Background: Goal of this study is to apply AI-based machine learning and advanced radiomic textures in Parkinson’s, to predict subthalamic-nucleus(STN) outcome of Parkinson disease patients’ stimulations done with deep brain stimulators by applying the radiomic techniques such that quantifiable image/signal-features’ or textures extrapolated.
Materials and Methods: Parkinson subjects (Parkinson diseased patients) were recruited in this study. Subjects with advanced PD of >5 years with good response to L-Dopa and H and Y score of<4>
Results: The highest mean prediction accuracy was obtained using normalized linear-regression (96.65±7.24%, AUC: 0.98±0.06) and DNN (87.25± 14.80%, AUC:0.87±0.18).
Conclusion: Findings reveal the potential power of radiomic-features achieved from hippocampus and amygdal? MRI within the prediction of STN-DBS motor outcomes for PD subjects.
Keywords: Amygdal?, Deep brain stimulation, Hippocampus, Motor outcome, Parkinson disease, Prediction, Radiomic features, Radiomics, Textures.
