Coarse gain recurrent integrator model for sensorimotor cortical command generation

The coarsely sampled inverse Jacobian, or "coarse gain" recurrent integrator command generator (CGRICG) model is a preliminary proposal for motor command generation in the sensorimotor cortex during unloaded horizontal reaching movements based significantly upon implicit coordinate transformation. The core of the model is an integrator that in negative feedback configuration performs approximate differentiation of a central movement trajectory command. Together with an accurate forward kinematic model and a distribution and scaling network that is only coarsely tuned in relation to movement direction, the feedback configuration achieves inverse kinematic transformation implicitly throughout the workspace. Simulations of the CGRICG directing motion of a six-muscle, two-joint arm model reproduce gross kinematic features of human and primate arm movement as well as plausible neurophysiological signals in internal sensorimotor cortical neurons. The CGRICG model suggests that cortical network connection strengths may be adapted to improve straightness of motion without affecting final target location.