Measuring Behavior 2005: Keijsers et al.

Automatic assessment of dyskinesia in Parkinson’s disease in daily life

N. Keijsers, M.W.I.M. Horstink and C.C.A.M. Gielen

Radboud University, Nijmegen, the Netherlands

During the first years of levodopa treatment, patients with Parkinson’s disease (PD) have a stable response to levodopa. However, after several years of levodopa treatment, an increasing number of patients show fluctuations in motor response (‘on’-’off’ fluctuations) and levodopa induced dyskinesias (abnormal involuntary movements). These complications constitute a major problem in the long-term management of PD and add substantially to the patient’s disability. New pharmacological and surgical treatments to reduce levodopa induced dyskinesias are becoming of more and more interest. Therefore, an automatic and portable device that can assess LID automatically and objectively in daily life is highly useful.

Methods

13 patients were continuously monitored in a home-like situation for a period of approximately 2.5 hours. During this 2.5-hour period, the patients performed about 35 functional daily-life activities. Behavior of the patients was measured using triaxial accelerometers, which were placed on 6 different positions of the body. A neural network was trained to assess the severity of LID using various variables of the accelerometer signals. Neural network scores were compared with the assessment by physicians, who evaluated the continuously videotaped behavior of the patients off-line.

Results

Neural network correctly classi.ed dyskinesia or the absence of dyskinesia in 15-minute intervals in 93.7, 99.7 and 97.0% for the arm, trunk and leg, respectively. In few cases of misclassi.cation, the rating by the neural network was in the class next to that indicated by the physician using the AIMS-score (0-4). The percentage of time that a segment was moving was the most important parameter used by the neural network. For the leg mainly parameters of both legs were important. For the arm and especially for the trunk, parameters, related to movements of other body segments, were relevant. Dyskinesia appeared to be more dominant in the lower frequencies than in higher frequencies.

Conclusions

The neural network could accurately assess the severity of LID and distinguish LID from voluntary movements in daily life situations. The results suggest that the method could be operating successfully in unsupervised ambulatory conditions.

Paper presented at Measuring Behavior 2005 , 5^th International Conference on Methods and Techniques in Behavioral Research, 30 August - 2 September 2005, Wageningen, The Netherlands.