Abstract
The adequacy of current models and research strategies applied to gait analysis is investigated. It is shown that erroneous results are obtained by using inadequate experimental methods and (or) oversimplified models of the human body in the computation of kinetic gait characteristics, especially joint moments. Therefore, such models and methods should not be used for clinical gait assessment. In addition to employing more appropriate models, it is suggested that more attention be paid to gait descriptors that provide a quantitative measure for the sensitivity of kinematic gait characteristics with respect to muscular or neural inputs.