본문 바로가기
HOME Research Research Archive

Research Archive

Professor Kyungho Yoon Mathematics and Computing

An AI Technology for Brain Stimulation Therapy Developed

An AI Technology for Brain Stimulation Therapy Developed


That accurately predicts the focal area of ultrasound and monitors the delivery situation in real-time and enhances the therapeutic effect and safety of transcranial focused ultrasound


Professor Kyungho Yoon’s research team from the School of Mathematics and Computing (Computational Science and Engineering) has recently developed an AI technology in collaboration with Harvard Medical School. This technology rapidly displays the internal ultrasound delivery status to the device user during brain stimulation therapy using transcranial focused ultrasound (tFUS).


tFUS is a technique which allows the stimulation of specific areas of the brain by concentrating ultrasound waves. This method is gaining significant attention in the field of brain stimulation therapy as it can be used in the non-invasive treatment of various brain diseases, including dementia, brain cancer, epilepsy, and Parkinson’s disease. However, during the transmission of ultrasound through the skull into the brain, significant distortions, such as reflection and refraction, can occur, leading to medical accidents by stimulating unintended brain areas. As such, monitoring the delivery status of focused ultrasound is crucial.

Figure 1. tFUSFormer process


Professor Yoon’s research team has developed a new super-resolution (SR) transformer model named the tFUSFormer. This model precisely visualizes the pressure field inside the skull generated by the ultrasound, displaying the delivery status of focused ultrasound in real-time.


According to their research, the tFUSFormer accurately predicted the focal area of ultrasound with an accuracy of 91% under the skull CT data conditions used for training and showed approximately 87% accuracy even with new data conditions.


This study represents groundbreaking research involving the teaching of an AI model about transcranial ultrasound delivery phenomena using mathematical modeling and computational science technology.


Professor Yoon stated, “This will lay the groundwork for the development of smart medical systems for patient-specific precision treatment beyond focused ultrasound therapy.” He further added, “The AI-based therapeutic assistant system developed here is expected to enhance the therapeutic effect and safety of tFUS, accelerating non-invasive treatment for various brain diseases.”


This research was published in IEEE Journal of Biomedical and Health Informatics, a top journal in the field of mathematics and computational biology.


Find out more 

Title of article: tFUSFormer: a physics-guided super-resolution transformer for the simulation of transcranial focused ultrasound propagation in brain stimulation

DOI: https://ieeexplore.ieee.org/abstract/document/10500878

Journal: IEEE Journal of Biomedical and Health Informatics

Contact corresponding author: Prof. Kyungho Yoon (yoonkh@yonsei.ac.kr)



Recommended Articles