An endogenous kappa-opioid agonist reduces the duration of phasic bursts in vasopressin cells. Non-synaptic post-spike depolarizing after-potentials underlie activity during bursts by increasing post-spike excitability and kappa-receptor activation reduces depolarizing after-potential amplitude in vitro. To investigate the effects of kappa-opioids on post-spike excitability in vivo, we analysed extracellular recordings of the spontaneous activity of identified supraoptic nucleus vasopressin cells in urethane-anaesthetized rats infused with Ringer's solution (n = 17) or the kappa-agonist, U50,488H (2.5 microg/h at 0.5 microl/h; n = 23), into the supraoptic nucleus over 5 days. We plotted the mean hazard function for the interspike interval distributions as a measure of the post-spike excitability of these cells. Following each spike, the probability of another spike firing in vasopressin cells recorded from U50,488H infused nuclei was markedly reduced compared to Ringer's treated vasopressin cells. To determine whether U50,488H could reduce post-spike excitability in cells that displayed spontaneous phasic activity, we infused U50,488H (50 microg/h at 1 microl/h, i.c.v.), for 1-12 h while recording vasopressin cell activity. Nine of 10 vasopressin cells were silenced by i.c.v. U50,488H 15 +/- 5 min into the infusion. Six cells exhibited spontaneous phasic activity before U50,488H infusion and recordings from three of these phasic cells were maintained until activity recovered; during U50,488H infusion, the activity of these three cells was irregular. Generation of the mean hazard function before and during U50,488H infusion revealed a reduction in post-spike excitability during U50,488H infusion. Thus, kappa-receptor activation reduces post-spike excitability in vivo; this may reflect inhibition of depolarizing after-potentials and may thus underlie the reduction in burst duration of vasopressin cells caused by an endogenous kappa-agonist in vivo.