Proceedings of The Physiological Society
University of Oxford (2011) Proc Physiol Soc 23, PC90
Optimal control of task-specific muscle synergies I: Feed-forward strategies
K. Nazarpour1, A. Jackson1
1. Newcastle University, Newcastle, United Kingdom.
Are muscle synergies hard-wired into neural circuits that reflect conserved biomechanical constraints  or can they be optimised to suit abstract task requirements ? We addressed this question using a myoelectric-controlled interface  which dissociates task requirements from limb biomechanics. 16 human subjects made repeated movements of a myoelectric cursor controlled by smoothed, rectified electromyogram (EMG) from two muscles. Elliptical target shapes imposed relevant and irrelevant dimensions within the task space, oriented such that movement variability should optimally be constrained along dimensions of either positive or negative EMG covariance. Muscle combinations included natural synergists (FDI-APB, ECR-FCR) and unnatural pairs (FDI-ADM, ECR-APB). Performance improved consistently over recording blocks, although scores were slightly lower for unnatural muscle pairs. An index of covariation (IoC) quantified the trial-to-trial cursor variability along axes of positive and negative EMG covariation. Multiple regression analysis of IoC revealed the main effect of target orientation for all the pairs, consistent with a buffering of movement variability into the appropriate task-irrelevant dimension. For natural synergists, the interaction between target shape and block number indicated a significant effect of training. The influence of target shape on IoC increased through the hold period consistent with a minimum intervention controller acting on visual feedback. However, even in the absence of the visual feedback, variability patterns were still modulated by target orientation although overall IoC values were shifted in the direction of positive covariation. A simple feed-forward model was used to predict optimal distributions of activity across a many-to-two cortico-motoneuronal projection in the presence of signal-dependent noise. The model explained (1) target-dependent modulation of IoC and (2) the overall positive covariation in the absence of visual feedback. A further prediction that common drive to muscles should vary with target orientation was verified by intermuscular coherence analysis. Enhanced beta-band coherence between FDI-APB was observed for targets that required positive EMG covariance. Rather than being limited to a small number of fixed synergies, the human hand can recruit a wide repertoire of co-ordinated muscle patterns appropriate for task demands. Convergence and divergence in cortico-spinal projections to distal motoneurons provide a rich neural substrate for flexible feed-forward minimisation of movement errors in task-relevant dimensions. In the presence of feedback, this strategy is enhanced by incorporation of minimum intervention control.
Where applicable, experiments conform with Society ethical requirements