Nonlinear time series analysis was used to estimate maximal Lyapunov exponents of select ankle and knee kinematics during three different conditions of treadmill walking: independent, side by side, and side by side with forced synchronization of stepping. Stride to stride variability was significantly increased for the condition in which individuals walked side by side and synchronized unintentionally when compared to the conditions of forced synchronization and independent walking. In addition, standard deviations of three kinematic variables of lower extremity movement were significantly increased during the condition in which unintentional synchronization occurred. No relationship was found between standard deviation and estimates of maximal Lyapunov exponents. An increase in kinematic variability during side by side walking for nonimpaired individuals who are not at risk of falling suggests that variability in certain aspects of performance might be indicative of a healthy system. Modeling this variability for an impaired individual to imitate may have beneficial effects on locomotor function. These results may therefore have implications for the rehabilitation of gait in humans by suggesting that a different functional outcome might be achieved by practicing side by side walking as opposed to more commonly used strategies involving independent walking.

Sensory motor synchronization can be used to alter gait behavior. This type of therapy may be useful in a rehabilitative setting, though several questions remain regarding the most effective way to promote and sustain synchronization. The purpose of this study was to describe a new technique for using synchronization to influence a person’s gait and to compare walking behavior under this paradigm with that of side by side walking. Thirty one subjects walked on a motorized treadmill that was placed on a platform that oscillated vertically at various frequencies and amplitudes. Synchronization with the platform and stride kinematics were recorded during these walking trials and compared with previously reported data from side by side walking. The results indicated that vertical oscillation of the treadmill surface at frequencies that matched subjects preferred stride or step frequency resulted in greater unintentional synchronization when compared with side by side walking data (up to 78.6±8.3% of the trial vs 59.2±17.4%). While intermittent phase locking was observed in all cases, periods of synchronization occurred more frequently and lasted longer while walking on the oscillating treadmill (mean length of periods of phase locking 11.85 steps vs 5.18 steps). Further, stride length, height and duration were altered by changing the frequency of treadmill oscillation. These results suggest that synchronization to a haptic signal may hold implications for use in a clinical setting.