Objective: To identify mobility-derived digital features that distinguish PD risk and track progression.
Background: Parkinson’s disease (PD) is marked by progressive motor decline, yet subtle motor alterations often precede clinical diagnosis by years. Wearable sensors objectively monitor daily movement, providing potential early PD risk stratification.
Method: Risk was defined using MDS-Prodromal Research Criteria likelihood ratios. Scores >70 indicated high risk. Accelerometry data from a 3-axis (Axivity AX6) worn for 24/7 was obtained. 244 features were extracted and reduced to six via recursive feature elimination with logistic regression: Diurnal mobility measures (transition jerk (TJ), sedentary bout time (SBT), percent of walking during the night (PWN), average active bout duration (AABD), standing time, and transitions from rest to active (TRA)). Data was split into training (70%) and test (30%) sets, stratified by target (LR≥70). A logistic-regression classifier was trained and evaluated by accuracy, F1-score, and AUC-ROC. Secondary analysis assessed motor change over 4–5 years using Jonckheere–Terpstra (JT) tests.
Results: The cohort included 170 participants (high risk; n=30 ; 54.5yrs+/-10.2, 26%F and low risk; n=140; 65.7yrs+/-10.2, 64%F). The six feature model discriminated between groups with AUC=0.84, accuracy=0.94, F1-score=0.93. Maximal sedentary bout was shorter in high-risk subjects (−22.6%; p<0.001), and TJ was lower (−7.8%; p<0.001). High-risk participants had more frequent (33.6% vs. 66.7%; p=0.0008) and longer walking bouts at night compared to low-risk(+10.2%; p=0.001). Progressors were more sedentary with longer yet less variable bout lengths and more bouts of walking at night (p=0.004; p=0.069 respectively).
Conclusion: Digital mobility measures can identify individuals at high risk for PD prior to diagnosis mainly reflecting reduced activity and fragmented nights. Mobiltiy features become more pronounced with progression to higher risk scores, reflecting the utility of digital monitoring in the preclinical stage.
To cite this abstract in AMA style:
A. Ziv, R. Vaknin Greenblatt, L. Yahimovich, Y. Lachberg, A. Thaler, JM. Hausdorff, A. Mirelman. Passive Lumbar Accelerometry Reveals Motor and Behavioral Signatures Associated with High Risk for Parkinson’s Disease [abstract]. Mov Disord. 2026; 41 (suppl 1). https://www.mdsabstracts.org/abstract/passive-lumbar-accelerometry-reveals-motor-and-behavioral-signatures-associated-with-high-risk-for-parkinsons-disease/. Accessed October 1, 2026.« Back to 2026 International Congress
MDS Abstracts - https://www.mdsabstracts.org/abstract/passive-lumbar-accelerometry-reveals-motor-and-behavioral-signatures-associated-with-high-risk-for-parkinsons-disease/