Interpretable Side-Aware Gait Representations for Stroke Rehabilitation

Rocco Salvatore Calabrò conducted a secondary analysis of a public multimodal gait dataset comprising 138 able-bodied adults and 50 adults with stroke. The study derived side-aware gait-state representations from 11 waveform domains, including kinematics and surface electromyography, for 43 stroke patients. The analysis was published on figshare in May 2026.

Use Cases

• Clustering post-stroke gait states based on fused kinematic and sEMG waveform data.
• Analyzing side-dependent gait asymmetry using paretic, non-paretic, and side-difference waveform views.
• Identifying key biomechanical contributors to gait states, such as ankle angle and vastus lateralis sEMG activity.
• Generating hypotheses for adaptive-assistance controller design in neurorobotics.
• Evaluating the stability of gait-state representations under different data scaling and dimensionality reduction scenarios.

Strengths

• Dataset includes 188 total participants (138 able-bodied, 50 with stroke).
• Analysis is based on 11 waveform domains, each represented by 1,001 time-normalized points.
• The retained fused solution organized 43 stroke patients into three distinct gait states.
• Strongest domain contributors were identified, including ankle angle and vastus lateralis sEMG.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the primary data is unknown, which may limit suitability assessment.
• The dataset file is a 38.6 KB DOCX document, suggesting limited raw data scope.

Provenance

Source

Secondary analysis of a public multimodal gait dataset.

Collection Method

Functional principal component analysis and clustering (Ward hierarchical, K-means) on waveform data.

Freshness

Last updated 2026-05-25 06:05:43; freshness should be verified.