Related Experiment Video
Updated: Mar 27, 2026
10:28
Dynamic Digital Biomarkers of Motor and Cognitive Function in Parkinson's Disease
Published on: July 24, 2019
16.5K
PERSONALIZED HYPOTHESIS TESTS FOR DETECTING MEDICATION RESPONSE IN PARKINSON DISEASE PATIENTS USING iPHONE SENSOR
Elias Chaibub Neto1, Brian M Bot, Thanneer Perumal
1Sage Bionetworks, 1100 Fairview Avenue North, Seattle, Washington 98109, USA*Corresponding author., elias.chaibub.neto@sagebase.org.
Pacific Symposium on Biocomputing. Pacific Symposium on Biocomputing
|January 19, 2016
Summary
Simple hypothesis tests can effectively detect dopaminergic medication response in Parkinson disease patients using smartphone sensor data. These personalized tests, ignoring data
Area of Science:
- Neurology and Biomedical Engineering
- Digital Health and Mobile Sensing
Background:
- Parkinson disease (PD) management requires objective measures of medication response.
- Smartphone-based sensors offer a scalable solution for collecting longitudinal patient data.
Purpose of the Study:
- To develop and validate hypothesis tests for detecting dopaminergic medication response in Parkinson disease patients.
- To assess the validity and performance of personalized testing approaches using smartphone sensor data.
Main Methods:
- Utilized longitudinal sensor data from smartphone-based active tapping tasks in PD patients.
- Developed two personalized hypothesis testing approaches: a union-intersection test and a classifier-based test.
- Evaluated statistical validity (Type I error) and power through simulations and real-world data analysis.
Main Results:
- Simple hypothesis tests, even when ignoring data's serial correlation, maintained nominal Type I error rates.
- Personalized union-intersection and classifier-based tests demonstrated statistical validity.
- The proposed personalized tests showed good performance in analyzing real-world data from the mPower study.
Conclusions:
- Personalized hypothesis tests ignoring longitudinal data structure are statistically valid and perform well for detecting medication response in Parkinson disease.
- These methods provide a robust baseline for evaluating more complex analytical approaches in digital health studies.
- Smartphone sensor data combined with appropriate statistical methods can facilitate objective and personalized assessment of treatment efficacy in Parkinson disease.
