Parkinson’s Disease Severity Index Based on Non-Motor Symptoms by Self-Organizing Maps

Abstract

Parkinson’s disease, a progressive neurodegenerative disorder of the motor system, shows non-motor symptoms up to 10 years before classic motor signs, highlighting the importance of early detection for effective treatment. This study proposes a severity index using an Artificial Neural Network (ANN) trained by the Self-Organizing Maps (SOM) algorithm, with data from the FOX Insight database. After pre-processing, 41,892 questionnaires were selected, covering 25 questions about non-motor symptoms, defined by a neurologist, and divided into four classes representing stages of the disease. The goal is to offer a tool to classify patients based on these symptoms, allowing for accurate monitoring and personalized interventions. Validation was carried out with data from patients responding to the questionnaire at spaced moments, simulating medical consultations. The study was successful in developing the severity index, highlighting the importance of gastrointestinal and urinary symptoms at different stages. The persistence of difficulty sleeping in group 3 indicates special attention must be paid to this symptom in the initial stages. These results highlight the clinical and practical relevance of the index, although more studies with real patients are needed for validation.