BackgroundToe-walking is one of the most common gait deviations (due to soleus and/or gastrocnemius muscle contractures), compromising the first (heel rocker) and second (ankle rocker) of the foot during walking.The aim of this study is to evaluate the effect of emulated artificially gastrocnemius and soleus contractures on the first and second rocker during walking. MethodAn exoskeleton was built to emulate contractures of the bilateral gastrocnemius and soleus muscles. Ten healthy participants were recruited to walk under the following conditions: without emulated contractures or with bilateral emulated contractures at 0°,10°, 20° and 30° of plantarflexion of the soleus or gastrocnemius in order to create an artificial restriction of dorsiflexion ankle movement. A linear regression from the ankle plantar-dorsiflexion angle pattern was performed on 0–5 % of the gait cycle (first rocker) and on 12–31 % of the gait cycle (second rocker) to compute the slope of the curve. The proportion of participants with the presence of the first and second rocker was then computed. A Statistical Parametric Mapping (SPM) analysis assessed the kinematic variations among different degrees of emulated contractures. FindingsThe first and second rockers are completely absent from 10° of plantarflexion emulated contracture. The data indicate there was a non-linear shift of the gait pattern of the ankle kinematics and an important shift toward plantarflexion values with the loss of the rockers. InterpretationThis study suggests that toe-walking in the experimental simulation situation is not necessarily due to a high emulated contracture level and can occur with a small emulated contracture by an adaptation choice.This study may improve interpretation of clinical gait analysis and shows that the link between the level of gastrocnemius/soleus emulated contracture and progression of toe-walking (increased plantarflexion during gait) is not linear.