PMC

A, the 10-90 % rise time and the width at half-amplitude of the EPSPs from the twelve connections decreased with temperature. B, the membrane time constants and neuronal input resistances obtained from a further sample of six pyramidal neurones decreased with temperature. C and D, EPSPs showed more effective temporal summation at low temperatures. Three spontaneous EPSPs arriving within an 80 ms period showed effective summation at 13 °C (C), while EPSPs at 36 °C, although generally having larger peak amplitudes and occurring at a higher frequency, showed less summation (D).

Values were calculated for epochs of 50 or, more commonly, 100 trials at each temperature. A, graph showing the reduction in the incidence of failures with increasing temperature for each of the twelve connections. B, as A, but showing group mean values. C, reduction in EPSP CV with temperature. D, the group mean skew of EPSP amplitude distributions decreases with temperature. E, histograms of EPSP amplitudes (100 trials in each case) for one connection recorded at different temperatures. At 13 °C (▪), the histogram shows pronounced positive skew (skew = 2.24), but at 36 °C (□) it is roughly symmetrical (skew = -0.22), consistent with a release probability of approximately 0.5 at the higher temperature. F, relationship between CV⁻² and mean amplitude for three example EPSPs. As the mean amplitude increased with temperature, CV⁻² also increased, and in these, and in all nine other cases, the trajectory was steeper than the line of identity (dotted), consistent with an increase in m.

A, the average frequency of spontaneous mEPSPs from ten individual pyramidal neurones recorded in TTX increased with temperature; 125 mEPSPs were analysed for each cell at each temperature. B, the mean peak amplitudes of the mEPSPs did not change with temperature. C, properties of the four connections apparently mediated by just a single release site. With increasing temperature, the mean proportion of successes (•) increased dramatically, but the mean potency of the connections (the mean amplitude of the successes; ▴) did not increase. The mean success rate at 36 °C was 0.64, implying a mean release probability of 0.64 for these connections at this temperature. D, examples of five consecutive trials from a strong, putative multi-site connection. At 23 °C this connection showed failures, latency jitter and substantial amplitude fluctuations (upper traces). At 36 °C, however (lower traces), it did not fail and showed little more amplitude variability than expected from the background noise. E, graph showing the fall in CV with temperature for the connection shown in D, calculated for epochs of 50 trials. At 36 °C, the CV was only 0.06, consistent with a release probability close to 1 at this temperature.

A, paired recording showing presynaptic action potentials (upper trace) and averaged postsynaptic responses (lower trace; averages of 50 trials); those shown by the thick lines were recorded at 23 °C, thin lines at 36 °C and dotted lines at 13 °C. B, examples of superimposed consecutive single trials from EPSPs of different mean amplitudes, recorded at 23 °C: upper traces, strong connection showing no failures; middle traces, connection of moderate strength, showing one failure; lower traces, weak connection. Note that the failure can be distinguished from the successes, even though the peak amplitude of the smaller success is only approximately 170 μV. C, graph showing the proportion of transmission successes as a function of the EPSP mean peak amplitude, recorded at 23 °C and showing that large EPSPs failed less often than small ones. Each point represents an individual connection and values were calculated for epochs of 100 trials. D, graph showing how the mean EPSP amplitude for the twelve connections tested increased with temperature. Because the EPSPs differed greatly in size, the mean amplitudes for each were normalized to their mean amplitude at 23 °C before pooling. Means were calculated for epochs of 50 or 100 trials. The continous line was fitted by linear regression (r= 0.996).