Biological Cybernetics

, Volume 106, Issue 11, pp 727–739

Dynamic primitives of motor behavior


DOI: 10.1007/s00422-012-0527-1

Cite this article as:
Hogan, N. & Sternad, D. Biol Cybern (2012) 106: 727. doi:10.1007/s00422-012-0527-1


We present in outline a theory of sensorimotor control based on dynamic primitives, which we define as attractors. To account for the broad class of human interactive behaviors—especially tool use—we propose three distinct primitives: submovements, oscillations, and mechanical impedances, the latter necessary for interaction with objects. Owing to the fundamental features of the neuromuscular system—most notably, its slow response—we argue that encoding in terms of parameterized primitives may be an essential simplification required for learning, performance, and retention of complex skills. Primitives may simultaneously and sequentially be combined to produce observable forces and motions. This may be achieved by defining a virtual trajectory composed of submovements and/or oscillations interacting with impedances. Identifying primitives requires care: in principle, overlapping submovements would be sufficient to compose all observed movements but biological evidence shows that oscillations are a distinct primitive. Conversely, we suggest that kinematic synergies, frequently discussed as primitives of complex actions, may be an emergent consequence of neuromuscular impedance. To illustrate how these dynamic primitives may account for complex actions, we brieflyreviewthree typesof interactivebehaviors: constrained motion, impact tasks, and manipulation of dynamic objects.


Discrete Submovement Rhythmic Oscillation Impedance Primitive 

Copyright information

© The Author(s) 2012

Authors and Affiliations

  1. 1.Department of Mechanical Engineering, Department of Brain and Cognitive SciencesMassachusetts Institute of TechnologyCambridgeUSA
  2. 2.Department of Biology, Department of Electrical and Computer Engineering, Department of PhysicsCenter for Interdisciplinary Research on Complex Systems, Northeastern UniversityBostonUSA

Personalised recommendations