Fig. 9.

A, Mean onset times of maximal firing in type 1 (T-1), type 2 (T-2), type II (T-II), and type III (T-III) neurons or offset of firing in pre-I neurons. Onset times in control cycles (white columns) were not significantly different from onset times with phasic inflation (gray columns). Statistically significant differences in onset times between cell types were found, however, and are indicated by arrows between pairs of cell types. These differences support inferences about connectivity.B, Diagram of inferred connectivity between respiratory neurons. SAR afferent input (SAR) causes brisk firing of biphasic neurons (biphasic) after lung inflation. Postinspiratory inhibition seen only in type 1 neurons (Type 1) is consistent with biphasic neuron inhibition of type 1 neurons. Because type 1 inspiratory onset coincides closely with cervical ventral root motor output and fires earliest, type 1 neurons are hypothesized to drive inspiratory premotoneurons, such as type III (Type III) and motoneurons (MN). Delay in onset, together with low cycle-to-cycle variability in burst onset times in type II neurons (Type II), suggests an efference copy pathway from motoneurons to type II neurons. Because onset of type II neuron firing precedes but is not significantly different from onset of pre-I inhibition, type II neurons are hypothesized to inhibit pre-I neurons. Peri-inspiratory inhibition in type III neurons (Type III) is consistent with inhibition from pre-I neurons; the absence of this inhibition in type II neurons suggests that reciprocal inhibition from pre-I neurons to type II is not present. Because type III firing onset is significantly later than pre-I inhibition, reciprocal inhibition from type III neurons to pre-I neurons is not inferred. Neuron classes proposed to be causal to respiratory rhythm generation include pre-I neurons (Onimaru et al.; 1997), type 1 neurons (Rekling et al., 1996; Gray et al., 1999), and type 2 neurons (Thoby-Brisson et al.; 2000).