ESTIMATION OF ANTERIOR TIBIAL TRANSLATION AND LIGAMENT LOADING IN HEALTHY AND ACL-DEFICIENT KNEES DURING WALKING 1 Qi Shao, 1 Toran D. MacLeod, 1 Kurt Manal and 1 Thomas S. Buchanan 1 Center for Biomedical Engineering Research University of Delaware, Newark, DE, USA email: shao@udel.edu , web: http:// www.cber.udel.edu INTRODUCTION Knowledge of internal knee-ligament loading is important for developing better surgical procedures and rehabilitation regimens of ACL-deficient patients. A few numerical models have been used to estimate ligament loading during walking [1,2]. However these models were not driven by measured EMGs, so they might have difficulty predicting the abnormal muscle activation patterns of ACL- deficient knees. In this paper we describe an EMG-driven model that incorporates a knee-ligament model, and we apply this approach to estimate anterior tibial translation (ATT), anterior shear forces, and ligament loading in the knee joints of an ACL- deficient subject and a healthy subject during walking. The results of the ACL-deficient gait will be compared with those of the healthy gait to explain how the ACL-deficient subject compensated for the loss of the ACL. METHODS One male healthy subject (mass 60.5 kg, height 1.70 m) and one male subject without an ACL (mass 74.0 kg, height 1.72 m; right leg was the affected leg) gave informed consent before participating in this study. The experimental protocol was approved by the Human Subjects Review Board of the University of Delaware. The subjects were required to finish four walking trials with right foot striking the force plate, and another four walking trials with left foot striking the force plate. EMG, joint position and force plate data were collected during the trials. EMGs were collected from nine muscles of the leg of interest using surface electrodes, including RF, VL, VM, SM, BFL, MG, LG, Sol, and TA [3]. In this study we simulated the stance phase of both knees of the ACL-deficient subject, and the right knee of the healthy subject. We calculated knee-ligament forces through a two- step procedure. First, an EMG-driven model [3,4] was used to estimate the muscle forces of the leg of interest to match the inverse dynamics calculated knee and ankle joint moments during stance phase of the walking trials. Figure 1: The knee joint model. Second, a knee model that incorporated knee- ligaments was developed to calculate ATT and ligament loading. The model included the tibiofemoral and patellofemoral joints in the sagittal plane, and three segments: the femur, tibia and patella (Fig. 1A). The muscle forces and joint reaction forces calculated from the previous step were used as inputs. The approach used to model the patellofemoral joint was similar to that used by Liu and Maitland [2]. The contact and geometric compatibility conditions were required to be satisfied at the tibiofemoral joint. Knee ligaments were modeled as nonlinear elastic elements [1]. The ATT, knee joint contact force and ligament forces were solved through iterations until the force equilibriums of the tibia were satisfied (Fig. 1B).