Skip to Main content Skip to Navigation

Identification and simulation of the motor command during the 3D non-contraints multi-articular movements

Abstract : The purpose of this thesis is to identify principles that could guide the planning of 3D upper-limb movements for different individuals. To this aim, the chosen approach combines novel experiments (namely, a “free reach-endpoint" motor task) with advanced computational techniques (here numerical inverse optimal control). Arm pointing movements without a prescribed final hand position are examined under different conditions of speed or load in order to let emerge various motor control strategies and assess the possible underlying motor planning principles. A core idea is to depart from classical point-to-point reaching paradigms (where the target is generally a dot, e.g. a spotlight target) to study a task in which the endpoint is left free to theparticipants in order to emphasize inter-individual differences as well as the selection process and motor decision that led to the observed strategies. This paradigm thus allows to better decipher the characteristics of the human motor controller. Empirical results are then modeled and interpreted in the inverse optimal control framework, hypothesizing that empirical arm trajectories derive from the minimization of a certain, possibly composite, cost function. This combined approach aims at revealing which principle or rule conceivably drives the planning process of these unrestrained upper-limb movements and to stablish a link between relevant motion parameters, cost functions and inter-individual peculiarities.The results show that subjects produced different motor strategies at both kinematic and dynamic levels depending on how they adapted to speed and/or load variations. Overall, significant motor adaptation of hand trajectories (e.g. location of reach endpoints) and motor commands (e.g. use of interaction torque) were found. Yet, some subjects exhibited stronger dependences than others who varied only little their reach strategies with respect to task-induced speed or load changes. When investigated from the optimal control viewpoint, these results could be accounted for by a composite cost essentially weighting kinematic and dynamic variables differentially at the motor planning stage. Such a composite model outperformed separate kinematic and dynamic ones in predicting the evolution of many important motion features and in explaining inter individual differences. Moreover, it allowed reconciling controversial findings of previous studies by showing that divergent adaptive behaviors can emerge depending on the weights of the elementary cost that may compose the total cost function. In sum, the present results suggest that motor planningof unrestrained3D arm movements necessarily mixes kinematic and kinetic variables and that this trade-off may be idiosyncratic and lead to subtle inter individual differences.
Document type :
Complete list of metadata

Cited literature [147 references]  Display  Hide  Download
Contributor : ABES STAR :  Contact
Submitted on : Monday, April 24, 2017 - 9:45:31 AM
Last modification on : Saturday, July 2, 2022 - 3:52:14 AM
Long-term archiving on: : Tuesday, July 25, 2017 - 2:00:19 PM


Version validated by the jury (STAR)


  • HAL Id : tel-01512591, version 1


van Hoan Vu. Identification and simulation of the motor command during the 3D non-contraints multi-articular movements. Neuroscience. Université Paris Saclay (COmUE), 2016. English. ⟨NNT : 2016SACLS509⟩. ⟨tel-01512591⟩



Record views


Files downloads