Representing Patterns Of Parkinsonian Movements Using Sequential Neuronal Networks

Our study is about a method of representing different states of central motor control in movement disorders using recursive neuronal networks with biological structure. Torques of limb movements computed from kinematic measurements, which were obtained from a 3-D TV based movement analysis system, were used as learning curves for a neuronal network with sequential structure (Jordan). Movements in Parkinsonian patients with ("on") and without ("off) medication led to different connections between the neuronal layers. Graphic representations of neuronal connections can be used to categorize Parkinsonian movements and they allow an insight into the motor control mechanisms of Parkinson's disease.