The importance of the cholinergic innervation of the neocortex in cognitive functions has been shown in a number of clinical and animal studies. Until recently, attempts to study the mode of action of acetylcholine in the neocortex have concentrated on muscarinic effects, whereas cholinergic actions mediated by nicotinic receptors have been difficult to demonstrate. The present work was undertaken to study the mechanism of action of nicotinic agents on cortical neurons and compare it to muscarinic effects by means of intracellular recordings in a slice preparation. The study was performed in the prelimbic area of the rat prefrontal cortex, a cortical region particularly involved in cognitive processes. Recordings were made from pyramidal cells located in layers II/III and synaptic potentials were evoked by stimulation of superficial cortical layers. Iontophoretic applications of nicotinic agonists (nicotine, dimethylphenylpiperazinium, cytisine) increased the amplitude of the monosynaptic excitatory postsynaptic potential mediated by non-N-methyl-D-aspartate glutamate receptors in 14% (22/159) of cells. This effect was abolished by the selective nicotinic blocker, neuronal bungarotoxin (IC50 = 0.6-0.7 microM) and by dihydro-beta-erythroidine (IC50 = 20-30 microM), whereas hexamethonium, mecamylamine, curare and alpha-bungarotoxin were ineffective. The nicotinic agonists did not change resting membrane potential, input resistance or current-voltage relationship. They also did not affect the depolarizations produced by glutamate applied by iontophoresis in the somatic or dendritic area. In contrast, the muscarinic agonists (muscarine, acetyl-beta-methylcholine) decreased the amplitude of the excitatory postsynaptic potential in 100% of the neurons tested. Atropine was more effective (IC50 = 0.08 microM) than pirenzepine (IC50 = 2 microM) to antagonize the muscarinic action. These effects were observed in the absence of any direct postsynaptic change in membrane potential or input resistance, provided that the site of the iontophoretic application was more than 100 microM distant from the soma. The muscarinic agonists did not influence the actions of iontophoretically applied glutamate. These results suggest that nicotinic and muscarinic agonists modulate excitatory synaptic transmission mediated at dendritic sites by non-N-methyl-D-aspartate glutamate receptors, possibly through a presynaptic action. Thus ascending cholinergic systems may take part in information processing in the prefrontal cortex through the control of ongoing excitation to pyramidal cells.