Quantitative comparison of explainable AI methods for interpreting deep learning-based classification of 3D gait kinematics.
Source check: Source checked. Not independent scientific confirmation.
Finding: Across three gait datasets, ResNet/LSTM/InceptionTime classifiers predict diagnostic group labels and four XAI methods identify recurring joint-angle features; Integrated Gradients is most robust and feature masking improves accuracy. This is a specific clinical gait-data result, but diagnostic cohort interpretation and exact reference standard are limited to existing datasets.
Technology: ResNet, LSTM and InceptionTime classify diagnostic groups from 3D gait kinematics; LIME, DeepLift, Integrated Gradients and sequential feature selection identify influential joint-angle features.
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- ResNet, LSTM and InceptionTime classify diagnostic groups from 3D gait kinematics; LIME, DeepLift, Integrated Gradients and sequential feature selection identify influential joint-angle features.
- Test and result
- Across three existing gait datasets, Integrated Gradients was the most robust explanation method in the reported comparisons, and feature masking improved classification accuracy. The findings are bounded to these datasets and reference labels.
- Limitation / still unknown
- The three gait datasets are existing clinical datasets with their own diagnostic labels; explanation robustness and feature masking do not establish external clinical validity or patient benefit.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://www.nature.com/articles/s41598-026-46243-0 · https://europepmc.org/article/MED/41917155
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.