FIGURE 10.

Linear regression between experimentally derived maximum isometric hip muscle torque measured from an isokinetic dynamometer and root‐mean‐squared errors of predicted isometric muscle torques in Generic elderly, Generic elderly optimised, Generic young, Generic young Optimised and Subject‐specific models through hip extension (a–c) and hip flexion (d–f). Adjusted R ² values are shown alongside associated p values. *Statistical significance (p < 0.05). The statistically significant positive correlations seen in the Generic elderly models through hip flexion suggest that there is a trend for these models to accurately predict muscle torques within individuals with low muscle force capabilities, but these muscle data may not be suitable for simulating muscle functions in stronger individuals. The statistically positive correlations and high root‐mean‐squared errors in the subject‐specific models through hip flexion suggest the framework of measuring muscle architecture from magnetic resonance images could underestimate the force‐generating capacities of these muscles, a deficiency which is exacerbated in models of stronger individuals