Neurotransmitters relaying ascending visceral information were examined by comparing the response of neurons in the insular cortex to vagal stimulation (0.8 Hz, 2 mA) before and after neurotransmitter antagonist injections (200 nl) in the ventroposterior parvocellular nucleus of the thalamus (VPpc). Cobalt (10 mM; presynaptic blocker) and kynurenate (100 microM; nonspecific excitatory amino acid antagonist) injections in the VPpc resulted in an attenuation (73-100 and 38-98%, respectively) of the evoked cortical response. Injections of the specific N-methyl-D-aspartate (NMDA) antagonist DL-2-amino-5-phosphonopentanoic acid (200 microM and 2 mM) did not affect the vagally evoked response, whereas the nonspecific non-NMDA antagonist L-glutamic acid diethylester (200 microM) attenuated the vagally evoked response by 66-100%. Three concentrations of the DL-alpha-amino-3-hydroxy-5-methylisoxazole-propionic acid (AMPA)-specific antagonist 6-cyano-7-nitroquinoxaline-2,3-dione (20 and 200 microM and 2 mM) attenuated the vagally evoked cortical response by 29 +/- 9, 31 +/- 10, and 59 +/- 8%, respectively. The more selective AMPA antagonist 6-nitro-7-sulphamoylbenzo(f)quinoxaline-2,3-dione (200 microM and 2 mM) inhibited the vagally evoked cortical response by 53 +/- 8 and 52 +/- 3%, respectively. Phentolamine (0.1 and 1.0 microM), a general alpha-adrenergic antagonist, and picrotoxin (0.1 and 1.0 microM), a GABA(A) antagonist, did not affect the vagally evoked response. Atropine, a muscarinic cholinergic antagonist, decreased the vagally evoked response by 40 +/- 2% at a concentration of 0.1 microM, but a higher concentration of 1.0 microM had no effect. These results indicate that the non-NMDA excitatory amino acid receptor is necessary for the relay of visceral information in the VPpc. Muscarinic receptors may modulate visceral neuronal excitability in the VPpc, although the exact interaction between the inhibitory (m2) and excitatory (m3 or m5) muscarinic receptor types found in the thalamus is not known.