Spikes recorded from single cells do not occur regularly. Interspike intervals show great variation. Since all spikes have the same amplitude, the intervals between them must carry any information they code for. Recently developed measures of spike activity have been shown to change in response to physiological stimuli. Cells in the suprachiasmatic nucleus show different activity at different times of day (Bhumbra et al. 2005a) and ADH cells alter their activity after osmotic stimulation (Bhumbra et al. 2005b). We investigated whether different interval patterns that potentially carried coded information altered the threshold for antidromic activation of the cells we recorded. Using the standard ventral approach to the hypothalamus and conventional recording techniques, we recorded extracellularly from neurohypophysial cells in urethane-anaesthetised (1.2 g/kg IP) rats and applied stimulus pulses to their axons. We found that trains of as few as 3 pulses with a 20 ms interval profoundly affected the threshold for stimulation. Short bursts of pulses lowered threshold, probably by depolarising the axon at the stimulus site. The decreased threshold was seen if stimulus interval was increased to 40 or 80 ms but was not seen if the interval between the stimuli was increased to 100 ms. The effect was seen in all 19 cells tested. In a representative experiment with trains of 5 stimuli, the 5th pulse evoked an antidromic spike 10/10 times with a 20 ms interval, 5/10 times with an 80 ms interval but never (0/10 times) with a 100ms interval. This suggests that, whatever the mechanism, the effect lasted only for 100 ms. By contrast, longer trains of pulses (10–20 pulses with a 20 ms interval delivered each second) raised threshold (in 15 of 20 cells tested). Under the conditions of the experiments, the effect lasted up to 30 s. The raised threshold occurred even if the interval between stimulus pulses was increased to 80 ms. Intracellular recordings (from the cell bodies of neurohypophysial cells) showed that brief bursts of (~10) spikes induced by intracellular injection of depolarising current lowered resting membrane potential (by 2-3 mV, probably by activating calcium gated potassium channels). It thus appears that the intervals between spikes profoundly affected the excitability of the axons of the recorded cells. The variability and patterning of spike interval distribution may thus carry information particularly if the changes in membrane potential alter the propagation of spikes across branch points in the region of axon terminals.