The primary footwear components of interest to improve performance are midsole material, weight, and longitudinal bending stiffness. Little is known about the effects of varied longitudinal bending stiffness across a range of running speeds. The purpose of this study was to identify changes in spatiotemporal variables, horizontal ground reaction forces, subjective comfort, and metabolic cost at different running speeds in response to varied longitudinal bending stiffness. Ten highly trained males ran at 14, 17 and 20 km/h in shoes with varying longitudinal bending stiffness (normal 5.9, stiff 10.5 and very stiff 17.0 N-m/rad). Ground reaction forces, metabolics and subjective comfort assessments were collected. There were significant changes (p < .05) in contact time, stride frequency, and stride length between shoe conditions at all three speeds. Peak propulsive force decreased with increased bending stiffness at all three speeds, but there was no change in braking or propulsive impulse. The patterns of changes in stride length and stride frequency were different between speeds. At 14 km/h, most participants elicited a minimum metabolic rate in the normal shoe. However, at 17 km/h an increased number of participants were more economical in the stiff shoe, despite it weighing an extra 50 g compared to the normal shoe. Running speed had an influence on subjective comfort, with participants tending to prefer the normal shoe at 14 km/h and the stiff shoe at 17 km/h. These results suggest that an optimal bending stiffness to reduce metabolic cost and improve comfort may be running speed dependent.

Background: Myoelectric control of upper extremity powered prostheses has been used clinically for many years, however this approach has not been fully developed for lower extremity prosthetic devices. With the advent of powered lower extremity prosthetic components, the potential role of myoelectric control systems is of increasing importance. An understanding of muscle activation patterns and their relationship to functional ambulation is a vital step in the future development of myoelectric control. Unusual knee muscle co-contractions have been reported in both limbs of trans-tibial amputees. It is currently unknown what differences exist in co-contraction between trans-tibial amputees and controls. This study compares the activation and co-contraction patterns of the ankle and knee musculature of trans-tibial amputees (intact and residual limbs), and able-bodied control subjects during three speeds of gait. It was hypothesized that residual limbs would have greater ankle muscle co-contraction than intact and able-bodied control limbs and that knee muscle co-contraction would be different among all limbs. Lastly it was hypothesized that the extent of muscle co-contraction would increase with walking speed.Methods: Nine unilateral traumatic trans-tibial amputees and five matched controls participated. Surface electromyography recorded activation from the Tibialis Anterior, Medial Gastrocnemius, Vastus Lateralis and Biceps Femoris of the residual, intact and control limbs. A series of filters were applied to the signal to obtain a linear envelope of the activation patterns. A co-contraction area (ratio of the integrated agonist and antagonist activity) was calculated during specific phases of gait.Results: Co-contraction of the ankle muscles was greater in the residual limb than in the intact and control limbs during all phases of gait. Knee muscle co-contraction was greater in the residual limb than in the control limb during all phases of gait.Conclusion: Co-contractions may represent a limb stiffening strategy to enhance stability during phases of initial foot-contact and single limb support. These strategies may be functionally necessary for amputee gait; however, the presence of co-contractions could confound future development of myoelectric controls and should thus be accounted for.

Electromyogram (EMG) signal representation is crucial in classification applications specific to locomotion and transitions. For a given signal, classification can be performed using discriminant functions or if-else rule sets, using learning algorithms derived from training examples. In the present work, a spectrogram based approach was developed to classify (EMG) signals for locomotion mode. Spectrograms for each muscle were calculated and summed to develop a histogram. If-else rules were used to classify test data based on a matching score. Prior knowledge of locomotion type reduced class space to exclusive locomotion modes. The EMG data were collected from seven leg muscles in a sample of able-bodied subjects while walking over ground (W), ascending stairs (SA) and the transition between (W-SA). Three muscles with least discriminating power were removed from the original data set to examine the effect on classification accuracy. Initial classification error was <20% across all modes, using leave one out cross validation. Use of prior knowledge reduced the average classification error to <11%. Removing three EMG channels decreased the classification accuracy by 10.8%, 24.3%, and 8.1% for W, W-SA, and SA respectively, and reduced computation time by 42.8%. This approach may be useful in the control of multi-mode assistive devices.

Wearable sensors and machine learning algorithms are becoming a viable alternative for biomechanical analysis outside of the laboratory. The purpose of this work was to estimate gait events from inertial measurement units (IMUs) and utilize machine learning for the estimation of ground reaction force (GRF) waveforms. Sixteen healthy runners were recruited for this study, with varied running experience. Force sensing insoles were used to measure normal foot-shoe forces, providing a proxy for vertical GRF and a standard for the identification of gait events. Three IMUs were mounted on each participant, two bilaterally on the dorsal aspect of each foot and one clipped to the back of each participant’s waistband, approximating their sacrum. Participants also wore a GPS watch to record elevation and velocity. A Bidirectional Long Short Term Memory Network (BD-LSTM) was used to estimate GRF waveforms from inertial waveforms. Gait event estimation from both IMU data and machine learning algorithms led to accurate estimations of contact time. The GRF magnitudes were generally underestimated by the machine learning algorithm when presented with data from a novel participant, especially at faster running speeds. This work demonstrated that estimation of GRF waveforms is feasible across a range of running velocities and at different grades in an uncontrolled environment.