The GABA(A) receptor system is implicated in a number of central nervous system (CNS) disorders, making GABA(A) receptor ligands interesting as potential therapeutic agents. Only a few different classes of structures are currently known as ligands for the GABA recognition site on the hetero-pentameric GABA(A) receptor complex, reflecting the very strict structural requirements for GABA(A) receptor recognition and activation. A large number of the compounds showing agonist activity at the GABA(A) receptor site are structurally derived from the GABA(A) agonists muscimol, THIP (Gaboxadol), or isoguvacine, which we developed at the initial stage of the project. Using recombinant GABA(A) receptors, functional selectivity has been shown for a number of compounds, including THIP, showing subunit-dependent potency and maximal response. The pharmacological and clinical activities of THIP probably reflect its potent effects at extrasynaptic GABA(A) receptors insensitive to benzodiazepines and containing alpha(4)beta(3)delta subunits. The results of ongoing clinical studies on the effect of the partial GABA(A) agonist THIP on human sleep pattern show that the functional consequences of a directly acting agonist are distinctly different from those seen after administration of GABA(A) receptor modulators, such as benzodiazepines. In the light of the interest in partial GABA(A) receptor agonists as potential therapeutics, structure-activity studies of a number of analogues of 4-PIOL, a low-efficacy partial GABA(A) agonist derived from THIP, have been performed. In this connection, a series of GABA(A) ligands has been developed showing pharmacological profiles ranging from low-efficacy partial GABA(A) agonist activity to selective antagonist effect.