{
  "id": 2210724,
  "title": "US teen builds AI tool for autism, ADHD screening",
  "url": "https://urgent.news/2026/08/20/us-teen-builds-ai-tool-for-autism-adhd-screening",
  "topic": "ai",
  "section": "AI",
  "published": "2026-08-20T18:51:18.000Z",
  "source": {
    "name": "The Economic Times",
    "slug": "the-economic-times",
    "url": "https://economictimes.indiatimes.com/news/international/global-trends/17-year-old-new-jersey-high-schooler-edward-kang-built-an-ai-tool-that-reads-the-retina-for-autism-and-adhd-screening-and-won-175000-for-his-groundbreaking-research/articleshow/133386541.cms"
  },
  "original_language": "en",
  "account": "A 17-year-old high school student named Edward Kang has developed an artificial intelligence tool called RetinaMind that could revolutionize the screening process for autism and ADHD. Kang, a senior at Bergen County Academies in Hackensack, New Jersey, created the AI-based screening prototype which analyzes retinal images to identify subtle patterns linked to autism spectrum disorder (ASD) and attention-deficit/hyperactivity disorder (ADHD). His innovative project earned him second place and $175,000 at the prestigious 2026 Regeneron Science Talent Search.\n\nThe inspiration behind RetinaMind stemmed from Kang's discovery of research suggesting that retinal images might hold clues about neurodevelopmental conditions. He noted the surprising yet intuitive connection between the retina and the brain, which develop from the same tissues. Previous research has also uncovered differences in retinal structures among individuals with neurodevelopmental disorders, albeit extremely subtle ones.\n\nKang trained AI models using retinal images from a vast public dataset to identify patterns that could differentiate neurotypical individuals from those with autism or ADHD. To enhance the system's transparency, he employed an explainable-AI technique called GradCAM, which generates heat maps highlighting the areas of retinal images that influenced the model's predictions. This feature allows researchers to gain insights into what the AI is perceiving.\n\nRetinaMind does not merely provide a result but also generates a heat map, enabling researchers to understand the AI's decision-making process. The model achieved an impressive accuracy rate of approximately 89%. Beyond computer modeling, Kang also constructed a retinal cell model to investigate gene changes that could potentially shed light on the biological differences associated with neurodevelopmental disorders.\n\nKang's work holds significant promise, as autism and ADHD are typically diagnosed through behavioral and developmental assessments rather than a straightforward physical test. This process can be time-consuming. Kang envisions RetinaMind enabling earlier diagnoses and more accessible screenings for neurodevelopmental disorders, potentially leading to earlier treatment and an improved quality of life for millions of patients worldwide.\n\nAlthough RetinaMind is currently a research prototype and screening tool, not an established clinical diagnostic test, it represents a groundbreaking approach to neurodevelopmental screening. Kang's achievement showcases his exceptional scientific talent and positions him among America's most promising young scientists. He will pursue his studies at MIT in the fall, where he can further delve into the intersection of artificial intelligence, neuroscience, and medicine.",
  "summary": null,
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}