Models of respiratory rhythm generation in the pre-Botzinger complex. III.Experimental tests of model predictions

We used the testable predictions of mathematical models proposed by Butera et al. to evaluate cellular, synaptic, and population-level components of t he hypothesis that respiratory rhythm in mammals is generated in vitro in t he pre-Botzinger complex (pre-BotC) by a heterogeneous population of pacema ker neurons coupled by fast excitatory synapses. We prepared thin brain ste m slices from neonatal rats that capture the pre-BotC and maintain inspirat ory-related motor activity in vitro. We recorded pacemaker neurons extracel lularly and found: intrinsic bursting behavior that did not depend on Ca2currents and persisted after blocking synaptic transmission; multistate beh avior with transitions from quiescence to bursting and tonic spiking states as cellular excitability was increased via extracellular K+ concentration ([K+](o)); a monotonic increase in burst frequency and decrease in burst du ration with increasing [K+](o); heterogeneity among different cells sampled ; and an increase in inspiratory burst duration and decrease in burst frequ ency by excitatory synaptic coupling in the respiratory network. These data affirm the basis for the network model, which is composed of heterogeneous pacemaker cells having a voltage-dependent burst-generating mechanism domi nated by persistent Na+ current (I-NaP) and excitatory synaptic coupling th at synchronizes cell activity. We investigated population-level activity in the pre-BotC using local "macropatch" recordings and confirmed these model predictions: pre-BotC activity preceded respiratory-related motor output b y 100-400 ms, consistent with a heterogeneous pacemaker-cell population gen erating inspiratory rhythm in the pre-BotC; pre-BotC population burst ampli tude decreased monotonically with increasing [K+](o) (while frequency incre ased), which can be attributed to pacemaker cell properties; and burst ampl itude fluctuated from cycle to cycle after decreasing bilateral synaptic co upling surgically as predicted from stability analyses of the model. We con clude that the pacemaker cell and network models explain features of inspir atory rhythm generation in vitro.