Motor primitives — new data and future questions
Introduction
Motor Primitives can be controversial. What is a ‘primitive’? The Free Dictionary online (http://www.thefreedictionary.com/primitive) states: 1. Not derived from something else; primary or basic. 2. a. Of or relating to an earliest or original stage or state; primeval. b. Being little evolved from an early ancestral type. 3. Characterized by simplicity or crudity. These definitions are close to what many of us consider ‘Motor Primitives’ to be in neuroscience, i.e., fundamental building blocks for constructing motion. Motor primitives (MPs) are thus defined here to be compositional elements for movement construction. Assuming more complex motion arises from rearranging and combining MPs, then a subset of fundamental motor primitives should be available to organize initial movements very early in development, and thereby contribute to the ontogenetic and adult construction of more complex motor acts, in one or more ways.
Compositional elements for movement could have various forms. Motor acts comprise overall kinematics, together with force interactions with the environment, and the underlying body kinetics, and regulation of impedance. These in turn depend on patterns of muscle activation and control. At what level should we look for compositional MPs as motor neuroscientist? Different authors have considered MPs at several levels. MPs have been defined as kinematic elements (i.e., patterns of motion without regard to force or mass, e.g., strokes [1, 2, 3•] or cycles [4, 5]), or as kinetic elements (force-based, e.g., viscoelastic force-field primitives [6, 7, 8, 9] and their associated muscle synergies [9, 10, 11]), or as neural drive circuits [12•]. Because the nervous system appears to organize both the executed motion as well as prepare the contingent acts and corrections required for any common or unexpected perturbations in any complex body motion, it follows that the most fundamental MPs need to be useful as compositional elements in each of these contexts, i.e., the fundamental MPs must be adequate building blocks to rapidly construct the kinematic, kinetic, impedance and contingent response aspects of a complex movement. Our current computational understanding of motor control likely has a place for both kinematic and kinetic elements [3•, 13, 14, 15]. However, reconciling the frameworks of kinematic MPs and kinetic/synergy MPs into a unified compositional scheme remains difficult. Explorations that enrich our understanding of movement in important ways continue in both areas as described below. Here I take the kinetic MPs as fundamental building blocks, and take kinematic MPs as more derived, according to the mechanical relations between them, but a satisfying unification of the two in motor neuroscience is eventually needed.
Section snippets
Evidence for kinematic movement primitives
Support for kinematic MPs derives from several methods. Primarily, researchers seek building blocks in complex kinematic patterns, decomposing motor acts at the kinematics level into their constituent elements. Various statistical techniques are available, each with pros and cons, to accomplish this task and break motion patterns into component elements. Early segmentation of kinematics by Viviani and Terzuolo, and by Flash and Hogan [2] has been elaborated today in several further ways.
Evidence for force-field motor primitives and spinal muscle synergies
Support for visco-elastic force fields as kinetic MPs and for their associated muscle synergies as MPs is strong (Figure 1, MP2). These two descriptions (force-field unit and muscle synergy) are equivalent to one another and largely interchangeable (see Figure 2, for more details). Support derives from physiological perturbations, neural recordings, statistical decomposition techniques in animal and human model system data, and more recently from molecular genetics methods and manipulations.
Circuit foundations of motor primitives
If MPs reflect an organizational principle in movement construction, we should expect the MP picture to give insight into the underlying circuits, e.g., spinal system architectures, in addition to forming motor pattern and biomechanical building blocks. Strong predictions of an MP framework are that the underlying neural circuitry should involve interneurons assembling the synergies and driving them as units. The simplest conceptions of these are as simple premotor projection neurons. In the
Computational frameworks for motor primitives
Both kinematic and force-field/muscle synergy MPs are currently viewed as rooted in some kind of optimization of movement or an approximation of it [2, 15, 23, 24]. The form of kinematic strokes in movement (i.e., MP1, Figure 1) have variously been attributed to the central nervous system (CNS) minimizing jerk, torque change, muscular effort or some weighted balance of these [2, 15, 23, 24]. Kinematic MPs are seen in a range of animal forms from soft-bodied muscular hydrostat effectors, to
Development, skill and plasticity in relation to Motor Primitives in humans
The concept of MPs raises a range of questions in human development. What collections of MPs are available in man, based on our current understanding, and do they constrain our options for human movement? Do skilled biped primates still continue to utilize the ‘built-in’ MPs seen in other specie? Are human and primate MPs constructed de novo in novel tasks, or for bipedal locomotion? Do neocortical representations replace or incorporate the evolutionarily older systems in skilled novel
Conclusions — Major issues for the future
Motor Primitives at a low level in the CNS can provide an animal with initial simplified control of movement, and can bootstrap development of movement using information captured and ‘baked in’ by many prior generations of evolutionary selection pressure, forming modularity that is closely matched to the biomechanics and standard movement repertoire. MPs may thereby almost immediately provide near optimal patterns of muscle activation in some species with high early motor demands. MPs also form
References and recommended reading
Papers of particular interest, published within the period of review, have been highlighted as:
• of special interest
•• of outstanding interest
Acknowledgments
Support from NIH NS072651, NS054894, NSF IIS 0827684 and the Craig Neilsen Foundation. Drs. Kim Dougherty, Ole Kiehn, and Brian Clark made very helpful comments on earlier versions of the manuscript.
References (76)
- et al.
Dynamical movement primitives: learning attractor models for motor behaviors
Neural Comput
(2013) - et al.
Parabolic movement primitives and cortical states: merging optimality with geometric invariance
Biol Cybern
(2009) - et al.
Modulation of muscle synergy recruitment in primate grasping
J Neurosci
(2008) - et al.
Spatial constant equi-affine speed and motion perception
J Neurophysiol
(2014) From basis functions to basis fields: vector field approximation from sparse data
Biol Cybern
(1992)- et al.
Merging of healthy motor modules predicts reduced locomotor performance and muscle coordination complexity post-stroke
J Neurophysiol
(2010) - et al.
Modular control of varied locomotor tasks in children with incomplete spinal cord injuries
J Neurophysiol
(2013) - et al.
Primitive reflexes in amyotrophic lateral sclerosis: prevalence and correlates
J Neurol
(2014) - et al.
Differences in adaptation rates after virtual surgeries provide direct evidence for modularity
J Neurosci
(2013) An organizing principle for a class of voluntary movements
J Neurosci
(1984)
The coordination of arm movements: an experimentally confirmed mathematical model
J Neurosci
Dynamic primitives of motor behavior
Biol Cybern
Transitions between discrete and rhythmic primitives in a unimanual task
Front Comput Neurosci
Dynamic primitives in the control of locomotion
Front Comput Neurosci
Computations underlying the execution of movement: a biological perspective
Science
Convergent force fields organized in the frog's spinal cord
J Neurosci
Linear combinations of primitives in vertebrate motor control
Proc Natl Acad Sci U S A
Rapid correction of aimed movements by summation of force-field primitives
Journal of Neuroscience
Modular premotor drives and unit bursts as primitives for frog motor behaviors
J Neurosci
Muscle synergies encoded within the spinal cord: evidence from focal intraspinal NMDA iontophoresis in the frog
J Neurophysiol
A neural basis for motor primitives in the spinal cord
J Neurosci
Motor primitives in vertebrates and invertebrates
Curr Opin Neurobiol
Computational motor control in humans and robots
Curr Opin Neurobiol
Compositionality in neural control: an interdisciplinary study of scribbling movements in primates
Front Comput Neurosci
When practice leads to co-articulation: the evolution of geometrically defined movement primitives
Exp Brain Res
Submovements grow larger, fewer, and more blended during stroke recovery
Motor Control
Kinematic decomposition and classification of octopus arm movements
Front Comput Neurosci
Individual premotor drive pulses, not time-varying synergies, are the units of adjustment for limb trajectories constructed in spinal cord,
J Neurosci
A simple experimentally based model using proprioceptive regulation of motor primitives captures adjusted trajectory formation in spinal frogs
J Neurophysiol
Simplified and effective motor control based on muscle synergies to exploit musculoskeletal dynamics
Proc Natl Acad Sci U S A
Do muscle synergies reduce the dimensionality of behavior?
Front Comput Neurosci
Deletions of rhythmic motoneuron activity during fictive locomotion and scratch provide clues to the organization of the mammalian central pattern generator
J Neurophysiol
Modelling spinal circuitry involved in locomotor pattern generation: insights from deletions during fictive locomotion
J Physiol
Development and functional organization of spinal locomotor circuits
Curr Opin Neurobiol
Muscle synergies during locomotion in the cat: a model for motor cortex control
J Physiol
Motor cortical regulation of sparse synergies provides a framework for the flexible control of precision walking
Front Comput Neurosci
Locomotor primitives in newborn babies and their development
Science
Control of reaching movements by muscle synergy combinations
Front Comput Neurosci
Cited by (152)
Physics-based character animation and human motor control
2023, Physics of Life ReviewsLearning complex upper-limb movements through practicing movement elements
2023, Human Movement ScienceIntra-muscle Synergies Stabilizing Reflex-mediated Force Changes
2022, NeuroscienceCitation Excerpt :The former function has been explored using various matrix factorization methods applied to correlation or covariance matrices in spaces of elemental variables (those produced by elements at the selected level of analysis) (d’Avella et al., 2003; Krishnamoorthy et al., 2003; Ivanenko et al., 2004; Ting and Macpherson, 2005; Tresch et al., 2006). These methods led to the identification of stable groups of elemental variables addressed with various terms such as modules, modes, primitives, and synergies (reviewed in Giszter, 2015; Latash, 2020b, 2021b); we will use the term mode for consistency with earlier studies. The latter function has been explored using the framework of the uncontrolled manifold (UCM) hypothesis (Scholz and Schöner, 1999; reviewed in Latash et al., 2007).
Cortical contributions to locomotor primitives in toddlers and adults
2022, iScienceCitation Excerpt :Walking involves spinal networks that accommodate locomotor primitives (Dominici et al., 2011; Grillner, 1985, 2006; Ivanenko et al., 2013). In vertebrate models, the spinal networks have been identified as populations of premotor interneurons located in the lumbosacral spinal cord and are often considered central pattern generators (CPGs) (Giszter, 2015; Takei et al., 2017). Locomotor primitives (or muscle synergies) comprise small sets of functional units generating the motor output to coordinate muscles (Dominici et al., 2011).