Parkinson's disease (PD) is a chronic and progressive neurological movement disorder, meaning that symptoms continue and worsen over time. Nearly 10 million people worldwide are living with Parkinson's disease. Finding cost-effective non-invasive monitoring techniques for detecting motor symptoms caused by Parkinson's disease are potentially of significant value for improving care. Of the PD symptoms, the motor symptoms are the most common and detectable signs that can be assessed unobtrusively for both diagnosis and for evaluating the effectiveness of the treatments. The goal of our study is to find methods for identifying and classifying the motor symptoms caused by Parkinson's disease. Focus of the study is on long-term motion tracking measurements conducted at home during normal everyday life. Both accelerometers connected to arm and leg and mobile phone inbuilt sensors carried in the belt are utilized in the study. The research has two main objectives / hypotheses: 1. Can the motor symptoms related to different levels of Parkinson's disease be identified using motion tracking sensors? The first objective includes extracting and screening the motion differences of patients in early stages of the diseases in comparison with the patients in developed stages (patients having hypokinesia, dyskinesia and state changes) of the diseases and their differences with healthy control elderly adults using advanced signal and data analytics. Data from questionnaires and walking test conducted in the hospital environment are utilized as comparison points. Goal is to test the hypothesis that the amount of motor symptoms can be detected and the three groups can be reliably separated using sensor data. 2. Can the time when the Parkinson medicine is taken be detected from the movement signals? A sample of 50 volunteer PD patients with early stage of the disease (no dyskinesia and state changes), plus 50 volunteer PD patients in the later stage of the disease (having dyskinesia and state changes), plus 50 volunteers who do not have Parkinson's disease will be recruited for the research. Study starts with a telephone screening and visit to the hospital. Background characteristics and stage of the Parkinson's disease is evaluated in the hospital using a UPDRS questionnaires (Unified Parkinson's Disease Rating Scale; Finnish version) and a standardized 20-step walking test. Before the walking test, accelerometer sensors are attached to the shank and on the nondominant wrist. In addition, the participant wears a smart mobile phone with embedded accelerometer and gyroscope sensors. Based on the questionnaires and walking test study physiotherapist classifies the participant into one of the three study groups. The major part of the study involves a 3-day motion screening in a free-living setting in which the subjects are wearing the abovementioned sensors for as long duration as they comfortably can and are willing. This 3-day study starts immediately after completion of the 20-step walking test in the hospital. During the 3-day study, subjects are free to live their lives without any additional tests. Subjects mark down the time when they take their Parkinson medication.
Study Type
OBSERVATIONAL
Enrollment
97
UPDRS (Unified Parkinson's Disease Rating Scale) questionnaires are utilized for the assessment of the disease stage.
20-step walking test is utilized either for assessing the disease stage (subjects having Parkinson disease) or for assessing the normal walking (subjects not having Parkinson disease)
Satakunta Central Hospital, Unit of Neurology
Pori, Finland
Accuracy of the classification of data from movement sensors in relation to the detected motor symptoms
Accuracy and consistency of the classification of the subjects in the 3 categories (early stage disease, developed stage of disease, no disease) based on movement signals recorded with accelerometers and gyroscopes. Sensitivity and specificity of the classification are analyzed. Several features and methods of classification are tested including time-domain features, time-frequency domain features and machine learning both from raw data and calculated feature sets.
Time frame: 3 days
Accuracy of the detection of the time when the Parkinson medicine was taken
Accuracy and consistency of detecting the time when the medicine is taken based on movement signals recorded with accelerometers and gyroscopes. Sensitivity and specificity of the detection are analyzed. Several features and methods of analysis are tested including time-domain features, time-frequency domain features and machine learning both from raw data and calculated feature sets.
Time frame: 3 days
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.