INTRODUCTION

Function Electrical Stimulation (FES) that can recruit multiple muscles (e.g., via intrafascicular multisite electrode stimulation (IFMS)) creates an over-determined system of actuators that requires complex control. Computer simulations that yield muscle forces for multijoint movements can reduce the experimental complexity of these large scale control problems. Thus, the purpose of this paper was to assess the validity of the musculoskeletal geometry and force generating capacity of a model developed to determine the timing and activation of muscles required for complex behaviors of the feline hindlimb.

METHODS

Surfaces of bones were reconstructed from CT scans from one left feline hindlimb. The kinematic structure of joints were measured experimentally or assumed to be revolute. Paths of 25 muscles were based on anatomic dissection and MRI sequences of the same specimen. Via points and wrapping structures were added to maintain muscle paths during joint rotations (SIMM, Musculographics). Architectural properties of each muscle were taken from the literature [2, 3, 4]. Reported moment arms [5, 6, 7, 8, 9, 10] were compared to model outputs. Simulated maximal muscle activation across a range of joint angles was used to estimate maximal force- and moment-generation capacities. Maximum isometric moment-generating production of the knee extensors was measured during maximal electrical stimulation experiments and compared to those predicted by the model.

RESULTS AND DISCUSSION

Model, experimental, and reported estimates of muscle moment arms compared favorably at each joint. Because moment arms are dependent on joint axis location and muscle path, these data suggest that the model appropriately represents joint structure and muscle paths. Maximum measured joint torques at the ankle [11] and knee compared well with maximal isometric model simulations. Differences between modeled and reported moment arms were largest for the hip extensors and knee flexors (the hamstrings muscles), but within the experimental variability of these measures (~1 cm) [5]. Thus, the model provides an accurate and meaningful representation of the feline hind limb anatomy and function.

REFERENCES

1. McDonnall D et al. Can J Physiol Pharmacol 82, 599-609, 2004

2. Sacks, R & R Roy, J Morphol 173, 185-95, 1982

3. Roy R et al, Acta Anat 159, 136-146, 1997

4. Burkholder T, Personal Comm. 2008

5. MacFadden L & N Brown, J Biomech 41, 3448-3457, 2007

6. Young R et al. Neurosci Lett 145, 137-140, 1992

7. Young R et al. Exp Brain Res 96, 141-151, 1993

8. Burkholder T & T Nichols J Morphol 261, 118-129, 2004

9. Boyd S & J Ronsky J Biomech 31, 279-283, 1998

10. Goslow G et al. J Morphol 141, 1-41, 1973

11. Lawrence J et al. J Neurophysiol 69, 282-285, 1993

12. Scovil C & J Ronsky J Biomech 39, 2055-2063, 2005

ACKNOWLEDGEMENTS

Supported by NIH R01-NS03967