A SELF-ORGANIZING NEURAL MODEL OF MOTOR EQUIVALENT REACHING AND TOOL USE BY A MULTIJOINT ARM

This paper describes a self-organizing neural model for eye-hand coord ination. Called die DIRECT model, it embodies a solution of the classi cal motor equivalence problem. Motor equivalence computations allow hu mans and other animals to flexibly employ an arm with more degrees of freedom than the space in which it moves to carry out spatially define d tasks under conditions that may require novel joint configurations. During a motor babbling phase, the model endogenously generates moveme nt commands that activate the correlated visual, spatial, and motor in formation that are used to learn its internal coordinate transformatio ns. After learning occurs, the model is capable of controlling reachin g movements of the arm to prescribed spatial targets using many differ ent combinations of joints. When allowed visual feedback, the model ca n automatically perform, without additional learning, reaches with too ls of variable lengths, with clamped joints, with distortions of visua l input by a prism, and with unexpected perturbations. These compensat ory computations occur within a single accurate reaching movement. No corrective movements are needed. Blind reaches using internal feedback have also been simulated. The model achieves its competence by transf orming visual information about target position and end effector posit ion in 3-D space into a body-centered spatial representation of the di rection in 3-D space that the end effector must move to contact the ta rget. The spatial direction vector is adaptively transformed into a mo tor direction vector, which represents the joint rotations that move t he end effector in the desired spatial direction from the present arm configuration. Properties of the model are compared with psychophysica l data on human reaching movements, neurophysiological data on the tun ing curves of neurons in the monkey motor cortex, and alternative mode ls of movement control.