Human Movement & Mobility Research
Qualisys Miqus

Comparison of Markerless and Marker-Based Motion Capture of Gait Kinematics in Individuals With Cerebral Palsy and Chronic Stroke: A Case Study Series

Research Square (preprint)
•
2023
•
Original research (preprint, not peer-reviewed; case study series)
Theia3D version:
v2022.1.0.2309
More Articles

Study at a glance

Three children with cerebral palsy and three adults with chronic stroke walked overground while marker-based and Theia3D markerless motion capture recorded gait kinematics simultaneously. Differences between systems were minimal and generally within established clinical measurement error, with transverse plane rotation in the cerebral palsy participants as a specific area of weaker agreement.

Summary

Disclaimer: The following summary details independent academic research. Theia3D is intended for research, sports performance and educational use only. It is not cleared as a medical device and is not intended to be used for the diagnosis, treatment, mitigation or prevention of any disease, injury or medical condition.

Editor’s note: This study used Theia3D version 2022.1.0.2309 (patch 7).

Study Overview

Design

A case series comparing marker-based and markerless motion capture, recorded simultaneously during overground walking, across two different clinical populations with varied gait impairment severity.

Participants

Three children with cerebral palsy (ages 7 to 16, Gross Motor Function Classification System levels 1 to 3, using ankle-foot orthoses and, in two cases, a walker or pushchair) and three adults with chronic stroke (ages 49 to 69, between roughly six months and eight years post-stroke, two using an ankle-foot orthosis and either a walker or cane).

Data Collection

Following a static calibration trial, participants walked at their self-selected pace through a capture volume while 10 marker-based cameras at 100 Hz and 10 markerless video cameras at 80 Hz recorded simultaneously. Three walking trials were completed, with two full strides per limb captured in each trial.

Analysis

Trunk, hip, knee, and ankle joint angles in the sagittal, frontal, and transverse planes were compared between systems using root mean square, maximum peak, minimum peak, and range of motion for discrete comparisons, and Pearson’s correlation and coefficient of multiple correlation for full waveform comparisons, with results interpreted against minimal clinically important difference (MCID) and minimal detectable change (MDC) values established in the stroke and cerebral palsy gait literature.

Key Findings

  • Very few statistically significant differences were found between systems across the full set of comparisons. Across root mean square joint angle values at the trunk, hip, knee, and ankle in all three planes for both limbs, only two reached statistical significance: frontal plane motion at the hip and ankle on the less affected limb.
  • At the trunk, agreement was strongest of any segment. Mean differences in maximum and minimum joint angles were under 1 degree in every plane and not statistically significant, leading the authors to highlight the trunk specifically as a strong candidate for reliable markerless tracking in future research with this population.
  • Where significant differences did occur, their magnitude generally stayed within or near established clinical thresholds. Hip and ankle sagittal plane differences, the primary plane clinicians rely on for stroke and cerebral palsy gait assessment, fell below the minimal detectable change values reported in the stroke literature for both limbs.
  • Ankle agreement in this clinical population was, in several respects, as good as or better than previously reported in healthy adults. Root mean square error values at the ankle across all three planes were smaller than those reported in prior healthy-adult comparisons, despite the participants’ wide range of ages, heights, and body types.
  • Transverse plane (rotational) agreement was the clear weak point, concentrated specifically in the cerebral palsy participants. While few individual transverse plane comparisons reached statistical significance, the authors report that waveform agreement in this plane was consistently poor for the cerebral palsy participants specifically, visible directly in the comparison plots, and speculate this may relate to foot deformities and the smaller anatomical features of a pediatric CP foot being harder for the system to identify from video.

What This Means for Researchers Working With Stroke and Cerebral Palsy Populations

This early evidence is encouraging for Theia3D’s potential as a research tool in stroke and cerebral palsy gait studies generally, and particularly promising for trunk kinematics specifically, where agreement with marker-based capture was strongest. It also comes with a specific, named caution: transverse plane kinematics in cerebral palsy, a population where foot and limb deformities are common, showed meaningfully weaker agreement and should not yet be treated as interchangeable with marker-based measurement. Given the small, heterogeneous sample, researchers should treat this as an encouraging early signal rather than a settled validation for either population.

The Role of Theia3D in Movement Analysis

Theia3D’s video-based approach allowed this study to assess gait in participants using ankle-foot orthoses, walkers, and a cane, assistive equipment and visual presentations that fall outside the healthy, able-bodied populations markerless systems are most often validated on. That made this study a genuine test of how the system performs when a participant’s anatomy and equipment differ meaningfully from typical training data, rather than a repeat of prior healthy-adult validation work.

Study Limitations and Future Directions

This is a preprint that has not yet undergone peer review, and the authors are explicit that this is a case series of six total participants across two distinct, individually heterogeneous populations, not a representative sample of either cerebral palsy or chronic stroke. Several participants used ankle-foot orthoses and assistive devices, which the authors note could have affected marker placement and visibility for both systems, and one cerebral palsy participant’s use of a rear-walker may have introduced additional visual occlusion. The authors also note that broader sensitivity of the markerless system to characteristics like anatomical deformities, clothing, and assistive device use has not yet been fully tested, and they call for further investigation across a wider range of clinical populations before drawing firmer conclusions, particularly regarding the transverse plane limitations identified here.

Citation: Steffensen, E.A., Magalhães, F., Knarr, B.A., Kingston, D.C. (2023). Comparison of markerless and marker-based motion capture of gait kinematics in individuals with cerebral palsy and chronic stroke: A case study series. Research Square preprint. https://doi.org/10.21203/rs.3.rs-2557403/v1

Link: Read the full preprint

Exploring Markerless Gait Analysis for Stroke or Cerebral Palsy Research?

This early case series found Theia3D produced gait kinematics closely comparable to marker-based capture in people with stroke and cerebral palsy, particularly at the trunk, with transverse plane rotation in cerebral palsy as a specific area still under investigation. If gait research in neurological populations is part of your work,talk to our team about where markerless capture fits today.

Recent Posts