Spatial Constraints on Visual Statistical Learning of Multi-Element Scenes