Overlapping neural networks for multiple motor engrams

The hypothesis was tested that learned movement trajectories of different shapes can be stored in, and generated by, largely overlapping neural networks. Indeed, it was possible to train a massively interconnected neural network to generate different shapes of internally stored, dynamically evolving movement trajectories using a general-purpose core part, common to all networks, and a special-purpose part, specific for a particular trajectory. The weights of connections between the core units do not carry any information about trajectories. The core network alone could generate externally instructed trajectories but not internally stored ones, for which both the core and the trajectory-specific part were needed. All information about the movements is stored in the weights of connections between the core part and the specialized units and between the specialized units themselves. Due to these connections the core part reveals specific dynamical behavior for a particular trajectory and, as the result, discriminates different tasks. The percentage of trajectory-specific units needed to generate a certain trajectory was small (2-5%), and the total output of the network is almost entirely provided by the core part, whereas the role of the small specialized parts is to drive the dynamical behavior. These results suggest an efficient and effective mechanism for storing learned motor patterns in, and reproducing them by, overlapping neural networks and are in accord with neurophysiological findings of trajectory-specific cells and with neurological observations of loss of specific motor skills in the presence of otherwise intact motor control.

Directional operations in the motor cortex modeled by a neural network of spiking neurons

A neural network with realistically modeled, spiking neurons is proposed to model ensemble operations of directionally tuned neurons in the motor cortex. The model reproduces well directional operations previously identified experimentally, including the prediction of the direction of an upcoming movement in reaching tasks and the rotation of the neuronal population vector in a directional transformation task.

Functional magnetic resonance imaging of motor cortex: hemispheric asymmetry and handedness

A hemispheric asymmetry in the functional activation of the human motor cortex during contralateral (C) and ipsilateral (I) finger movements, especially in right-handed subjects, was documented with nuclear magnetic resonance imaging at high field strength (4 tesla). Whereas the right motor cortex was activated mostly during contralateral finger movements in both right-handed (C/I mean area of activation = 36.8) and left-handed (C/I = 29.9) subjects, the left motor cortex was activated substantially during ipsilateral movements in left-handed subjects (C/I = 5.4) and even more so in right-handed subjects (C/I = 1.3).

Motor cortical activity preceding a memorized movement trajectory with an orthogonal bend

Two monkeys were trained to make an arm movement with an orthogonal bend, first up and then to the left ([symbol: see text]), following a waiting period. They held a two-dimensional manipulandum over a spot of light at the center of a planar working surface. When this light went off, the animals were required to hold the manipulandum there for 600-700 ms and then move the handle up and to the left to receive a liquid reward. There were no external signals concerning the "go" time or the trajectory of the movement. It was hypothesized that during that period signs of directional processing relating to the upcoming movement would be identified in the motor cortex. Following 20 trials of the memorized movement trajectory, 40 trials of visually triggered movements in radially arranged directions were performed. The activity of 137 single cells in the motor cortex was recorded extracellularly during performance of the task. It was found that 62.8% of the cells changed activity during the memorized waiting period. During the waiting period, the population vector (Georgopoulos et al. 1983, 1984) began to grow approximately 130 ms after the center light was turned off; it pointed first in the direction of the second part of the memorized movement (<--) and then rotated clockwise towards the direction of the initial part of the movement (increases). These findings indicate processing of directional information during the waiting period preceding the memorized movement. This conclusion was supported by the results of experiments in ten human subjects, who performed the same memorized movement ([symbol: see text]). In 10% of the trials a visual stimulus was shown in radially arranged directions in which the subjects had to move; this stimulus was shown at 0, 200, and 400 ms from the time the center light was turned off. We found that as the interval increased the reaction time shortened for the visual stimulus that was in the same direction as the upward component of the memorized movement.

Cortical cell types from spike trains

The patterns of cell activity recorded extracellularly in the motor cortex of behaving monkeys were classified into the following three groups using a combination of cluster and discriminant analyses of 1925 spike trains: (a) cells with low discharge rate and low bursting (67.1%), (b) cells with low discharge rate but bursting (20.2%), and (c) cells with high discharge rate and low bursting (12.7%). The percentage of directionally tuned cells and of cells engaged during a memorized delay task were very similar in all three cell groups.

Cognitive neurophysiology of the motor cortex

A major challenge of current neuroscience is to elucidate the brain mechanisms that underlie cognitive function. There is no doubt that cognitive processing in the brain engages large populations of cells. This article explores the logic of investigating these problems by combining psychological studies in human subjects and neurophysiological studies of neuronal populations in the motor cortex of behaving monkeys. The results obtained show that time-varying psychological processes can be visualized in the time-varying activity of neuronal populations. Moreover, the functional interactions between cells in the motor cortex are very similar to those observed in a massively interconnected artificial network performing the same computation.

Common processing constraints for visuomotor and visual mental rotations

Naive human subjects were tested in three different tasks: (1) a visuomotor mental rotation task, in which the subjects were instructed to move a cursor at a given angle from a stimulus direction; (2) a visual mental rotation task, in which the subjects had to decide whether a displayed letter was normal or mirror image regardless of its orientation in the plane of presentation; and (3) a visuomotor memory scanning task, in which a list of two to five stimuli directions were presented sequentially and then one of the stimuli (test stimulus), except the last one, was presented again. Subjects were instructed to move a cursor in the direction of the stimulus that followed the test stimulus in the previous sequence. The processing rate of each subject in each task was estimated using the linear relation between the response time and the angle (mental rotation tasks) or the list length (memory scanning task). We found that the processing rates in the mental rotation tasks were significantly correlated but that neither correlated significantly with the processing rate in the memory scanning task. These results suggest that visuomotor and visual mental rotations share common processing constraints that cannot be ascribed to general mental processing performances.

A dynamical neural network model for motor cortical activity during movement: population coding of movement trajectories

As a dynamical model for motor cortical activity during hand movement we consider an artificial neural network that consists of extensively interconnected neuron-like units and performs the neuronal population vector operations. Local geometrical parameters of a desired curve are introduced into the network as an external input. The output of the model is a time-dependent direction and length of the neuronal population vector which is calculated as a sum of the activity of directionally tuned neurons in the ensemble. The main feature of the model is that dynamical behavior of the neuronal population vector is the result of connections between directionally tuned neurons rather than being imposed externally. The dynamics is governed by a system of coupled nonlinear differential equations. Connections between neurons are assigned in the simplest and most common way so as to fulfill basic requirements stemming from experimental findings concerning the directional tuning of individual neurons and the stabilization of the neuronal population vector, as well as from previous theoretical studies. The dynamical behavior of the model reveals a close similarity with the experimentally observed dynamics of the neuronal population vector. Specifically, in the framework of the model it is possible to describe a geometrical curve in terms of the time series of the population vector. A correlation between the dynamical behavior of the direction and the length of the population vector entails a dependence of the "neural velocity" on the curvature of the tracing trajectory that corresponds well to the experimentally measured covariation between tangential velocity and curvature in drawing tasks.

Functional imaging of human motor cortex at high magnetic field

1. We used conventional gradient echo magnetic resonance imaging (MRI) at high field strength (4 Tesla) to functionally image the right motor cortex in six normal human subjects during the performance of a sequence of self-paced thumb to digit oppositions with the left hand (contralateral task), the right hand (ipsilateral task), and both hands (bilateral task). 2. A localized increase in activity in the lateral motor cortex was observed in all subjects during the task. The area of activation was similar in the contralateral and bilateral tasks but 20 times smaller in the ipsilateral task. The intensity of activation was 2.3 times greater in the contralateral than the ipsilateral task.

Motor cortical activity in a memorized delay task

Two rhesus monkeys were trained to move a handle on a two-dimensional (2D) working surface in directions specified by a light at the plane. They first captured with the handle a light on the center of the plane and then moved the handle in the direction indicated by a peripheral light (cue signal). The signal to move (go signal) was given by turning off the center light. The following tasks were used: (a) In the non-delay task the peripheral light was turned on at the same time as the center light went off. (b) In the memorized delay task the peripheral light stayed on for 300 ms and the center light was turned off 450-750 ms later. Finally, (c) in the non-memorized delay task the peripheral light stayed on continuously whereas the center light went off 750-1050 ms after the peripheral light came on. Recordings in the arm area of the motor cortex (N = 171 cells) showed changes in single cell activity in all tasks. In both delay tasks, the neuronal population vector calculated every 20 ms after the onset of the peripheral light pointed in the direction of the upcoming movement, which was instructed by the cue light. Moreover, the strength of the population signal showed an initial peak shortly after the cue onset in both the memorized and non-memorized delay tasks but it maintained a higher level during the memorized delay period, as compared to the non-memorized task.(ABSTRACT TRUNCATED AT 250 WORDS)

Three-dimensional drawings in isometric conditions: planar segmentation of force trajectory

Normal human subjects grasped an isometric handle with an unrestrained, pronated hand. They were asked to exert forces continuously to draw lemniscates (figure eights) in specified or self-chosen planes and in the presence or absence of a three-dimensional visual feedback cursor and a visual template. In every condition, the mean plane orientation in the force space differed appreciably between the two loops of the figure, as described previously by Soechting and Terzuolo (1987a) for free drawing arm movements. These findings suggest that the planar segmentation of the motor trajectory is not a consequence of joint motion but arises from central constraints related to the production of motor trajectory in space.

The motor cortex and the coding of force

The relation of cellular activity in the motor cortex to the direction of two-dimensional isometric force was investigated under dynamic conditions in monkeys. A task was designed so that three force variables were dissociated: the force exerted by the subject, the net force, and the change in force. Recordings of neuronal activity in the motor cortex revealed that the activity of single cells was directionally tuned and that this tuning was invariant across different directions of a bias force. Cell activity was not related to the direction of force exerted by the subject, which changed drastically as the bias force changed. In contrast, the direction of net force, the direction of force change, and the visually instructed direction all remained quite invariant and congruent and could be the directional variables, alone or in combination, to which cell activity might relate.

Three-dimensional drawings in isometric conditions: relation between geometry and kinematics

Normal human subjects grasped a 3-D isometric handle with an otherwise unrestrained, pronated hand and exerted forces continuously to draw circles, ellipses and lemniscates (figure-eights) in specified planes in the presence or absence of a 3-D visual force-feedback cursor and a visual template. Under any of these conditions and in all subjects, a significant positive correlation was observed between the instantaneous curvature and angular velocity, and between the instantaneous radius of curvature and tangential velocity; that is, when the force trajectory was most curved, the tangential velocity was lowest. This finding is similar to that obtained by Viviani and Terzuolo (1982) for 2-D drawing arm movements and supports the notion that central constraints give rise to the relation between geometric and kinematic parameters of the trajectory.

Cognitive spatial-motor processes. 7. The making of movements at an angle from a stimulus direction: studies of motor cortical activity at the single cell and population levels

Two rhesus monkeys were trained to move a handle on a two-dimensional (2-D) working surface either towards a visual stimulus ("direct" task) or in a direction orthogonal and counterclockwise (CCW) from the stimulus ("transformation" task), depending on whether the stimulus appeared dim or bright, respectively. Thus the direction of the stimulus (S, in polar coordinates) and the direction of the movement (M) were the same in the direct task but differed in the transformation task, such that M = S + 90 degrees CCW. The task (i.e. brightness) condition (k = 2, i.e. direct or transformation) and the direction of the stimulus (m = 8, i.e. 8 equally spaced directions on a circle) resulted in 16 combinations (k x m = 16 "classes") that were varied from trial to trial in a randomized block design. In 8 of these combinations the direction of the stimulus was the same for both tasks, whereas the direction of the movement was the same in the remaining 8 cases. The electrical signs of cell activity (N = 394 cells) in the arm area of the motor cortex (contralateral to the performing arm) were recorded extracellularly. The neural activity was analyzed at the single cell and neuronal population levels, and a modeling of the time course of single activity during the transformation task was carried out. We found the following. (a) Individual cells were active in both tasks; no cells were found that were active exclusively in only one of the two tasks. The patterns of single cell activity in the transformation task frequently differed from those observed in the direct task when the stimulus or the movement were the same. More specifically, cells could not be consistently classified as "movement"-or "stimulus"-related for frequently the activity of a particular cell would seem "movement-related" for a particular stimulus-movement combination, "stimulus-related" for another combination, or unrelated to either movement or stimulus for still another combination. Thus no real insight could be gained from such an analysis of single cell activity. (e) In a different analysis, we explored the idea that a changing directional signal could be detected in the time course of single cell activity during the reaction time. For that purpose we modeled the time course of single activity observed in the transformation task as a linear, weighted combination of influences from the direct task, taking the time patterns of cell activity during the stimulus, intermediate and movement directions in the direct task as estimates of the postulated directional influences.(ABSTRACT TRUNCATED AT 400 WORDS)

Cortical control of motor behavior at the cellular level

The studies reviewed in this paper describe the relations of single-cell activity in central motor structures to complex visuomotor tasks and document the fact that various cortical areas process visuomotor information in parallel. Moreover, the studies provide clear evidence that the map in the motor cortex is modifiable and dynamically maintained.

Cognitive spatial-motor processes. 5. Specification of the direction of visually guided isometric forces in two-dimensional space: time course of information transmitted and effect of constant force bias

The effects of an external constant force bias on the information transmitted (Ti) by the direction of isometric force exerted in 2-dimensional (2-D) space by human subjects were studied using an isometric manipulandum and random dot stereograms generated in a color display (Massey et al. 1988, Massey et al. 1990). Subjects exerted force on the manipulandum such that a visual force-feedback cursor would move in the direction of a visually defined stimulus in the stereo display. The time course of force development and the gain of directional information during increasing force intensity were also studied. We found the following. (a) When no bias force was applied, the force exerted by the subject increased from near zero to greater than 200 gram-force at the end of a trial and was close to the visually defined direction. When a constant bias force of 110 gram-force was applied in various directions in blocks of trials, the force exerted by the subject increased in time, as above; however, its direction also changed in time so that the instantaneous vector sum of the bias force and the force exerted by the subject pointed close to the visually defined direction. The Ti and the reaction time (RT) did not differ significantly in the two experimental conditions. These results suggest that the directional control of isometric forces is very efficient, especially in relation to visuomotor coordination. (b) The Ti was calculated at various levels of force intensity, as the latter increased from approximately 50 gram-force to 200 gram-force.(ABSTRACT TRUNCATED AT 250 WORDS)

Cognitive spatial-motor processes. 6. Visuomotor memory scanning

Fourteen human subjects performed in a modified Sternberg memory-scanning task. First, they made a series of 2-6 movements in different directions from a central point towards peripheral lights on a planar working surface ("list trials"). Then, after a warning signal, one of the previous list stimuli, except the last, was presented again ("test trial"). Subjects were instructed to move in the direction of the stimulus which was presented next in sequence in the list. The mean reaction time (RT) in the test trials increased as a linear function of the number of movements, S, in the list: Mean RT (ms) = 105 + 205.8S (2 less than or equal to S less than or equal to 6). This finding suggests that the task involves memory scanning of visuomotor list items.

Cognitive spatial-motor processes. 4. Specification of the direction of visually guided isometric forces in two-dimensional space: information transmitted and effects of visual force-feedback

The information transmitted (Ti) by the direction of two-dimensional (2-D) isometric forces at different stereoscopic depths was studied in 50 naive human subjects using an isometric manipulandum and random dot stereograms generated in a color display (Massey et al. 1988). Subjects viewed the display through appropriate color filters and perceived the image of a disk rotated about a horizontal axis on the frontal plane; the top of the disk was rotated around that axis by 15, 45, 60 and 80 degrees away from the subject. Each of these disks involved a different amount of stereoscopic depth perception which was lowest for the 15 degrees and highest for the 80 degrees tilt. Subjects were instructed to exert force in the direction of a visual target presented on the disk in a reaction time task. The instantaneous force exerted by the subjects on the manipulandum was shown on the disk in the form of a feedback cursor. Information transmitted, reaction time (RT) and systematic directional deviations were calculated. We found the following. (a) Ti increased with input information but at a lower rate; at the highest level of input information studied (5.91 bits), Ti was 4.1 bits at the 15 degrees tilt. This high value of Ti suggests that directional information for isometric force is processed very efficiently. However, this Ti was consistently lower than that transmitted by the direction of movement (Georgopoulos and Massay, 1988). (b) Ti did not differ significantly among the 15-60 degrees tilt but was 0.19 bits less for the 80 degrees tilt. RT did not differ among the 15-80 degrees tilts.(ABSTRACT TRUNCATED AT 250 WORDS)

Parietal cortex neurons of the monkey related to the visual guidance of hand movement

A class of neurons specifically related to hand movements was studied in the posterior parietal cortex while the monkeys manipulated different types of objects. We examined the neuronal activity during manipulation of objects by the hand in the light and in the dark. Fifty-five neurons were active during manipulation in the dark and were classified as "hand-movement-related" neurons. Of these, 38/55 (69%) cells were also influenced by the visual stimulus. Most of the hand-movement-related neurons were selective in the type of objects manipulated. Moreover, some of these cells were selective in the axis of orientation of the object. These results suggest that the hand-movement-related neurons of the parietal cortex are concerned with the visual guidance of the hand movement, especially in matching the pattern of movement with the spatial characteristics of the object to be manipulated.

Visuomotor coordination in reaching and locomotion

Locomotion and reaching have traditionally been regarded as separate motor activities. In fact, they may be closely connected both from an evolutionary and a neurophysiological viewpoint. Reaching seems to have evolved from the neural systems responsible for the active and precise positioning of the limb during locomotion; moreover, it seems to be organized in the spinal cord. The motor cortex and its corticospinal outflow are preferentially engaged when precise positioning of the limb is needed during locomotion and are also involved during reaching and active positioning of the hand near objects of interest. All of these motor activities require visuomotor coordination, and it is this coordination that could be achieved by the motor cortex and interconnected parietal and cerebellar areas.