Journal of NeuroEngineering and Rehabilitation

positive work performed by the foot and ankle unit was on average 2.2 times higher compared to work per- formed by the SACH foot, but remains about half of the intact leg ankle foot and ankle power [ 9 ] . Similar results have been found in previous studies into the energy re- turn of ESAR and SACH feet [ 28 , 29 ], although the mag- nitude of energy return of prosthetic feet reported in literature varies considerably. This is not only due to differences between prosthetic feet type, but also due to methodological differences. Specifically, the way re- searchers deal with the non-rigidity of the foot segment and the lack of a well-defined ankle center of rotation in the calculation of power generated by the compliant prosthetic foot and ankle system [ 8 , 30 ] . These issues obscure the comparison of energetic properties between different feet presented in literature. The increased push-off power of the ESAR foot likely results in a higher velocity of the center of mass at toe-off. In fact, center of mass velocity decreased during the double support phase for both prosthetic feet, but this decrease was attenuated more in the ESAR condi- tion compared to the SACH condition. This difference in the change of center of mass velocity during the spe- cific phase of double support indicates that increased push-off power with ESAR likely accounts for the ob- served effects, as push off power is predominantly gener- ated during this double support phase. This relation between ankle push-off power and center of mass pro- pulsion is corroborated by previous simulation studies, which mathematically showed a direct relation between the amount of energy return of prosthetic feet and the propulsion of the body ’ s center of mass [ 31 , 32 ] . Alterna- tively, it might be argued that the difference in roll-over shape [ 33 ] between the two types feet, might contribute to the difference in center of mass velocity at toe off. It has been demonstrated previously that the roll-over shape of the ESAR foot used in this study has a larger arc length, providing a longer lever to the foot segment [ 9 ] . In passive feet a proper roll-over shape could also enhance the step-to-step transition, apart from ankle push off power, and attenuate deceleration of the center of mass double support [ 34 ] . It is however not known how roll-over shape and push off power interact in dy- namic ESAR feet, and how the effect on center of mass velocity could be partitioned over both. This should be explored in the future studies. Because of the larger center of mass velocity at toe-off with the ESAR foot, the extrapolated center of mass pro- jects more anterior compared to the SACH condition. This allows the prosthetic user to make a larger step with the intact leg, without compromising the backward margin of stability. This increase in intact step length, observed with the ESAR foot, resulted in increased step length symmetry when walking with the ESAR foot. Although step length symmetry, on its own, is not ne- cessarily a functional benefit [ 22 ] , often gait symmetry is considered a goal in gait training [ 35 , 36 ] . From a cos- metic point of view patients might prefer a close to normal symmetric gait pattern as to not stand out in the crowd. Furthermore, it has been speculated that gait symmetry might indirectly provide functional benefits as it could reduce mechanical overload on the intact and residual leg and on the low back, which are both common co-morbidities in people with a lower limb amputation [ 21 , 37 ] . This difference in mechan- ical loading of the intact leg has indeed been presented previously (in terms of external mechanical work) for this data set [ 9 ] . Hence, improving gait symmetry might be considered a relevant functional benefit of ESAR feet. The backward margin of stability at toe off was similar between foot condition. Nevertheless, this can be inter- preted as a positive effect of the ESAR foot. Without the additional push off power of the ESAR foot, as in the SACH foot, an increased intact step length would have resulted in reduced margins of stability given the con- straints outlined in Fig. 1 . This can further be substanti- ated when analyzing the change in backward margin of stability during double support (Fig. 4 ) , the phase during which push-off occurs. At heel strike of the intact leg margin of stability is smaller in the ESAR condition compared to the SACH condition. This is due to the fact that participants walk with a larger intact step length with the ESAR foot, while center off mass velocity at heel strike is similar between conditions. During double support the center of mass velocity decreases in both conditions but this decrease is smaller with the ESAR foot. The concomitant higher center of mass velocity at toe off with the ESAR foot results in similar margins of stability between feet, making up for the initial negative effect of increased step length. A similar effect of ankle push-off on the control of the backward margin of sta- bility can also be seen when comparing the intact and prosthetic step in prosthetic gait [ 22 ] . Such enhanced control over the backward margin of stability might affect balance confidence of the prosthetic user, which has been indicated as an important predictor of self-reported mobility performance and social activity [ 38 , 39 ] . When the prosthetic user is perceptive to this change in control over the backward margin of stability, this effect might contribute to the preference of many users for energy storing and return prosthetic feet over solid feet. This study, designed to investigate a specific mechan- ical consequence of the constraints of prosthetic feet, is subject to several limitations. The experiments were per- formed at a fixed walking speed between conditions (prosthetic foot). This was done to avoid the Houdijk et al. Journal of NeuroEngineering and Rehabilitation 2018, 15 (Suppl 1):76 Page 46 of 72