A research team led by Professor Wonhee Lee from Kyung Hee University’s Department of Software Convergence has successfully developed an AI-based brain age prediction model. Their research, published in the international journal Computers in Biology and Medicine (IF 7.0), moves beyond conventional single-modality imaging approaches by utilizing a multimodal neuroimaging technique to analyze brain aging with greater accuracy.

Sung-hwan Moon, a lead researcher on the project, highlighted the significance of the study, stating, "Our research explores how brain age prediction can help assess individual neurological conditions and contribute to the early diagnosis of brain disorders."

Enhancing Accuracy Through Multimodal AI Analysis

Brain age, distinct from an individual’s chronological age, refers to the biological state of the brain, determined through neuroimaging data. It is a crucial indicator of overall brain health and is particularly valuable for the early detection of neurodegenerative diseases, such as Alzheimer’s and dementia, as well as mental health disorders.

Previous studies primarily relied on single MRI datasets for brain age prediction, which limited the precision of analyses. In contrast, Kyung Hee University’s research team implemented a multimodal approach, integrating multiple imaging techniques to enhance the accuracy and comprehensiveness of brain age assessments.

Moon explained, "Structural MRI alone has limitations in evaluating white matter integrity, which plays a critical role in brain function. Since the brain is highly complex, a more holistic analysis that combines different types of imaging data is necessary to obtain precise measurements."

Professor Lee further emphasized the growing importance of multimodal data in AI-driven medical imaging, stating, "The current trend in AI research involves integrating multiple data sources to provide more comprehensive and reliable analyses. In the field of medical AI imaging, this approach is proving to be particularly valuable."

Linking Emotional States to Accelerated Brain Aging

The study also examined how emotional well-being influences brain aging, revealing that individuals with higher levels of loneliness and stress tend to experience faster brain aging.

Findings showed that participants exhibiting elevated levels of anger or depression had brain ages that exceeded their chronological ages, suggesting that emotional factors significantly impact the brain’s biological aging process.

Moon noted, "Our results confirm that psychological factors can accelerate brain aging. This highlights the potential for brain age prediction to serve not only as a biomarker for neurological health but also as an essential tool for evaluating mental well-being."

Advancing AI-Driven Early Diagnosis in Healthcare

For brain age prediction to be fully established as a clinical biomarker, its correlation with actual disease progression must be thoroughly validated. As a next step, the research team plans to further investigate brain age as a diagnostic marker and explore how AI can be used for early detection of brain disorders.

Professor Lee emphasized the importance of ensuring AI models perform reliably in real-world applications, stating, "AI should not only function effectively with training data but also produce consistent results when applied in clinical settings. Improving the model’s generalizability and reliability will be a key focus moving forward."

A Student Researcher’s Journey into AI and Neuroscience

Moon’s interest in AI emerged during his studies at Kyung Hee University’s International College, leading him to pursue a double major in Software Convergence. Through coursework in Capstone Design and undergraduate research projects, he became involved in the brain age prediction study.

Reflecting on his journey, he shared, "At first, I had only a vague curiosity about how AI could be applied to healthcare. However, with the guidance of my professor, I was able to explore the field in depth and refine my research interests."

Developing the machine learning model required proficiency in linear algebra, probability, and statistics, concepts that initially posed a challenge. However, consistent study and hands-on experience helped him build expertise in AI-driven neuroscience research. Moving forward, he aims to integrate global perspectives from his international education background with advanced AI technology, fostering an interdisciplinary approach to innovation in neuroscience and medical AI.