During transport of hand-held objects, the grip force is modulated in parallel with the load force changes. The control scheme underlying this grip-load force coupling involves subtle interplay between feedforward and feedback mechanisms. Based on internal models of the motor system and object properties, the load force can be predicted and the GF motor command can be specified in a feedforward manner. Moreover, during the course of arm movement, the CNS is informed by sensory feedback about mechanical events such as the lift-off of the object, slippage or excessive grip force. This information is used to correct the motor commands and to update the internal model of the motor apparatus and object. In this thesis, three experiments were conducted to examine the relative contributions of sensory-driven and anticipatory control of GF adjustments during cyclic vertical movement with a hand-held load. The main point was to assess whether internal models underlying the grip-load force coupling are robust when the environmental context was changed or when the sensory feedback was suppressed. Two experiments in parabolic flight were conducted to study the effects of a change in gravity on the dynamics of prehension. The main perturbation was that the novice subjects applied unnecessarily high safety margins during their first trial at 0 and 1.8 g in order to secure the grasp insofar as the gravitational component of the load force was unpredictable. By contrast, the temporal coupling between GF and LF was maintained regardless of the gravity conditions because the inertial component of the load could be still predicted from the arm motor command (efference copy). In the second study performed during parabolic flight, we have observed that the subjects were able to exert the same grip force for equivalent load generated either by a change of mass, gravity or acceleration despite the fact that it requires different arm motor commands. These two experiments brought further evidence that the predictive mechanisms largely contribute to the GF adjustment. Static forces such gravity are taken into account in the motor plan allowing adequate motor command and precise prediction of the incoming load force change. The GF output would depend on the precision of this prediction that can be evaluatedonly after the movement onset through sensory information about the actual state of the system. The third experiment performed in this thesis studied the role of cutaneous afferents in object manipulation by anesthetizing the thumb and index finger. In addition to their phasic slip-detection function, the cutaneous afferents are required for setting the background level of the grip force. Actually, in absence of tactile feedback, the temporal coupling between the grip and load forces is maintained but the mean magnitude of GF progressively decreases leading to object slipping. It is hypothesized that accumulating error occurred in the LF prediction leading to inadequate level of GF. Cutaneous afferents are thus required to correct and maintain the internal model of the arm-hand object system.

During transport of hand-held objects, the grip force is modulated in parallel with the load force changes. The control scheme underlying this grip-load force coupling involves subtle interplay between feedforward and feedback mechanisms. Based on internal models of the motor system and object properties, the load force can be predicted and the GF motor command can be specified in a feedforward manner. Moreover, during the course of arm movement, the CNS is informed by sensory feedback about mechanical events such as the lift-off of the object, slippage or excessive grip force. This information is used to correct the motor commands and to update the internal model of the motor apparatus and object. In this thesis, three experiments were conducted to examine the relative contributions of sensory-driven and anticipatory control of GF adjustments during cyclic vertical movement with a hand-held load. The main point was to assess whether internal models underlying the grip-load force coupling are robust when the environmental context was changed or when the sensory feedback was suppressed. Two experiments in parabolic flight were conducted to study the effects of a change in gravity on the dynamics of prehension. The main perturbation was that the novice subjects applied unnecessarily high safety margins during their first trial at 0 and 1.8 g in order to secure the grasp insofar as the gravitational component of the load force was unpredictable. By contrast, the temporal coupling between GF and LF was maintained regardless of the gravity conditions because the inertial component of the load could be still predicted from the arm motor command (efference copy). In the second study performed during parabolic flight, we have observed that the subjects were able to exert the same grip force for equivalent load generated either by a change of mass, gravity or acceleration despite the fact that it requires different arm motor commands. These two experiments brought further evidence that the predictive mechanisms largely contribute to the GF adjustment. Static forces such gravity are taken into account in the motor plan allowing adequate motor command and precise prediction of the incoming load force change. The GF output would depend on the precision of this prediction that can be evaluatedonly after the movement onset through sensory information about the actual state of the system. The third experiment performed in this thesis studied the role of cutaneous afferents in object manipulation by anesthetizing the thumb and index finger. In addition to their phasic slip-detection function, the cutaneous afferents are required for setting the background level of the grip force. Actually, in absence of tactile feedback, the temporal coupling between the grip and load forces is maintained but the mean magnitude of GF progressively decreases leading to object slipping. It is hypothesized that accumulating error occurred in the LF prediction leading to inadequate level of GF. Cutaneous afferents are thus required to correct and maintain the internal model of the arm-hand object system.