Generation of limb trajectories with a sequential network

The authors of this study focus on the representation and generation of unconstrained aiming movements of a limb by means of a neural network architecture. The goal is that of producing a time trajectory of a limb from a starting posture toward a target specified as a sensory stimulus. The velocity profile along the trajectory is bell-shaped as in most movements performed by biological systems. The generalization capabilities of the network are investigated as well as its internal organization. Two experiments are performed on the trained network to test its robustness to noise and its dynamical properties.<>