Close Menu
Versa AI hub
  • AI Ethics
  • AI Legislation
  • Business
  • Cybersecurity
  • Media and Entertainment
  • Content Creation
  • Art Generation
  • Research
  • Tools

Subscribe to Updates

Subscribe to our newsletter and stay updated with the latest news and exclusive offers.

What's Hot

AI Watermark 101: Tools and Techniques

July 5, 2025

AI makes science simple, but does that make it right? Research warns that LLMS is oversimplifying important research

July 5, 2025

AI Art Generation Using Primo Models: Unlock Creative Business Opportunities in 2024 | AI News Details

July 5, 2025
Facebook X (Twitter) Instagram
Versa AI hubVersa AI hub
Saturday, July 5
Facebook X (Twitter) Instagram
Login
  • AI Ethics
  • AI Legislation
  • Business
  • Cybersecurity
  • Media and Entertainment
  • Content Creation
  • Art Generation
  • Research
  • Tools
Versa AI hub
Home»Research»Korean researchers have developed an AI that can determine attention deficit hyperactivity disorder.
Research

Korean researchers have developed an AI that can determine attention deficit hyperactivity disorder.

versatileaiBy versatileaiApril 21, 2025No Comments2 Mins Read
Share Facebook Twitter Pinterest LinkedIn Tumblr Reddit Telegram Email
#image_title
Share
Facebook Twitter LinkedIn Pinterest Email
I’m not a fan
Photos to help you understand the article. (Photo source = chatgpt)

Korean researchers have developed an AI that can determine attention deficit hyperactivity disorder (ADHD) by looking at photos of the fundus of the retina.

The research team led by Chun Geun-Ah, Professor Choi Hang-Nyeong of Pediatric Psychiatry at a retired hospital, and Professor Park Yoo-Rang of the Medical School Life Systems Information Research Class at Yonsei University, said the accuracy of AI in ADHD reached 96.9% after seeing photos of the retinal base.

Attention deficit hyperactivity disorder (ADHD) is a neurodevelopmental disorder found in 5-8% of school-age children. Attention deficit, impulsivity, and hyperactivity are major symptoms, with delayed diagnosis and treatment affecting academic, social relationships and emotional development.

Because ADHD diagnosis is conducted through interviews and questionnaire assessments, it is more likely that examiner subjectivity will intervene. The boundaries between normal behavior and symptoms are unknown, and it is not easy to diagnose so quickly that the diagnosis is inconsistent.

The research team has developed an AI that allows for objective and rapid screening of ADHD by looking at photos of the retinal fundus. This AI development used 1,108 retinal fundus photographs, four learning algorithm models, and Autolph pipeline technology. Automov Pipeline is a research tool that morphologically analyzes retinal blood vessels.

The predictive performance is excellent, with the area below the receiver operating characteristic curve (AUROC) value being the graph value comparing sensitivity (true positive rate) with specificity (false positive rate) and up to 0.969 (accuracy 96.9%). The closer to 1, the better the performance.

Shapley Additive Description (SHAP) analysis explained how predictive results from AI models emerged and derived important retinal functions related to ADHD. Typical symptoms were increased vascular density, decreased arterial vessel width, and altered structure of the optic disc.

Furthermore, AI measured the accuracy of predicting the extent of visually selective attentional damage by looking at photos of the retinal fundus in patients with ADHD. Visual selective attention is the ability to focus on a particular area, and ADHD patients are less capable. The AI ​​accuracy reached 87.3%.

Professor Cheon Geun-Ah confirmed that “the study not only predicts the likelihood that retinal fundus photographs will be used as important biomarkers for ADHD diagnosis, but also predicts defects in executive functioning, such as visual selective attention.”

This study was conducted with the support of the Korean Information and Information Association.

author avatar
versatileai
See Full Bio
Share. Facebook Twitter Pinterest LinkedIn Tumblr Email
Previous ArticlePowerful embedded models tailored for patents and IPs are supported by face-hugging experts
Next Article AI enhances Genflow’s drug research capabilities – ICYMI
versatileai

Related Posts

Research

AI makes science simple, but does that make it right? Research warns that LLMS is oversimplifying important research

July 5, 2025
Research

In the midst of intense AI talent races, Meta’s active recruitment target open-rai researcher

June 30, 2025
Research

Lossless compression tailored to AI

June 30, 2025
Add A Comment
Leave A Reply Cancel Reply

Top Posts

New Star: Discover why 보니 is the future of AI art

February 26, 20252 Views

Impact International | EU AI ACT Enforcement: Business Transparency and Human Rights Impact in 2025

June 2, 20251 Views

Presight plans to expand its AI business internationally

April 14, 20251 Views
Stay In Touch
  • YouTube
  • TikTok
  • Twitter
  • Instagram
  • Threads
Latest Reviews

Subscribe to Updates

Subscribe to our newsletter and stay updated with the latest news and exclusive offers.

Most Popular

New Star: Discover why 보니 is the future of AI art

February 26, 20252 Views

Impact International | EU AI ACT Enforcement: Business Transparency and Human Rights Impact in 2025

June 2, 20251 Views

Presight plans to expand its AI business internationally

April 14, 20251 Views
Don't Miss

AI Watermark 101: Tools and Techniques

July 5, 2025

AI makes science simple, but does that make it right? Research warns that LLMS is oversimplifying important research

July 5, 2025

AI Art Generation Using Primo Models: Unlock Creative Business Opportunities in 2024 | AI News Details

July 5, 2025
Service Area
X (Twitter) Instagram YouTube TikTok Threads RSS
  • About Us
  • Contact Us
  • Privacy Policy
  • Terms and Conditions
  • Disclaimer
© 2025 Versa AI Hub. All Rights Reserved.

Type above and press Enter to search. Press Esc to cancel.

Sign In or Register

Welcome Back!

Login to your account below.

Lost password?