AI-designed ultrasound lens targets multiple brain regions at once, cutting pain

South Korean researchers have developed a technology capable of delivering personalized brain-stimulation treatment to patients with neurological conditions such as Parkinson's disease and depression.

A research team led by Professor Hwang Jae-yoon of the Department of Electrical Engineering and Computer Science at DGIST (Daegu Gyeongbuk Institute of Science and Technology), working jointly with a team led by Professors Jeong Eui-heon and Kwon Hyeok-sang of the Department of Biomedical Engineering at GIST (Gwangju Institute of Science and Technology), developed a technology called Thickness-Only Acoustic Hologram, or TOAH. The AI-based system directly optimizes the thickness of a 3D-printed lens to precisely stimulate multiple brain regions at the same time.

Conventional ultrasound brain-stimulation technology has a critical limitation: ultrasound waves lose focus as they pass through the hard, irregular skull, and the energy delivered to multiple targets becomes uneven. Overcoming this problem previously required complex, expensive multi-channel equipment fitted with numerous ultrasound transducers.

Hwang's team solved the problem by combining AI with physics-based optimization. The AI directly designs the three-dimensional thickness structure of a lens to be fabricated by a 3D printer, enabling ultrasound waves to travel past skull refraction and reach multiple brain targets accurately — dramatically reducing error.

The key advantage is that multi-site brain stimulation requires only a single thin 3D-printed lens and a single ultrasound transducer. Without expensive equipment, ultrasound can be focused simultaneously and evenly across multiple brain regions.

Experiments on actual mouse skulls and simulations showed that the new technology forms focal points far more accurately than conventional methods and delivers energy uniformly to multiple targets. It also significantly improved the suppression of unnecessary energy concentration on the skull, reducing the side effect of skull heating.

When the technology was used to simultaneously stimulate both sides of the thalamus in mice with neuropathic pain, excessive neural activity decreased and pain responses improved markedly. Simulations using human skull data confirmed the same gains in focal accuracy and multi-target energy balance.

The most immediate application is in noninvasive brain disease research and treatment. The technology can alleviate pain by simultaneously stimulating the sites responsible for neuropathic pain, and it can be extended to multi-target neuromodulation for conditions involving multiple brain regions, such as Parkinson's disease and depression.

"The key is that AI directly designs the lens structure, enabling ultrasound to be delivered accurately to multiple brain regions beyond the skull," Hwang said. "By achieving precise multi-site brain stimulation with just a single ultrasound transducer and a 3D-printed lens, we expect this can be expanded into a patient-tailored, noninvasive treatment platform that safely delivers energy — without surgery or incision — for pain as well as degenerative brain diseases and psychiatric and neurological disorders."

The research, supported by the Ministry of Science and ICT and the National Research Foundation of Korea, was published in the international journal Brain Stimulation in July.