1. Whole cell recordings were performed on the somata of CA1 pyramidal neurons in the rat hippocampal slice preparation Remote synaptic events were evoked by electrical stimulation of Schaffer collateral/commissural fibers in outer stratum radiatum. To isolate non-N-methyl-D-aspartate (NMDA)-mediated excitatory postsynaptic potentials (EPSPs), bath solutions contained the NMDA receptor antagonist, D-2-amino-5-phosphonovaleric acid (D-APV; 30 microM), the gamma-aminobutyric acid-A (GABAA) receptor antagonist, bicuculline (10 microM), and the GABAB receptor antagonists, CGP 35348 (30 microM) or, in some experiments, saclofen (100 microM). 2. Local application of tetrodotoxin (TTX; 0.5-10 microM) into the proximal region of the apical dendrite reduced the peak amplitude of somatically recorded EPSPs by 28% on average. In contrast to dendritic TTX application, injection of TTX into the axosomatic region of the recorded neuron reduced EPSP amplitude by only 12% on average. 3. Spill-over of dendritically applied TTX into stratum pyramidale or into outer stratum radiatum was ruled out experimentally: somatic action potentials and field EPSPs recorded near the stimulation site in outer stratum radiatum remained unaffected by local TTX application. 4. Variations of somatic membrane potential revealed a strong voltage dependence of EPSP reduction after dendritic TTX application with the effect increasing substantially with membrane depolarization. Together with the field recordings from stratum radiatum, this finding argues strongly against a predominantly presynaptic site of TTX action. 5. We therefore ascribe the EPSP decrease after local TTX application to the proximal dendrite to suppression of dendritic Na+ channels, which we assume to give rise to a noninactivating (persistent) Na+ current (INaP) in the subthreshold voltage range. Our data suggest that presumed dendritic INaP produces considerable elevation of remote excitatory signals, thereby compensating for much of their electrotonic attenuation. 6. The experimental findings were related to computer simulations performed on a reduced compartmental model of the CA1 neuron. Because the experimental evidence available so far yields only indirect clues on the strength and distribution of INaP, we allowed considerable variations in these parameters. We also varied both size and location of synaptic input. 7. The major conclusions drawn from these simulations are the following: somatic INaP alone produces little EPSP enhancement; INaP density at the axon hillock/initial segment has to be at least twice the density at the soma to produce substantial EPSP amplification; depending on the density and distribution of dendritic INaP, < or = 80% of a remote synaptic potential arrives at the soma (compared with only 52% in a passive dendrite); synaptic potentials receive progressively more elevation by dendritic INaP the stronger they are; even if restricted to the proximal segment of the apical dendrite, INaP also affects dendritic processing at more distal segments; and spatial distribution rather than local density appears to be the most important parameter determining the role of dendritic INaP in synaptic integration.