Explainable FoG Detection from Wearable Accelerometry — Analysis Code

Zenodo (CERN European Organization for Nuclear Research)

Abstract

Analysis pipeline (signal preprocessing and windowing, feature engineering, subject-independent leave-one-subject-out cross-validation, cross-dataset external validation, and SHAP explainability) accompanying the manuscript "Explainable Deep Learning for Freezing-of-Gait Detection from Wearable Accelerometry: A Cross-Dataset External Validation Study." Subject-independent data splits and training-fold-only standardisation are enforced in the released scripts.

Authors 3

1. Qiufeng Cheng

   Shandong First Medical University · Liaocheng People's Hospital

   Affiliation as printed

   Liaocheng People's Hospital and Liaocheng Hospital Affiliated to Shandong First Medical University

2. Guifeng Zhang Aachen

   RWTH Aachen University · Universitätsklinikum Aachen

   Affiliation as printed

   University Hospital RWTH Aachen

3. jingjing zhang

   Shandong First Medical University · Liaocheng People's Hospital

   Affiliation as printed

   Liaocheng People's Hospital and Liaocheng Hospital Affiliated to Shandong First Medical University