Is There a Medical Test for Autism? Blood, Brain Scans and AI

No. In 2026 there is still no medical test for autism that can diagnose it on its own. The CDC confirms there is no blood test for autism. Diagnosis is based on a child’s development and behavior, assessed by trained clinicians. Some FDA-authorized devices and many research tests can help, but none replaces a full assessment [1][2].

Key points

  • No blood, urine, hair, or stool test can diagnose autism today.
  • Brain scans and EEGs are not routine. Doctors use them only for specific medical reasons, such as possible seizures.
  • Genetic testing can sometimes find a cause, but it cannot tell you whether your child is autistic.
  • A few FDA-authorized tools (using eye-tracking or AI) can help clinicians, but they are aids, not stand-alone tests.
  • Be cautious of companies that sell a quick autism test directly to parents, especially if a treatment is sold too.

Why there is no single medical test for autism

Autism is defined by differences in social communication and by restricted or repetitive behaviors and interests. It is a behavioral diagnosis, so clinicians diagnose it by looking at history and behavior, not a lab value [1][3]. Autism is also very varied. Many different genes and early-life factors are involved, which makes a single biological marker hard to find [5]. Learn more in what causes autism.

A large systematic review of 280 studies found 940 different possible biomarkers. Most studies were small, results often did not repeat in other studies, and the authors described a “replication crisis.” They concluded no biomarker yet has enough evidence for clinical use in autism trials [4]. A 2025 review of very early markers reached a similar view: some findings are promising, but they are preliminary and need validation before clinical use [5].

Autism blood test and urine test: is a medical test for autism on the way?

Researchers have looked for patterns in blood proteins, metabolism, immune markers, and more. Some small studies report high accuracy, but they have not yet been confirmed in large, independent groups of children.

  • Blood proteins: a 2024 study of 154 boys found a panel of 12 serum proteins that separated autistic and non-autistic boys with about 83% sensitivity and 85% specificity. The authors say it still needs testing in new groups [6]. An earlier 2021 version of this paper was retracted [7], which shows how unsettled this field still is.
  • Blood and urine chemistry: a 2018 study of 69 children reported that markers of protein damage in blood and urine could separate the groups well. The authors noted that no clinical chemistry test for autism was available, and the study was small [8].
  • Maternal antibodies: researchers are developing blood tests for antibodies in some mothers that may relate to a subset of autism. A 2026 lab study validated the test method only. Its accuracy for predicting autism still needs clinical testing [28].

Doctors may still order blood tests for good reasons, such as genetic testing, or lead and iron levels in a child who often eats non-food items (pica) [12]. These tests look for causes or health problems. They do not diagnose autism.

Brain scans: is MRI or EEG a medical test for autism?

NICE advises doctors not to routinely do medical investigations as part of an autism assessment. It suggests an EEG only if epilepsy is suspected [2]. A summary of the AAP guidance also reserves EEG for suspected seizures, because screening EEGs are often abnormal even in children without seizures [12].

Research is interesting, though. In a 2017 study of babies with an autistic older sibling, MRI scans at 6 to 12 months helped predict which of them would be diagnosed with autism at 24 months [10]. A 2018 EEG study in a similar group also reported high accuracy [11]. These were research studies in high-likelihood families. They have not become clinical tests for the general population.

Eye-tracking and AI tools: the medical test for autism that comes closest

Eye-tracking

Eye-tracking measures where a child looks while watching videos of social scenes. In a study of 475 toddlers aged 16 to 30 months at specialty clinics, it had a sensitivity of 71% and specificity of 81% compared with expert diagnosis [13]. This research led to the EarliPoint device. In March 2026 the FDA cleared the EarliPoint Assessment for children aged 16 months through 7 years, as an aid for qualified clinicians and not a stand-alone diagnosis [14].

AI-based diagnosis aid

Canvas Dx (originally the Cognoa ASD Diagnosis Aid) was authorized by the FDA in 2021. It combines a parent questionnaire, home videos, and a health care provider questionnaire [15]. In its main study of 425 children, it gave a definite result for only about 32% of children. In those children, sensitivity was 98% and specificity was 79% [15][16]. Its 2025 clearance again states it is not a stand-alone diagnostic [17].

Screening apps and other AI

  • SenseToKnow: an app that records a toddler’s responses to short movies and a game using a phone or tablet camera. Studies in 2023 and 2024 found promising accuracy in pediatric clinics and when parents used it at home [18][19]. It is a screening tool under study, not a diagnosis.
  • Retinal photos: a 2023 study reported perfect accuracy for spotting autism from eye photographs in one hospital in Korea [20]. Results that look perfect in one center need careful checking in other settings before anyone should use them.
  • Gut microbes: a 2024 study of 1,627 children found a panel of stool microbiome markers that separated autistic and non-autistic children with good accuracy [9]. This is early research, not a test you can use.

In October 2026 the FDA cleared a third aid, Neurora, which uses AI to analyse video and audio recorded during a clinician’s behavioural assessment of children aged 18 months to 12 years. It is not a stand-alone test either [29].

Genetic testing for autism: what it can and cannot tell you

Genetic testing is the medical test for autism that doctors most often recommend, but it answers a different question. It looks for a genetic cause or a related condition. It does not decide whether a child is autistic.

A summary of the AAP guidance recommends chromosomal microarray and fragile X testing for children with autism, with exome sequencing as an option if these are negative [12]. ACMG recommends exome or genome sequencing as a first- or second-tier test for children with developmental delay, intellectual disability, or birth differences [21]. In a meta-analysis of children with neurodevelopmental conditions, including autism, exome sequencing found a genetic cause in about 36%, compared with about 15 to 20% for microarray in earlier studies [22].

A genetic result can help with medical care and family planning [21]. Many autistic children have no genetic finding [22], and that does not change their diagnosis or support needs.

Hype and red flags: tests sold to parents

Some companies sell autism-related tests directly to families or through clinics. A few are early versions of real research. Others have little or no evidence.

  • Hair tests: one company launched a hair-based test in 2025 that it says can help providers rule out autism in children aged 1 month to 36 months. It is run in the company’s own certified lab. The company’s related product has an FDA “Breakthrough” designation but is listed as still in development [23]. A Breakthrough designation is not FDA clearance or approval.
  • Lab-developed tests: these are tests designed and used within a single lab. A 2025 court ruling set aside the FDA’s rule for overseeing them, and the FDA returned to its earlier rules [24]. A test being available does not mean the FDA has reviewed its accuracy.
  • Hair mineral analysis: when one hair sample was sent to six commercial labs, results differed by more than tenfold for many minerals, and the labs gave conflicting advice. Researchers called the results unreliable [25].

Red flags to watch for

  • Any medical test for autism that claims to diagnose it from a single sample of blood, urine, hair, or stool.
  • Words like “breakthrough,” “new discovery,” “quick fix,” or “miraculous,” which the FDA lists as health fraud warning signs [26].
  • Testimonials instead of published studies, or results that sound too good to be true [26].
  • A test result that leads to buying the same company’s supplements, diets, or treatments. The FDA warns that products claiming to cure or treat autism are misleading and can cause serious harm [27]. See our guide to unproven autism treatments.

What you can do

  • Start with your child’s doctor and ask for developmental screening. Our M-CHAT guide explains a common first step.
  • Ask for a referral for a full assessment if concerns continue. See autism diagnosis tests around the world for how this works in different countries.
  • Ask your doctor whether genetic testing or a hearing test makes sense for your child.
  • Before paying for any medical test for autism, ask: Is it FDA-authorized? Is it published in a peer-reviewed journal? Will it change my child’s care?

When to talk to a professional

Talk to your child’s doctor if you have concerns about your child’s communication, play, or social development, or before you pay for any medical test for autism. If a doctor suggests an MRI or EEG, ask what medical question it is meant to answer. You can also read our medical disclaimer and how we check our information.

Frequently asked questions

Can a blood test diagnose autism?

No. The CDC states there is no medical test for autism, such as a blood test, that can diagnose it. Several research teams have reported promising blood markers, but these studies are mostly small and have not been confirmed in large, independent groups of children. Your doctor may order blood tests for genetic causes or other health problems, which is different from diagnosing autism.

Can an MRI or EEG show autism?

Not in everyday practice. Research studies have found brain differences in groups of infants who were later diagnosed, but these are not clinical tests. NICE advises against routine medical investigations in autism assessments, and EEGs are generally used only when seizures are suspected. A doctor may order a scan for a specific medical reason, not to diagnose autism.

Will genetic testing tell me if my child is autistic?

No. Genetic testing looks for a genetic cause or a related condition, such as fragile X syndrome. It can help with health planning and family decisions. But autism is diagnosed by behavior and development. A normal genetic result does not rule autism out, and many autistic children have no genetic finding at all.

Is the EarliPoint eye-tracking test a real autism diagnosis?

EarliPoint is FDA-cleared, but as an aid for qualified clinicians, not a stand-alone diagnosis. It measures where a child looks while watching social videos. In its main study, it correctly identified many but not all autistic toddlers. A clinician still needs to look at your child’s full history and development before making a diagnosis.

References

  1. Centers for Disease Control and Prevention. Screening for Autism Spectrum Disorder. Last reviewed October 1, 2026. Link
  2. National Institute for Health and Care Excellence. Autism spectrum disorder in under 19s: recognition, referral and diagnosis. Clinical guideline CG128. Published 2011; recommendations dated 2011 and 2017; last reviewed June 2021. Link
  3. Hyman SL, Levy SE, Myers SM; Council on Children With Disabilities, Section on Developmental and Behavioral Pediatrics. Identification, Evaluation, and Management of Children With Autism Spectrum Disorder. Pediatrics. 2020;145(1):e20193447. Link
  4. Parellada M, Andreu-Bernabeu Á, Burdeus M, San José Cáceres A, Urbiola E, Carpenter LL, et al. In Search of Biomarkers to Guide Interventions in Autism Spectrum Disorder: A Systematic Review. Am J Psychiatry. 2023;180(1):23-40. Link
  5. Wang BM, Mills Z, Jones HF, Montgomery JM, Lee KY. Presymptomatic Biological, Structural, and Functional Diagnostic Biomarkers of Autism Spectrum Disorder. J Neurochem. 2025;169(5):e70088. Link
  6. Hewitson L, Mathews JA, Devlin M, Schutte C, Lee J, German DC. Blood biomarker discovery for autism spectrum disorder: A proteomic analysis. PLoS One. 2024;19(12):e0302951. Link
  7. Hewitson L, Mathews JA, Devlin M, Schutte C, Lee J, German DC. Blood biomarker discovery for autism spectrum disorder: A proteomic analysis. PLoS One. 2021;16(2):e0246581. (Retracted publication.) Link
  8. Anwar A, Abruzzo PM, Pasha S, Rajpoot K, Bolotta A, Ghezzo A, et al. Advanced glycation endproducts, dityrosine and arginine transporter dysfunction in autism – a source of biomarkers for clinical diagnosis. Mol Autism. 2018;9:3. Link
  9. Su Q, Wong OWH, Lu W, Wan Y, Zhang L, Xu W, et al. Multikingdom and functional gut microbiota markers for autism spectrum disorder. Nat Microbiol. 2024;9(9):2344-2355. Link
  10. Hazlett HC, Gu H, Munsell BC, Kim SH, Styner M, Wolff JJ, et al.; IBIS Network. Early brain development in infants at high risk for autism spectrum disorder. Nature. 2017;542(7641):348-351. Link
  11. Bosl WJ, Tager-Flusberg H, Nelson CA. EEG Analytics for Early Detection of Autism Spectrum Disorder: A data-driven approach. Sci Rep. 2018;8(1):6828. Link
  12. Sanchack KE. Autism Spectrum Disorder: Updated Guidelines from the American Academy of Pediatrics. Am Fam Physician. 2020;102(9):629-631. Link
  13. Jones W, Klaiman C, Richardson S, Aoki C, Smith C, Minjarez M, et al. Eye-Tracking-Based Measurement of Social Visual Engagement Compared With Expert Clinical Diagnosis of Autism. JAMA. 2023;330(9):854-865. Link
  14. U.S. FDA 510(k) K253442: EarliPoint Assessment (EarliTec Diagnostics). Decision March 5, 2026. Summary via FDA device database mirror (Innolitics). Link
  15. U.S. Food and Drug Administration. FDA authorizes marketing of tool to aid in diagnosis of autism spectrum disorder (Cognoa ASD Diagnosis Aid; De Novo DEN200069, June 2, 2021). FDA update published in AAP News, 2021. Link
  16. Megerian JT, Dey S, Melmed RD, Coury DL, Lerner M, Nicholls CJ, et al. Evaluation of an artificial intelligence-based medical device for diagnosis of autism spectrum disorder. NPJ Digit Med. 2022;5(1):57. Link
  17. U.S. FDA 510(k) K243558: Canvas Dx (Cognoa, Inc.). Decision April 11, 2025. Summary via FDA device database mirror (Innolitics). Link
  18. Perochon S, Di Martino JM, Carpenter KLH, Compton S, Davis N, Eichner B, et al. Early detection of autism using digital behavioral phenotyping. Nat Med. 2023;29(10):2489-2497. Link
  19. Krishnappa Babu PR, Di Martino JM, Aiello R, Eichner B, Espinosa S, Green J, et al. Validation of a Mobile App for Remote Autism Screening in Toddlers. NEJM AI. 2024;1(10). Link
  20. Kim JH, Hong J, Choi H, Kang HG, Yoon S, Hwang JY, Park YR, Cheon KA. Development of Deep Ensembles to Screen for Autism and Symptom Severity Using Retinal Photographs. JAMA Netw Open. 2023;6(12):e2347692. Link
  21. Manickam K, McClain MR, Demmer LA, Biswas S, Kearney HM, Malinowski J, et al.; ACMG Board of Directors. Exome and genome sequencing for pediatric patients with congenital anomalies or intellectual disability: an evidence-based clinical guideline of the American College of Medical Genetics and Genomics (ACMG). Genet Med. 2021;23(11):2029-2037. Link
  22. Srivastava S, Love-Nichols JA, Dies KA, Ledbetter DH, Martin CL, Chung WK, et al.; NDD Exome Scoping Review Work Group. Meta-analysis and multidisciplinary consensus statement: exome sequencing is a first-tier clinical diagnostic test for individuals with neurodevelopmental disorders. Genet Med. 2019;21(11):2413-2421. Link
  23. LinusBio. Pipeline: ClearStrand-ASD and StrandDx. Company web page (accessed October 2026). Link
  24. U.S. Food and Drug Administration. Laboratory Developed Tests. Content current as of September 19, 2025. Link
  25. Seidel S, Kreutzer R, Smith D, McNeel S, Gilliss D. Assessment of commercial laboratories performing hair mineral analysis. JAMA. 2001;285(1):67-72. Link
  26. U.S. Food and Drug Administration. Don’t Be a Victim of Health Fraud Scams. Consumer brochure. Link
  27. U.S. Food and Drug Administration. Medication Health Fraud for Specific Diseases and Conditions. Content current as of January 28, 2026. Link
  28. McInerney M, Hurley B, Barkow J, Menning K, Nicolace J, Schauer J, Van de Water J, Wassman ER. Analytical Validation of an ELISA Assay for Maternal Autoantibody-Related Autism. Diagnostics (Basel). 2026;16(11):1665. Link
  29. US Food and Drug Administration. 510(k) premarket notification K260064: Neurora (Argus Cognitive). October 2026. Link

Professional review: this guide is waiting for review by a qualified clinician. It was written and checked against the published sources listed above. How we check our guides.

This guide is general education, not medical advice. Always talk to your child’s doctor or care team. Last checked: October 2026.