Single cell spike rate and secretion response to increasing synaptic stimulation.

The spiking model (parameters in Table 1) is coupled to the fitted secretion model (Table 2). Non-phasic spiking is generated by setting g_L = 0. (A) Examples of the four different modes of spike patterning generated by the same phasic spiking model with varied input rates. (B) The non-phasic model (left) shows a non-linear increase in spike rate with increased input. The phasic spiking model (right), after very little response at low input rates, shows a more linear increase in spike rates. The rate of increase is initially steep as the patterning transitions from irregular firing to full phasic spiking, but this is followed by a wide range of very linear increase as bursts lengthen and intraburst firing rate increases. (C) The non-phasic model (left) shows a similarly non-linear secretion response to increasing spike rate, showing a slow increase in secretion at low frequencies, followed by a facilitation driven rapid increase, which then slows along with the reduced spike rate response. The phasic secretion profile (right) shows an initial steep increase which slows as the intra-burst spike rate reaches the optimal response frequency, and longer bursts allow less recovery from fatigue. (D) Examining secretion against spike rate the non-phasic model shows a progressive increase in secretion with increasing frequency. By contrast the phasic model shows a very linear response up to ∼4 Hz, after which secretion becomes relatively independent of mean spike rate. (E) Secretion per spike shows that secretion from phasic cells is achieved with much greater efficiency at synaptic input rates exceeding 200 Hz with optimal efficiency at about 400–600 Hz, corresponding to mean firing rates of 2–4 spikes/s (comprising relatively short sparse bursts). The non-phasic cells show a less variable, but also less efficient response.