Changes in Predicted Muscle Coordination with Subject-Specific Muscle Parameters for Individuals after Stroke

Hindawi Publishing Corporation Stroke Research and Treatment Volume 2014, Article ID 321747, 7 pages http://dx.doi.org/10.1155/2014/321747 Research Article Changes in Predicted Muscle Coordination with Subject-Specific Muscle Parameters for Individuals after Stroke Brian A. Knarr,1 Darcy S. Reisman,2 Stuart A. Binder-Macleod,2 and Jill S. Higginson3 1 Delaware Rehabilitation Institute, STAR Health Sciences Complex, University of Delaware, 540 S. College Avenue, Newark, DE 19716, USA 2 Department of Physical Therapy, STAR Health Sciences Complex, University of Delaware, 540 S. College Avenue, Newark, DE 19716, USA 3 Department of Mechanical Engineering, STAR Health Sciences Complex, University of Delaware, 540 S. College Avenue, Newark, DE 19716, USA Correspondence should be addressed to Brian A. Knarr; Received 3 April 2014; Accepted 6 June 2014; Published 25 June 2014 Academic Editor: Steve Kautz Copyright © 2014 Brian A. Knarr et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Muscle weakness is commonly seen in individuals after stroke, characterized by lower forces during a maximal volitional contraction. Accurate quantification of muscle weakness is paramount when evaluating individual performance and response to after stroke rehabilitation. The objective of this study was to examine the effect of subject-specific muscle force and activation deficits on predicted muscle coordination when using musculoskeletal models for individuals after stroke. Maximum force generating ability and central activation ratio of the paretic plantar flexors, dorsiflexors, and quadriceps muscle groups were obtained using burst superimposition for four individuals after stroke with a range of walking speeds. Two models were created per subject: one with generic and one with subject-specific activation and maximum isometric force parameters. The inclusion of subject-specific muscle data resulted in changes in the model-predicted muscle forces and activations which agree with previously reported compensation patterns and match more closely the timing of electromyography for the plantar flexor and hamstring muscles. This was the first study to create musculoskeletal simulations of individuals after stroke with subject-specific muscle force and activation data. The results of this study suggest that subject-specific muscle force and activation data enhance the ability of musculoskeletal simulations to accurately predict muscle coordination in individuals after stroke. 1. Introduction Musculoskeletal simulations have the potential to provide insight into muscle coordination and function for individuals with gait deficits. Previous musculoskeletal simulations have shown how muscle coordination can be altered based on changes in muscle properties [1–4]. A current limitation of musculoskeletal simulations, however, is that the appropriate muscle properties to use for a specific individual are unknown. For a particular subject or population (e.g., stroke), muscle parameters may differ greatly from default model values, and it has been suggested that selection of muscle parameters can have a relevant impact on simulation results [5–7]. Muscle weakness, characterized by lower forces during a maximal volitional contraction, is a major limiting factor affecting performance of poststroke gait [8]. The two main causes of poststroke muscle weakness are disuse atrophy [9] and impaired muscle activation by the central nervous system [10]. Studies have shown a reduction in skeletal muscle mass and an increase in intramuscular fat in the paretic limb of stroke survivors [9, 11]. Additionally, electromyography (EMG) has been used to demonstrate activation impairment in stroke survivors, with measured EMG amplitude lower on the paretic side muscles compared to the nonparetic side [12]. More recently, studies have used the burst superimposition technique, which applies electrical stimulation superimposed over a volitional contraction, to measure subject-specific 2 maximum force generation ability and volitional activation ratio of muscles for healthy and poststroke populations [13– 16]. A study by Xiao and Higginson (2010) explored the sensitivity of a musculoskeletal model to changes in muscle parameters, showing that predicted muscle forces are sensitive to values of tendon slack length, optimal fiber length, and differences greater than 10% in maximum isometric force [3]. Strength deficits seen after stroke are often in excess of 10%, with previous studies reporting paretic side voluntary moment 80% less than nonparetic force for some individuals [15–18]. Since variation of muscle properties influences muscle force and coordination, it is possible that the inclusion of relevant muscle parameters in musculoskeletal models will lead to more accurate and meaningful results in persons after stroke. It has been shown in previous work that model predictions of muscle coordination are altered when muscle weakness is simulated [1, 4]; however, these studies only involved randomly imposed weakness to healthy simulations. To date, no studies have built subject-specific musculoskeletal models which include experimentally measured values for muscle weakness from a clinical population such as individuals after stroke. Therefore, the objective of this study was to examine the effect of subject-specific muscle force and activation deficits on muscle coordination when using musculoskeletal models for individuals after stroke. Threedimensional subject-specific musculoskeletal models were built using experimental gait data from subjects after stroke. Two simulations were created per subject, one using generic and one using subject-specific isometric force and maximum volitional activation model parameters based on experimentally measured data. We hypothesized that subject-specific activation and muscle force data would result in altered predicted muscular control patterns that are consistent with muscle compensation strategies that have been reported in both modeling and clinical studies. Additionally, we hypothesized that the timing of the subject-specific activations predicted by the musculoskeletal model would agree better with the timing of experimentally recorded electromyography measured during gait when subject-specific model parameters were used. 2. Methods Four individuals after stroke (65 ± 8 yrs, 9 ± 4 months after stroke) were recruited to participate in this study. Subjects were included in this study if they met the following criteria: 6 months after a stroke involving cerebral cortical regions, being able to walk for 5 minutes at self-selected speed without a brace or assistive device, passive paretic ankle dorsiflexion range of motion to reach at least 5∘ of plantar flexion with the knee flexed, and presence of deficits in walking (...truncated)