Is statistical learning of a salient distractor's color implicit, inflexible and distinct from inter-trial priming?