Dataset_Effect of isokinetic eccentric resistance training on strength, flexibility and muscle architecture.xlsx

Data collection

Functional and structural parameters of the right shoulder were assessed using diffusion-weighted MRI scans and isokinetic dynamometry three to five days before the first and after the final eccentric intervention session.

Isokinetic Testing

IsoMed 2000 (D&R Ferstl GmbH, Hemau, Germany) was used to test the ROM and strength of the right shoulders external and internal rotators. Each subject was tested in the supine position with fixed shoulder-arm joints (Figure 1) as for the eccentric training described (Table 1). In addition, the shoulder was fixed ventrally to prevent shoulder elevation in an individually standardized manner.

Flexibility tests

Active stretching testing consisted of 15 trials alternating between internal and external rotation. Volunteers were instructed to move the apparatus slowly and with ease until aROMmax. The turning point were reached when the subject was unable to continue the antagonist stretch generated by the agonist muscles. Passive stretching tests were performed at an isokinetic speed of 5°/s. The passive stretch involved 15 trials of alternating internal and external rotation with eyes closed and muscles subjectively relaxed until pROMmax, characterized by 8 Nm of passive torque, was reached.

Strength tests

The voluntary maximum strength tests were performed separately for concentric and eccentric rotation, also alternating between internal and external rotation. Strength measurements were performed over an amplitude of 160° (70° internal rotation and 90° external rotation) at an isokinetic speed of 60/°s, 180°/s and 30°/s. All tests were performed with a rest period of 60 seconds. . Each subject's position and dynamometer adapter gravity correction value and settings were individualized at baseline and used for post-tests.

Magnetic Resonance Imaging

A 3-Tesla Siemens MAGNETOM Prisma scanner (Siemens Healthcare, Erlangen, Germany) with a shoulder coil (XL, 16-channel) was used for MRI scans. The participants lay in a head first supine position with the right arm in the neutral position and the hand supinated. The MR protocol consisted of a T1-weighted and a diffusion-weighted sequence. The T1-weighted sequence consisted the following parameter settings: repetition and echo time TR/TE = 492/20 ms, slice thickness = 0.7 mm, flip angle = 120°, field of view = 180 x 180 mm², matrix = 256x256 mm². Diffusion-weighted sequence were acquired as following: repetition and echo time TR/TE = 6100/69 ms, slice thickness = 5.2 mm, flip angle = 90°, field of view = 240 x 240 mm², matrix = 122 x 122 mm², 48 diffusion sampling directions with b = 400 s/mm². Total scan time reached 12 minutes.

Data processing

MATLAB v.R2022a (MathWorks. Natick. USA) was used to process the isokinetic dynamometer data. Flexibility calculations were based on the average of the third through seventh trials in each test, with the first two trials excluded and defined as familiarization. pROMmax was calculated when subjects reached the end point characterized by either 8Nm or 100° internal rotation and 130° external rotation. For further analysis of changes in the morphology of the passive torque-angle curves from baseline to posttest, a three-parametric e-function (Figure 3) was fitted from the instant of minimum torque (internal rotation) to maximum torque (external rotation) or to the end of the torque-angle curve for each trial. Then, the evaluated e-function was used to calculate the angle at the instant of 0.01 Nm/degree (pROMsub) and to calculate the fit until the defined endpoints using SPM1d (See the method evaluation in the Appendix).

For strength analysis, the mean of three maximum consecutive repetitions were used. Raw data were filtered using a 6Hz cutoff frequency (Alt et al., 2018). From each test, the acceleration and deceleration phases were cut, leaving only the interval with the desired isokinetic velocity. The maximum torque was then identified in the isokinetic phase and the mean torque was calculated over this phase. Descriptive and inferential statistics in SPSS v.27 (IBM. Armonk. New York. USA) used absoulte torque. For comparison of curve morphology using SPM1d, data were normalized to body mass[24,25].

MRI data were first processed using Mimics Materialise v.24.0 (Leuven, Belgium) for manually segment the supraspinatus and infraspinatus muscles to extract volume of interest (VOI). Then, the diffusion-weighted images were processed and corrected using DSI Studio (v. 3^rd of December 2021. http://dsi-studio.labsolver.org). Based on VOI-based deterministic fiber tractography muscle FL and FV were calculated using DSI Studios tool for statistics. More details for the MRI and segmentation can be found in a previous methods paper (Vetter, 2023).

Statistics

MATLAB v.R2022a (MathWorks. Natick. USA) and SPSS v.27 (IBM, Armonk, New York, USA) were used for statistics. Descriptive results were based on mean and standard deviation (±). Participants were excluded from further analysis if the data distribution showed outliers and z-transformed values exceeded 2.5. For statistical analysis, missing variables were calculated in SPSS. Repeated measures multivariate analysis of variance (MANOVA) was used to show main effects and interactions between different factors. The factors were time (pre and post), mode (eccentric and concentric), and speed (30°/s, 60°/s and 180°/s). For further post-hoc mean comparisons, repeated measures univariate analysis of variances (ANOVA) and paired t-tests were used. The p-value was set at 0.05. Differences between the torque-angle curves were calculated using SPM1d integrated in MATLAB.

Abbreviations: