Restitution and compensation in the recovery of function in the lower extremities of stroke survivors: design of evaluation and training methods

Since recovery of walking and balance control is an important aspect in the rehabilitation of stroke patients, insight in the responsible recovery mechanisms is of great importance. Recovery can be achieved through restitution of original movement patterns and function in the affected leg and/or through compensation of the non affected leg. The first aim of this thesis was to develop and evaluate methods which can be used to distinguish between restitution and compensation in the recovery of function in the lower extremities of stroke survivors. We developed two methods to assess the function of each separate leg in balance control. One method was based on identifying the underlying stabilizing mechanisms of each leg, whereas the other one was based on quantifying the stabilizing effect of the generated corrective torques of each leg. Evaluation of these methods in a cross sectional study in chronic stroke survivors showed that the affected leg contributed significantly less to balance control than the non affected leg. A small paretic contribution to balance control did not keep the stroke survivors from achieving a good function balance, which indicated that the non paretic leg can compensate for the loss of function in the paretic leg.
The second aim of this thesis was to provide a solid scientific basis for using assist-as-needed algorithms in robot-aided gait training with the new robotic gait trainer LOPES, which we developed. In assist-as-needed algorithms the provided robotic assistance is adapted to the capabilities of the subject. Ultimately this kind of algorithm can be used in gait training with LOPES that specifically emphasizes the use of original (restitution) or alternative (compensation) movement strategies. Based on the effectiveness of these interventions the importance of the different recovery mechanisms can be inferred. In this thesis, we provided a basis for using these algorithms. First, we showed the need to adapt the provided assistance to achieve optimal motor learning. Assisting forces reduced the rate and amount of learning of a new motor task. Second, we demonstrated that near-to-normal walking was possible in the LOPES gait training device, which is a requisite for implementing assist-as-needed algorithms.