Use of musculoskeletal modeling
As described above, musculoskeletal models can be used to estimate muscle-tendon lengths during gait, which allows evaluating gait at the level of impairments. Several musculoskeletal models exist in the literature that can be used to calculate muscle-tendon lengths during dynamic activity such as walking, using 3D kinematic data as input. These models can be incorporated in commercial packages such as SIMM (Delp et al., 1990), AnyBody (www.AnyBodyTech.com), or more recently in open-source software packages such as OpenSim (Delp et al., 2007).
The SIMM musculoskeletal model (shown in Figure 1.4) is one of the most wide-spread models used for calculation of muscle-tendon lengths in the literature. This 3-dimensional model has been developed by Delp et al. (1990; 1995) and used previously for the calculation of muscle-tendon lengths in CP. The full body model contains 86 degrees of freedom, 117 joints, and 344 muscle-tendon actuators. The joints have anatomically accurate kinematics, for example the knee model includes the sliding and rolling of the tibia and patella on the femur. Muscle paths are modeled using anatomical via-point and wrapping surfaces. The
General introduction
21 model can be scaled to individual subject sizes, and can
be made to match measured patient gait kinematics, in order to calculate muscle-tendon length and velocity during gait.
In Chapter 2 of this thesis, three different models for the calculation of hamstrings length will be compared in a validation study. Their accuracy to calculate peak hamstrings length will be evaluated at a range of combinations of hip and knee angles. In Chapter 3 to 6 of this thesis, the SIMM lower extremity model, scaled to individual subjects sizes, will be used to calculate semitendinosus, biceps femoris, psoas, gastrocnemius, and soleus lengths during gait.
Use of healthy subjects as a model
One method to investigate specific aspects of pathological gait in isolation is to use healthy subjects as a model, by letting them simulate one specific aspect of pathological gait, or by imposing one specific ‘impairment’. This allows studying these particular aspects in isolation, not including any other impairments or gait deviations. Several studies have investigated the effects of voluntary toe-walking (Davids et al., 1999; Riley and Kerrigan, 2001; Romkes and Brunner, 2007), voluntary crouch walking (Harlaar, 2003), or of imposed shortened hamstrings length (Matjacic and Olensek, 2007; Whitehead et al., 2007) in healthy subjects.
Chapter 3 of this thesis will adopt this approach in a study on voluntary crouch gait. This chapter will address the question whether crouch gait per se coincides with short muscle-tendon length or slow stretch velocity of hamstrings muscles during gait. This could give indirect evidence for the possible effect of contractures (short peak length) and spasticity (slow peak velocity) on crouch gait. This chapter also addresses the relative effect of crouch gait and variation of walking speed on hamstrings length and velocity.
Modulating the effect of spasticity by varying walking speed
Another way to study the effect of a specific impairment on gait is to modulate this impairment and investigate the effect. Intervention studies use this approach, for example by reducing muscle excitation with botulinum toxin (Scholtes et al., 2007b), or increasing muscle strength with a strength training program (Dodd et al., 2002), and evaluating the effect on the gait pattern. However, these studies are time-consuming, and it is often difficult to interfere in only one specific impairment. Since spasticity is defined as a velocity-dependent phenomenon, the effects of spasticity in particular can also be modulated by imposing different walking speeds. Increasing walking speed could be expected to increase the velocity with which muscles are stretched, thereby enhancing the effects of spasticity.
&ŝŐƵƌĞϭ͘ϰ͘'ƌĂƉŚŝĐĂůĚŝƐƉůĂLJŽĨƚŚĞ
^/DD ŵŽĚĞů ƵƐĞĚ ŝŶ ŚĂƉƚĞƌ ϮͲϲ ŽĨƚŚŝƐƚŚĞƐŝƐ͘
Chapter 4 to 6 of this thesis will use this approach to try and gain a better understanding of the specific effects of spasticity on gait, as well as of the effects of walking speed itself.
Chapter 4 evaluates the relationship between hamstrings spasticity as measured during a clinical spasticity test and peak hamstrings length and lengthening velocity during gait, for a range of walking speeds. Chapter 5 evaluates muscle-tendon length and lengthening velocity during gait for the gastrocnemius and soleus muscles. Spastic calf muscles with and without contractures in children with CP will be compared to muscles in typically developing children.
Measuring spasticity directly during gait
Instead of correlating spasticity as measured during passive testing to gait parameters, it is also possible to study spasticity effects directly during gait. This can be done by studying the
‘velocity-dependent increase in muscle tone’ during gait, by relating muscle activity to muscle-tendon stretch velocity. This method, termed dynamic spasticity, has been proposed by Crenna (1998) and will be investigated in Chapter 6 of this thesis. In this chapter, the dynamic spasticity of the plantar flexor muscles is investigated, by relating phases of muscle-tendon stretch to muscle activity.
Forward dynamic modeling
A different approach that is particularly suitable to study specific aspects of pathological gait in isolation is the use of forward dynamic simulation of gait. Forward dynamic simulation follows the natural way of causality: the gait pattern is ‘synthesized’ by applying forces or moments to a biomechanical model of the human body and observing the resulting movement. This allows answering hypothetical ‘what if’ questions, by changing one or more of the parameters of the model and evaluating the resulting output. In the literature, two main approaches exist to forward dynamic modeling of gait: one using complex musculoskeletal models (similar to the abovementioned models for muscle-tendon length calculation), in which the gait data of subjects are ‘tracked’ to achieve a forward dynamic simulation based on inverse analysis, (e.g. Delp et al., 2007). Opposed to this more complex approach is the so-called dynamic walking approach, which uses simpler, conceptual models that allow studying basic principles of human gait in a more fundamental manner. Since these models synthesize a gait pattern that is repeatable (i.e. it can produce perpetual, stable gait), they are also called limit cycle walking models. In Chapter 7 of this thesis, the latter approach will be applied to study factors that may lead to a stiff-knee gait pattern. This chapter presents a forward dynamic model of normal and crouch gait. Using this model, the effects of a crouched posture, as well as the effects of push-off strength and hip torque on the dynamics of the swing leg are studied.
In Chapter 8 the main findings of this thesis will be summarized and discussed. This chapter provides an evaluation of the methods used, reflects on the fundamental and clinical implications of this research, and gives some recommendations for further study.