Exploring the torque- velocity relationship in postmenopausal women: Analyzing the influence of data processing

TL;DR

This study compares different models for fitting torque-velocity data in postmenopausal women and finds that data processing methods significantly affect the results.

Contribution

The study introduces a comparison of linear and nonlinear models for torque-velocity relationships and highlights the impact of data processing in postmenopausal women.

Findings

The goodness of fit and estimated parameters differ significantly between actual and target data for both knee extensors and elbow flexors.

Quadratic polynomial models provided the best fit for elbow flexors, while Hill’s model was best for knee extensors but produced unrealistic velocity estimates.

The study recommends using quadratic polynomial models for torque-velocity data in postmenopausal women.

Abstract

The main aims of this study were to compare the goodness of fit and derived parameters of linear and non-linear models for fitting the torque-velocity (TV) relationship in postmenopausal women, and to examine the influence of data processing on the results obtained.

Sixteen physically active postmenopausal women completed the experiment. Knee extensor (KE) and elbow flexor (EF) muscle strength was evaluated in the dominant limb using an isokinetic dynamometer. Isometric and isokinetic tests were conducted at 30, 60, 120, 180, 240, and 300°/s. Peak torque and the corresponding joint angles were recorded for each test. TV data were fitted using linear, quadratic polynomial (PM), and Hill’s (HM) regression models. TV relationships were analyzed using both actual data (i.e., the velocity achieved and its associated torque; TVA) and target data (i.e., the velocity preset on the dynamometer and the torque reported; TVT). TV parameters derived from each model and their goodness of fit were calculated for both TVA and TVT relationships.

The goodness of fit and the estimated TV parameters derived from the regression models differed significantly between TVA and TVT for both KE and EF (P < 0.05). For TVA, the models with the best fit were HM for KE and PM for EF. However, HM yielded unrealistically high theoretical maximum velocity values (6764.69 ± 11619.09°/s) for KE. Parameter estimates for TVA differed significantly between models (P < 0.001).

Caution is advised when performing isokinetic assessments at high velocities in middle-aged women. The obtained data should be carefully examined, as TVA and TVT should not be used interchangeably. The choice of model can influence the estimated parameters. We recommend using quadratic polynomial models to fit TV data for both KE and EF in postmenopausal women.

Figures