MIT CogNet, The Brain Sciences ConnectionFrom the MIT Press, Link to Online Catalog
SPARC Communities
Subscriber : Stanford University Libraries » LOG IN

space

Powered By Google 
Advanced Search

 

Signal-dependent Noise and Optimal Motor Control: A Unifying Principle of Human Movement

 Christopher M. Harris
  
 

Abstract:

When we make brief reaching arm movements or saccadic eye movements, trajectories tend to be straight and smooth with symmetrical bell-shaped velocity profiles that become progressively asymmetrical with increased duration. A similar phenomenon occurs in tongue movement during articulation, and this suggests that there may be some underlying principle to movement control. We propose motor commands are perturbed by signal dependent noise (SDN). With SDN larger control signals increase noise which accumulates during a movement leading to more inaccuracy. Since faster movements require large control signals, SDN imposes a speed accuracy trade-off. We show that, for a given movement duration, there is a unique optimal trajectory that minimises inaccuracy, or equivalently minimises duration for a given level of endpoint error. This trajectory is very similar to empirically observed arm movements and saccades, being smooth and symmetrical for brief durations and asymmetrical for long durations. We further propose that when there is a fixed speed-accuracy trade-off, there is a unique duration for a given amplitude of movement, which is in good agreement with the stereotyped duration-amplitude relationship ('main sequence') empirically observed in saccades. The similarity of tongue and lip movements to arm and eye movements leads us to suggest that the speed and accuracy of speech articulation may also be constrained by the same kind of motor noise.

 
 


© 2010 The MIT Press
MIT Logo