An always-updating digest. Research subagents read new studies, trials, and reporting around the clock and rewrite them in clear, honest language — with sources and caveats attached.
Artificial intelligence tools are being studied as potential supports for autism identification processes, particularly in contexts with limited specialist access. While some systems demonstrate moderate accuracy in research settings, significant questions remain about real-world reliability, bias mitigation, and the need for human oversight in clinical decision-making.
Multiple studies indicate that variations in eye contact patterns in infants as young as 2-6 months may correlate with later autism identification. A [2024 study in Nature](https://www.nature.com/articles/s41598-024-55643-z) (n=112 infants) found that 2-minute eye-tracking protocols showed promise for early identification, while [UCSD research](https://neurosciences.ucsd.edu/centers-programs/autism/early/eye-tracking-research.html) explores how early support might begin when developmental differences are noted.
Recent studies highlight health inequities faced by older autistic adults, particularly those without formal diagnoses. UK research suggests 89-97% of autistic adults over 40 may be undiagnosed, potentially affecting their access to appropriate healthcare and support. While biological aging patterns may differ in autistic individuals, researchers emphasize the need to consider systemic barriers and individual variability when interpreting these findings.
Recent analyses suggest autism identification data in low-resource settings may not fully represent actual support requirements due to systemic barriers, service disparities, and inconsistent data collection. Studies from Ethiopia to Brazil reveal needs for healthcare provider education and inclusive learning environments, while methodological differences complicate global comparisons.
Recent studies indicate that 50-70% of autistic individuals may also meet criteria for ADHD, with shared genetic factors and overlapping traits influencing diagnostic practices. However, adults and non-stereotypical presentations remain under-researched.
Researchers are investigating potential biological markers associated with autism, including gut microbiome patterns and eye-tracking behaviors, that might one day help identify developmental differences earlier. While these approaches show promise in controlled studies, experts emphasize they currently serve as research tools rather than clinical diagnostics.
Recent studies indicate that autism diagnosis rates between genders approach parity by early adulthood, challenging the traditional 4:1 male-to-female childhood ratio. This points to systemic barriers in recognizing autism in girls, including assessment tool limitations and societal expectations, with documented impacts on mental health and access to accommodations.
The CDC's latest report finds 3.2% of U.S. 8-year-olds were identified as autistic in 2022, up from 2.8% in 2020. The increase reflects reduced disparities in diagnosis access, particularly for historically underserved groups, though boys remain 3.4 times more likely to be identified than girls.
Artificial intelligence is being investigated for potential applications in autism identification and support systems, with studies indicating varying accuracy levels in experimental screening tools and adaptive assistance approaches. Researchers stress these technologies require thorough validation, inclusive design practices, and integration with clinical expertise to address systemic care disparities.
Emerging research suggests eye-tracking measures may help identify infants with higher likelihood of developing autistic traits as early as 2-6 months old, though current protocols remain investigational. These tools could eventually help reduce evaluation wait times while emphasizing support over diagnosis.
Between 89-97% of autistic adults over 40 in the UK may lack formal diagnoses according to a [King's College London review](https://www.kcl.ac.uk/news/up-to-90-of-middle-aged-and-older-autistic-adults-are-undiagnosed-in-the-uk-new-review-finds), creating barriers to healthcare access. Peer-reviewed studies show this population experiences disproportionate mental health challenges and physical health disparities when systems fail to recognize neurodivergence.
While global understanding of autism expands, significant disparities in identification and accommodation access remain, particularly in regions with limited healthcare infrastructure. Barriers include provider knowledge gaps, cultural stigma, and systemic inequities, leaving many autistic people without needed supports.
Growing evidence shows autism and ADHD frequently co-occur, with 50-70% of autistic people meeting ADHD criteria. Shared genetic and neural mechanisms suggest these neurotypes aren't as separate as once thought, leading to calls for more integrated assessment and support approaches.
New studies examine how objective measures like eye-tracking and microbiome analysis might enhance early autism identification alongside behavioral observation. While some approaches show potential for recognizing developmental differences before age 2.5—when supportive services are most impactful—experts emphasize these tools require further validation across diverse groups and contexts.
Autistic women and girls are often underidentified due to differences in symptom presentation, clinician biases, and diagnostic criteria historically based on male presentations. While diagnosis rates equalize by adulthood, delayed recognition can create barriers to understanding and support. Intersectional factors further complicate identification for Black women and girls.
The CDC's 2025 report reveals autism prevalence among 8-year-olds increased to 1 in 31 (3.2%) in 2022, up from 1 in 36 in 2020. This rise reflects improved screening and diagnosis, particularly for Black, Hispanic, and female children, with the male-to-female ratio decreasing from 4.2:1 to 3.8:1. Significant state-level variations persist, from 1 in 19 in California to 1 in 103 in Texas.
New AI systems analyzing speech patterns and behavioral data show promise for assisting autism identification and support needs assessment, particularly in areas with specialist shortages. While not replacements for comprehensive evaluations, these tools may help prioritize cases and adapt supports to individual communication styles. Significant validation, equity, and ethical challenges remain before widespread implementation.
Peer-reviewed research suggests that differences in eye contact and visual attention in infants aged 2-12 months may correlate with later autism identification. A [2024 Nature study](https://www.nature.com/articles/s41598-024-55643-z) demonstrated the feasibility of brief eye-tracking protocols, while systematic reviews note consistent findings across studies. Experts emphasize these markers indicate developmental differences, not diagnoses, and require further validation in diverse populations.
Studies indicate 89-97% of autistic adults over 40 in the UK lack formal diagnoses, often due to historical gaps in autism recognition. While these individuals face higher risks for certain health conditions, emerging research also highlights adaptive strategies and community support that can improve outcomes.
While global autism recognition continues to improve, diagnosis and access to accommodations remain uneven worldwide due to workforce shortages, cultural differences in understanding neurodiversity, and healthcare system limitations. Community-based approaches and carefully validated technologies may help bridge gaps when adapted locally.
Long-separated diagnostic categories of autism and ADHD now show extensive overlap, with studies indicating 40-70% of autistic individuals also meet ADHD criteria. Genetic and clinical research points to shared biological pathways while revealing unique support needs at their intersection. The term 'AuDHD' reflects community-driven recognition of this co-occurrence, though it remains absent from diagnostic manuals.
