ReviewCentral processes for the multiparametric control of arm movements in primates
Introduction
Almost 20 years ago, the paradigm of correlating primary motor cortical neuronal firing to movements about an isolated joint [1] changed to correlating neuronal firing to whole arm movements [2]. This experimental shift was paralleled by a conceptual shift. The correlations once sought between neurons and parameters of muscle activity or single-joint movements 3., 4. were now sought between neurons and parameters of the hand trajectory 2., 5.. The focus turned to whether global or executive features of movement were encoded [6]. The debates intensified as to whether movements/kinematics or muscles/dynamics were controlled 7., 8., 9••.. Moreover, questions concerning the coordinate system in use became prominent. Neuronal correlates to joint or muscle parameters imply the use of an intrinsic (relative to the arm) coordinate system. Correlations to hand position or path in space imply an extrinsic coordinate system 8., 10., 11..
This paradigm shift added an experimental richness not possible for single joint movements. Yet the debate remained largely focused on elucidating the one crucial parameter of movement encoded or the one crucial coordinate system used in a motor area. In contradistinction, this review examines the central processing of motor signals during multi-joint limb movements and emphasizes four points.
First, the search for a single, dominant, encoded movement parameter is unlikely to be fruitful. All available evidence suggests that neurons in motor structures signal many aspects of limb movements and are modulated by various motor and non-motor factors. Second, the correlations with specific motor parameters are unlikely to be invariant but rather change in time or over a behavioral sequence as required by the task. Third, the search for a single coordinate frame in which a neuron encodes these parameters is also unlikely to be fruitful. Processing motor, sensory and cognitive parameters will involve information represented in several coordinate systems. These coordinate systems are thought to evolve sequentially during the sensorimotor transformations occurring with limb movements 12., 13., 14., 15.. Last, the parameters represented in the discharge of neurons reflect the tasks and paradigms being investigated. Task-specific processing may provide the flexibility needed to generate the vast repertoire of movements that human and non-human primates exhibit.
Section snippets
Directional tuning
Pivotal to the shift in the framework for studying the central control of arm movements was the observation by Georgopoulos et al. 2., 5. that the discharge of primary motor cortex (M1) neurons is broadly ‘tuned’ to the direction of arm movement for reaching in two and three dimensions. Characterization of the directional tuning typically includes fitting firing to a cosine-tuning function to define the preferred direction (PD) at which the firing is greatest. Directional tuning is a prominent
Encoding scalar parameters of arm movement
Accurate arm movements require the control of not only direction of movement but also other scalar parameters including speed, amplitude and accuracy. Human psychophysics point to the independent or serial processing of direction as well as other movement parameters 12., 41., 42., 43., 44., 45., 46.. Not surprisingly several non-directional scalars including amplitude 31., 47., 48., 49••., speed 9••., 16., 19., 22., 32. and accuracy [29•] also modulate the discharge of motor cortical and/or
Encoding other directionally sensitive parameters of arm movement
When neurons in the PMd and M1 signal both direction and a scalar, the directional tuning is not altered. The directional tuning of PMd and M1 neurons changes over time and during a behavioral sequence, consistent with the encoding of multiple directional parameters or vectors. In PMd and M1, the PD is influenced by arm posture 51., 52., 53. and starting position [54]. Eye position can shift the PD of these neurons 55., 56., as can visual target location 14., 15., 57. and selective attention
Coordinate systems and transforms
Whether intrinsic (e.g. joint angles or muscles) or extrinsic (e.g. hand- or workplace-centered) coordinate systems are used by M1 has been extensively debated 7., 8., 10., 40.. Because motor cortical neurons signal kinematic, dynamic, postural, visual and oculomotor information, the neural substrate exists for representing signals in multiple coordinate systems. Altering the required pronation–supination (forearm rotation) and flexion–extension movements of the wrist to move a cursor to a
Higher cortical processes also modulate PMd and M1 neurons
The relations of motor cortical neuronal firing to arm movements are not rigid; they depend not only on task-specific behavioral but also cognitive content 57., 67.. The changes in PD that occur with learning to move in a force field [60••], or with altered visuomotor relationships, illustrate how complex coordinate transformation alters the directional signal. Other higher cortical processes modulate motor cortical neuronal activity. The rotation of the PD of individual cells and the
Conclusions
The firing of PMd, M1 and cerebellar neurons contain many motor and non-motor signals. The most studied of these, directional tuning, is a robust phenomenon that has proved useful not only for understanding the planning and execution of movements but also for evaluating occult cognitive processing. But there is no single directional signal; instead, directional kinematic and kinetic variables contribute to the directional tuning. Within a behavioral context, the directional signal may vary in
Acknowledgements
The work was supported in part by NIH grants 5R01-NS31530 and 2R01-NS18338, and NSF grant IBN-9873478.
References and recommended reading
Papers of particular interest, published within the annual period of review, have been highlighted as:
•of special interest
••of outstanding interest
References (75)
Do neurons in the motor cortex encode movement direction? An alternative hypothesis
Neurosci Lett
(1988)- et al.
A method for detecting the time course of correlation between single-unit activity and EMG during a behavioral task
J Neurosci Methods
(1995) - et al.
Processing of multiple kinematic signals in the cerebellum and motor cortices
Brain Res Rev
(2000) - et al.
Cortical control of reaching movements
Curr Opin Neurobiol
(1997) Current issues in directional motor control
Trends Neurosci
(1995)- et al.
Directional tuning profiles of motor cortical cells
Neurosci Res
(2000) Representation of movements and muscles by pyramidal tract neurons of the precentral motor cortex
- et al.
On the relations between the direction of two-dimensional arm movements and cell discharge in primate motor cortex
J Neurosci
(1982) - et al.
Relation of pyramidal tract activity to force exerted during voluntary movement
J Neurophysiol
(1968) - et al.
Functional classes of primate corticomotoneuronal cells and their relation to active force
J Neurophysiol
(1980)