How common is it for a manic or depressive phase to only last a few days? My reading on the subject and experience with people who have this issue would suggest it's measured in months. Although I guess I have also heard of terms like "rapid cycling".
I've been successfully managing a bipolar disorder for over ten years - and unsuccessfully for years prior to that.
In my _personal_ experience, manic phases last 2 weeks to 2 months, followed by 4-10 months of catatonia.
It is approximately the same for my aunt and grandmother. In their cases, lithium and electroconvulsive therapy were the only things that work. For me, lamotrigine and modafinil are life-changing wonder drugs.
That said, people in my group therapy reported cycles as short as days to weeks - rapid cycling, as you say.
I'd guess, like another commenter, that some people might get this type of oscillating behaviour due to feedback mechanisms: working for 48 hours straight (enabled by mania) could trigger a recession, requiring intensive recovery, which enables a new visible episode of mania.
Note that the "off time" need not be matched by a depressive episode (and in fact I find that unlikely, since depression is not very restful) -- it could simply be exhaustion hiding the symptoms of the underlying (long) manic episode.
It depends. It can last a day, a week, a month, or even longer.
Consider that it's "just" a mood disorder, where mood management can be affected by random stimuli. So, their mood might be stable until some stimulus comes by. And that stimulus could just be a random memory, which can create a feedback loop. It's "fun".
