Objective: To describe wearable sensor-derived measures in Parkinson’s disease (PD) research.

Background: Many commonly used outcome measures of PD are subjective, rater dependent, and infrequently used. Digital tools allow for continuous and passive characterization of PD inside and outside of the clinic setting.

Method: In a two-year observational study, participants used wearable sensors, and completed visits at baseline, months 6, 12, and 24. Visits included the MDS-UPDRS. Sensors were worn on the chest and most affected (or dominant) arm and leg for one week following each visit, during waking hours on days 1-6 and for an overnight period on day 7. Using the accelerometer data from sensors during the baseline visit and home monitoring period, lying, sitting, standing, and walking activity states were determined. Additionally, tremor assessments (rhythmicity index (RI), tremor proportion [1]) were performed. Interaction between tremor and activity states and correlation of tremor and UPDRS scores were analyzed.

Results: 27 participants with PD (mean age 66.2 years (SD=7.5), 8 female) completed a baseline visit. Over full duration of sensor wear, PD participants walked for 8.6% (SD=3%) of the time, spent 17.7% (SD=11.5%) of their time lying down, and sat for 52.2% (SD=9.2%) of the time.

Over non-walking durations of sensor wear, PD participants exhibited rhythmic movements associated with tremor for 12.5% (SD=12.0%) of their day (approximately 3 hours/day). Sitting intervals had greater average tremor proportion (14.8%) than when standing (11.5%) or lying (6.1%). On average, rhythmic tremor-motions had the highest amplitudes while sitting and standing and the lowest amplitudes while lying.  A sample activity and rhythmicity plot is provided [Figure1].

Sensor-derived tremor proportion correlated strongly with investigator-rated rest tremor constancy, such that higher MDS-UPDRS 3.18 ON-state scores were associated with heightened tremor proportion (r = 0.91, p < 0.001). There was also a moderate correlation between tremor proportion and participant-reported tremor severity (MDS-UPDRS item 2.10) (r = 0.47, p = 0.01).

Conclusion: Sensor-derived data correlates with clinical measures and enables the objective collection of detailed tremor data remotely. Longitudinal results of this study may help us identify digital markers of disease progression.

References: [1] Adams, J.L., Dinesh, K., Snyder, C.W. et al. A real-world study of wearable sensors in Parkinson’s disease. npj Parkinsons Dis. 7, 106 (2021). https://doi.org/10.1038/s41531-021-00248-w

« Back to 2022 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/remote-assessment-of-tremor-using-wearable-sensors-in-parkinsons-disease-research/