Comparing Visual Attention with Leading and Following Virtual Agents in a Collaborative Perception-Action Task in VR