Uncoupling the correlates of locomotor costs: a factorial approach

Although the metabolic cost of running has been found to vary systematically with running speed and body mass in animals, no mechanistic model of cost determination has been widely accepted. Because two suggested cost determinants--mechanical power output and the rate of force application--covary with running speed when gait (stride frequency) is unconstrained, examining the influence of each factor on cost has been problematic. We experimentally uncoupled the usual relationships between three readily manipulated cost correlates--speed, stride frequency, and mass distribution--to study the determinants of the cost of running in humans along a section of the cost response surface. Two levels of each cost correlate were used in a factorial design in which subjects ran at all eight combinations of factor levels while rate of oxygen consumption (VO2), mechanical power, and contact time were measured as response variables. Contact time and mechanical power were measured from high-speed cinefilm. Mechanical power accounted for 88%, and contact time (rate of force application) 57%, of speed-related VO2 variation. Variation in cost produced by changes in stride frequency and mass distribution were largely explained by mechanical power. Contact time was not significantly correlated with overall variation in VO2, while mechanical power output explained 80% of that variation. For the range of conditions tested, we conclude that the mechanical power derived from muscular contractions is an important determinant of the cost of running. While we have shown that the rate at which muscles shorten is not the sole determinant of the cost of running, the extent to which this factor influences cost remains unclear.