Frequency dependence of EPSC following repetitive stimulation of the DVC neurons. A, EPSC traces in response to repetitive stimulation at 0.1 Hz (top of each panel) and 20 Hz (bottom). Top to bottom: representative responses of type I and type II DMX neurons and a dm-cNTS (NTS) neuron. Traces in-between the responses were removed so that the changes in the EPSCs at different frequencies could be directly compared on the same time-scale. B, normalized EPSC amplitude at each stimulation (stimulus number 1-15) at 0.1 Hz and 20 Hz. Each marker and vertical bar represent the mean and SE of the values in 39 (type I DMX), 42 (type II DMX) and 18 (dm-cNTS) neurons. The length of the vertical SE bars was often smaller than the marker size, which made them masked behind the markers; this suggests fairly similar trajectories among neurons. Note the marked frequency-dependent depression at 20 Hz, especially in type II DMX and dm-cNTS neurons. Two vertical broken lines indicate the range of the 5th to 9th responses, in which typical depression patterns of the EPSC amplitude at 20-Hz stimulation were observed and used for evaluating the fdi. C, the depression of EPSC amplitude during the course of stimulation depended on the stimulation frequency. Abscissa, stimulus frequency; ordinate, the degree of EPSC amplitude depression as evaluated by the mean amplitude of EPSC5 to EPSC9 (5 EPSC responses), which was normalized to that of EPSC1. **, p< 0.01; *, p < 0.05; Kruskal-Wallis test to compare type I, type II DMX and dm-cNTS neurons; +, p < 0.05; ++, p < 0.01; Mann-Whitney test to compare the (EPSC5-9)/EPSC1 values at different stimulation frequencies with that at 0.1-Hz stimulation in each neuron group. D, distribution of fdi values estimated from recordings in type I DMX, type II DMX and dm-cNTS neurons. E, relationship between the PPR (abscissa) and fdi. Summary of data obtained for all cell types. Note the strong positive correlation between fdi and PPR (Spearman's ρ = 0.90, p < 0.01, n = 99).