| Abstrakt: |
A recent report discusses research on Parkinson's disease and its cognitive symptoms of impulse control disorders (ICDs) in patients who are treated with dopamine agonist therapy. Researchers from Chang Gung University in Taiwan propose using an electroencephalogram (EEG)-driven machine-learning scenario to assess ICD comorbidity in Parkinson's disease. They conducted a study using a low-cost, custom LEGO-like headset to record EEG activity during a cognitive task, and optimized a support vector machine (SVM) and support vector regression (SVR) pipeline to detect ICD comorbidity and estimate its severity. The results showed promising accuracy in differentiating subjects with ICD from those with Parkinson's disease, and the researchers suggest that their findings could facilitate the development of a wearable computer-aided diagnosis system for assessing the risk of cognitive comorbidity in Parkinson's disease patients. [Extracted from the article] |
