Adaptation of the preferred human bouncing pattern toward the metabolically optimal frequency.

Humans often appear to prefer movement patterns that minimize the metabolic energy expenditure of performing a task. However, it is not clear whether this preference is dependent on adaptation to feedback or results from a previously learned motor plan. We recently found that for a bouncing task with an identifiable neuromechanical resonant frequency, humans do not initially prefer to bounce at the resonant frequency despite its presumed metabolic benefits. The purpose of the present study was to determine whether humans adapt their preferred bouncing frequency over time to approach the metabolic optimum. Subjects (n = 12) performed a series of 6-min reclined bouncing trials while we quantified bounce frequency, metabolic rate, and rate of positive mechanical work performed on the body. In one trial, subjects bounced at their preferred frequency. In five other trials, subjects bounced at frequencies prescribed by a metronome to match specific percentages of their resonant frequency (80-120%). Positive mechanical work rate was held constant across trials by having subjects match real-time visual feedback to a target. The metabolic rate was lowest during prescribed frequency trials near resonance, not during the preferred frequency trial when subjects were free to choose the bouncing frequency. While the initial preferred frequency was lower than the resonant frequency, the preferred frequency gradually approached resonance over the course of 6 min. These results provide evidence that humans do not choose their preferred movement pattern based on an unchanging learned motor plan, but instead adapt their preferred frequency in response to feedback. Our findings may have implications for clinical populations, as reduced sensory acuity could prevent identification of the metabolically optimal movement pattern.