Arguing for the psychological uses of narrative, Bill Benzon at The Valve distinguishes the "autobiographical self" (i.e. identity over time) from the "core self" (i.e. one's integrated psycho-physiological state at any given time). He claims that the "core self," influenced as it may be by intense situational and physical factors (he uses hunger and sexual desire as examples), not to mention its transient nature, threatens to disrupt the autobiographical self. He suggests that narrative (he specifically mentions "play-acting" and "storytelling") usefully provides an overarching frame within which to understand and evaluate our dispositions and behaviors over time.
The account leaves out a lot of course (for instance, it would seem that temperament straddles both kinds of self). And his case seems a bit extreme--as if even a starving man would look back on his life as having been little more than an ultimately unsuccessful quest for food--but there may be something to it. After all, someone in a deep depression may view much of his past "through a glass darkly" in a way that lightens considerably when the episode relents. And obviously the two selves affect each other reciprocally and continuously.
Staying with Benzon's schema, it would seem that psychological distress occurs in two varieties. Unhappiness is a malady of the autobiographical self, a dismayed sense that one's story has somehow gone awry through vicisitudes of sensibility or circumstance. One seeks in a therapist a kind of narrative catalyst that will open up unimagined possibilities, including the often profound possibility of actually being listened to and perhaps even understood. Dysfunction of the core self manifests as symptoms that may actively impede functioning. There is considerable overlap between the two, but arguably we resort to psychotropic medication inasmuch as symptoms appear to be beyond the power of narrative to reframe. But nothing is more frustrating than to try to treat unhappiness with meds or to tackle narrative-resistant symptoms with more narrative. Diagnostic confusions and controversies arise from the difficulty of distinguishing symptoms from unhappiness.
It occurs to me that like certain other phenomena such as religion and even music, narrative broadly considered (that is, interest in all stories whether contained in books, film, gossip, or hearsay) is hard to explain because it is very widespread but not truly universal. Some faculties, such as hunger and thirst, are obviously ubiquitous because their absence is not compatible with life. Others, such as basic senses or sexuality, are not imperative for individual life but are so typical of the species that their absence is uncontroversially deemed pathological.
Inasmuch as existence is necessarily temporal, some interest in narrative is presupposed, even if only speculation as to where the next meal will come from. But sophisticated narrative--that is, at least at the level of communal folk tales--has, like religion, been found to exist in virtually every human society. And yet just as there is a reliable minority of individuals who are irreligious, there are of course people here and there who are relatively free of the narrative bug, who may be more invested in other domains of experience (facts, ideas, bodily experience, etc.). If religion and narrative truly are central to (individual and species) human identity, then how is it that even a (non-pathological) small minority more or less escape their purview? Perhaps diversity itself is such a powerful evolutionary engine that it constantly throws out alternatives to the prevailing cultural trajectory, suggesting of course that those faculties we view as indispensable are actually contingent.