1. Retraction: Hybrid deep learning and feature selection approach for autism detection from rs-fMRI data. PLoS One. 2026; 21(7): e0354652.

Lien vers le texte intégral (Open Access ou abonnement)

2. Failla C, Minutoli R, Chilà P, Doria G, Scarcella I, Marraffa C, Corpina F, Roccaforte G, Crifò N, Meduri A, Pioggia G, Marino F. Immersive wearable virtual reality for autism: a systematic review of current evidence. Front Psychiatry. 2026; 17: 1771573.

INTRODUCTION: Immersive and wearable virtual reality (VR) is an emerging technology with growing potential to support assessment and intervention ifor autistic people. The methodological heterogeneity of existing studies limits the interpretation and generalization of current evidence. METHODS: A systematic review with a narrative synthesis was conducted in accordance with the PRISMA guidelines. Electronic searches were performed in PubMed, Scopus, IEEE Xplore, Web of Science, and Google Scholar, identifying studies published between 2015 and August 2025. Twenty-two studies investigating wearable and immersive VR interventions in children and adults with ASD met the eligibility criteria. RESULTS: The included studies demonstrated that wearable VR interventions may improve social communication, joint attention, emotional regulation, daily living skills, executive functioning, and user engagement. Innovative technologies, including eye-tracking and artificial intelligence-based systems, also enabled objective assessment of gaze behaviour, social interaction, and physiological responses. Nevertheless, the evidence was characterized by considerable methodological heterogeneity, predominantly small sample sizes, limited use of randomized controlled designs, and scarce long-term follow-up, reducing the generalizability of the findings. DISCUSSION: Wearable VR represents a promising tool for personalized assessment and intervention in ASD. Based on the current evidence, we propose a structured pre-intervention assessment integrating sensory, cognitive, emotional, and VR tolerance profiles to support individualized intervention planning. Future research should prioritize standardized outcome measures, rigorous study designs, and longitudinal investigations to strengthen the clinical translation of VR-based interventions in autism.

Lien vers le texte intégral (Open Access ou abonnement)

3. Han Y, Li X, Gu R, Ding Q, Qu X, He Y, Chen J, Luo YLL. Genetic and environmental origins of the co-occurrence of autistic traits, anxiety and depression: Evidence from meta-analysis and twin study. J Affect Disord. 2026; 414: 122317.

BACKGROUND: Prior twin studies have shown genetic and environmental overlap between autistic traits and internalising symptoms. However, it remains unknown whether this overlap reflects a common latent liability or similar aetiological influences operating through independent pathways. We investigated the genetic and environmental origins of the co-occurrence of autistic traits and internalising symptoms with a meta-analysis and an empirical twin study. METHODS: Study 1 meta-analysed six twin studies examining autistic traits and internalising symptoms, predominantly in Western child and adolescent samples. Study 2 applied multivariate genetic modelling to a Chinese community sample of 178 twin pairs (109 monozygotic and 69 dizygotic, aged 24-35). RESULTS: Study 1 showed genetic (rA = 0.42) and non-shared environmental (rE = 0.15) overlap between autistic traits and internalising symptoms, although substantial heterogeneity was observed across estimates. Study 2 was broadly consistent with the findings of Study 1. It further suggested that the co-occurrence of autistic traits, anxiety and depressive symptoms may be partly explained by a common latent factor shaped by genetic and non-shared environmental influences, although the common pathway model was only tentatively favoured over the independent pathway model. CONCLUSION: Across meta-analytic and empirical twin studies, the co-occurrence of autistic traits, anxiety and depressive symptoms appears to be shaped in part by overlapping genetic and non-shared environmental influences across these three traits. This generally aligns with the transdiagnostic framework of psychopathology. It also highlights that individuals presenting autistic traits may benefit from screening for internalising symptoms, and vice versa.

Lien vers le texte intégral (Open Access ou abonnement)

4. Hejazi A, Hajisoltani R, Farbib M, Sadr H, Mehrabi F, Baluchnejadmojarad T, Mehrabi S. Therapeutic potential of edaravone and flurbiprofen in valproic acid-induced autism: Targeting oxidative stress and neuroinflammation. IBRO Neurosci Rep. 2026; 21: 342-51.

BACKGROUND AND OBJECTIVES: Autism Spectrum Disorder (ASD) presents significant therapeutic challenges, with growing evidence implicating neuroinflammation and oxidative stress in its pathophysiology. This study aimed to investigate the protective effects of edaravone and flurbiprofen, administered alone and in combination, in a valproic acid (VPA)-induced autism model in male rats. MATERIALS AND METHODS: An autism model was established through prenatal VPA exposure. Social behavior, anxiety, and memory were assessed using a battery of standardized behavioral tests. Pro-inflammatory and oxidative stress markers were evaluated via ELISA, and hippocampal neuronal density was assessed by Nissl staining. RESULTS: All treatment groups (edaravone, flurbiprofen, and their combination) showed significant improvements in autism-related behaviors, including enhanced social interaction, reduced memory deficits, and decreased anxiety. The treatments lowered pro-inflammatory cytokines (IL-6, TNF-α) and MDA levels, while restoring the activity of antioxidant enzymes GPX and SOD. CONCLUSIONS: Edaravone and flurbiprofen effectively improved autism-like behaviors by targeting shared pathways of oxidative stress and neuroinflammation. However, the combination therapy offered no additive benefits over monotherapy. This finding suggests that monotherapy with either drug is sufficient to achieve maximal therapeutic effects, a point with significant implications for future treatment strategies in ASD. SIGNIFICANCE STATEMENT: This study advances neurodevelopmental neuroscience by demonstrating that independently targeting oxidative stress or neuroinflammation achieves maximal behavioral and hippocampal rescue in an autism model, revealing a convergent therapeutic ceiling that challenges combinatorial polypharmacy and refines our understanding of how overlapping pathological pathways govern neural circuit dysfunction and functional recovery.

Lien vers le texte intégral (Open Access ou abonnement)

5. Kilmer M, Patton S, Randolph D, Manges K, Milone G, Pamungkas L. Primary care providers’ beliefs, intentions, behaviors, and capabilities: Implications for workforce readiness in pediatric autism care. J Am Assoc Nurse Pract. 2026.

BACKGROUND: Workforce shortages, training gaps, and role ambiguity among primary care providers (PCPs) contribute to delays in pediatric autism identification and management. PURPOSE: This study’s aim is to examine PCPs’ beliefs, intentions, behaviors, and capabilities related to pediatric autism care and identify factors associated with workforce readiness. METHODOLOGY: Guided by the Reasoned Action Approach, this cross-sectional survey included 107 PCPs nationwide. The questionnaire assessed Beliefs, Intentions, Behaviors, and Capabilities. Multivariate analyses tested differences by provider type, years in practice, and practice setting, with follow-up analysis of variance used to identify contributing items. Linear discriminant analysis and path modeling examined interdomain relationships. RESULTS: Provider type was the most consistent predictor of differences across domains. Significant multivariate effects were observed for Beliefs and Capabilities, whereas fewer differences emerged for Intentions and Behaviors. Nurse practitioners demonstrated greater alignment across domains compared with physicians and physician assistants but reported lower diagnostic confidence. Practice setting influenced referral behaviors, with rural providers reporting greater reliance on specialists. Path analysis indicated that Capabilities mediated the relationship between Beliefs and both Intentions and Behaviors. CONCLUSIONS: Primary care providers readiness for autism care varied primarily by provider type, with additional influence from practice setting and limited effects of years in practice. Although nurse practitioners demonstrated a consistent readiness profile, gaps in diagnostic confidence may constrain independent management. IMPLICATIONS: Targeted training, decision-support tools, and tiered diagnostic models may strengthen PCP capacity and support earlier autism identification and management in primary care.

Lien vers le texte intégral (Open Access ou abonnement)

6. Liu X, Lu X, Lei S, Shang J, Wang W, Liu Y, Liu N, Xiang Y, Liu X. Rapid glycomic analysis of serum EVs reveals altered N-glycosylation patterns in ASD. Anal Bioanal Chem. 2026.

Objective laboratory diagnostics for autism spectrum disorder (ASD) are lacking, necessitating rapid clinical screening tools. Because serum extracellular vesicle (EV) N-glycosylation captures critical neurodevelopmental signatures, we developed a fast, biologically interpretable diagnostic strategy. EVs from ASD patients with language impairment and neurotypical controls were isolated using a rapid extra-polyethylene glycol precipitation/filtration (EPF) workflow, benchmarked against ultracentrifugation. Following MALDI-TOF/MS profiling, machine learning was re-evaluated using repeated nested cross-validation to reduce optimistic bias and potential information leakage. Among five classifiers, Random Forest (RF) showed the best overall balance across discrimination, calibration, and classification metrics. RF-based SHAP analysis provided transparent interpretation, highlighting key discriminative glycans, including H4N3S1F1, H5N5S1F1, and H3N5F1. To elucidate molecular mechanisms, we integrated public EV transcriptomic data. This revealed significant dysregulation of N-glycosylation machinery genes (e.g., MAN1A1, NEU1, OSTC, RPN2), whose expression directionally aligned with observed glycan shifts in synaptic pathways. Collectively, this rapid serum EV N-glycomic workflow, combined with leakage-controlled RF-based interpretation, provides a promising foundation for non-invasive ASD biomarker discovery and future multicenter validation.

Lien vers le texte intégral (Open Access ou abonnement)

7. Pan J, Zhang H, Zhai Y, Zhang J, Deng H. Exploring atypical spatial-functional coupling in adolescent autism spectrum disorder: insights from neurodevelopment and transcriptomic architecture. Front Neurosci. 2026; 20: 1780430.

Autism Spectrum Disorder (ASD) is associated with atypical large-scale brain network organization, yet how spatial-functional dependencies relate to clinical features and molecular reference maps remains incompletely understood. To quantify spatial functional heterogeneity (Sill) and coherence persistence (Range), we analyzed resting-state fMRI data from 162 ASD and 175 TD adolescents, all aged 12-18. Compared with TD, adolescents with ASD exhibited significantly increased Sill within higher-order association networks, including the left Language and right Posterior Multimodal networks, whereas no group differences in Range survived multiple-comparison correction. Within the ASD group, elevated Sill was selectively associated with greater social-affective symptom severity but not restricted and repetitive behaviors. To explore potential biological correlates, we integrated cortical gene expression reference data and identified transcriptomic patterns associated with regional Sill differences. These genes showed enrichment for synaptic signaling, mitochondrial processes, and glial-related functions, highlighting multiscale correspondence between spatial-functional organization and molecular reference maps. Together, these results demonstrate statistical associations among altered spatial-functional properties, clinical severity, and transcriptomic profiles related to synaptic signaling, mitochondrial processes, and glial-related functions in ASD, providing a complementary spatial perspective on large-scale functional organization.

Lien vers le texte intégral (Open Access ou abonnement)

8. Pecukonis M, Geffin A. Socioeconomic Status and Quantity of Shared Book Reading, But Not Screen Time, are Positively Associated With Language Skills in a Nationally Representative Sample of Autistic Children. Autism Res. 2026: e70321.

Studies have shown that both distal and proximal environmental factors influence language development in neurotypical children. More specifically, higher socioeconomic status (SES) and more frequent shared book reading have been linked to higher language skills in neurotypical children, while higher screen time has been linked to lower language skills. However, little research has focused on understanding how the environment shapes language development in autistic children, who often experience challenges using and/or understanding spoken language. The present study used data from the 2018-2022 United States National Survey of Children’s Health to investigate the associations between SES, quantity of shared book reading and screen time experienced at home, and expressive and receptive language skills in a nationally representative sample of two to five year old autistic children. Structural equation modeling revealed that autistic children from higher SES households, and who experienced shared book reading more frequently at home, had higher expressive and receptive language skills. Shared book reading also mediated the association between SES and children’s language skills. Contrary to what has been found in previous studies of neurotypical children, screen time was not associated with autistic children’s expressive or receptive language skills. In all, findings suggest that distal (SES) and proximal (shared book reading) environmental factors play an important role in autistic children’s language development. Findings have important implications for clinical practice, as interventions that promote frequent shared book reading may support autistic children’s language development.

Lien vers le texte intégral (Open Access ou abonnement)

9. Rashid N, Peckham A, Yakkala VK, Cosand L, Rajagopalan K. Characteristics of adult individuals with Rett syndrome treated or untreated with trofinetide in the United States. Front Neurol. 2026; 17: 1856312.

BACKGROUND: Rett syndrome (RTT) is a rare, progressive MECP2-related neurodevelopmental disorder with substantial lifelong morbidity that persists into adulthood. Although survival has improved, adults often experience evolving multisystem complications and fragmented transition care. Trofinetide (TROF) is approved for ages ≥2 years old, yet data on adults with RTT in the real-world setting remains limited; this study evaluated demographics and characteristics of adults >20 years of age who are treated vs. untreated with TROF. METHODS: A retrospective analysis of individuals with RTT diagnosis (ICD-10-CM: F84.2) from a linked medical claims and specialty pharmacy database from 01/01/2021 to 09/30/2024 was conducted. RTT individuals were categorized into two groups based on treatment status: treated group (index date: 1st TROF prescription (RX) 04/01/2023 to 03/31/2024) and untreated group (index date: assigned date using a risk set sampling method). Individuals who were ≤20 years of age at index date or had diagnosis for brain trauma or cerebrovascular disease prior to RTT diagnosis were excluded. RTT individuals were required to have continuous enrollment for ≥6 months pre-index and post-index. Demographics and clinical characteristics were assessed during pre-index among the treated and untreated groups. Continuous variables were summarized as means and SD; categorical variables as counts and percentages. RESULTS: There were 1,820 adult RTT individuals (>20 year old) eligible for the analysis: treated group (n = 290 [15.9%]) and untreated group (n = 1,530 [84.1%]). Mean age (SD) at index date was 30.9 (9.9) years (treated group) vs. 33.5 (10.0) years (untreated); and 5.5% vs. 6.7% were males in the treated vs. untreated groups, respectively. Treated group had higher rates of differential diagnoses, but similar rates of baseline comorbidities among both. Treated group also had higher rates of RTT related clinical features vs. untreated group. CONCLUSION: In this real-world analysis, only 16% of eligible adult RTT individuals were initiated on TROF, while 84% remain untreated. There is a high unmet need for adult individuals with RTT to initiate treatment with TROF. Treated group had higher rates of RTT related clinical features and differential diagnoses; however, the observation that TROF is being used in adults with greater to similar baseline complexity compared to untreated may provide reassurance for prescribers to consider TROF in adults who were untreated.

Lien vers le texte intégral (Open Access ou abonnement)

10. Romanos-Sirakis E. Seeing the Full Spectrum: A Personal Reflection on Autism. Ann Fam Med. 2026; 24(4): 364-5.

The struggles faced by individuals and families affected by autism can vary tremendously. This, in turn, means that each family’s reality is quite different. The extreme differences in the experiences faced can lead to different needs and different perceptions of autism. In this essay, a physician shares her experiences as a pediatric physician and a mother of a child with autism, and describes a dichotomy of experiences noted within the autism community. As medical professionals, we must understand the differences in perceptions and experiences to properly care for patients with autism and their families.Abstract also available in عربي (Arabic); Deutsch (German); Español (Spanish); Francais (French); हिन्दी (Hindi); Indonesia (Indonesian); (Chinese); (Japanese); Portugues (Portuguese).

Lien vers le texte intégral (Open Access ou abonnement)

11. Salomone E, de Leonardis G, Ferrante C, Startari F, Gila L, Scattoni ML. Scaling Up Caregiver Support for Autism in Public Health Care: Adoption, Implementation Pace and Sustainability of a Train-the-Trainer Model for the WHO Caregiver Skills Training in Italy. Autism. 2026: 13623613261469910.

Train-the-Trainer (TTT) models can scale parent-mediated interventions (PMIs) in public health care, yet evidence on large-scale implementation remains limited. This study reports the implementation of an adapted TTT model for the WHO Caregiver Skills Training (CST) within the Italian National Health Service, evaluating adoption, acceptability, feasibility and sustainability. Twenty-eight clinicians from 10 Regions undertook the four-phase TTT. A mixed-methods design combined surveys with focus groups, analyzed using the Consolidated Framework for Implementation Research (CFIR). Twenty trainees (71%) qualified as Master Trainers (MTs), and 17 of these (85%) trained facilitators; 60% of MTs achieved full fidelity within the expected timeline. Satisfaction ratings were high, whereas feasibility ratings were mid-range. Implementation inhibitors clustered in the CFIR Inner setting (available resources, leadership engagement), Outer setting (lack of formal recognition), and Process (training model’s intensity structure) domains. Accelerators included Characteristics of individuals (motivation, readiness to change), Intervention characteristics (responsiveness to families’ needs, accessibility) and Inner Setting features (integration within existing care pathways). Overall, the adapted TTT was acceptable and supported early adoption and fidelity; however, lower feasibility for select capacity-building components and multilevel barriers indicate that sustained integration will require stronger organizational support and policy endorsement. Proposed adaptations offer an adaptable framework for global efforts to expand caregiver support within public health systems.Lay AbstractFamilies of autistic children often benefit from programs that teach caregivers practical skills to support their child’s development. However, these programs can be hard to implement in public health systems due to a shortage of trained professionals and because interventions may not align with existing service structures. To overcome these challenges, the Italian National Institute of Health has implemented the World Health Organization’s Caregiver Skills Training (CST) program through initiatives funded by the National Autism Fund. To ensure the program could be effectively adopted within the Italian public health context, we applied a « Train-the-Trainer » model. This approach trains experienced clinicians to become « Master Trainers, » who in turn train other professionals to deliver the program to families, supporting scalability and sustainability within the system. In our study, 28 clinicians from 10 regions of Italy took part in a four-step training process. Most (20 trainees) completed the program successfully, and many (17 trainees) went on to train facilitators. Surveys and focus groups examined how well the program was taken up, how acceptable it felt, how doable it was in routine services, and what might help it last. Trainers reported high satisfaction with the program, although some faced difficulties such as limited institutional support. Factors that helped implementation included strong motivation and the program’s perceived usefulness for families. Barriers included organizational challenges and the need for formal recognition of the program. Overall, this approach looks promising for bringing CST into public services, but durable, wider use will require stronger organizational support and policy endorsement. The proposed adjustments offer a practical path to expand caregiver support and offer methods that can be transferred to public health systems in other countries.

Lien vers le texte intégral (Open Access ou abonnement)

12. Su C, Liu Z, Zhang S, Yi A, Xiao Y. Altered white matter network topology is associated with more profound language impairment in children with autism spectrum disorder. Neuroimage Clin. 2026; 51: 104042.

Children with autism spectrum disorder (ASD) often exhibit language deficits, yet the influence of varying language deficits on global white matter networks remains underexplored. In this study, diffusion-weighted imaging data were collected from a cohort of Chinese children with ASD (n = 67, 54 boys) and typically developing (TD) children (n = 36, 23 boys) aged 20 to 93 months. K-means clustering divided the ASD sample into higher language (ASD-HL, n = 29) and lower language (ASD-LL, n = 38) subgroups. We examined topological properties of brain networks and compared global and nodal characteristics across ASD subgroups and TD children. Relationships between autism symptom severity, language abilities, and brain network characteristics were also assessed. The ASD-LL subgroup showed reduced global efficiency (Eg), fewer hubs, decreased fiber connectivity, and fewer inter-hemispheric connections, compared to both ASD-HL and TD groups. Decreased Eg was associated with more severe autism symptoms in the ASD-LL subgroup, but not in the ASD-HL subgroup. Moderation analysis revealed that language ability significantly moderated the link between symptom severity and Eg: higher symptom severity was significantly associated with lower Eg in children with lower language ability, but not in those with higher language ability. These findings suggest that language deficits contribute to alterations in white matter networks and may modulate the impact of autism symptoms on brain structural inefficiency in children with ASD. This study underscores the importance of targeting language skills in interventions and using language ability as a key stratification factor in ASD research.

Lien vers le texte intégral (Open Access ou abonnement)

13. Turabee ZB, Brown DJ, Mahmud M, Oikonomou A, Rahman MA, Burton A, Shopland N, Clarke D, Gray F. ASC-emotion: A privacy-aware dataset for analysing emotional dysregulation and engagement in children with Autism(). MethodsX. 2026; 17: 104073.

Predicting emotional dysregulation events in children with Autism is essential for timely mitigation of triggering events and prevention of further escalation of the situation. However, there is a scarcity of accessible and standarised datasets for use in AI-based research associated with challenging behaviours in children with ASC. To address this gap, we have curated a novel privacy-preserved dataset as part of an Erasmus+ funded project (AI-TOP-2020-1-UK01-KA201-079,167). The dataset was validated using three machine learning architectures which gave an accuracy of 96% in detecting the affective states related to learning in children with Autism. The key contributions of this study are: 1. Development of an Autism meltdown dataset exemplifies methodological rigor, advances an urgent clinical challenge through early detection and intervention, and enables broad impact by promoting reproducibility, benchmarking and translational health outcomes. 2. Implementation of privacy preserving measures addresses ethical concerns regarding the use of video data with this vulnerable population as part of the machine learning pipeline. 3. High levels of accuracy are demonstrated via empirical validation of the dataset through three machine learning models (BiLSTM, Graphical Neural Network – EdgeConv, PointCNN+LSTM) for detecting affective states related to learning and physiological arousal in children with Autism.

Lien vers le texte intégral (Open Access ou abonnement)

14. Yamashita M, Shou Q, Hamatani S, Matsuzaki H, Fujieda M, Hirano Y, Kagitani-Shimono K, Okazawa H, Mizuno Y. Transdiagnostic monoamine-based subtyping for attention-deficit/hyperactivity disorder and autism spectrum disorder via unsupervised machine learning. J Neural Transm (Vienna). 2026.

Neuroimaging and molecular studies have examined the etiology of attention-deficit/hyperactivity disorder (ADHD) and autism spectrum disorder (ASD). However, their findings remain inconsistent because of within-disorder heterogeneity and cross-disorder phenotypic overlap. We sought to identify monoamine-based subtypes across ADHD and ASD and clarify their brain structural characteristics. In 83 children with ADHD and/or ASD, we applied unsupervised machine learning (NbClust with K-means) to identify neurodevelopmental disorder (NDD) phenotypes using urinary monoamine metabolite (MM) profiles. Behavioral symptoms, cognitive performance, cortical surface area, and gray matter volume (GMV) were evaluated for each NDD phenotype and for 83 typically developing (TD) children as controls. Clustering identified two urinary MM-defined NDD phenotypes: NDD-A (n = 18, including 5 ADHD, 2 ASD, and 11 ADHD + ASD cases), characterized by high levels of 4-hydroxy-3-methoxyphenylglycol, 5-hydroxyindoleacetic acid, and homovanillic acid, and NDD-B (n = 65, including 16 ADHD, 19 ASD, and 30 ADHD + ASD cases), characterized by low levels of these metabolites. Urinary 4-hydroxy-3-methoxyphenylglycol levels correlated positively with social communication difficulties in NDD-A. NDD-B showed significantly lower cognitive control, cognitive flexibility, and inhibitory control than TD. Structurally, compared with TD, NDD-A showed significant surface area enlargement in the isthmus cingulate gyrus, whereas NDD-B exhibited significant GMV reductions primarily in fronto-opercular/orbitofrontal regions, with additional reductions in the superior parietal lobule and supramarginal gyrus. These findings suggest that urinary MM-defined NDD-A and NDD-B phenotypes are associated with distinct cognitive and brain structural characteristics. Such phenotype specificity may provide a novel framework for understanding within-disorder heterogeneity and cross-disorder phenotypic overlaps.

Lien vers le texte intégral (Open Access ou abonnement)