Journal of NeuroEngineering and Rehabilitation

Procedure Participants visited the prosthetic center on two separate days to assess their gait pattern while using their pre- scribed ESAR foot (for all participants this was the Vari-Flex, Össur, Iceland) and a SACH foot (1D10, Ottobock, Germany). On the first day gait analysis was performed while participants walked with their prescribed ESAR foot to which they were already accustomed. At the end of this measurement session participants were fitted with the SACH foot below their existing socket, which was aligned by a certified prosthetist. Participants used this foot in their daily life during the next 24 h to get accustomed to it, before returning to the clinic for a sub- sequent gait analysis using this SACH foot. During the gait analysis participants walked up and down a 10-m walkway at a fixed walking speed of 1.2 m ∙ s − 1 , which was controlled online using photocells (Microgate RaceTime2, Italy). A fixed speed over all par- ticipants and conditions was selected, since the outcome measures of this study are highly affected by walking speed, and potential speed differences between condi- tions would obscure the direct dependence of the analyzed gait parameters on foot type. The speed of 1.2 m ∙ s − 1 was selected as it was expected that partici- pants would be able to walk comfortably at this speed with both types of feet. Self-selected walking speed of the participants was measured before the experiment with the ESAR foot as a reference. This self-selected speed appeared on average to be slightly but significantly higher (1.27 m ∙ s − 1 , p = 0.03). Data from a minimum of 3 strides were collected for both intact and prosthetic leg while walking with the two different prosthetic feet. Data collection Kinematic data was collected using a 10-camera opto-electronic system at 100 Hz (VICON; Oxford, United Kingdom). Markers were attached bilaterally on the anterior and posterior iliac spines, lateral epicondyles of the femur, lateral malleolus of the fibula. For the pros- thetic side, lateral malleolus location was approximated as the distal end of the rigid pylon. Ground reaction forces were measured at 1000 Hz using two force plates (0.60 × 0.40 m. Kistler: Winterthur, Switzerland) embed- ded in the middle of the walkway. Gait speed while crossing the force plates was monitored using two pho- tocells (Microgate Racetime 2; Bolzano, Italy). Data analysis Force plate data was filtered at 100 Hz using a fourth order zero lag Butterworth low pass filter. All analyses were performed in the sagittal plane of progression. Force plate data was used to identify initial contact and toe-off based on a threshold vertical force of 25 N. Pros- thetic step length (SL prosthetic ) was calculated as the dis- tance between the malleolus marker of the prosthetic leading and intact trailing leg at the moment of initial con- tact. Intact step length (SL intact ) was calculated in a similar method at the time of initial contact of the intact leg. Step length symmetry (SL symm ) was defined as the difference between prosthetic step length and intact step length: SL symm ¼ SL prosthetic − SL intact ð 1 Þ Power generated by the prosthetic foot and ankle dur- ing stance was calculated using the method outlined by Prince et al. [ 8 , 26 ] , summing both the translational power and rotational power transferred from the foot to the shank: P ankle ¼ F dist ∙ v dist þ M dist ∙ ω shank ð 2 Þ where the subscript ‘ dist ’ represents the distal point of Fig. 1 During walking, forward progression is maintained when the extrapolated center of mass (X CoM ) projects anterior to the posterior border of the base of support at toe-off, i.e. when the backward margin of stability (MoS BW ) at toe off is positive. In prosthetic gait, control of MoS BW is affected by reduced push-off power of the prosthetic foot. a Depicts the prosthetic step for which no problem occurs. b Depicts the intact step. Due to the reduced push-off power of the prosthetic foot the center of mass velocity is reduced and hence X CoM projects less far anteriorly. With normal step length, the MoS BW would be reduced causing a treat for a loss of progression or a backward fall. c When the intact leg step length is reduced, MoS BW is restored but at the expense of step length asymmetry (i.e intact step is shorter compared to the prosthetic step) Houdijk et al. Journal of NeuroEngineering and Rehabilitation 2018, 15 (Suppl 1):76 Page 43 of 72