Neural Signal Processing Breakthrough Targets Parkinson’s Symptom Control
A research collaboration led by Professor Ji-Woong Choi of the Daegu Gyeongbuk Institute of Science and Technology (DGIST) has developed a new deep-learning technology. The system, dubbed “NONOS,” addresses a hurdle in neuroengineering: the inability to distinguish between the brain’s rhythmic oscillations—which can signal disease progression—and the background “noise” that often obscures them. The findings were published in the international journal NeuroImage.
Deep brain stimulation (electrical stimulation therapy) involves implanting electrodes to deliver targeted electrical pulses to specific areas of the brain. However, devices often struggle to adjust because they cannot cleanly separate the pathological “beta waves”—rhythms linked to movement slowness—from the background electrical activity of the brain. Until now, conventional analysis methods have relied on filtering techniques that often mistake background shifts for rhythmic ones, or windowing methods that lose critical information by averaging signals over time.
Isolating Brain Rhythms with High Temporal Resolution
The NONOS technology operates on a specific mathematical characteristic of brain signals: the power of background electrical components decreases exponentially as frequency increases. By training an AI model to identify and subtract this predictable background pattern, the researchers were able to isolate pure rhythmic components. This is analogous to isolating a single human voice in a crowded, humming café, allowing for finer control over how a device interprets neural data.
The impact of this resolution is significant. In tests involving Parkinson’s disease patients, the research team demonstrated that NONOS could predict which foot a patient would step forward with an accuracy score of 0.98, a significant improvement over the 0.54 to 0.79 range achieved by conventional methods. In experiments using mice modeling Parkinson’s disease, the technology proved capable of detecting the rapid decrease in beta-wave strength immediately following electrical stimulation—a feat that traditional filtering systems frequently miss.
The significance of this study lies in demonstrating that rhythms and background signals mixed together in brain signals can be separated without losing temporal resolution. If we further develop NONOS to operate in real time on implantable brain devices by reducing its computational requirements, it will contribute to improving the accuracy of personalized deep brain stimulation therapy, which adjusts electrical stimulation according to each patient’s condition. — Professor Ji-Woong Choi, DGIST
Broader Applications Beyond Parkinson’s
While the initial application focuses on Parkinson’s disease, the research team suggests the underlying technology can be applied to other conditions. Given that NONOS was created to distinguish between periodic and aperiodic elements within biological signals that change over time, it has the potential to be applied to a broad spectrum of time-varying data beyond the specific cases detailed in the study.
The project, which included collaboration from the Kumoh National Institute of Technology, Korea University, and the Korea Brain Research Institute, was supported by the National Research Foundation of Korea’s STEAM Research Program and the Ministry of Education’s Glocal Lab program. The research team is now looking toward reducing the computational load of the NONOS model so that it can be integrated into implantable brain devices.
Should the NONOS framework be effectively optimized to function in real time on brain implants by lowering its processing demands, it holds the potential to enhance the precision of customized deep brain stimulation, which modifies electrical output based on the unique state of each patient.
