A vertical ground reaction force-based gait deviation index using PCA and phase specific features
Authors:
- Helisyah Nur FadhilahHFHelisyah Nur Fadhilah
- Dimas AdiputraDADimas Adiputra
- Rifdatun NimahRNRifdatun Nimah
- Tanzilal MustaqimTMTanzilal Mustaqim
- Dedi IrawanDIDedi Irawan
- Wahyu Dwi LestariWLWahyu Dwi Lestari
- Amrul FaruqAFAmrul Faruq
- Kadek Heri SanjayaKSKadek Heri Sanjaya
- Ammar Sulthoni AdliAAAmmar Sulthoni Adli
- Ni Luh Putu Sri DiantariNDNi Luh Putu Sri Diantari
- Gloria Belinda RandaGRGloria Belinda Randa
- Muslimah MuslimahMMMuslimah Muslimah
Preprint
- 20 October 2026
- preprint
- Published by Elsevier in SSRN Electronic Journal
Abstract
The Gait Deviation Index (GDI) provides a single measure of gait deviation by quantifying the distance from a normative gait pattern. However, conventional GDI requires three-dimensional motion-capture data, limiting its application when only kinetic measurements are available. This study investigates whether a compact vertical ground reaction force (vGRF)-based index can capture gait deviation comparable to that kinematic GDI. At a walking speed of 1.0 m/s, vGRF signals were normalized to body weight and gait-cycle time, segmented into four clinically interpretable phases, and represented using phase-specific kinetic and bilateral asymmetry features. Principal Component Analysis (PCA) was applied to the control group (P01-P30) to derive the normative feature space. The resulting distance-based vGRF deviation index (vGRF-DI) was scaled to a mean of 100 and standard deviation of 10 in the control population. A kinematic GDI was independently calculated from the same participants and used as the reference measure. The current processed dataset contained 41 participants, comprising 30 young healthy controls and 11 older healthy participants. The subject-level vGRF-DI showed a significant Pearson correlation with the kinematic GDI (r = 0.400, p = 0.0096), while the Spearman correlation was ρ = 0.270. Among the phase-specific measures, the strongest association was observed during p3 (push-off), with r = 0.448. The findings provide preliminary evidence that a compact vGRF-based feature representation can capture part of the gait-deviation variation represented by the kinematic GDI while providing phase-specific information. The vGRF-DI could therefore be considered a complementary measurement rather than an equivalent replacement for the kinematic GDI.Keywords
This publication has 29 references indexed in Scilit:
- The Gait Deviation Index as an indicator of gait abnormality among degenerative spinal pathologiesEuropean Spine Journal, 2019
- Gait symmetry measures: A review of current and prospective methodsBiomedical Signal Processing and Control, 2018
- The effect of weight-bearing asymmetry on dynamic postural stability in people with chronic strokeGait & Posture, 2017
- Reliability and Minimum Detectable Change of the Gait Deviation Index (GDI) in post-stroke patientsGait & Posture, 2017
- Gait Deviation Index Correlates With Daily Step Activity in Children With Cerebral PalsyArchives of Physical Medicine and Rehabilitation, 2015
- Repeatability and validation of Gait Deviation Index in children: Typically developing and cerebral palsyGait & Posture, 2014
- Further evidence of validity of the Gait Deviation IndexGait & Posture, 2010
- The Gait Profile Score and Movement Analysis ProfileGait & Posture, 2009
- Application of principal component analysis in vertical ground reaction force to discriminate normal and abnormal gaitGait & Posture, 2009
- The gait deviation index: A new comprehensive index of gait pathologyGait & Posture, 2008