The effect of the different training regimes and histories on the spatiotemporal characteristics of human running was evaluated in four groups of subjects who had different histories of engagement in running-specific training; sprinters, distance runners, active athletes, and sedentary individuals. Subjects ran at a variety of velocities, ranging from slowest to fastest, over 30 trials in a random order. Group averages of maximal running velocities, ranked from fastest to slowest, were: sprinters, distance runners, active athletes, and sedentary individuals. The velocity-cadence-step length (V-C-S) relationship, made by plotting step length against cadence at each velocity tested, was analyzed with the segmented regression method, utilizing two regression lines. In all subject groups, there was a critical velocity, defined as the inflection point, in the relationship. In the velocity ranges below and above the inflection point (slower and faster velocity ranges), velocity was modulated primarily by altering step length and by altering cadence, respectively. This pattern was commonly observed in all four groups, not only in sprinters and distance runners, as has already been reported, but also in active athletes and sedentary individuals. This pattern may reflect an energy saving strategy. When the data from all groups were combined, there were significant correlations between maximal running velocity and both running velocity and step length at the inflection point. In spite of the wide variety of athletic experience of the subjects, as well as their maximum running velocities, the inflection point appeared at a similar cadence (3.0 ± 0.2 steps/s) and at a similar relative velocity (65-70%Vmax). These results imply that the influence of running-specific training on the inflection point is minimal.

PURPOSE: The purpose of this study was to investigate the relationship between spatiotemporal variables of running and onset/offset timing of rectus femoris (RF) and biceps femoris (BF) muscle activities in both legs. METHODS: Eighteen male well-trained athletes (age = 20.7 ± 1.8 yr) were asked to run 50 m at maximal speed. The spatiotemporal variables (running speed, step frequency, and step length) over the distance from 30 to 50 m were measured. In addition, RF and BF muscle activities were obtained from both legs using wireless EMG sensors. To quantify the onset and offset timing of muscle activity, the band-pass filtered (20-450 Hz) EMG signal was processed using a Teager-Kaiser energy operator filter. We calculated RF and BF onset/offset timings (%) in both legs (e.g., ipsilateral leg RF [iRF] and contralateral leg BF [cBF]) during running cycle. Based on those timings, we obtained the EMG timing variables (%) as follows: "Switch1 (iBF-offset to iRF-onset)," "Switch2 (iRF-offset to iBF-onset)," "Scissors1 (cBF-onset to iRF-onset)," and "Scissors2 (iRF-offset to cBF-offset). RESULTS: We found that "Switch2" had positive (r = 0.495, P = 0.037), "Scissors1" had negative (r = -0.469, P = 0.049), and "Scissors2" had positive (r = 0.574, P = 0.013) correlations with step frequency. However, these variables had no significant correlations with running speed or step length. CONCLUSIONS: These results indicate that higher step frequency would be achieved by smoother switching of the agonist-antagonist muscle activities and earlier iRF activation relative to the cBF activity. To improve sprint performance, athletes and coaches should consider not only muscle activities in one leg but also coordination of muscle activities in both legs.

This paper proposes a method that can classify foot strike types using a deep learning model and can extract unknown factors, which enables to evaluate running motions without being influenced by biases of sports experts, using the contribution degree of input values (CDIV). Accelerometers are attached to the runner's body, and when the runner runs, a fixed camera observes the runner and acquires a video sequence synchronously with the accelerometers. To train a deep learning model for classifying foot strikes, we annotate foot strike acceleration data for RFS (Rearfoot strike) or non-RFS objectively by watching the video. To interpret the unknown factors extracted from the learned model, we calculate two CDIVs: the contributions of the resampling time and the accelerometer value to the output (foot strike type). Experiments on classifying unknown runners' foot strikes were conducted. As a common result to sport science, it is confirmed that the CDIVs contribute highly at the time of the right foot strike, and the sensor values corresponding to the right and left tibias contribute highly to classifying the foot strikes. Experimental results show the right tibia is important for classifying foot strikes. This is because many of the training data represent difference between the two foot strikes in the right tibia. As a conclusion, our proposed method could extract unknown factors from the classifier and could interpret the factors that contain similar knowledge to the prior knowledge of experts, as well as new findings that are not included in conventional knowledge.