Because smooth-pursuit eye movements (SPEM) can be executed only in the presence of a moving target, it has been difficult to attribute the neuronal activity observed during the execution of these eye movements to either sensory processing or to motor preparation or execution. Previously, we showed that rhesus monkeys can be trained to perform SPEM directed toward an “imaginary” target defined by visual cues confined to the periphery of the visual field. The pursuit of an “imaginary” target provides the opportunity to elicit SPEM without stimulating visual receptive fields confined to the center of the visual field. Here, we report that a subset of neurons [85 “ imaginary” visual tracking (iVT)-neurons] in area MST of 3 rhesus monkeys were identically activated during pursuit of a conventional, foveal dot target and the “imaginary” target. Because iVT-neurons did not respond to the presentation of a moving “imaginary” target during fixation of a stationary dot, we are able to exclude that responses to pursuit of the “imaginary” target were artifacts of stimulation of the visual field periphery. Neurons recorded from the representation of the central parts of the visual field in neighboring area MT, usually vigorously discharging during pursuit of foveal targets, in no case responded to pursuit of the “imaginary” target. This dissociation between MT and MST neurons supports the view that pursuit responses of MT neurons are the result of target image motion, whereas those of iVT-neurons in area MST reflect an eye movement–related signal that is nonretinal in origin. iVT-neurons fell into two groups, depending on the properties of the eye movement–related signal. Whereas most of them (71%) encoded eye velocity, a minority showed responses determined by eye position, irrespective of whether eye position was changed by smooth pursuit or by saccades. Only the former group exhibited responses that led the eye movement, which is a prerequisite for a causal role in the generation of SPEM.