Dataset and code for lasso regression model that identifies kinematic changes related to reductions in the knee adduction moment when adopting a toe-in gait modification. Uses selectiveInference package to compute p-values on the selected features. Further description is available in the following manuscript: .
One hundred seven individuals with medial knee osteoarthritis walked naturally and with 10° toe-in. The features in the X_TIdiff represent the difference in a kinematic variable from natural to toe-in walking. Details on how these covariates were computed can be found in the manuscript. y_TIP1diff is each subject's change in average the first peak knee adduction moment (KAM) from natural to toe-in walking. A reduction from natural walking is considered positive.
Feature order in X_TIdiff matrix:
- trunk sway angle (max during stance)
- knee flexion angle (at time of first peak KAM)
- pelvic list angle (at time of first peak KAM)
- knee adduction angle (at time of first peak KAM)
- step width
- medio-lateral center of pressure position in foot frame (time of first peak KAM)
- anterior-posterior center of pressure position in foot frame (time of first peak KAM)
- frontal-plane tibia angle (time of first peak KAM)
- foot progression angle
- medio-lateral distance between centers of pressure over consecutive steps (time of first peak KAM)
- medio-lateral distance between knee joint centers over consecutive steps (time of first peak KAM)
- pelvic rotation angle (time of first peak KAM)
Download lassoRegression.r and the two csv files. Change working directory in code.