The use of extrinsic stimulation to control activity in neuronal networks i.e., neurocontrol, is a key problem in control engineering and neuroscience. Here, we study the general problem of selective spiking in a population of neurons. The goal is to use an input stimulus in order to induce a spike in a specific neuron of a population while keeping all others suppressed. We formulate a strict version of this problem for the class of Integrate-and-Fire neuron models, which amounts to an optimal control problem with state constraints. While possible to solve in low dimensions, the strict problem is harder to handle for larger networks. Thus, we relax the problem via regularization and derive the ensuing optimal controls for selective spiking. The properties of the solution are highlighted through several examples. The results provide a tractable, scalable solution for a baseline neurocontrol problem.