Journal of NeuroEngineering and Rehabilitation

the rigid part of the prosthetic leg at approximately the level of the malleoli of the intact leg. F dist and v dist are the reaction forces and linear velocity of this distal point, M dist represent the net moment at the distal point and ω shank the angular velocity of the shank. Ankle push-off work (W ankle , J ∙ kg − 1 ) was determined as the time inte- gral of the positive power burst prior to toe-off. Center of mass position (CoM) was calculated from the average of the four iliac markers. Center of mass vel- ocity (v CoM ) was calculated as the time derivative of the CoM position. Following the description of Hof et al. (2005, 2008) the extrapolated center of mass (X CoM ), represents the predicted position of the center of mass after the natural cycle time of the pendular motion of the leg, and was calculated as: X CoM ¼ CoM þ v CoM ffiffiffiffiffiffiffiffi l . g r ð 3 Þ with l representing leg length (distance from floor to tro- chanter major), g representing gravitational acceleration and ffiffiffiffiffiffiffiffi l . g r representing the natural frequency of the leg pendulum. The backward margin of stability was defined accord- ing to Hak et al. [ 23 ] : MoS BW ¼ X CoM − BoS ð 4 Þ with the posterior border of the base of support (BoS) of the leading leg represented by the malleolus. Hence, a positive MoS BW indicates a stable condition in which the CoM passes the leading stance foot. This definition is in line with previous studies of Hak et al. [ 22 , 27 ] . However, it is the reverse from the original definition of Hof et al. [ 25 ] , which was postulated for upright stand- ing during which X CoM should not pass the border of the base of support. Therefore, contrary to the current definition, Hof et al. defined MoS positive when it did not exceed the border of the BoS. All primary outcomes were assessed at the instant of toe-off of the trailing prosthetic leg, as this is the instant that the trailing leg can no longer generate push-off power to accelerate CoM. Hence, at toe-off the condi- tion for dynamic stability (i.e MoS BW > 0) needs to be satisfied. However, v CoM and MoS BW were also analyzed at heel strike of the intact leading leg (occurring prior to toe-off ), to test whether differences in these parameters between feet originate primarily during the double sup- port phase, during which push off power is predomin- antly generated. Except from step length and step length symmetry, outcomes were only analyzed for the step in which the prosthetic leg is the trailing push-off leg and the intact leg is the leading leg (i.e. the intact step). All parameters were separately analyzed for each of the three strides collected, after which outcomes were averaged to obtain a mean score for each subject and prosthetic foot type. Statistics The differences in push-off work of the prosthetic foot, step length, step length symmetry, v CoM and MoS BW at toe-off between walking with ESAR and SACH foot were analyzed using paired samples t-tests. Differences in the changes in v CoM and MoS BW during double support, from heel contact to toe-off, between ESAR and SACH were analyzed using two-way ANOVA. Significance level was set a-priori at p -value < 0.05. Results All participants succeeded to walk comfortably on both ESAR and SACH foot. Walking speed in both foot condi- tions was on average 1.22 ± 0.02 m ∙ s − 1 . Stride length did not differ between condition (1.38 ± 0.06 vs. 1.37 ± 0.07 for ESAR vs. SACH). Push-off power of the prosthetic foot was significantly higher while walking with the an ESAR foot compared to a SACH foot (Fig. 2 ) . This resulted in an increase of push-off work of 120% when walking with the ESAR foot compared to SACH (0.11 ± 0.03 vs. 0.05 ± 0.02 J ∙ kg − 1 for ESAR and SACH resp., p < 0.001)(Fig. 3 ) , as was previ- ously reported by Wezenberg et al. [ 9 ] . Step length of the intact step was larger when walking with the ESAR foot compared to the SACH foot (0.68 ± 0.03 vs. 0.66 ± 0.04 m, p = 0.004) (Fig. 3 ) . This increase in intact Fig. 2 Push-off power of the prosthetic foot as a function of normalized stance time. The ESAR foot (red) generates negative power, storing elastic energy, in midstance and generates a higher positive push-off power, returning, more elastic energy during push-off compared to the SACH foot (green). The coloured surface below the power profile indicates the amount of work delivered during push-off (Figure amended from Wezenberg et al. 2014 [ 9 ] ) Houdijk et al. Journal of NeuroEngineering and Rehabilitation 2018, 15 (Suppl 1):76 Page 44 of 72