Measured and predicted mechanical internal work in human locomotion.

Predictive methods estimating mechanical internal work (W(int), i.e., work to accelerate limbs with respect to BCOM during locomotion) are needed in absence of experimental measurements. A previously proposed model equation predicts such a parameter based upon velocity, stride frequency, duty factor, and a compound critical term (q) accounting for limb geometry and inertial properties. That first predicted W(int) estimate (PW(int)) has been validated only for young males and for a limited number of horses. The present study aimed to extend the comparison between model predictions and experimentally measured W(int) (MW(int)) data on humans with varying gender, age, gait, velocity, and gradient. Seventy healthy subjects (males and females; 7 age groups: 6-65years) carried out level walking and running on treadmill, at different velocities. Moreover, one of the subject groups (25-35years) walked and ran also at several uphill/downhill gradients. Reference values of q represent the main important results: (a) males and females have similar q values; (b) q is independent on velocity and gradient. Also, different data filtering depth was found to affect MW(int) and, indirectly, PW(int), thus also the reference q values here obtained (0.08 in level, 0.10 in gradient) suffer a - 20\% underestimation with respect to the previous predicting model. Despite of this effect, the close match between MW(int) and PW(int) trends indicates that the model equation could be satisfactorily applied, in various locomotion conditions.