There's a problem with Ramachandran's logic. Let's walk through it.

1. He says that there are collections of neurons in the brain that map specifically to places on your skin, so that when you are touched at that place on your skin, the corresponding neurons fire in the brain. That's true.
2. He says that people who have lost a limb can still feel that limb. That is also true; the phenomenon is called Phantom Limb. Phantom Limb occurs because those neurons that are mapped to the missing limb can still fire, even though there is no limb to send signals to the neurons.
3. He says that people who watch someone else getting pricked by a pin cause the same neurons to fire in their brain that would have fired had they been pricked by the pin themselves, but that you don't actually feel the pain that they are feeling because the brain can tell that the signals are not coming from your limb. Ramachandran calls this empathy. That makes perfect sense; the brain is always modelling the outside world, and it can do that through either the skin or the eyes.
4. He says that people with missing limbs can look at someone being pricked with a pin and feel pain in their phantom limb, because (essentially) the brain cannot tell the difference between their imagined limb, and someone else's imagined limb. This also makes sense to me.
5. He concludes that, since the person with the phantom limb cannot tell the difference between his own phantom pain, and the pain that he imagines someone else is having, that this proves that people's minds are connected together in some metaphysical way.
   

This last point (#5) is where his logic begins to break down.

Remember, Ramachandran already stated that the reason the sensory neurons fire in someone else's brain when they see someone else being pricked is that the observer's brain is constructing a model of the event from what they saw. No unseen metaphysical connection is required to accomplish this, and therefore it's not possible to draw any conclusions about what might happen if the "observer" doesn't actually see the event with his own eyes.

This of course, doesn't disprove that mirror neurons could somehow be involved in telepathy or remote viewing, but Ramachandran's explanations don't provide any evidence either. He himself describes the purpose of these neurons as having a modeling function that works through the ordinary senses, not some form of telepathy.

Note that Ramachandran himself clarifies the issue in the same talk, when one of the audience members asks him how mirror neurons are related to telepathy. His answer: they aren't:

The important thing about mirror neurons: there's nothing psychic or spooky going on. You're taking a visual input and transforming it into somatosensory sensations. When you meditate, you can't synchronize [these sensations solely] in meditation, but you can look at body language, facial expressions, beautific expressions and so forth, and that could transmit to mirror neuron systems. The emotional aspects of mirror neurons have not been adequately studied.

To hear Ramachandran respond to the audience member, go to this page, scroll down the list of topics in the center of the page until you get to "Nothing Psychic About Mirror Neurons", and click on the link provided.

What about pineal gland? Have you ever consider that as an option for communicating without the "physical eyes"