Classifying the State of Parkinsonism by Using Electronic Force Platform Measures of Balance

Several measures of balance obtained from quiet stance on an electronic force platform are described. These measures were found to discriminate patients with Parkinson’s disease (PD) from normal control subjects. First-degree relatives of patients with PD show greater variability on these measures. A primary goal is to develop sensitive measures that would be capable of identifying impaired balance in early stages of non-clinical PD. We developed a trinomial logistic model that classifies a subject as either normal, pre-parkinsonian, or parkinsonian taking as input the measures developed from the platform data.