Prediction of FoG in Parkinson's Disease: A Systematic Review

This study presents a systematic literature review on the application of FoG prediction models in patients with Parkinson’s disease. To this end, surveys were conducted in four scientific databases, resulting in 202 publications, with only original primary studies being considered. Sixteen publications were selected for the study, which used inertial and plantar pressure sensors and applied classification algorithms, both machine learning and deep learning, either alone or in combination. The models’ performance improved; however, some challenges remain regarding class imbalance, computational capacity, and the sensors’ usability in everyday use.