Research on the Post-Operative Outcomes of Parkinson's Disease Using a Neural Network Optimized With a Genetic Algorithm

Parkinson's disease is a progressive neurodegenerative pathology characterized by impairment of motor function. Dynamic monitoring and neurological assessment of the patient's clinical condition after surgical interventions used for treatment, especially after deep brain stimulation procedures, are of particular importance. This study proposes a hybrid approach that integrates a recurrent neural network with a stochastic model for analyzing post-operative outcomes related to Parkinson's disease. Through a mathematical model constructed based on the Ornstein-Uhlenbeck process, changes in tremor, motor functions, and dopamine levels over time have been imitated. Clinical biomarkers, medication protocols, and biosignal data have been analyzed using a neural network optimized with a genetic algorithm to identify prognostic indicators. Simulation and test results show that the proposed approach is capable of discriminating between stable and deteriorating post-operative patient conditions with high accuracy. This model can also serve as an effective tool in developing personalized therapy strategies by playing a supportive role in neurological decision-making.