Elsevier

Journal of Physiology-Paris

Volume 98, Issues 4–6, July–November 2004, Pages 331-348
Journal of Physiology-Paris

Specificity of sensorimotor learning and the neural code: Neuronal representations in the primary motor cortex

https://doi.org/10.1016/j.jphysparis.2005.09.005Get rights and content

Abstract

Human studies show that the learning of a new sensorimotor mapping that requires adaptation to directional errors is local and generalizes poorly to untrained directions. We trained monkeys to learn new visuomotor rotations for only one target in space and recorded neuronal activity in the primary motor cortex before, during and after learning. Similar to humans, the monkeys showed poor transfer of learning to other directions, as observed by behavioral aftereffects for untrained directions. To test for internal representations underlying these changes, we compared two features of neuronal activity before and after learning: changes in firing rates and changes in information content. Specific elevations of firing rate were only observed in a subpopulation of cells in the motor cortex with directional properties corresponding to the locally learned rotation; namely cells only showed plasticity if their preferred direction was near the training one. We applied measures from information theory to probe for learning-related changes in the neuronal code. Single cells conveyed more information about the direction of movement and this specific improvement in encoding was correlated with an increase in the slope of the neurons’ tuning curve. Further, the improved information after learning enabled a more accurate reconstruction of movement direction from neuronal populations. Our findings suggest a neural mechanism for the confined generalization of a newly acquired internal model by showing a tight relationship between the locality of learning and the properties of neurons. They also provide direct evidence for improvement in the neural code as a result of learning.

Introduction

During performance of visually guided movements, the brain transforms visuospatial information into appropriate motor commands [6], [28], [68]. Psychophysical studies suggest that when humans learn a novel visuomotor transformation, an internal model of limb dynamics and kinematics is modified. These internal models allow the motor system to achieve the desired outcome—reaching towards a visible goal—under the new conditions [30], [33], [65], [80]. Learning a new motor skill that requires adaptation to directional errors generalizes poorly across movement directions [16], [32], workspace [19] and posture [3]. This is similar to sensory systems [1], [20], where practice can induce behavioral improvement that is specific to the situation experienced during the practice sessions. These findings suggest a reliance on neuronal elements with localized spatial fields [13], [19], [20]; namely that changes occur in neurons with fine selectivity (i.e., receptive fields, tuning curves) for the stimuli experienced or the movements made during training.

Neuronal elements with selectivity to movement direction are common in many parts of the motor system, including the primary motor cortex (M1) where the majority of cells show directional tuning [17], [26]. M1 participates in the planning and execution of reaching movements [28], [59] and has been shown to be involved in motor learning on the cellular level [51], in human studies [25], [29], [38], [64], and in electrophysiological studies [21], [35], [43], [79]. Only recently, however, an electrophysiological study of M1 linked neuronal selectivity to movement direction with the limited transfer of learning [46].

Although many studies indicate that learning can induce specific changes in brain activity [20], [53], [71], [72], this finding does not necessarily imply that newly learned skills are “better” represented in the brain. The crucial question is: Do neurons encode task parameters, such as movement direction, any better after learning? In the motor system, such improved encoding [9] can be used for decoding by downstream areas and as an efference copy for further computation [69], [80]. It can also be used by an external observer to allow for more accurate prediction of behavior [34], [73]. In a recent study [45], we examined two questions. First, do learning-induced changes in firing rates provide more information on the task? Second, what aspect of cell activity contributes the most to this improvement?

To address the first question, we employed an information-theory analysis [10], [50] to calculate the mutual information (Fig. 7) between cell activity and direction of movement. Informational measures have two relevant advantages. First, they use the full distribution (estimated from the data) of neuronal activity and do not assume any specific shape of the tuning curve or noise distribution. This allows for a more fine-tuned examination of learning-related changes. Second, they provide a measure as to how well different directions can be differentiated, based on neuronal activity. To address the second question, we examined two features of the neuronal response that could contribute to the increase in information: response variability and the slope of the tuning curve. Finally, to demonstrate that the observed increase in information can be extracted, we used the neuronal activity to decode the actual movement direction.

Section snippets

Behavioral paradigm

Monkeys moved a manipulandum to control the movement of a cursor on a video screen located 50 cm from their torso and eyes with the goal of moving the cursor from a starting point at the center of the screen (origin) to a visual target in a delayed go-signal paradigm (Fig. 1a); this required the monkey to hold (as verified by hand velocity and EMG) the cursor in the origin circle for a random 750–1500 ms after the target onset. The disappearance of the origin indicated the go-signal. In each

Results

This report combines and discusses findings described in two recent publications [45], [46].

Discussion and conclusions

The present findings demonstrate a modification in the activity and spatial tuning functions of neurons in the primary motor cortex as a result of learning a local visuomotor skill. We suggest that this altered activity reflects learning and retention of the newly acquired internal model, one that converts visuospatial cues into the motor commands required for hand movements and/or joint rotations to achieve the goal. In addition, we show that these changes affect the efficiency of the neural

Acknowledgements

We thank Chen Nathan and Thomas Boraud for participating in the recordings, Gal Chechick and Amir Globerson for helpful discussions of Information theory and Hagai Bergman and Steven P. Wise for constant help and many fruitful discussions. This study was supported in part by a Center for Excellence grant (8006/00) administered by the ISF, by a BMBF-DIP grant and by grant 2001073 administered by the BSF. We are also thankful for the contributions by the Golden charitable trust and by Avraham and

References (82)

  • T.D. Sanger

    Neural population codes

    Curr. Opin. Neurobiol.

    (2003)
  • S.H. Scott

    The role of primary motor cortex in goal-directed movements: insights from neurophysiological studies on non-human primates

    Curr. Opin. Neurobiol.

    (2003)
  • N. Suga et al.

    Plasticity and corticofugal modulation for hearing in adult animals

    Neuron

    (2002)
  • M. Sur et al.

    Cortical plasticity: time for a change

    Curr. Biol.

    (2002)
  • M. Ahissar et al.

    Task difficulty and the specificity of perceptual learning

    Nature

    (1997)
  • G.E. Alexander et al.

    Preparation for movement: neural representations of intended direction in three motor areas of the monkey

    J. Neurophysiol.

    (1990)
  • P. Baraduc et al.

    Adaptation to a visuomotor shift depends on the starting posture

    J. Neurophysiol.

    (2002)
  • M.J. Berry et al.

    The structure and precision of retinal spike trains

    Proc. Natl. Acad. Sci. USA

    (1997)
  • T. Brashers-Krug et al.

    Consolidation in human motor memory

    Nature

    (1996)
  • C.A. Buneo et al.

    Direct visuomotor transformations for reaching

    Nature

    (2002)
  • L.L. Chen et al.

    Evolution of directional preferences in the supplementary eye field during acquisition of conditional oculomotor associations

    J. Neurosci.

    (1996)
  • L.L. Chen et al.

    Conditional oculomotor learning: population vectors in the supplementary eye field

    J. Neurophysiol.

    (1997)
  • T.M. Cover et al.

    Elements of Information Theory

    (1991)
  • D.J. Crammond et al.

    Modulation of preparatory neuronal activity in dorsal premotor cortex due to stimulus–response compatibility

    J. Neurophysiol.

    (1994)
  • M.R. DeWeese et al.

    How to measure the information gained from one symbol

    Network

    (1999)
  • O. Donchin et al.

    Quantifying generalization from trial-by-trial behavior of adaptive systems that learn with basis functions: theory and experiments in human motor control

    J. Neurosci.

    (2003)
  • J.R. Flanagan et al.

    Composition and decomposition of internal models in motor learning under altered kinematic and dynamic environments

    J. Neurosci.

    (1999)
  • F. Gandolfo et al.

    Motor learning by field approximation

    Proc. Natl. Acad. Sci. USA

    (1996)
  • A.P. Georgopoulos et al.

    On the relations between the direction of two-dimensional arm movements and cell discharge in primate motor cortex

    J. Neurosci.

    (1982)
  • A.P. Georgopoulos et al.

    Primate motor cortex and free arm movements to visual targets in three-dimensional space. II. Coding of the direction of movement by a neuronal population

    J. Neurosci.

    (1988)
  • Z. Ghahramani et al.

    Generalization to local remappings of the visuomotor coordinate transformation

    J. Neurosci.

    (1996)
  • P.L. Gribble et al.

    Overlap of internal models in motor cortex for mechanical loads during reaching

    Nature

    (2002)
  • P. Haggard et al.

    Localising awareness of action with transcranial magnetic stimulation

    Exp. Brain Res.

    (1999)
  • N.G. Hatsopoulos et al.

    Information about movement direction obtained from synchronous activity of motor cortical neurons

    Proc. Natl. Acad. Sci. USA

    (1998)
  • H. Imamizu et al.

    Internal representations of the motor apparatus: implications from generalization in visuomotor learning

    J. Exp. Psychol. Hum. Percept. Perform.

    (1995)
  • I.H. Jenkins et al.

    Motor sequence learning: a study with positron emission tomography

    J. Neurosci.

    (1994)
  • S. Kakei et al.

    Muscle and movement representations in the primary motor cortex

    Science

    (1999)
  • S. Kakei et al.

    Direction of action is represented in the ventral premotor cortex

    Nat. Neurosci.

    (2001)
  • A. Karni et al.

    Functional MRI evidence for adult motor cortex plasticity during motor skill learning

    Nature

    (1995)
  • J.W. Krakauer et al.

    Independent learning of internal models for kinematic and dynamic control of reaching

    Nat. Neurosci.

    (1999)
  • J.W. Krakauer et al.

    Learning of visuomotor transformations for vectorial planning of reaching trajectories

    J. Neurosci.

    (2000)
  • Cited by (15)

    • A stable brain from unstable components: Emerging concepts and implications for neural computation

      2017, Neuroscience
      Citation Excerpt :

      An intuitively simple scheme for maintaining neural stability is to ensure that stationary outputs are not influenced by underlying dynamic components at all - rather, only stable elements at any level of the system give rise to phenomena at the next level (Fig. 2A). This scheme would ensure that dynamic elements of a synapse, circuit, or network are weighted differentially from stable elements, such that their contributions are ignored or at least minimized (Panas et al., 2015). The cells with frequent changes in afferent input or intrinsic excitability, could contribute substantially less, if at all, to the overall output of the circuit.

    • Are fast/slow process in motor adaptation and forward/inverse internal model two sides of the same coin?

      2013, Medical Hypotheses
      Citation Excerpt :

      A number of these studies are summarized as follows: Some of neurophysiological studies have highlighted changes in the primary motor cortex (M1) during learning of force field or visuomotor rotation studies [30,31]. They have found that changes in M1 do not happen during the initial phase of adaptation (i.e., when errors are large and the fast process is the dominant component).

    • Motor Learning with Unstable Neural Representations

      2007, Neuron
      Citation Excerpt :

      Neural recordings in behaving animals have revealed much about the mechanisms underlying motor learning. Changes in single-unit activity have been correlated with learning sensorimotor associations (Mitz et al., 1991; Ojakangas and Ebner, 1992; Paz et al., 2003; Paz and Vaadia, 2004; Wise et al., 1998), learning movement sequences and skills (Cohen and Nicolelis, 2004; Nakamura et al., 1998), and adapting to novel mechanical environments (Gandolfo et al., 2000; Li et al., 2001; Padoa-Schioppa et al., 2002, 2004; Xiao et al., 2006). One assumption implicit in many of these studies is that there is an underlying stable neural representation for familiar behavior, and thus changes in the neural representation necessarily reflect motor learning.

    View all citing articles on Scopus
    View full text