Mode
Text Size
Log in / Sign up

AI models show high correlation with physics-based methods to predict peak knee adduction momentArtificial intelligence predicts knee movement patterns during walking

AI-generated summary of the cited source, checked by automated accuracy review. How we work

Key Takeaway
Note that AI models show high correlation with physics-based methods for predicting key knee biomechanics.

This meta-analysis synthesized data from 4 included studies to evaluate the performance of artificial intelligence (AI) in predicting knee biomechanics during walking for patients with knee osteoarthritis. The analysis focused on comparing AI-predicted values against traditional physics-based methods.

Key findings indicate that AI-predicted peak knee adduction moments were highly correlated with physics-based methods, showing a mean difference of 0.03%BW*Ht (95% CI: -0.08 to 0.14). Additionally, AI-predicted knee flexion time-series showed RMSE ranging from 8.39 to 8.81 degrees across the gait cycle. For peak knee contact forces, correlations ranged from R = 0.17 to 0.92 with NRMSE values between 0.21 and 0.70.

The authors noted several limitations, including limited data for spatiotemporal parameters and a small number of studies providing direct physics-based comparators. The study quality ranged from very low to high. While AI approaches show potential for predicting specific knee biomechanics, the results indicate that further refinement and clinical validation are necessary before these models can be fully integrated into routine practice.

How this fits prior evidence

This meta-analysis addresses a gap in the technological assessment of diagnostic tools for knee osteoarthritis. While prior evidence has established the efficacy of non-pharmacological interventions such as exercise therapy, TENS, and specific mind-body exercises like Tai Chi and Wuqinxi to manage pain and function, this study evaluates the technical accuracy of AI models in predicting biomechanical markers like peak knee adduction moment.

Living with knee osteoarthritis often means dealing with constant wear and tear. To better understand this, researchers looked at how well artificial intelligence (AI) can predict the physical forces and movements happening inside the knee joint during walking.

The study compared AI predictions against traditional physics-based methods. The results showed that AI was highly correlated with standard measurements for peak knee adduction moments, which are key indicators of how much stress is placed on the joint. It also successfully predicted knee flexion time-series and various contact forces across different parts of the knee.

While these findings show promise for using technology to map out joint health, there are still hurdles to clear. The study was based on a small number of reports, and some data points were less consistent than others. More research is needed to fully validate these AI models in everyday clinical settings before they can be used routinely by doctors.

What this means for you:
AI shows promise in predicting how knee joints handle stress during walking for people with osteoarthritis.

Common questions

How accurate is the AI at predicting knee movement?

The AI showed a high correlation with physics-based methods when measuring peak knee adduction moments. It also successfully predicted knee flexion time-series and various contact forces, though researchers note that accuracy across all variables is not yet fully established.

What specific parts of the knee did the AI track?

The AI was used to predict several factors, including peak knee adduction moments, flexion time-series, and contact forces for the medial, lateral, and total areas of the knee.

Is this technology ready to be used in clinics today?

While the AI shows potential to predict specific biomechanics, more refinement and validation are needed. The study was based on only four included studies, so further testing is required before it can be used in clinical settings.

Study Details

Study typeMeta analysis
EvidenceLevel 1
PublishedJul 2026
View Original Abstract ↓
This review synthesizes current literature on the use of artificial intelligence (AI) to predict knee biomechanics during walking in people with knee osteoarthritis (OA). Four databases were searched from inception to 22/01/2025. Risk of bias was assessed using a modified Newcastle-Ottawa Scale. Study quality was assessed using Grading of Recommendations, Assessment, Development, and Evaluations. Gait spatiotemporal parameters, knee kinematics, knee kinetics, and knee internal biomechanics calculated with both AI and physics-based methods were compared using root mean squared error (RMSE), normalized RMSE (NRMSE), mean absolute error (MAE) with standard deviation (SD), or correlation coefficients (R), and pooled for reporting. Of 883 studies screened, 8 were included for review, and four provided appropriate data for meta-analysis. Studies ranged from very low to high quality. Limited data were available for spatiotemporal parameters, with few studies including direct physics-based comparators. AI-predicted knee flexion time-series had RMSE ranging from 8.39 ± 4.13° to 8.81 ± 4.25° across the gait cycle. Meta-analysis found AI-predicted peak knee adduction moment was highly correlated with its physics-based counterpart (R: 0.86 and 0.60) with moderate errors (MAE: 0.37 and 0.45) and mean differences 0.03%BW*Ht [95% CI: -0.08 to 0.14]). AI-predicted peak knee contact forces (medial, lateral, and total) had correlations ranging from R = 0.17 to 0.92, and NRMSE varied between 0.21 (0.01) and 0.70 (0.05) relative to physics-based values. Overall, AI approaches have potential to predict specific knee biomechanics, but refinement and validation are needed to improve prediction accuracy across all knee biomechanical variables.
Free Newsletter

Clinical research that matters. Delivered to your inbox.

Join thousands of clinicians and researchers. No spam, unsubscribe anytime.