I don't know what you do to assess their state of consciousness, but it doesn't sound very nice so far.
No nice at all. In the UK one of the favorites is nipple twisting. In the States we usually just apply intense nailbed pressure. The point is to inflict pain to see some form of response for diagnostic and predictive purposes.
What is the difference between this state and motor neuron disease? People with MND "do not respond" - at least not much. But they are certainly conscious.
In no way, shape, or form even close. Motor neuron diseases, in their various forms, affect motor output only. The brain is left relatively unaffected (in ALS there is some evidence of a mild dementing process in some people) so the EEG remains completely normal. We see normal wake-sleep cycles and normal background rhythms when they are awake. There are also certain nerves not affected by the process so we can continue communicating with people.
It sounds to me from your description that the people you are talking about are experiencing something like a dream.
Nope. Nothing like a dream state. We know what dream states look like on an EEG -- clear correlations. This is nothing like that. The only similarity is with spindle coma and that is not like a sleep state either.
You see, hear, feel and are conscious during a lucid dream.
Using proper clinical terms lucind dreaming is a hallucination. It is internally generated, but seeing, hearing, and feeling certainly occur.
During normal REM, the same parts of the brain are active as during waking perception but the "thresholds are lower" (or something like that) so the activity is less intense, accounting for the relative "dimness" of normal dreams compared to veridical perception.
Not completely correct. We see a very superficially similar pattern on the EEG during REM sleep. The areas of the brain that are activated are largely in the visual cortex. One of the very interesting aspects of dreaming is that smells and tastes are rarely (but sometimes) reported.
If thresholds were lower, then the activity would be more intense, but that is just a technical use of the terms and not important for the discussion. The real difference is that the experience is possibly muted to a certain extent for very unclear reasons (we don't really know since we must rely on reports through memory) and the sense of it all is very strange and certainly not organized along daily waking rules. We are also paralyzed during the whole thing.
During lucid dreams we are still "disconnected from the physical world" (the experiences are "wild", like dreams) but the level of activity in the brain is the same as being fully awake.
Based on what evidence? Where is the cut-off for a lucid rather than a "normal" dream? There are dream states that actually begin in the waking state. Are you referring to those?
So in your coma patients, is the brain state closer to that during REM or that during lucid dreaming and full-waking states?
It's not in any way like any of those. If you go on the EEG evidence it is wholly other.