By studying how net joint moments and individual muscle forces contribute to the control of human movement, we intend to link the capabilities of a physically challenged person to his or her ability to perform activities of daily living. Measuring these contributions is critical for defining adaptive movement control strategies, understanding the process by which movement control strategies are selected and implemented, and determining the contribution that assistive technologies play in enhancing function. We have extended existing movement analysis methodologies that estimate the influence of joint moments or muscle forces on the motion of individual joints and body segments, as well as their contribution to overall task performance. Recently this work has concentrated on five project areas: 1) clinical application of joint-moment-based induced acceleration analysis, 2) induced acceleration analysis for model verification, 3) implementation of induced acceleration techniques within commercial software, 4) application of muscle-based induced acceleration analysis in a clinical setting, and 5) expansion of the current PDB models. Earlier efforts to determine the contribution of net muscular joint moments and individual muscle forces to human movement control have been limited by the assumption that the relative contribution of muscular effort to the control of movement is independent of the position of the body segments. A generalized coordinate formulation of the equations of motion shows that this assumption is incorrect and can lead to errors in interpreting human movement analysis data. Our past research revealed that the generation of forward velocity during normal gait arises primarily from ankle joint plantar flexor push-off, rather than through a controlled fall. We found that the ankle joint plantar flexor muscle group produces forward acceleration of the upper body during periods in which the same muscle group acts eccentrically. We examined the relative contribution of the lower extremity joint moments to the maintenance of upright posture (support) and found that, during the single limb support phase, ankle moments generate the greatest contribution to support. We also examined the relative sensitivity of the joint accelerations to the joint moments during the foot-flat phase of gait and found that the ankle, knee and hip joint accelerations are almost twice as sensitive to moments generated at the knee than to the moments at any other lower extremity joint. During the heel-off phase of gait, the acceleration sensitivities at both the ankle and hip joints increase in response to moments at their own joint. It is clear from these data that there is significant redundancy in controlling the motion of the stance limb. Application of the model to patients with severe muscle weakness has shown how they can exploit this redundancy to develop a variety of alternative movement control strategies during gait to compensate for lost muscle function. We have further extended the model used in these analyses to estimate the sources of mechanical energy exchange between segments, and have used it to determine the muscle groups responsible for the inter-segment transfer of mechanical energy during gait. We found that the energy transfer depends on the sign of the joint moment, rather than on the sign of the joint power, and that pairs of net joint moments function in combination to balance energy flow through the leg and trunk segments during normal gait. Present efforts are also focused on implementing a 3D biomechanical model to determine the contributions of individual muscles to controlling pathologic mootion during functional tasks. The model, developed by Rick Neptune at the University of Texas, is being incorporated into the Visual3D software (C-Motion, Inc,) under a joint grant between C-Motion, Inc., the University of Texas and the Physical Disabilities Branch. The first phase of the grant has involved restructuring the Visual3d software to separate the induced acceleration analysis computations into dynamic link libraries (dlls) that offer the analysis based on any one of three separate dynamics engines (SD/Fast from PTC Software, ADAMS from MSC software, and MAMBO developed by Harry Dankowiez at the Virginia Tech). The Mambo dll is an entirely new option which, in addition to reducing processing time compared to the ADAMS dll, will require no additional software licenses when supplied by C-Motion to clinical laboratories. The work on the induced acceleration dlls, which is nearing completion, allows users to select the dynamics engine which best suits their computer needs and financial constraints. Further work by the PDB on this project has included the development and implementation of a global optimization method for computing the position and orientation of anatomical segments during movement in a manner that will be mechanically consistent with the Neptune model. The Physical Disabilities Branch has also begun working with the Neptune model to test its clinical applicability. This work has been focused in three areas: 1) testing on a single clinical case to determine an appropriate cost function for estimating the muscle forces associated with the subject?s measured movement, 2) preparation of ten clinical cases for input into the model and 3) development of a Windows-based computer cluster to reduce the computer processing time associated with the model. Early testing of the model on a clinical case has shown significant improvement in the ability of the Neptune model to reproduce the motion pattern of a patient whose gait was measured in the PDB motion capture laboratory. This project is ongoing and work with a new 3D model is expected to begin shortly. Additionally, although the implementation of a Windows-based computation cluster has proven to be more difficult than originally anticipated, the PDB is aiming to get the cluster fully operational during the fall of 2005. The foot traditionally has been modeled as a single segment for most analyses of gait, although it is composed of 26 bones. As a result, the usefulness of traditional models for clinical research has been limited. We have expanded our motion capture and analysis methods to create a multi-segment foot model and tested its applicability and reproducibility in healthy feet as well as those that have been affected by rheumatoid arthritis. Future research aims to combine multi-segmental foot motion capture data along with plantar pressure assessment and customized foot skeletal models to create unique three-dimensional visualization and analysis tools to assess normal and impaired foot function during gait.
Showing the most recent 10 out of 12 publications
