Introduction: The present study aimed at creating and
verifying a workflow to perform subject-specific strength-scaling of
musculoskeletal models, and validating the strength-scaled models using
isometric joint torque measurements. Methods: The participants
consisted of 21 males and 7 females. A field strength assessment across 10
exercises was used to estimate the participants’ one-repetition-maximum (1RM).
The 1RM measures were implemented in an optimization routine, calculating a set
of strength factors capable of scaling all included muscles in the 10 different
exercise-specific musculoskeletal models. Further, peak joint torques were
investigated using dynamometer obtained isometric strength measurements for
elbow flexion and extension, knee flexion and extension, and ankle plantar
flexion. Results: The optimization based strength-scaled models showed
an improvement of mean normalized root mean square error from 48.39 (+/-
22.99) % to 28.13 (+/- 15.47) % compared to the standard-scaled models.
Discussion: The optimization routine was faster than previously used
methods and showed a higher accuracy than the standard strength-scaling of
musculoskeletal models. Issues in the simple muscle models wrapping around the
knee and ankle joints made the comparison with the dynamometer data infeasible.
The present study shows an improvement when applying the optimization routine
for whole body musculoskeletal models, and other or more exercises could easily
be implemented for scalability. However, utilizing simple musculoskeletal muscle
models cannot readily be used to estimate and compare peak joint torque for near
end range of motion angles.