We present a novel manipulandum for understanding the sensorimotor processes involved in object grasping. We have developed a closed-loop prosthetic hand, with 2 degrees of control and 32 channels of vibrotactile feedback of fingertip force and finger positions. In order to understand this model sensorimotor circuit we first tackle two sub-problems: (Q1) Do we integrate artificial sensory feedback (vibrotactile) with our other modalities (vision, proprioception) in a statistically optimal manner based on sensory uncertainty? We run subjects through a pursuit tracking task with noisy visual and vibrotactile cues to cursor location, and describe the resulting trajectories with a Kalman filter model; and (Q2) Are grasp trajectories and temporal force profiles a predictable function of the actuation commands that control the hand and the available feedback? We run subjects through a new tracking task where the grasp size and force on an object are modulated, and compare the resulting trajectories to those predicted by the optimal feedback control (OFC) framework.