Furthermore, these state transitions are almost entirely learned reflexes. Consciousness tends to merely reflect on what happened and judge whether it was a mistake or not.
That people are Markov should be the guiding paradigm of psychology, yet psychologists keep talking about cycles rather than state transitions, and of moods like depression in terms of months not moments. Odd.
If people are Markov then:
The way to avoid an unwanted mental state is to un-learn reflexive transitions to that state, and to learn as many as possible different transitions out of that state.
The way to increase the likelyhood of a good mental state is to learn as many as possible transitions into that state, and to learn not to transition out of it.
A manic-depressive person's balanced mental state has many transitions out, either to mania or depression. The maintenance of a balanced state should require finess but not a lot of effort.
It may require some effort to get a depressive person out of a depressive state, but once that is achieved keeping them out requires finess rather than brute force -- finding ways they transition to depression and tweaking them just a little (some similarity to Cognitive Behavioural Therapy here).
Reducing post-traumatic stress is a matter of avoiding transitions into remembering the event, learning ways to stop remembering it, and un-learning physiological transitions induced by the memories.
etc