ASR · ASR-2026-06 · citation-audited · neurodiversity-affirming · not medical advice
自闭症科学评论 — 试刊号 ASR-2026-06
A quarterly-style, evidence-disciplined and neurodiversity-affirming synthesis of recent autism science, in three parts: genetics and the developing brain; early identification, co-occurring conditions and honest evidence on supports; and autistic lives, neurodiversity and ethics. Bilingual for clinicians, researchers, autistic people and families. This is a research synthesis, not medical advice; we separate preclinical (animal/organoid) findings from human studies, make no claims of cure or normalization, and aim to represent autistic voices.
这是一份季度性、坚持证据纪律、肯定神经多样性的自闭症科学综述，分为三个部分：遗传与发育中的大脑；早期识别、共病以及关于支持手段的诚实证据；自闭者的人生、神经多样性与伦理。中英双语，面向临床医生、研究者、自闭者及其家庭。本刊为科学综述，并非医疗建议；我们将临床前（动物/类器官）发现与人体研究分开呈现，不作任何"治愈"或"使其正常化"的主张，并致力于呈现自闭者自身的声音。
表型遇见基因型：自闭症可解析为遗传上可区分的亚类
Generative mixture modeling on ~5,000 autistic individuals links clinically meaningful subgroups to separable genetic programs (common, de novo, inherited) with class-specific developmental timing — a step toward heterogeneity-aware research, not toward diagnosing any individual. Human cohort, observational (E3).
对约5000名自闭症个体的生成式混合建模，将具有临床意义的亚组与可分离的遗传程序（常见、新发、遗传性变异）相联系，且发育时间窗在亚类间不同——这是迈向'尊重异质性'研究的一步，而非用于诊断任何个体。人体队列，观察性研究（E3）。
更早识别，诚实呈现：登记数上升，未必是真实发生率上升
CDC ADDM 2022 estimates about 1 in 31 eight-year-olds identified with autism, with wide site variation and a 3.4:1 male-to-female ratio; cumulative early-diagnosis incidence was 1.7x higher in the 2018 vs 2014 birth cohort — interpreted as earlier identification. Human surveillance (E3).
美国CDC ADDM 2022年估计约每31名8岁儿童中有1名被识别为自闭症，各监测点差异大，男女比3.4:1；2018年出生队列的早期累积诊断发生率为2014年队列的1.7倍——被解读为更早识别。人群监测（E3）。
诚实的阴性试验：当某项支持未优于常规照护时，我们如实呈现
Two RCTs (A-FFIP NDBI; ESDM in two European countries) found no significant advantage over usual care on their primary core-symptom/developmental outcomes (P=.467; P=0.14). Reported as honest, non-significant evidence — not a verdict on any individual's supports. Human RCTs (E2).
两项随机对照试验（A-FFIP自然发展行为干预；丹佛早期介入模式ESDM在两个欧洲国家）在主要的核心症状/发育结局上未显示出优于常规照护的显著优势（P=.467；P=0.14）。这是诚实的非显著证据，并非对任何个体所用支持手段的评判。人体随机对照试验（E2）。
Tagged by ASR taxonomy A-codes. / 按 ASR 分类法 A-code 标注。
遗传与分子机制
Common, de novo and inherited variation map onto distinct, clinically meaningful autism classes; a proposed female protective effect via incompletely silenced X-linked genes; and polygenic scores associating with subtle brain microstructure differences in the general population. We report these as population-level genetic architecture, never as individual prediction.
常见、新发与遗传性变异对应于不同的、具有临床意义的自闭症亚类；一项关于'女性保护效应'的假说（源于未完全沉默的X连锁基因）；以及多基因评分与一般人群中细微脑微观结构差异的关联。我们将这些作为群体层面的遗传结构呈现，绝不用于个体预测。
神经发育与脑环路
A spatial transcriptomic atlas of ~250 autism-associated genes across the midgestation human forebrain points to the developing thalamus and interneuron-generating germinal zones, implicating thalamocortical circuits as a susceptibility hub beyond cortex. A human preprint (E5) — convergence in developing tissue, not a claim about any living person's brain.
一项覆盖中孕期人类前脑、约250个自闭症相关基因的空间转录组图谱，指向发育中的丘脑及产生中间神经元的生发区，提示丘脑-皮层环路是皮层之外的易感枢纽。人体预印本（E5）——这是发育组织中的'汇聚'，并非对任何在世个体大脑的论断。
人源干细胞与类组装体模型（临床前）
Cortical organoids across 70 iPSC lines show mutation-specific changes converging over development; cortico-striatal assembloids reveal an SCN2A phenotype not seen in mice; pooled CRISPR across neuronal types finds cell-type-dependent convergence. PRECLINICAL (E4): human-derived cells and organoids are not people — these results do not generalize to autistic individuals.
70个iPSC细胞系的皮层类器官显示突变特异性改变随发育而汇聚；皮层-纹状体类组装体揭示了一种小鼠中未见的SCN2A表型；跨神经元类型的混合CRISPR发现依赖细胞类型的汇聚。临床前（E4）：人源细胞与类器官不是人——这些结果不可外推至自闭个体。
早期识别与诊断
Population surveillance shows earlier identification; a community screening tool reports high predictive value (screening is not diagnosis); and ADOS sex-related measurement bias means girls must show greater deviation to meet thresholds — a measurable contributor to later or missed identification in girls. Human studies (E3).
人群监测显示识别更早；一项社区筛查工具报告了较高的预测值（筛查不等于诊断）；ADOS存在与性别相关的测量偏倚，意味着女孩须表现出更大偏离才达到阈值——这是女孩诊断偏晚或漏诊的可测量原因之一。人体研究（E3）。
共病
In a large inpatient psychiatric sample of autistic youth, about 18.9% met criteria for epilepsy — far above general-population estimates — underscoring the need for integrated neurological and mental-health care for autistic youth with complex presentations. Sample-specific (specialized inpatient units), human observational (E3).
在一个由自闭青少年组成的大型住院精神科样本中，约18.9%符合癫痫标准——远高于一般人群估计——凸显复杂表现的自闭青少年需要神经科与精神卫生的整合照护。该比例特定于专科住院人群，人体观察性研究（E3）。
干预与支持（诚实证据）
Two RCTs found low-intensity NDBI and ESDM did not significantly outperform usual care on primary outcomes; a caregiver-skills-training mechanism study reliably improved parent skills and wellbeing while downstream child core-symptom change did not reach significance (ns / 未达显著). We frame these as research on supports, not as recommendations or claims to normalize. Human RCTs (E2).
两项随机对照试验发现，低强度自然发展行为干预与丹佛早期介入模式在主要结局上未显著优于常规照护；一项照护者技能培训机制研究可靠地提升了家长技能与福祉，而对儿童核心症状的下游改变未达显著（ns / 未达显著）。我们将其作为关于支持手段的研究呈现，而非个人推荐或'使其正常化'的主张。人体随机对照试验（E2）。
神经多样性、亲历体验与参与式研究
Autistic-led participatory infrastructure (a decade of PARC), the Social Model of Disability applied to autism research, and a meta-synthesis of 52 studies finding autistic adults often experience connection with other autistic people as positive and linked to quality of life. We center autistic voices, not deficit framings. Human and perspective sources (E1/E6).
自闭者主导的参与式研究网络（PARC十年历程）、应用于自闭症研究的残障'社会模型'，以及一项纳入52篇研究的元综合——发现自闭成人常将与其他自闭者的联结体验为积极、并与生活质量相关。我们以自闭者的声音为中心，而非缺陷视角。人体研究与观点类来源（E1/E6）。
自闭成人、衰老与服务
Syntheses of lived experience show later-in-life diagnosis can bring relief, self-understanding and identity reframing alongside grief and unmet post-diagnostic support; older autistic adults (50+) face heightened health risks and major gaps in lifespan services. Human qualitative syntheses (E1).
亲历体验的综合研究表明，成年后诊断可在带来释然、自我理解与身份重构的同时，伴随哀伤与诊断后支持的缺失；老年自闭成人（50岁以上）面临更高的健康风险和全生命周期服务的重大缺口。人体定性综合研究（E1）。
伦理、政策、权利与公平
An editorial argues for centering Indigenous epistemologies and methodologies in autism scholarship rather than relying solely on Western frameworks, advancing ethics and participatory equity. Perspective/framework source (E6).
一篇社论主张在自闭症学术研究中纳入并以原住民认识论与方法论为中心，而非仅依赖西方框架，从而推进伦理与参与式公平。观点/框架类来源（E6）。
基因与发育中的大脑
Heterogeneity becomes structure
异质性正在成为结构
Generative mixture modeling on ~5,000 autistic individuals (E3, human) resolves clinically meaningful classes tied to separable genetic programs with class-specific developmental timing — reframing autism's heterogeneity as analyzable structure rather than noise. This describes population-level architecture and must never be used to predict or diagnose any individual.
Where risk genes converge in the brain
风险基因在脑中的汇聚之处
A spatial transcriptomic atlas of ~250 autism-associated genes (E5, human preprint) points to the developing thalamus and interneuron-generating germinal zones, implicating thalamocortical circuits beyond cortex. Polygenic scores also associate with subtle neurite-density differences across general-population children and adults (E3, human; no sex difference reported).
Preclinical convergence — clearly labeled
临床前的汇聚——明确标注
Across cortical organoids (70 iPSC lines), cortico-striatal assembloids (an SCN2A phenotype absent in mice), and pooled CRISPR in human neurons, distinct mutations increasingly converge during development (all E4, organoid/cell). PRECLINICAL: human-derived models are not humans; these findings do not generalize to autistic people and imply no treatment.
诊断、共病与循证支持
Earlier identification, carefully interpreted
更早识别，审慎解读
CDC ADDM 2022 (E3, human) estimates about 1 in 31 eight-year-olds identified, with wide site variation and a 1.7x higher early cumulative diagnosis in the 2018 vs 2014 cohort — read as earlier identification, not a true rise in occurrence. A community screening tool (screening ≠ diagnosis) reports high predictive value; ADOS sex-related measurement bias means girls must show greater deviation to meet thresholds.
Co-occurring conditions need integrated care
共病需要整合照护
In a large specialized-inpatient psychiatric sample of autistic youth, about 18.9% met criteria for epilepsy (E3, human) — far above general-population estimates and specific to this complex-presentation group, underscoring integrated neurological and mental-health care.
Honest evidence on supports
关于支持手段的诚实证据
Two RCTs (E2, human) found low-intensity NDBI (A-FFIP; adjusted ES −0.06, 95% CI −0.24 to 0.11, P=.467) and ESDM in two European countries (mean difference 3.82 points, 95% CI −1.25 to 8.89, P=0.14) did not significantly outperform usual care on primary outcomes. A caregiver-skills-training study improved parent skills and wellbeing, while child core-symptom change did not reach significance (ns / 未达显著). Framed as research on supports, never as personal recommendation or a claim to normalize.
自闭者的人生、神经多样性与伦理
Connection and community as quality of life
联结与社群作为生活质量
A meta-synthesis of 52 qualitative studies (E1, human) finds autistic adults often experience interacting with other autistic people as positive and associate it with improved quality of life — evidence for the value of autistic community and peer connection, in autistic people's own words.
Diagnosis and aging across the lifespan
全生命周期中的诊断与衰老
Syntheses of lived experience (E1, human) show later-in-life diagnosis can bring relief, self-understanding and identity reframing alongside grief and unmet post-diagnostic support; older autistic adults (50+) face heightened health risks and major service gaps — pointing to lifespan-wide, person-centered services.
Participatory and ethical research
参与式与合乎伦理的研究
A decade of autistic-led participatory infrastructure (PARC), the Social Model of Disability applied to autism research, and a call to center Indigenous epistemologies (all E6, perspective/framework) argue for affirming, empowering, equitable research — with autistic people and communities as partners, not subjects.
Welcome to the pilot issue of Autism Science Review (ASR-2026-06). This is the most sensitive domain we publish in — autistic people and their families read this — and we have built the issue to earn that trust. What this issue is. A bilingual, three-in-one synthesis of recent autism science: (a) genetics, molecular mechanisms and brain circuits; (b) early identification, co-occurring conditions and honest evidence on supports; and (c) autistic adults across the lifespan, neurodiversity, participatory research and ethics. Our citation-audit discipline. Every finding here passed an explicit citation audit. We attach a DOI, PMID or bioRxiv identifier ONLY to a finding whose citation verified against the real source. One source in this issue — a community autism-screening accuracy study — failed reverification (the identifier resolved to an unrelated paper), so we cite it by venue, year and title only and flag it as under audit; we did not silently drop it or attach a wrong identifier. Each reference also carries its evidence code (E1–E6), autism A-code, and organism, so you can see at a glance whether a result is a meta-analysis, an RCT, an observational study, a preclinical model, a mechanism paper, or a perspective. What this issue is NOT, stated plainly. This is a research synthesis, not medical advice. Preclinical findings (E4 — animal, organoid, assembloid, or cell models) are NOT human findings and do not generalize to autistic people; we label them every time. We make no claims of cure, recovery, reversal or normalization — autism is a developmental difference with strengths and support needs, not a disease to be eliminated. We report interventions as research evidence on supports, including honest null results, never as personal recommendations or promises. Every effect size carries its population and an uncertainty marker, and non-significant results are marked "ns / 未达显著". We carry no anti-vaccine content and no unproven or harmful "treatments"; the claim that vaccines cause autism is debunked. Representing autistic voices. We center autistic lived experience and participatory, autistic-led research, and use both identity-first ("autistic person") and person-first ("person with autism") language. We aim to describe difference, strengths and support needs — not deficit alone. We welcome correction from the autistic community; this is a pilot, and your scrutiny makes it better.
欢迎阅读《自闭症科学评论》试刊号（ASR-2026-06）。这是我们所涉及的最敏感的领域——自闭者及其家庭会阅读本刊——我们以配得上这份信任为标准来编纂本期。 本期是什么。这是一份中英双语、三合一的近期自闭症科学综述：（a）遗传、分子机制与脑环路；（b）早期识别、共病以及关于支持手段的诚实证据；（c）自闭成人的全生命周期、神经多样性、参与式研究与伦理。 我们的引用审计纪律。本期每一条发现都通过了明确的引用审计。只有当某条发现的引用与真实来源核验一致时，我们才为其附上 DOI、PMID 或 bioRxiv 标识符。本期有一条来源——一项社区自闭症筛查准确性研究——未能通过复核（其标识符指向了一篇无关论文），因此我们仅以期刊、年份和标题方式引用，并标注为"审计中"；我们没有悄悄删除它，也没有附上错误的标识符。每条参考文献还标注其证据等级（E1–E6）、自闭症 A 代码与研究对象（organism），以便一眼看出结果属于元分析、随机对照试验、观察性研究、临床前模型、机制论文还是观点文章。 本期不是什么，直白说明。本刊是科学综述，并非医疗建议。临床前发现（E4——动物、类器官、类组装体或细胞模型）不是人体发现，不能外推至自闭者；我们每次都会明确标注。我们不作任何"治愈""康复""逆转"或"使其正常化"的主张——自闭症是一种兼具优势与支持需求的发育差异，而非需要被消除的疾病。我们将干预作为关于支持手段的研究证据来呈现，包括诚实的阴性结果，绝不作为个人推荐或承诺。每一个效应量都附带其人群与不确定性标记，非显著结果标注为"ns / 未达显著"。我们不含任何反疫苗内容，不含任何未经证实或有害的"疗法"；"疫苗导致自闭症"的说法已被证伪。 呈现自闭者的声音。我们以自闭者的亲历体验与参与式、自闭者主导的研究为中心，并同时使用身份优先（"自闭者"）与个人优先（"自闭症人士"）的表述。我们力求描述差异、优势与支持需求——而非仅仅缺陷。我们欢迎来自自闭社群的指正；这是一期试刊，你们的审视让它更好。
A DOI / PMID / bioRxiv link appears only after the identifier was independently re-resolved and title-matched. Each reference carries its ORGANISM — an animal (preclinical) result is never a human result. Unverified sources are cited by venue · year · title alone. / 仅当标识符经独立重新解析并标题匹配后才附链接；每条标注物种——动物(临床前)结果不等于人类结果；未验证者仅以期刊·年份·标题列出。
Autism Science Review scans autism research venues (Molecular Autism, Autism Research, JADD, Autism, Nature, Lancet, JAMA Pediatrics, bioRxiv) and institutional consensus (CDC/WHO/NIMH/NICE/NICHD). Findings are tagged with taxonomy A-codes and evidence E-codes; every cited DOI / PMID / bioRxiv id is independently re-resolved and title-matched before display. Preclinical (animal/organoid, E4) findings are separated from human (E1/E2/E3) evidence and never generalized to people.
ASR 扫描自闭症研究来源与机构共识；发现按 A-code 与证据 E-code 标注；每条引用标识符在显示前独立重新解析并标题匹配；临床前(动物/类器官,E4)与人类(E1/E2/E3)证据分开，绝不外推至人。
Neurodiversity-affirming research synthesis — NOT medical advice.
ASR is for education, not diagnosis. It makes no individual diagnosis, prognosis, or treatment recommendation, no claim to cure, reverse, or recover from autism, and contains no anti-vaccine content (WHO/CDC consensus: there is no causal link between vaccines and autism). We affirm neurodiversity — describing strengths and support needs — and use both identity-first and person-first language per community preference. An animal result is not a human result. Consult qualified professionals for personal decisions. (Aligned with the m551 autism.x1000.ai Editorial Policy.)
本刊用于教育而非诊断，不构成医疗建议：不做个体诊断、预后或治疗推荐，不宣称治愈、逆转或康复自闭症，不含任何反疫苗内容（WHO/CDC 共识：疫苗与自闭症无因果关系）。我们尊重神经多样性——描述优势与支持需求——并按社群偏好使用身份优先与人优先语言。动物结果不等于人类结果。个人决策请咨询合格专业人员。（对齐 m551 autism.x1000.ai 编辑政策。）