This thesis investigates how adults, infants, and children use visual information to control reaching movements. To do this, the kinematics of a reaching movement were recorded using a 6 camera motion tracking system. Adult reaching movements were investigated in three different experiments. The first experiment looked at the effect of visual information about the reach space and the target on reaching movements. Adult reaches are significantly affected by removal of visual information about the reach space and the target, e.g. peak speed decreases as distance information is occluded. Adult reaching movements are also affected by the length of the delay between viewing an object and then reaching for it in complete darkness. Experiment 2 reveals that there is a linear increase in movement duration, decrease in peak speed, and increase in maximum grip aperture with increasing temporal delay. This is due to the decay of dorsal visual information. Experiment 3 found that a cautious reaching movement can be defined as a reach where duration increases, average speed decreases, and peak timing is proportionally earlier in the reach. The three developmental experiments investigated the changes in reaching in infancy, in typically developing 5-year-old children, and in children with Williams syndrome (WS). Between 9 and 16 months of age, infants develop better coordination of reaching movements. Improved dark reaching behaviour in 16-month-olds may be associated with more mature processing in the dorsal visual stream. Reaching movements in 5-year-old children are straighter and faster than infant reaches but behaviour is not yet stereotyped like adults. WS children generate reaches that are slower and incorporate more movement units than 5-year-old children, and WS behaviour may reflect poor processing of dorsal visual information rather than poor control of arm movement.

The objects with which the hand interacts with may significantly change the dynamics of the arm. How does the brain adapt control of arm movements to this new dynamic? We show that adaptation is via composition of a model of the task's dynamics. By exploring generalization capabilities of this adaptation we infer some of the properties of the computational elements with which the brain formed this model: the elements have broad receptive fields and encode the learned dynamics as a map structured in an intrinsic coordinate system closely related to the geometry of the skeletomusculature. The low--level nature of these elements suggests that they may represent asset of primitives with which a movement is represented in the CNS.

To coordinate visual perception and motor control in daily life where we are constantly surrounded by motion, we are dependent on normal visuomotor capacity. One essential prerequisite for normal visuomotor capacity is smooth pursuit eye movements (SP). Infants born very preterm (VPT = born <32 gestational weeks) are at high risk of developing disabilities in higher brain functions i.e. perception, cognition, concentration and coordination. In this thesis visuomotor capacity was investigated in a cohort of VPT infants (n = 113) and compared to control groups of full term (FT) infants. Levels of SP were measured at 2 and 4 months’ corrected age (CA). At 8 months’ CA reaching capacity toward a moving object was evaluated as this represents an executive activity guided by vision that develops at an early age. Lower levels of SP were found in the VPT infants compared to FT controls. The VPT boys showed higher levels of SP compared to the VPT girls. In VPT infants without major neonatal morbidities lower levels of SP was found compared to the FT controls. No difference in total capacity of gaze tracking was found, although the VPT infants lagged the object more at 4 months’ CA and used more saccades at 2 months’ CA. With age the VPT infants’ SP levels increased, but with a wider dispersion compared to the FT controls, and the levels of SP at 4 months’ CA corresponded to the levels of the FT infants at 2 months.  A number of perinatal risk factors were found to be negatively associated to lower levels of SP, and this effect was more pronounced in VPT infants with multiple risk factors,.  When evaluating the capacity to reach a moving object at 8 months’ CA, the VPT infants showed significantly more bimanual reach and more curved reaching paths to catch the object as compared to the FT control group. In conclusion, a delayed visuomotor capacity was found in VPT infants compared to FT control infants at 2, 4 and 8 months’ CA. Some VPT infants with perinatal risk factors did not develop in levels of SP between 2 and 4 months’ CA.

Human sensorimotor control is remarkably adept at utilising contextual information to learn and recall systematic sensorimotor transformations. Here, we investigate the motor representations that underlie such learning, and examine how motor memories acquired based on different contextual information interact. Using a novel three-dimensional robotic manipulandum, the 3BOT, we examined the spatial transfer of learning across various movement directions in a 3D environment, while human subjects performed reaching movements under velocity-dependent force field. The obtained pattern of generalisation suggested that the representation of dynamic learning was most likely defined in a target-based, rather than an extrinsic, coordinate system. We further examined how motor memories interact when subjects adapt to force fields applied in orthogonal dimensions. We found that, unlike opposing fields, learning two spatially orthogonal force fields led to the formation of separate motor memories, which neither interfered with nor facilitated each other. Moreover, we demonstrated a novel, more general aspect of the spontaneous recovery phenomenon using a two-dimensional force field task: when subjects learned two orthogonal force fields consecutively, in the following phase of clamped error feedback, the expression of adaptation spontaneously rotated from the direction of the second force field, towards the direction of the first force field. Finally, we examined the interaction of sensorimotor memories formed based on separate contextual information. Subjects performed reciprocating reaching and object manipulation tasks under two alternating contexts (movement directions), while we manipulated the dynamics of the task in each context separately. The results suggested that separate motor memories were formed for the dynamics of the task in different contexts, and that these motor memories interacted by sharing error signals to enhance learning. Importantly, the extent of interaction was not fixed between the context-dependent motor memories, but adaptively changed according to the task dynamics to potentially improve overall performance. Together, our experimental and theoretical results add to the understanding of mechanisms that underlie sensorimotor learning, and the way these mechanisms interact under various tasks and different dynamics.

Choisir quel mouvement effectuer est une fonction primordiale du système nerveux central. Comment ces décisions sont prises est encore sujet à débats. Une hypothèse traditionnelle pose qu’elles sont prises de façon sérielle, à l’aide de processus perceptifs qui alimentent un exécutif central, qui communique ensuite au système moteur quel mouvement effectuer. L’hypothèse alternative préférée par notre équipe est que les mouvements potentiels commencent à être préparés en parallèle et entrent en compétition pour les effecteurs. Dans le but de tester ces hypothèses, notre équipe a enseigné à un macaque une tâche de prise de décision motrice. Le sujet y est placé devant un écran où deux cibles apparaissent. Chacune a une valeur qui découle de deux caractéristiques : sa luminosité (BU pour bottom-up, information ascendante) et l’orientation d’une ligne qui la coupe comme une aiguille d’horloge (TD pour top-down, information descendante.) Le sujet choisit une des deux à l’aide d’un mouvement d’atteinte, et reçoit une récompense proportionnelle à sa valeur. Cette tâche permet de comparer plusieurs types d’essais : certains présentent une seule cible, une absence de choix, ou deux cibles identiques, un choix sans conséquence. D’autres ont une cible plus valable que l’autre, le choix est alors facile. On peut alors faire varier la caractéristique (BU ou TD) qui donne une plus grande valeur à la meilleure cible. Finalement, on peut montrer deux cibles de valeur égale, mais dont une tire sa valeur d’un bon score TD et l’autre, d’un bon score BU. Le sujet doit alors, en quelque sorte, choisir entre les caractéristiques. Pendant que le sujet exécute la tâche, on enregistre ses potentiels de champ locaux (LFP) à l’aide de deux réseaux d’électrodes déplaçables individuellement, l’un placé dans le cortex pariétal postérieur (PPC) et l’autre, dans le cortex prémoteur dorsal (PMd). L’analyse de ces données à l’aide de spectrogrammes, et une discussion des réactions spécifiques dans les bandes de fréquences alpha, bêta et gamma, sont présentées ici. / Choosing which movement to make is a primary function of the central nervous system. How these decisions are made is still a matter of debate. A traditional hypothesis posits that such decisions are made in a serial fashion: perceptual processes feed into a central executive, which then communicates to the motor system which movement to make. The alternative hypothesis preferred by our team is that potential movements begin to be prepared in parallel, and compete for effectors until a consensus forms in brain areas related to controlling the movements. In order to test these hypotheses, our team taught a macaque to perform a reach-based decision-making task. The subject is placed in front of a screen on which two targets appear. Each target has a value derived from two features: its brightness (BU, bottom-up information) and the orientation of a line that crosses it like a clock hand (TD, top-down information.) The subject freely chooses one of the two targets by reaching it, and then receives a reward proportional to its value. This task compares several types of trials: some show a single target, therefore no choice, or show two identical targets, which means the choice has no consequences. Other trials have one target that is more valuable than the other, which makes the choice easy. The feature which gives that better target a greater value can be either BU or TD. Finally, some trials show two targets of equal value, but one of them derives its value from a good TD score while the other derives its value from a good BU score. The subject must then choose between the features. While the subject performs the task, local field potentials (LFP) are recorded using two individually movable electrode arrays. One array is placed in the posterior parietal cortex (PPC) and the other, in the dorsal premotor cortex (PMd). The data thus obtained is analyzed using spectrograms, and a discussion of specific responses in the alpha, beta, and gamma frequency bands is presented here.