By definition STDP is full learning alghoritm that makes connections between neurons stronger or weaker. If that taken in consideration then you dont need anything else. BUT, if you make enuph examples of some particular input you will have all you need to make "weight" dimension on that neuron. If you use same analogy on neuron connections maybe you can get satisfactorily outcome. Result you are looking for is set of inputs that accure in same time. Basicly what that means, you need set of inputs that are close enuph (custom delta error) who gives you general maximum of examined neuron.