a few notes: i'm more or less a materialist, and i accept some version of a token-token identity thesis: that is, for example, there might not be an interesting identity between, let's say, anger, and a discernible family of material brain states, but in any given case of anger, there is some material state that corresponds to or is that event. my only qualification would be that the state is not necessarily exclusively a brain state, but could encompass a state of the nervous system, the organism as a whole, and the external world (i am a "content externalist").
also, on the concept of disease: we usually think of disease as the operation of an external agent on an organism, like a virus, e.g. but this is a difficult question, and some ways that the organism might degenerate or be deranged could count as diseases, as in some caused by malnutrition or vitamin definciencies etc: the notion of disease is not well-defined in common use. alzheimer's is an interesting case, for example.
but i don't think that bipolar is likely to be well-described in either of these senses. rather, there might be a family of characteristic brain conditions correlated to some extent with a constellation of symptoms. but that's the very best you will get, moving from an extraordinarily rough symptomatic constellation, generalized as a disease. since every brain-state resembles every other brain-state in indefinitely many respects (indeed, every thing resembles every other thing in indefinitely many respetcs) it's definitely not surprising if we could roughly identify some resemblances between the token brain states correlated with bipolar mental/emotional states. and notice, there might also be correlations to environmental factors and so on that are probably more or less exactly as good.
then direction of explanation introduces a whole host of further complications, not seriously present even in a case like alzheimer's: you're not liable to be able to temporally arrange the relevant brain states as prior to the emotional symptoms and so on.
and then you have to ask: what do you get out of believing that bipolar is a disease? who's making the money and devising the models? etc. i don't do all this to fuck up your treatment, but what one wants or needs to believe to get better is not the same as what is true; indeed, that you want or need to believe it or benefit from believing it is a good reason to be sceptical of it.