; Hoogman, M., et al., Individual differences v. the average patient: mapping the heterogeneity in ADHD using normative models [J]. Psychol Med, 2020, 50 (2): 314-323.\u003c/li\u003e\n\u003cli\u003eTang, S.; Sun, N.; Floris, D. L., et al., Reconciling Dimensional and Categorical Models of Autism Heterogeneity: A Brain Connectomics and Behavioral Study [J]. Biological psychiatry, 2020, 87 (12): 1071-1082.\u003c/li\u003e\n\u003cli\u003eGuo, X.; Zhang, X.; Chen, H., et al., Exploring the heterogeneity of brain structure in autism spectrum disorder based on individual structural covariance network [J]. Cereb Cortex, 2023, 33 (12): 7311-7321.\u003c/li\u003e\n\u003cli\u003eKashyap, R.; Kong, R.; Bhattacharjee, S., et al., Individual-specific fMRI-Subspaces improve functional connectivity prediction of behavior [J]. NeuroImage, 2019, 189: 804-812.\u003c/li\u003e\n\u003cli\u003eShan, X.; Uddin, L. Q.; Ma, R., et al., Disentangling the Individual-Shared and Individual-Specific Subspace of Altered Brain Functional Connectivity in Autism Spectrum Disorder [J]. Biological psychiatry, 2024, 95 (9): 870-880.\u003c/li\u003e\n\u003cli\u003eZhou, G.; Zhao, Q.; Zhang, Y., et al., Linked Component Analysis From Matrices to High-Order Tensors: Applications to Biomedical Data [J]. Proceedings of the IEEE, 2016, 104 (2): 310-331.\u003c/li\u003e\n\u003cli\u003eZhou, G.; Cichocki, A.; Zhang, Y., et al., Group Component Analysis for Multiblock Data: Common and Individual Feature Extraction [J]. IEEE transactions on neural networks and learning systems, 2016, 27 (11): 2426-2439.\u003c/li\u003e\n\u003cli\u003eKashyap, R.; Bhattacharjee, S.; Yeo, B. T. T., et al., Maximizing dissimilarity in resting state detects heterogeneous subtypes in healthy population associated with high substance use and problems in antisocial personality [J]. Human brain mapping, 2020, 41 (5): 1261-1273.\u003c/li\u003e\n\u003cli\u003eZhang, A.; Liu, L.; Chang, S., et al., Connectivity-Based Brain Network Supports Restricted and Repetitive Behaviors in Autism Spectrum Disorder Across Development [J]. Frontiers in psychiatry, 2022, 13: 874090.\u003c/li\u003e\n\u003cli\u003eDi Martino, A.; O\u0026apos;Connor, D.; Chen, B., et al., Enhancing studies of the connectome in autism using the autism brain imaging data exchange II [J]. Scientific data, 2017, 4: 170010.\u003c/li\u003e\n\u003cli\u003eArnett, J. J., Emerging adulthood. A theory of development from the late teens through the twenties [J]. The American psychologist, 2000, 55 (5): 469-80.\u003c/li\u003e\n\u003cli\u003eOrganization, W. H., ICD-10 : international statistical classification of diseases and related health problems : tenth revision [J]. Acta Chirurgica Iugoslavica, 2010, 56 (3): 65-9.\u003c/li\u003e\n\u003cli\u003eLord, C.; Risi, S.; Lambrecht, L., et al., The autism diagnostic observation schedule-generic: a standard measure of social and communication deficits associated with the spectrum of autism [J]. Journal of autism and developmental disorders, 2000, 30 (3): 205-23.\u003c/li\u003e\n\u003cli\u003eRachel; S.; Ball, The Gesell developmental schedules: Arnold Gesell (1880\u0026ndash;1961) [J]. Journal of Abnormal Child Psychology, 1977, 5 (3): 233-239.\u003c/li\u003e\n\u003cli\u003eWoolger, C., Wechsler Intelligence Scale for Children-Third Edition (wisc-iii) [J]. Springer US.\u003c/li\u003e\n\u003cli\u003eJia, X. Z.; Wang, J.; Sun, H. Y., et al., RESTplus: an improved toolkit for resting-state functional magnetic resonance imaging data processing [J]. Science bulletin, 2019, 64 (14): 953-954.\u003c/li\u003e\n\u003cli\u003eAshburner, J., A fast diffeomorphic image registration algorithm [J]. NeuroImage, 2007, 38 (1): 95-113.\u003c/li\u003e\n\u003cli\u003eDosenbach, N. U.; Nardos, B.; Cohen, A. L., et al., Prediction of individual brain maturity using fMRI [J]. Science (New York, N.Y.), 2010, 329 (5997): 1358-61.\u003c/li\u003e\n\u003cli\u003eZou, H.; Hastie, T., Regularization and variable selection via the elastic net [J]. Journal of the Royal Statistical Society, 2005, 67 (5): 768-768.\u003c/li\u003e\n\u003cli\u003eFinn, E. S.; Shen, X.; Scheinost, D., et al., Functional connectome fingerprinting: identifying individuals using patterns of brain connectivity [J]. Nature neuroscience, 2015.\u003c/li\u003e\n\u003cli\u003eYang, B.; Wang, M.; Zhou, W., et al., Disrupted network integration and segregation involving the default mode network in autism spectrum disorder [J]. Journal of affective disorders, 2023, 323: 309-319.\u003c/li\u003e\n\u003cli\u003eZhou, H. X.; Chen, X.; Shen, Y. Q., et al., Rumination and the default mode network: Meta-analysis of brain imaging studies and implications for depression [J]. NeuroImage, 2020, 206: 116287.\u003c/li\u003e\n\u003cli\u003eKing, S.; Mothersill, D.; Holleran, L., et al., Early life stress, low-grade systemic inflammation and weaker suppression of the default mode network (DMN) during face processing in Schizophrenia [J]. Translational psychiatry, 2023, 13 (1): 213.\u003c/li\u003e\n\u003cli\u003eGattuso, J. J.; Perkins, D.; Ruffell, S., et al., Default Mode Network Modulation by Psychedelics: A Systematic Review [J]. The international journal of neuropsychopharmacology, 2023, 26 (3): 155-188.\u003c/li\u003e\n\u003cli\u003eWing, L.; Gould, J., Severe impairments of social interaction and associated abnormalities in children: epidemiology and classification [J]. J Autism Dev Disord, 1979, 9 (1): 11-29.\u003c/li\u003e\n\u003cli\u003eSimon; Baron-Cohen; Alan, et al., Mechanical, behavioural and Intentional understanding of picture stories in autistic children [J]. British Journal of Developmental Psychology, 1986.\u003c/li\u003e\n\u003cli\u003eFrith, C., Is autism a disconnection disorder? [J]. The Lancet. Neurology, 2004, 3 (10): 577.\u003c/li\u003e\n\u003cli\u003eGeschwind, D. H.; Levitt, P., Autism spectrum disorders: developmental disconnection syndromes [J]. Current opinion in neurobiology, 2007, 17 (1): 103-11.\u003c/li\u003e\n\u003cli\u003eKana, R. K.; Uddin, L. Q.; Kenet, T., et al., Brain connectivity in autism [J]. Frontiers in human neuroscience, 2014, 8: 349.\u003c/li\u003e\n\u003cli\u003eIacoboni, M.; Lieberman, M. D.; Knowlton, B. J., et al., Watching social interactions produces dorsomedial prefrontal and medial parietal BOLD fMRI signal increases compared to a resting baseline [J]. NeuroImage, 2004, 21 (3): 1167-73.\u003c/li\u003e\n\u003cli\u003eSpreng, R. N.; Grady, C. L., Patterns of brain activity supporting autobiographical memory, prospection, and theory of mind, and their relationship to the default mode network [J]. Journal of cognitive neuroscience, 2010, 22 (6): 1112-23.\u003c/li\u003e\n\u003cli\u003eSpreng, R. N., The fallacy of a \u0026quot;task-negative\u0026quot; network [J]. Frontiers in psychology, 2012, 3: 145.\u003c/li\u003e\n\u003cli\u003eDemetriou, E. A.; Lampit, A.; Quintana, D. S., et al., Autism spectrum disorders: a meta-analysis of executive function [J]. Mol Psychiatry, 2018, 23 (5): 1198-1204.\u003c/li\u003e\n\u003cli\u003eXie, R.; Sun, X.; Yang, L., et al., Characteristic Executive Dysfunction for High-Functioning Autism Sustained to Adulthood [J]. Autism research : official journal of the International Society for Autism Research, 2020, 13 (12): 2102-2121.\u003c/li\u003e\n\u003cli\u003eSadozai, A. K.; Sun, C.; Demetriou, E. A., et al., Executive function in children with neurodevelopmental conditions: a systematic review and meta-analysis [J]. Nature human behaviour, 2024, 8 (12): 2357-2366.\u003c/li\u003e\n\u003cli\u003eLin, H. Y.; Perry, A.; Cocchi, L., et al., Development of frontoparietal connectivity predicts longitudinal symptom changes in young people with autism spectrum disorder [J]. Transl Psychiatry, 2019, 9 (1): 86.\u003c/li\u003e\n\u003cli\u003eSeeley, W. W.; Menon, V.; Schatzberg, A. F., et al., Dissociable intrinsic connectivity networks for salience processing and executive control [J]. The Journal of neuroscience : the official journal of the Society for Neuroscience, 2007, 27 (9): 2349-56.\u003c/li\u003e\n\u003cli\u003eEckert, M. A.; Menon, V.; Walczak, A., et al., At the heart of the ventral attention system: the right anterior insula [J]. Human brain mapping, 2009, 30 (8): 2530-41.\u003c/li\u003e\n\u003cli\u003eAttanasio, M.; Mazza, M.; Le Donne, I., et al., Salience Network in Autism: preliminary results on functional connectivity analysis in resting state [J]. European archives of psychiatry and clinical neuroscience, 2024.\u003c/li\u003e\n\u003cli\u003eLim, Y. H.; Partridge, K.; Girdler, S., et al., Standing Postural Control in Individuals with Autism Spectrum Disorder: Systematic Review and Meta-analysis [J]. J Autism Dev Disord, 2017, 47 (7): 2238-2253.\u003c/li\u003e\n\u003cli\u003eCook, J., From movement kinematics to social cognition: the case of autism [J]. Philosophical transactions of the Royal Society of London. Series B, Biological sciences, 2016, 371 (1693).\u003c/li\u003e\n\u003cli\u003eSchmitt, L. M.; Cook, E. H.; Sweeney, J. A., et al., Saccadic eye movement abnormalities in autism spectrum disorder indicate dysfunctions in cerebellum and brainstem [J]. Mol Autism, 2014, 5 (1): 47.\u003c/li\u003e\n\u003cli\u003eBojanek, E. K.; Wang, Z.; White, S. P., et al., Postural control processes during standing and step initiation in autism spectrum disorder [J]. J Neurodev Disord, 2020, 12 (1): 1.\u003c/li\u003e\n\u003cli\u003eDziuk, M. A.; Gidley Larson, J. C.; Apostu, A., et al., Dyspraxia in autism: association with motor, social, and communicative deficits [J]. Developmental medicine and child neurology, 2007, 49 (10): 734-9.\u003c/li\u003e\n\u003cli\u003eGao, Y.; Linke, A.; Jao Keehn, R. J., et al., The language network in autism: Atypical functional connectivity with default mode and visual regions [J]. Autism research : official journal of the International Society for Autism Research, 2019, 12 (9): 1344-1355.\u003c/li\u003e\n\u003cli\u003eGuo, Z.; Tang, X.; Xiao, S., et al., Systematic review and meta-analysis: multimodal functional and anatomical neural alterations in autism spectrum disorder [J]. Mol Autism, 2024, 15 (1): 16.\u003c/li\u003e\n\u003cli\u003eJiang, 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.\u003c/li\u003e\n\u003cli\u003eHorien, C.; Floris, D. L.; Greene, A. S., et al., Functional Connectome-Based Predictive Modeling in Autism [J]. Biol Psychiatry, 2022, 92 (8): 626-642.\u003c/li\u003e\n\u003cli\u003eGuo, X.; Zhai, G.; Liu, J., et al., Inter-individual heterogeneity of functional brain networks in children with autism spectrum disorder [J]. Mol Autism, 2022, 13 (1): 52.\u003c/li\u003e\n\u003cli\u003eVieira, S.; Bolton, T. A. W.; Sch\u0026ouml;ttner, M., et al., Multivariate brain-behaviour associations in psychiatric disorders [J]. Transl Psychiatry, 2024, 14 (1): 231.\u003c/li\u003e\n\u003cli\u003eKoutsouleris, N.; Meisenzahl, E. M.; Davatzikos, C., et al., Use of neuroanatomical pattern classification to identify subjects in at-risk mental states of psychosis and predict disease transition [J]. Archives of general psychiatry, 2009, 66 (7): 700-12.\u003c/li\u003e\n\u003cli\u003eScheinost, D.; Noble, S.; Horien, C., et al., Ten simple rules for predictive modeling of individual differences in neuroimaging [J]. NeuroImage, 2019, 193: 35-45.\u003c/li\u003e\n\u003cli\u003eRosenberg, M. D.; Casey, B. J.; Holmes, A. J., Prediction complements explanation in understanding the developing brain [J]. Nature communications, 2018, 9 (1): 589.\u003c/li\u003e\n\u003cli\u003eFinn, E. S.; Todd Constable, R., Individual variation in functional brain connectivity: implications for personalized approaches to psychiatric disease [J]. Dialogues in clinical neuroscience, 2016, 18 (3): 277-287.\u003c/li\u003e\n\u003cli\u003eCui, W.; Ma, Y.; Ren, J., et al., Personalized Functional Connectivity Based Spatio-Temporal Aggregated Attention Network for MCI Identification [J]. IEEE transactions on neural systems and rehabilitation engineering : a publication of the IEEE Engineering in Medicine and Biology Society, 2023, 31: 2257-2267.\u003c/li\u003e\n\u003cli\u003eKazeminejad, A.; Sotero, R. C., Topological Properties of Resting-State fMRI Functional Networks Improve Machine Learning-Based Autism Classification [J]. Frontiers in neuroscience, 2018, 12: 1018.\u003c/li\u003e\n\u003cli\u003eUddin, L. Q.; Supekar, K.; Menon, V., Reconceptualizing functional brain connectivity in autism from a developmental perspective [J]. Frontiers in human neuroscience, 2013, 7: 458.\u003c/li\u003e\n\u003cli\u003ePaus, T.; Keshavan, M.; Giedd, J. N., Why do many psychiatric disorders emerge during adolescence? [J]. Nature reviews. Neuroscience, 2008, 9 (12): 947-57.\u003c/li\u003e\n\u003cli\u003eNomi, J. S.; Uddin, L. Q., Developmental changes in large-scale network connectivity in autism [J]. Neuroimage Clin, 2015, 7: 732-41.\u003c/li\u003e\n\u003cli\u003ePadmanabhan, A.; Lynch, C. J.; Schaer, M., et al., The Default Mode Network in Autism [J]. Biol Psychiatry Cogn Neurosci Neuroimaging, 2017, 2 (6): 476-486.\u003c/li\u003e\n\u003cli\u003eMay, K. E.; Kana, R. K., Frontoparietal Network in Executive Functioning in Autism Spectrum Disorder [J]. Autism research : official journal of the International Society for Autism Research, 2020, 13 (10): 1762-1777.\u003c/li\u003e\n\u003cli\u003eIshikawa, T.; Shimuta, M.; H\u0026auml;usser, M., Multimodal sensory integration in single cerebellar granule cells in vivo [J]. eLife, 2015, 4.\u003c/li\u003e\n\u003cli\u003eRonconi, L.; Casartelli, L.; Carna, S., et al., When one is Enough: Impaired Multisensory Integration in Cerebellar Agenesis [J]. Cerebral cortex (New York, N.Y. : 1991), 2017, 27 (3): 2041-2051.\u003c/li\u003e\n\u003cli\u003eXiao, L.; Scheiffele, P., Local and long-range circuit elements for cerebellar function [J]. Current opinion in neurobiology, 2018, 48: 146-152.\u003c/li\u003e\n\u003cli\u003eModi, M. E.; Sahin, M., Translational use of event-related potentials to assess circuit integrity in ASD [J]. Nature reviews. Neurology, 2017, 13 (3): 160-170.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003e\u003cstrong\u003eTable 1\u003c/strong\u003e Participant Demographics of ABIDE\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"633\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eASD (M \u0026plusmn; SD)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTD (M \u0026plusmn; SD)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e\u003cstrong\u003et/\u0026chi;\u0026sup2;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003ep\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCohen\u0026rsquo; d\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAll age\u003c/strong\u003e\u003cstrong\u003e（\u003c/strong\u003e\u003cstrong\u003e6-39 years\u003c/strong\u003e\u003cstrong\u003e）\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003eN=208\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003eN=229\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e-\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e-\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e-\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eAge\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003e15.29\u0026plusmn;6.99\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003e15.87\u0026plusmn;7.14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e-0.868\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e0.386\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e-0.083\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eSex（M/F）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003e192/16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003e191/38\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e7.975\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e0.005**\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e-\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 161px;\"\u003e\n \u003cp\u003eIQ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e102 \u0026plusmn; 17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 107px;\"\u003e\n \u003cp\u003e113\u0026plusmn;14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 77px;\"\u003e\n \u003cp\u003e-7.249\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 95px;\"\u003e\n \u003cp\u003e\u0026lt; 0.001**\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e-0.698\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e6-11 years\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003eN= 85\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003eN= 83\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eAge\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003e9.60\u0026plusmn;1.61\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003e9.59\u0026plusmn;1.54\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e0.039\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e0.969\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e0.006\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eSex（M/F）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003e78/7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003e69/14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e2.861\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e0.091\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eIQ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003e104 \u0026plusmn;18\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003e115\u0026plusmn;14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e-4.056\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e\u0026lt;0.001**\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e-.630\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e12-17 years\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003eN=72\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003eN=76\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eAge\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003e14.77\u0026plusmn;1.69\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003e14.35\u0026plusmn;1.59\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e1.562\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e0.120\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e0.257\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eSex（M/F）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003e69/3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003e61/15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e8.390\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e0.004**\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eIQ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003e99\u0026plusmn;16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003e110\u0026plusmn;13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e-4.721\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e\u0026lt;0.001**\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e-1.703\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e18-25 years\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003eN=31\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003eN=42\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eAge\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003e21.75\u0026plusmn;2.15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003e21.35\u0026plusmn;2.24\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e0.757\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e0.451\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e0.179\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eSex（M/F）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003e27/4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003e37/5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e0.016\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e0.898\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eIQ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003e106\u0026plusmn;16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003e114\u0026plusmn;11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e-2.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e0.024*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e-0.550\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 161px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e26-39 years\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003eN=20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003eN=28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eAge\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e31.35\u0026nbsp;\u0026plusmn; 4.25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 107px;\"\u003e\n \u003cp\u003e30.47\u0026nbsp;\u0026plusmn; 3.67\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 77px;\"\u003e\n \u003cp\u003e0.774\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 95px;\"\u003e\n \u003cp\u003e0.443\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e0.227\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eSex（M/F）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 109px;\"\u003e\n \u003cp\u003e18/2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003e24/4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e0.196\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 95px;\"\u003e\n \u003cp\u003e0.658\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eIQ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 109px;\"\u003e\n \u003cp\u003e103\u0026nbsp;\u0026plusmn; 15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 107px;\"\u003e\n \u003cp\u003e117\u0026nbsp;\u0026plusmn; 15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 77px;\"\u003e\n \u003cp\u003e-3.218\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 95px;\"\u003e\n \u003cp\u003e0.002**\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 85px;\"\u003e\n \u003cp\u003e-0.942\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e*\u003cem\u003ep\u003c/em\u003e\u0026lt;0.05, **\u003cem\u003ep\u003c/em\u003e\u0026lt;0.01. Abbreviations: IQ, intelligence quotient; M, male; F, female; M: mean; SD: standard deviation\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2\u0026nbsp;\u003c/strong\u003eParticipant demographics of validation database\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"687\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eASD (M \u0026plusmn; SD)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTD (M \u0026plusmn; SD)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 86px;\"\u003e\n \u003cp\u003e\u003cstrong\u003et/\u0026chi;\u0026sup2;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003ep\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCohen\u0026rsquo; d\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAll age (2-17years)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003eN=146\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003eN=99\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 86px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e-\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e-\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e-\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eAge\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e7.16 \u0026plusmn; 3.69\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e7.63 \u0026plusmn; 3.31\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 86px;\"\u003e\n \u003cp\u003e-1.035\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.302\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e-0.135\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eSex（M/F）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e127/19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e70/29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 86px;\"\u003e\n \u003cp\u003e9.925\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.002*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e-\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2 - 5 years\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003eN= 66\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003eN= 30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 86px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eAge\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e4.06 \u0026plusmn; 1.04\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e4.26 \u0026plusmn; 0.91\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 86px;\"\u003e\n \u003cp\u003e-0.925\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.358\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e-0.204\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eSex（M/F）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e54/12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e17/13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 86px;\"\u003e\n \u003cp\u003e6.774\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.009*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eDQ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e67.40 \u0026plusmn; 14.05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e102.87\u0026plusmn;11.09\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 86px;\"\u003e\n \u003cp\u003e-12.132\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e\u0026lt;0.001**\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e-2.691\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e6 \u0026ndash; 11 years\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003eN=61\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003eN=55\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 86px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eAge\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e8.26 \u0026plusmn; 1.60\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e7.91\u0026plusmn;1.57\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 86px;\"\u003e\n \u003cp\u003e1.197\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.234\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e0.223\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eSex（M/F）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e55/6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e43/12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 86px;\"\u003e\n \u003cp\u003e3.168\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.075\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eIQ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e82.58\u0026plusmn;24.54\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e116.34\u0026plusmn;13.28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 86px;\"\u003e\n \u003cp\u003e-9.009\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e\u0026lt;0.001**\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e-1.703\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e12 \u0026ndash; 17 years\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003eN=19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003eN=14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 86px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eAge\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e14.37\u0026plusmn;1.80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e13.79\u0026plusmn;1.76\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 86px;\"\u003e\n \u003cp\u003e0.937\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.356\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e0.430\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eSex（M/F）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e17/2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e12/4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 86px;\"\u003e\n \u003cp\u003e1.281\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.258\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003eIQ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 142px;\"\u003e\n \u003cp\u003e90.53 \u0026plusmn; 22.48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e110.71 \u0026plusmn; 9.47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 86px;\"\u003e\n \u003cp\u003e-3.150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 81px;\"\u003e\n \u003cp\u003e0.004**\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 94px;\"\u003e\n \u003cp\u003e-1.110\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eAbbreviations: ASD: autism spectrum disorder; TD: typical development; IQ: intelligence quotient; DQ: development quotient; M, male; F, female; M: mean; SD: standard deviation.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-psychiatry","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bpsy","sideBox":"Learn more about [BMC Psychiatry](http://bmcpsychiatry.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bpsy/default.aspx","title":"BMC Psychiatry","twitterHandle":"@BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Autism spectrum disorder, Age, Individual-specific, Functional connectivity, Brain network","lastPublishedDoi":"10.21203/rs.3.rs-8832166/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8832166/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eAutism spectrum disorder (ASD) is a group of highly heterogeneous neurodevelopmental disorders with onset in early childhood. Functional magnetic resonance imaging (fMRI) studies have revealed that ASD is related to altered functional connectivity (AFC), and that individual-specific change isolated from AFC with data-driven method of individuals with ASD has significantly enhanced the predictive ability for behavioral symptoms. Although few studies have incorporated age as a factor, it is critically important for ASD, a neurodevelopmental disorder, as age significantly influences the disorder's onset and progression.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eIn this study, we analyzed 437 participants (208 ASD, 229 typical development) from the Autism Brain Imaging Data Exchange, employed the common orthogonal basis extraction (COBE) algorithm to isolate age-related individual-specific features and examine their predictive abilities for clinical behaviors. A validation analysis was performed in an independent sample.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eWe found that the age-related, individual-specific feature set significantly improved behavioral prediction. The most substantial improvement was observed in predicting social behavior among adolescents with ASD, which showed a peak increase of 133% and an average increase of 41% compared to AFC. These findings were replicated in an independent validation dataset.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eThe age-related individual-specific subspace demonstrates superior predictive power for clinical symptoms. This underscores the critical importance of incorporating both inter-individual variability and the developmental perspective into ASD biomarker exploration and targeted intervention research.\u003c/p\u003e","manuscriptTitle":"Age-related individual-specific subspace of autism spectrum disorder based on common orthogonal basis extraction algorithm improves the accuracy of clinical symptoms prediction","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-24 08:19:14","doi":"10.21203/rs.3.rs-8832166/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-03-30T13:29:42+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-25T18:10:09+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-24T09:17:44+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"330676403422529832563537742408131717965","date":"2026-03-16T02:27:44+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"328064229209123108856400151040473820562","date":"2026-03-11T13:55:04+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"219559479815663666284736094188688216384","date":"2026-02-18T05:21:55+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-02-17T11:48:39+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-02-12T09:55:35+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-02-11T04:57:27+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-02-11T04:56:52+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Psychiatry","date":"2026-02-09T15:26:50+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-psychiatry","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bpsy","sideBox":"Learn more about [BMC Psychiatry](http://bmcpsychiatry.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bpsy/default.aspx","title":"BMC Psychiatry","twitterHandle":"@BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"dd5aa7f8-0c1d-4651-bdef-7cd0fb21ef0b","owner":[],"postedDate":"February 24th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-05-05T10:40:10+00:00","versionOfRecord":[],"versionCreatedAt":"2026-02-24 08:19:14","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8832166","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8832166","identity":"rs-8832166","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}