Yandex has unveiled a groundbreaking AI-powered solution capable of assessing brain development in infants — marking the world’s first neural network designed specifically for this purpose. Developed by Yandex B2B Tech, the Yandex School of Data Analysis, and St. Petersburg State Pediatric Medical University, the technology automates MRI analysis for babies under 12 months old, reducing evaluation time from several days to just minutes.

The system, freely available on GitHub, is designed for use by hospitals and research institutions worldwide. It holds the potential to transform how clinicians detect cerebral palsy (CP) and other central nervous system disorders, enabling faster diagnosis and more effective rehabilitation planning.

Tackling a Global Pediatric Challenge

Cerebral palsy remains one of the most prevalent causes of childhood disability, affecting roughly 2–3 in every 1,000 live births, according to the World Health Organization (WHO). Early diagnosis is crucial for improving patient outcomes, yet identifying CP in infants under one year has been a longstanding clinical challenge.

MRI scans offer valuable insights into an infant’s brain structure but are notoriously complex to interpret due to the low contrast between gray and white matter — critical tissues that govern cognitive and motor development. While scanning takes 20–40 minutes, image analysis often consumes hours or days of a radiologist’s time, particularly when tracking multiple scans for developmental monitoring.

AI Steps In: How the Neural Network Works

Yandex’s new AI model automates the segmentation and classification of brain tissues in MRI scans, achieving more than 90% accuracy in distinguishing gray and white matter. The tool drastically reduces turnaround time and improves the precision of developmental assessments.

The project was inspired in part by prior global research efforts, such as the 2019 MICCAI Grand Challenge, which focused on infant brain segmentation. However, Yandex researchers faced the same obstacle that stymied earlier attempts — a lack of annotated data.

To overcome this, the team collaborated closely with pediatric radiologists to create new segmentation masks and train the neural network using a combination of machine learning experiments and clinical validation. The result is a purpose-built model optimized for real-world medical use.

"Although many commercial radiology tools exist, none had tackled newborn MRI analysis," said Anna Lemyakina, Head of the Yandex Cloud Center for Technologies and Society. "By combining AI expertise with clinical knowledge, we’ve built a solution that helps radiologists diagnose faster and plan rehabilitation earlier."

Practical Benefits for Global Healthcare

Yandex’s open-source approach means any medical institution or research organization can deploy the tool without licensing restrictions. Its potential benefits include:

• Improved accuracy and objectivity: The model quantifies the ratio of gray-to-white matter, ensuring consistent and data-driven assessments.

• Faster diagnosis: Processing times are reduced from days to minutes, enabling timely intervention for at-risk infants.

• Higher clinical throughput: Automating MRI segmentation allows radiologists to dedicate more time to complex diagnoses and patient consultations.

• Support for junior specialists: The AI can assist less-experienced clinicians in interpreting infant brain scans, enhancing confidence and consistency.

Accelerating Early Detection and Research

By cutting analysis time and boosting diagnostic accuracy, Yandex’s innovation may play a key role in accelerating early cerebral palsy detection and improving global child healthcare outcomes. The solution also supports longitudinal research by standardizing brain development measurements across large datasets.

In an era where AI-driven medicine continues to evolve, Yandex’s open-source contribution underscores the transformative potential of AI in clinical diagnostics — not to replace human expertise, but to enhance precision, speed, and accessibility.

Get in touch with our MarTech Experts.