Sit-to-Walk predicts Freezing of Gait status over than cognitive load: an experimental study with Linear Discriminant Analysis

Freezing of Gait (FoG) is one of the most debilitating symptom of patients with Parkinson Disease (PwP) as it could leads to fall and loss of independence. We evaluated the motor behavior in simple and complex gait movements with three groups of participants: PwP with and without FoG in ON state condition, and participants without neurologic and gait disorders. All participants were required to walk, to begin to walk (Step Initiation), and raise from a chair for walking (Sit-to-Walk). Analysis of variance helped to identify task-specific kinematic and kinetic variables which differentiate among the three groups. Classification accuracy of such variables was obtained by training a Linear Discriminant Analysis – LDA algorithm on reduced sub-samples, and by testing it on the left out subjects.