Simulations of a ventrolateral medullary neural network for respiratory rhythmogenesis inferred from spike train cross-correlation

Abstract

Connections among ventrolateral medullary respiratory neurons inferred from spike train analysis were incorporated into a model and simulated with the program SYSTM11 (MacGregor 1987). Inspiratory (I) and expiratory (E) neurons with augmenting (AUG) and decrementing (DEC) discharge patterns and rostral I-E/I neurons exhibited varying degrees of adaptation, but no endogenous bursting properties. Simulation parameters were adjusted so that respiratory phase durations, neuronal discharge patterns, and short-time scale correlations were similar to corresponding measurements from anesthetized, vagotomized, adult cats. Rhythmogenesis persisted when the strength of each set of connections was increased 100% over a smaller effective value. Changes in phase durations and discharge patterns caused by manipulation of connection strengths or population activity led to several predictions. (a) Excitation of the I-E/I population prolongs the inspiratory phase. (b) Rhythmic activity can be reestablished in the absence of I-E/I activity by unpatterned excitation of I-DEC and I-AUG neurons. (c) An increase in I-DEC neuron activity can cause an apneustic respiratory pattern. (d) A decrease in I-DEC neuron activity increases the slope of the inspiratory ramp and shortens inspiration. (e) Excitation of the E-DEC population prolongs the expiratory phase or produces apnea; inhibition of E-DEC neurons reduces expiratory time. (f) Excitation of E-AUG cells causes I-AUG neurons to exhibit a step rather than a ramp increase in firing rate at the onset of their active phase. The results suggest mechanisms by which the duration of each phase of breathing and neuronal discharge patterns may be regulated.