The question as to whether specific antimanic drugs differ in their action profile from nonspecific drugs is addressed in regard to symptomatic, nosographic, regulatory and physiopathological issues. Results from clinical studies have shown that mood stabilizers and typical neuroleptics differ as regards improvement of manic symptoms: the former appear to act more evenly on all symptoms of mania, showing a more total normalization of affect, ideation and behaviour whereas the latter tend to sedate patients or to cause a psychomotor retardation, leaving the core manic symptoms unaffected. This has been many times underlined, in particular for lithium, notwithstanding the fact that rating scales employed in clinical trials have often been charged to fall far short of being sensitive enough to pick up the qualitative changes in manic psychopathology. Antimanic drugs may also be more or less specific in their capacities to treat all facets of the manic episode (psychotic, depressive, irritable) whatever the bipolar subtype (bipolar I, II, rapid and non-rapid cycling, secondary bipolar disorder) or the disease stage (early and late episodes). In this respect divalproate seems to have a broader spectrum of efficacy than other available agents. Newer antipsychotics such as olanzapine are promising too. From a regulatory point of view, the current European requirements for a specific antimanic drug are more stringent than the US requirements of the Food and Drug Administration (FDA). Efficacy must be demonstrated in short-term studies showing an effect in acute mania; moreover it has to be shown that efficacy is to be maintained during the episode. So far, three armed randomized controlled trials are required, in which the test product is compared both with placebo and with a standard treatment. A possible design is a comparison of test product, placebo and active control for 3 weeks followed by a two-arm phase for the remaining 9 weeks, comparing only test product and active control. In addition, a specific antimanic has to demonstrate that it does not cause switching to depression. As regards physiopathology, integrative models of bipolar disorder, ie kindling and behavioural sensitization, offer an exciting perspective on the specificity issue; agents active in these models initialize a cascade of intracellular signaling that leads to changes in the expression of immediate early genes as well as late effector genes in corticolimbic structures: the former may contribute to acute symptomatology whereas the latter give rise to neuroanatomical reorganization which could underlie more stable changes in mood and cognition. Due to their action on intracellular messaging systems, dopamine D(1) receptors, serotonin 5HT(1a) or 5HT(2a) receptors, especially in orbitofrontal circuit, antimanic agents may exhibit a more specific activity than other drugs. This specificity could concern a whole spectrum of bipolarity which might be characterized by impulsivity.