Hi gmanatl
This puzzle is generally considered impossible to solve - indeed, there are mathematical proofs of this. However, Brian doesn't fully explain the rules, which means, I would argue, that it is solvable the way he explains it. This caused me some confusion when I thought I'd solved it.
Brian doesn't explain that the lower middle room (room 4 in the video) is taken to have two upper walls, one either side of where it that touches the wall above it; in other words, you need to assume that there is a wall between rooms 4 and 1 and another wall between rooms 4 and 2. Now, conventionally, this would be considered one wall, not two - any normal person standing in that room would say it had four walls! - but that's not how it is for this puzzle. Of course, if you're going to introduce an unconventional interpretation of a word, you really need to make that clear, which Brian doesn't.
If you take room 4 to have only 4 walls, you can indeed solve the puzzle (I did), but if you take it to have 5 walls then you can't. Sadly, however, the former version is not the famous insoluble 5 room problem whereas the latter is, so there's little fame or glory to be had!
Anyway, have a look a this webpage for a clear explanation of the rules and a helpful diagram, and then see if you've actually solved the puzzle. They also explain it in terms of doors between rooms which I found to be a more helpful way to think of it.
http://www.archimedes-lab.org/How_to_Solve/5_rooms.html
The wrong solution shown at the top of this webpage is the one that I would think is actually correct if we take the conventional meaning of the word 'wall' - and its exactly what I came up with after watching that episode. Is that the one you have?
Anyway, hope all this is relevant ... ! Maybe you've come up with something completely different?
All the best,
Matt