Are running speeds maximized with simple-spring stance mechanics?

Are the fastest running speeds achieved using the simple-spring stance mechanics predicted by the classic spring-mass model? We hypothesized that a passive, linear-spring model would not account for the running mechanics that maximize ground force application and speed. We tested this hypothesis by comparing patterns of ground force application across athletic specialization (competitive sprinters vs. athlete nonsprinters, n = 7 each) and running speed (top speeds vs. slower ones). Vertical ground reaction forces at 5.0 and 7.0 m/s, and individual top speeds (n = 797 total footfalls) were acquired while subjects ran on a custom, high-speed force treadmill. The goodness of fit between measured vertical force vs. time waveform patterns and the patterns predicted by the spring-mass model were assessed using the R-2 statistic (where an R-2 of 1.00 = perfect fit). As hypothesized, the force application patterns of the competitive sprinters deviated significantly more from the simple-spring pattern than those of the athlete, nonsprinters across the three test speeds (R-2 <0.85 vs. R-2 >= 0.91, respectively), and deviated most at top speed (R-2 = 0.78 +/- 0.02). Sprinters attained faster top speeds than nonsprinters (10.4 +/- 0.3 vs. 8.7 +/- 0.3 m/s) by applying greater vertical forces during the first half (2.65 +/- 0.05 vs. 2.21 +/- 0.05 body wt), but not the second half (1.71 +/- 0.04 vs. 1.73 +/- 0.04 body wt) of the stance phase. We conclude that a passive, simple-spring model has limited application to sprint running performance because the swiftest runners use an asymmetrical pattern of force application to maximize ground reaction forces and attain faster speeds.