A Postural Control Model to Assess the Improvement of Balance Rehabilitation in Parkinson's Disease

Abstract

Studies have shown that balance and mobility in people with Parkinson's disease (PD) can improve through rehabilitation interventions. However, until now no quantitative method investigated how these patients improve their balance control. In this study, a single inverted pendulum model with PID controller was used to describe the improvement of forty PD patients after a 12-session therapy program, and to compare their balance with twenty healthy subjects. The Center of Pressure (COP) data were recorded in seven sensory conditions - on rigid and foam surface, each with eyes open and closed, and with visual disturbance; and stance on rigid surface with attached vibrator to the Achilles tendons. From COP data four Stabilogram Diffusion Function (SDF) measures were extracted. In order to find the appropriate model parameters (three control parameters and a noise gain) from the SDF measures, first model simulations were performed to tune an artificial neural network (ANN) which relates the SDF measures to the PID parameters, and second the trained ANN was used to find the suitable PID model parameters from the experimentally recorded SDF measures. Statistical analysis revealed that patients had lower control parameters and noise gain than healthy subjects; confirming reduced control ability and sensory information in PDs. Balance rehabilitation improved the patients' clinical scores, which is reflected in the increased control parameters (particularly in foam tasks), and noise gain (in tasks on rigid surface). The presented method provides a good and sensitive measure to describe functional balance and mobility in PD.