Crouch gait, one of the most common movement abnormalities among children with cerebral palsy, is characterized by persistent flexion of the knee during the stance phase. Short hamstrings are thought to be the cause of crouch gait; thus, crouch gait is often treated by surgical lengthening of the hamstrings. In this study, a graphics-based model of the lower extremity was used in conjunction with three-dimensional kinematic data obtained from gait analysis to estimate the lengths of the hamstrings and psoas muscles during normal and crouch gaits. Only three of 14 subjects with crouch gait (four of 20 limbs with knee flexion of 20 degrees or more throughout stance) had hamstrings that were shorter than normal by more than 1 SD during walking. Most (80%) of the subjects with crouch gait had hamstrings of normal length or longer, despite persistent knee flexion during stance. This occurred because the excessive knee flexion was typically accompanied by excessive hip flexion throughout the gait cycle. All of the subjects with crouch gait had a psoas that was shorter than normal by more than 1 SD during walking. These results emphasize the need to consider the geometry and kinematics of multiple joints before performing surgical procedures aimed at correcting crouch gait.

We used multivariate kinematics and joint torque measurements during dynamic posturography to determine the relative contributions of changes in overall control gain, relative weighting of sensors, and noise-like effects on posture control in the elderly. Our results show that sway coordination and amplitude both change with age, but that changes in overall feedback gains do not explain these differences. We propose that increased sway of elderly subjects in platform sway-referenced conditions is due to sensory noise or decreased ability to detect small motions of the platform, while increased sway during visual sway-referencing is due to re-weighting of the various sensors.

Multivariate descriptors of sway were used to test whether altered sensory conditions result not only in changes in amount of sway but also in postural coordination. Eigenvalues and directions of eigenvectors of the covariance of shnk and hip angles were used as a set of multivariate descriptors. These quantities were measured in 14 healthy adult subjects performing the Sensory Organization test, which disrupts visual and somatosensory information used for spatial orientation. Multivariate analysis of variance and discriminant analysis showed that resulting sway changes were at least bivariate in character, with visual and somatosensory conditions producing distinct changes in postural coordination. The most significant changes were found when somatosensory information was disrupted by sway-referencing of the support surface (P = 3.2 x 10(-10)). The resulting covariance measurements showed that subjects not only swayed more but also used increased hip motion analogous to the hip strategy. Disruption of vision, by either closing the eyes or sway-referencing the visual surround, also resulted in altered sway (P = 1.7 x 10(-10)), with proportionately more motion of the center of mass than with platform sway-referencing. As shown by discriminant analysis, an optimal univariate measure could explain at most 90% of the behavior due to altered sensory conditions. The remaining 10%, while smaller, are highly significant changes in posture control that depend on sensory conditions. The results imply that normal postural coordination of the trunk and legs requires both somatosensory and visual information and that each sensory modality makes a unique contribution to posture control. Descending postural commands are multivariate in nature, and the motion at each joint is affected uniquely by input from multiple sensors.

No abstract