Combining Oculus and perception neuron integrations

Making the camera a child of the head bone is probably going to double-count your head's motion. The head bone's transform is controlled by the Perception Neuron and the camera's transform is controlled by the Rift. You're probably right that the Rift's tracking system will give a more accurate head pose, but we still need to know the relationship between the head and the rest of the body. I would suggest making the camera the root GameObject of your rig and writing a script that moves the body rig each frame so that the head bone's world pose matches the camera's world pose.