C.,
My son started having sleep terrors at the same age. He had surgery prior to it, and I've heard that either the trauma of the surgery at such a young age when they can't understand and/or the anesthesia can be a trigger. He really DID look awake, and on some level could recognize we were there but "where" he thought we all were was not reality, but his dream state. He would even say "mama". He was Dx'd with the terrors and they continued off and on for another 2 years. Now they are few and far between. Our doctor was adamant about not leaving our son to 'cry it out' because if he was in a terror he could very seriously hurt himself by trying to 'escape' whatever he was afraid of and attempting to get out of his crib/bed. It IS like a sleep walking stage. We also were told to absolutely not use the Ferber method with him because if he fell asleep in fear or upset, it was only making the whole situation worse guaranteed a night of terrors.
Only YOU know what feels right for your family. The medical community knows very little about sleep terrors. There are a lot of alternative methods but they seem a bit out there. I've tried most of them. Some help, some not. Keeping a routine and never letting your child get over tired are crucial.
My child did sleep in his own crib in his own room almost immediately coming home from the hospital and still does, of course at almost 5,(but in a bed!)...however, I did always go to him to comfort him in the night if he needed it. But we were not into cosleeping as a practice. I prefer to sleep in my own bed, w/o my child, frankly. Here is food for thought: I find it sort of hypocritical that as parents, WE sleep not alone but usually with a spouse/partner, and if WE have a bad dream in the night we roll over to our spouse for comfort, or get up turn on the light and get a drink of water, none of which our child can do. Yet we expect them to just lay in the dark, alone, with no comfort until morning when they have a bad dream. For our convenience.
Best wishes and follow your heart,
L.