The results can be predicted by analyzing the microelectrode measurement signal with artificial intelligence during Parkinson’s disease deep brain stimulator implantation. As more data and experience are accumulated in the future, it is expected to be of great help in treatment.
Seon-ha Paik, Hee-chan Kim, Seok-gyu Seon, Seoul National University Hospital, and Professor Gwang-Hyun Park of Sejong Chungnam National University Hospital analyzed the microelectrode measurement records of 34 Parkinson’s disease patients who underwent deep brain stimulation under general anesthesia using artificial intelligence deep learning techniques to predict clinical outcomes after surgery. Announced on the 22nd.
Parkinson’s disease occurs when dopamine neurons located in the midbrain are lost about 70% compared to normal people without knowing the cause. It is the most common senile degenerative brain disease after Alzheimer’s disease, and about 2 out of 100 people over 65 are observed.
Tremors, stiffness, postural anxiety, and walking disorders are symptoms of Parkinson’s disease. A typical treatment performed to suppress the onset of symptoms is deep brain stimulation. It regulates neural circuits by putting electrodes in abnormal areas of the brain and stimulating them. Finding the correct and appropriate target is paramount.
In the operating room, the patient’s skull is punctured, microelectrodes are placed in the brain area determined by MRI, and then the position is moved little by little to measure electrical signals generated from the brain. At this time, the recorded electrical signal is analyzed and the actual stimulation electrode is inserted in the position where the effect is expected to be the best.
The research team analyzed the signal obtained through the microelectrode through artificial intelligence deep learning to predict the result. After that, the patient’s condition after the actual surgery was divided according to the degree of improvement and compared with the artificial intelligence prediction.
Deep brain stimulation is performed on both sides, but considering that the effects of each electrode on the left and right of the body will be different, the ratio of left and right is applied differently using multiple structures within the artificial intelligence algorithm.
The ratio of 5:1 and 6:1 showed the highest prediction accuracy, reaching a maximum of 80.21%. The research team said that it showed similarity to the functional structure of the actual basal ganglia of the cranial nerve.
Professor Sunha Paik said, “It will be a new paradigm for finding the optimal target when performing deep brain stimulation in Parkinson’s disease patients.”
Professor Kim Hee-chan said, “This is a new attempt by applying deep learning techniques to predict the prognosis of deep brain stimulation transplantation,” and predicted that “in the future, more clinical decision support systems using artificial intelligence techniques will be developed.”
This study, which first attempted to predict surgical outcomes using artificial intelligence for microelectrode measurement signals, was published in the latest issue of the international journal’PLOS ONE’.