RESUMO
ObjectiveTo compare the retest reliability and discriminant validity of dynamic postural stability indices for functional ankle instability (FAI) obtained by different algorithms based on acceleration signals at different positions of human body. MethodsFrom April to June, 2021, 21 subjects with unilateral FAI and 21 subjects with normal ankle were recruited. Three inertial sensors were attached to the waist points, knee and ankle positions. The ground reaction force (GRF) and kinematics data of the subjects in multi-direction single leg landing test were collected synchronously by 3D force plate and inertial sensors. The unbounded third order polynomial (UTOP) fitting method was used to calculate the stability time, and the root mean square was used to caculate the stability index. ResultsMost of the indicators calculated based on acceleration signal correlated with that based on GRF with low coefficient (|r| = 0.116 to 0.368, P < 0.05). The stability time and stability index based on the acceleration signals of different positions of human body showed low to high retest reliability (CMC 0.30 to 0.91). For the females, among the stability time based on acceleration signal, eleven indexes achieved average to very high discriminant validity (AUC = 0.702 to 0.942, P < 0.05); eight of the stability indexes reached general level of discriminant validity (AUC = 0.717 to 0.782, P < 0.05). No algorithms achieved good discriminant effect in male subjects. ConclusionBased on the acceleration signal of waist point in single-leg landing stability test, the stability time calculated by UTOP algorithm can evaluate the dynamic postural stability of female FAI patients with high discriminant validity and medium to high retest reliability.