Deep learning-enabled accurate assessment of gait impairments in Parkinson’s disease using smartphone videos - Takeaways - MDSpire

Deep learning-enabled accurate assessment of gait impairments in Parkinson’s disease using smartphone videos

  • By

  • Jianda Han

  • Zhihua Tian

  • Jialing Wu

  • Kai Zhang

  • Shaohua Li

  • Fahd Baig

  • Peipei Liu

  • Ravi Vaidyanathan

  • Francesca Morgante

  • Weiguang Huo

  • December 13, 2025

  • 0 min

Share

  • 1

    A deep learning framework was developed to assess gait impairments in Parkinson's Disease using smartphone-recorded videos.

  • 2

    The framework achieved a micro-average AUC of 0.87 and an F1 score of 0.806, comparable to clinical specialists.

  • 3

    It effectively identified medication-induced gait changes with a precision of 73.68%, surpassing traditional assessment methods.

  • 4

    The interpretable framework extracted traditional motion markers and discovered novel digital biomarkers for disease progression.

  • 5

    This approach has potential for routine assessment of gait impairments in clinical and home settings, enhancing personalized therapies.

Original Source(s)

Deep learning-enabled accurate assessment of gait impairments in Parkinson’s disease using smartphone videos

  • by Jianda Han, Zhihua Tian, Jialing Wu, Kai Zhang, Shaohua Li, Fahd Baig, Peipei Liu, Ravi Vaidyanathan, Francesca Morgante, Weiguang Huo

  • December 13, 2025

Related Content

Japan Clears First iPSC Therapies for Clinical Use 

Two iPSC-derived cell therapies receive time-limited authorization following early clinical trials

  • March 25, 2026

  • 2 min
Nocebo Effects in Long-term Health Conditions: A Systematic Review of Experimental Studies

Clinical AI is Not (Yet) Trustworthy-But It Could Be