Objective: To determine the utility of context-awareness algorithms based on sensors in smartphones to improve ambulatory Parkinson’s disease (PD) monitoring.

Background: Previous research shows that optimal motion sensor-based ambulatory PD monitoring requires multiple sensors to reduce the influence of external factors like environment and activity. Additional sensors included in a smartphone can provide detailed contextual information about patient activity and environment such as mode of travel, human interactions, and time in specific locations. Integrating this information with current motion sensor algorithms may improve output and provide simpler use.

Method: Eight adults with PD (5M, 3F, age 58-70 years, disease duration 3-12 years) were recruited to wear motion sensors on the wrist and ankle for four days while awake. Data from the motion sensors was processed into metrics quantifying tremor, dyskinesia, slowness, and gait using previously validated algorithms. Subjects also carried a smartphone that included an open-source app for recording GPS location, accelerometry, and microphone audio. The app processed the recorded data into activity being performed, location, travel time, ambient noise level, and detection of conversation. Relationships between the new measures, PD monitoring recordings, and pre- and post-study patient surveys were examined.

Results: Based on the motion-sensor system, five participants experienced tremor (3%-21% of day), four had dyskinesia (2%-39% of day), all eight had slowness (30%-52% of day), and the eight spent 2%-16% of the day walking. Based on the smartphone activity detection, participants averaged 2.4 car trips per day, 86% of the time at home, 5.8% of the day in conversation, and 14.7% of the day walking. Symptom scores during vehicular travel demonstrated changes in tremor, dyskinesia, and slowness scores (p<0.005), likely due to the restraint of the arms and legs and the natural vibration of the car. Activity measures from the smartphone significantly correlated with various subjective quality of life survey measures (p<0.05).

Conclusion: Synchronization of symptom scores and smartphone outputs demonstrated new methods of using context awareness to improve PD symptom scoring. Combining smartphone sensors with motion sensors provides additional information that may increase accuracy of PD monitoring, reduce the required numbers of wearable sensors, and provide quality of life metrics.

« Back to MDS Virtual Congress 2020

MDS Abstracts - https://www.mdsabstracts.org/abstract/smartphone-based-context-awareness-to-improve-ambulatory-parkinsons-monitoring/