Study at a glance
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: Theia3D version v2023.1.0.310(patch:1), as stated in the paper.
Study Overview
Design
A cross-sectional construct validity study comparing 3D gait kinematics and kinetics from marker-based and markerless (Theia3D) motion capture systems, recorded in randomized order in healthy older adults.
Participants
Thirty community-dwelling healthy older adults (16 men, 14 women; mean age 75.0 years), screened for cognitive status and physical function and excluded for conditions affecting walking, medications affecting walking, cognitive impairment, or a body mass index above 30.
Data Collection
Gait data were collected using marker-based (46 reflective markers) and markerless (8 Qualisys Miqus cameras) motion capture systems in randomized order, both at 85 Hz. Markerless trials were recorded with participants wearing their own usual clothing rather than tight-fitting minimal attire; a supplementary analysis in the paper found only small differences between minimal and usual clothing. Marker placement was performed by an experienced member of the research team. Walking speed did not differ significantly between the two conditions.
Analysis
Eight time-normalized right gait cycles were randomly selected and averaged per participant per system. Waveform similarity was assessed using the cross-correlation coefficient (Rxy); range of motion and peak ankle plantarflexion moment used Pearson's correlation. Bland-Altman analysis quantified bias and limits of agreement, and root-mean-square differences (RMSD) were compared against informal thresholds of 5 degrees for kinematic angles and range of motion, and 10% of signal amplitude for kinetics.
Key Findings
- Sagittal-plane kinematics showed strong agreement at the hip, knee, and ankle. Cross-correlation coefficients were 0.9 or higher at all three joints in the sagittal plane, with the single exception of the pelvis, which showed weak agreement (Rxy = 0.17) and a bias exceeding 5 degrees across roughly half of the gait cycle.
- Frontal- and transverse-plane kinematics showed weaker, more variable agreement. Correlations in these planes ranged from 0.17 to 0.49 at most joints, with functional limits of agreement exceeding the study's 5-degree threshold across the entire gait cycle for all transverse-plane hip, knee, and ankle motion and for pelvic frontal-plane motion; ankle frontal-plane motion was a notable exception, showing strong agreement (Rxy = 0.77).
- RMSD values ranged from 2.9 degrees at the ankle in the sagittal plane up to 7.7-7.8 degrees for hip flexion/extension and transverse-plane knee and ankle rotation, and the marker-based system showed higher trial-to-trial variability than the markerless system overall, particularly in the frontal and transverse planes.
- Joint kinetics showed strong agreement across most measures, with the clear exceptions of hip internal/external rotation moment and ankle frontal-plane moment, both of which showed poorer agreement and wider limits of agreement during the stance phase of gait; peak ankle plantarflexion moment showed a strong correlation (r = 0.76) with bias and agreement limits within the study's threshold.
- Range-of-motion agreement was strongest for the knee and ankle in the sagittal plane (r = 0.70 and 0.76, respectively) and moderate for hip flexion/extension (r = 0.60), with most other range-of-motion comparisons not reaching statistical significance; marker-based range-of-motion values were generally higher than markerless values, with some exceptions in the sagittal and transverse planes.
What This Means for Movement Science Researchers Working With Older Adults
For researchers and lab methodologists considering markerless capture for older-adult gait assessment, this study offers a plane-specific validation picture consistent with findings in other populations: sagittal-plane lower-limb kinematics and most kinetics showed good construct validity, supporting markerless capture's use for the gait measures most central to everyday walking analysis in this group, while frontal- and transverse-plane motion, and pelvic kinematics specifically, warrant more caution. The authors' explicit real-world design choice, collecting markerless data in participants' own clothing, is also a practically relevant finding for anyone planning field- or community-based gait studies with older adults, since it suggests clothing variability is unlikely to meaningfully affect results.
The Role of Theia3D in Movement Analysis
Theia3D was the markerless system evaluated throughout this study, generating 3D lower-limb joint kinematics and kinetics from gait trials recorded while participants wore their own clothing, without the marker placement time or minimal-clothing requirement of the marker-based comparison system. This let the researchers assess construct validity under conditions closer to how markerless capture would likely be used in practice with an older-adult population.
Study Limitations and Future Directions
The authors note that no gold-standard reference exists for 3D gait analysis, so this study assesses agreement between two methods rather than the absolute accuracy of either one against a true measure of movement. They point to a "cross-talk" effect, previously documented in marker-based systems as a result of anatomical landmark placement error, that appeared to affect both systems similarly during mid-swing knee frontal-plane kinematics, consistent with prior work. The authors also caution that markerless software is updated regularly and that their findings are specific to the Theia3D version used in this study. Given the weaker agreement observed in the frontal and transverse planes and for pelvic kinematics, the authors recommend further validation work before markerless and marker-based systems are used interchangeably for full three-dimensional gait analysis in this population.
Citation: Carvalho, A., Vanrenterghem, J., Cabral, S., d'Assunção, A. M., Carnide, F., Veloso, A. P., & Moniz-Pereira, V. (2025). Construct validity of markerless three-dimensional gait biomechanics in healthy older adults. Gait & Posture, 120, 217–225. https://doi.org/10.1016/j.gaitpost.2025.04.022
Link: Read the full study
How Well Does Markerless Capture Hold Up in an Older-Adult Population?
This study found that Theia3D measures sagittal-plane lower-limb gait kinematics and most kinetics in healthy older adults with good construct validity, while frontal- and transverse-plane motion and pelvic kinematics still show gaps worth accounting for. If your work involves gait assessment in older adults and you want to understand how markerless capture performs for this specific population, talk to our team about what this validation could mean for your own research.