The largest Lyapunov exponent (LyE) is a metric that quantifies the stride-to-stride fluctuations that naturally occur in walking, where a larger value indicates greater fluctuations or a less consistent gait pattern. The LyE has previously been found to be increased in transtibial amputees and strongly related to prosthesis preference. However, the effect of prosthesis type and adaptation on stride-to-stride fluctuations is unknown. The purpose of this study was to determine the effects of prosthesis type and adaptation on stride-to-stride fluctuations as measured by the LyE.
Method:
21 transtibial amputees consented to participate in this IRB approved randomized, crossover design study. Subjects were randomly assigned to begin a 3 week adaptation period with either an appropriate or less appropriate prosthesis according to their K-level. Subjects walked on a treadmill at their self-selected pace for 3 minutes (12 cameras, 60 Hz; Motion Analysis Corp., Santa Rosa, CA). Subjects then returned at the middle and at the end of the adaptation period for similar testing. This was then repeated with the other prosthesis design for another 3 week period. Joint angle time series were analyzed using the LyE. Main effects for leg (prosthetic vs. sound), prosthesis (more vs. less appropriate), and adaptation (visit 1 vs. visit 2 vs. visit 3) at the hip, knee, and ankle were tested through a 2x2x3 fully repeated ANOVA (α=0.05) with Fisher’s LSD for post-hoc. An analysis of trend was performed for adaptation effects over the 3 weeks.
Analysis/Results:
At the ankle, the LyE was significantly reduced for the sound leg compared to the prosthetic leg (F1,20=10.52, p=0.004; Fig 1). The more appropriate prosthesis resulted in a reduced LyE (F1,20=7.357, p=0.013). The effect of adaptation was also significant (F2,40=3.640, p=0.035). Post-hoc tests showed increased LyE at visit 1 compared to visit 2 (p=0.047) and at visit 3 compared to visit 2 (p=0.014). This resulted in a significant quadratic trend (p=0.016). No differences were found at the knee and hip.
Conclusions:
The more appropriate prosthesis design resulted in a more consistent gait pattern. Interestingly, adaptation seemed to have an effect similar to other studies investigating motor skill development with an initial freezing of the available degrees of freedom during learning followed by a release of these degrees of freedom. LyE at initial fitting was not different from the LyE following an adaptation period. The LyE is able to distinguish between an appropriate and less appropriate prosthesis design.
See more of: Research