Modeling and Simulation of Normal and Pathological Gait

The diagnosis and treatment of gait abnormalities in children with cerebral palsy is challenging. A combination of several factors, including muscle spasticity, muscle weakness, bony malalignment, and neurological impairment may contribute to a patient’s movement abnormality. In theory, correcting these factors with the appropriate treatment will improve the patient’s gait pattern. Identifying the set of factors to target with treatment is difficult, however, since the abnormal gait pattern and set of contributing factors differ between patients. Further, the human body is a complex 3-dimensional linkage, and consequently muscles often have non-intuitive roles during locomotion that are difficult to discern from examining electromyographs (EMG) and joint motions (Fig. 3.6.1). Modeling and simulation of the musculoskeletal system is a powerful tool for quantifying muscle function during pathological gait, which can in turn help identify why a specific patient walks with an abnormal gait, and enable us to design an appropriate treatment plan. ‘Modeling’, in the context of gait, is a term that often conjures images of 3-dimensional musculoskeletal models and complex dynamic simulations. It is important to recognize that the term ‘model’ simply refers to a set approximations used to represent a system of interest – in this case, the human body. For example, a model of the musculoskeletal system is needed to perform a conventional gait analysis. The process of inverse dynamics, which calculates joint angles and moments during gait (Davis et al. 1991), requires a model of the body that represents the rotational axes of the joints and the inertial properties of the body segments (i.e. masses and moments of inertia). Recent advances in the field of biomechanics allow us to extend this model in several ways to rigorously define the roles of skeletal alignment, muscle activations, muscle–tendon dynamics, and other factors in generating a normal gait pattern. We can first extend the model used for inverse dynamics with a representation of 3-dimensional musculoskeletal geometry (Fig. 3.6.1, orange box). We represent each muscle as a path or set of paths between the muscle’s origin and insertion, possibly with wrapping surfaces or via points to approximate the more complex geometries of tendon sheaths or overlapping muscles (Delp et al. 1990, Van der Helm et al. 1992). With this addition to the musculoskeletal model, we can calculate muscle moment arms, lengths and velocities as a function of a subject’s joint kinematics (Hoffinger et al. 1993, Delp et al. 1996, Schutte et al. 1997, Thompson et al. 1998). This type of model enables us to quantify, for example, the lengths and velocities of a patient’s hamstrings muscles to determine if the muscles are