Objective: To determine whether smartwatch sensor data correlates with clinician-measured Parkinson’s disease symptoms.

Background: Data collected with smartwatches could be used for long-term, continuous assessment of a patient’s PD motor symptoms outside of the clinic. Short-term objective assessments of PD symptom severity in a lab or home environment using data collected from multiple body sensors is feasible (Patel et al., 09). With a single accelerometer, smartwatches can continuously acquire large amounts of objective wearable data, but their ability to monitor PD motor symptoms accurately and their correlation to a neurologist’s clinical assessment remains unclear. Our group has developed algorithms that process accelerometer data into objective measures of symptom severity that could be employed to assist in treatment and care.

Methods: Patients with PD (n=31; age: 62±8.8 yrs; disease duration: 8±4.5 yrs) wore a GENEActiv watch (3D acceleration, sampled at 50Hz) on their most clinically affected side while repeating a set of motor task 12-16 times during different stages of the L-Dopa medication wearing-off cycle. Symptom severity scores were assigned by qualified clinician for each task using clinically validated scales. Accelerometer data was analyzed to estimate severity of bradykinesia during pronation/supination, dyskinesia during walking and tremor during static tasks. Time and frequency-based measures were extracted to describe the intensity of movement, level of tremor, number of cycles and frequency of periodic activities. Tree-based models were used to estimate the symptoms scores.

Results: For each symptom, the correlation of the algorithm score to the clinical score was determined. Bradykinesia in pronation/supination was described by intensity, cycles count and rotation angle (r=0.59). Dyskinesia in walking was described by step count, intensity and frequency coverage (r=0.66). Tremor was described by the intensity of typical tremor frequencies (r=0.47).

Conclusions: Initial findings suggest that objective measures can be extracted from a continuously worn smartwatch to describe PD symptom severity. This capability can support continuous tracking of the disease outside the clinic with minimal burden. A follow-up trial with a larger cohort is required to further validate these results and to evaluate the assessment of symptoms over time.

« Back to 2016 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/objective-measures-of-parkinsonian-motor-symptoms-using-a-continuously-worn-smartwatch/