Detection and assessment of motor disorders in Parkinson's disease.

Abstract

Dopaminergic drugs, such as levodopa and dopamine agonists, are highly effective in the treatment of Parkinson's disease (PD). However, with each year of levodopa treatment, the number of PD patients who suffer from fluctuations in motor response increases. In addition to these fluctuations, patients can suffer from abnormal involuntary movements (dyskinesai). Dyskinesia is a disabling and distressing complication of chronic levodopa therapy in patients with PD. The actual emergence of dyskinesias throughout the day depends mainly on the timing and quantity of each individual dose levodopa. Self-assessment of the motor state by a patient is often unreliable. Therefore, a portable device that can assess dyskinesia automatically in daily life would be highly useful. The major purpose of this thesis was the development of a method for automatic detection and assessment of dyskinesias in daily life of a patient with PD. Movements of patients were measured using accelerometers placed at different body segments. Neural networks were trained using variables from the accelerometer signals and physicians rating to assess the severity of dyskinesia. The results indicated that the neural network can accurately assess the severity of LID and could distinguish between dyskinesia from voluntary movements in daily life situations. Analysis of the neural network revealed several new variables, which are relevant for assessing the severity of dyskinesia