Assessment of Bradykinesia in Parkinson's disease patients through a multi-parametric system.

The aim of this paper is to describe and present the results of the automatic detection and assessment of bradykinesia in motor disease patients using wireless, wearable accelerometers. The current work is related to a module of the PERFORM system, a FP7 project from the European Commission, that aims at providing an innovative and reliable tool, able to evaluate, monitor and manage patients suffering from Parkinson's disease. The assessment procedure was carried out through a developed C# library that detects the activities of the patient using an activity recognition algorithm and classifies the data using a Support Vector Machine trained with data coming from previous test phases. The accuracy between the output of the automatic detection and the evaluation of the clinician both expressed with the Unified Parkinson's disease Rating Scale, presents an average value of [68.3 ± 8.9]%. A meta-analysis algorithm is used in order to improve the accuracy to an average value of [74.4 ± 14.9]%. Future work will include a personalized training of the classifiers in order to achieve a higher level of accuracy.

Gait assessment in Parkinson's disease patients through a network of wearable accelerometers in unsupervised environments.

Parkinson's disease (PD) predominantly alters the motor performance of the affected individuals. In particular, the loss of dopaminergic neurons compromises the speed, the automaticity and fluidity of movements. As the disease evolves, PD patient's motion becomes slower and tremoric and the response to medication fluctuates along the day. In addition, the presence of involuntary movements deteriorates voluntary movement in advanced state of the disease. These changes in the motion can be detected by studying the variation of the signals recorded by accelerometers attached in the limbs and belt of the patients. The analysis of the most significant changes in these signals make possible to build an individualized motor profile of the disease, allowing doctors to personalize the medication intakes and consequently improving the response of the patient to the treatment. Several works have been done in a laboratory and supervised environments providing solid results; this work focused on the design of unsupervised method for the assessment of gait in PD patients. The development of a reliable quantitative tool for long-term monitoring of PD symptoms would allow the accurate detection of the clinical status during the different PD stages and the evaluation of motor complications. Besides, it would be very useful both for routine clinical care as well as for novel therapies testing.

A comprehensive motor symptom monitoring and management system: the bradykinesia case.

The current work describes a methodology to automatically detect the severity of bradykinesia in motor disease patients using wireless, wearable accelerometers. This methodology was tested with cross validation through a sample of 20 Parkinson's disease patients. The assessment of methodology was carried out through some daily living activities which were detected using an activity recognition algorithm. The Unified Parkinson's Disease Rating Scale (UPDRS) severity classification of the algorithm coincides between 70 and 86% from that of a trained neurologist depending on the classifier used. These severities were calculated for 5 second segments of the signal with 50% of overlap. A bradykinesia profiler is also presented in this work. This profiler removes the overlap of the segments and calculates the confidence of the resulting events. It also calculates average severity, duration and symmetry values for those events. The profiler has been tested with a bogus dataset. Future work includes better training for the severity classifier with a larger sample and testing the profiler with real, longterm patient data in a projected pilot phase in three European hospitals.