REGULARITY ANALYSIS IN A COMPARTMENTAL MODEL OF CHOPPER UNITS IN THE ANTEROVENTRAL COCHLEAR NUCLEUS

We investigate the discharge patterns of chopper units in the anteroventral cochlear nuclear (AVCN) by developing an equivalent cylinder compartmental model of AVCN stellate cells, which are the sources of the chopper response pattern. The model consists of a passive dendritic tree connected to somatic and axonal compartments with voltage-sensitive channels. Synaptic inputs to the model are simulated auditory nerve fiber responses to best-frequency tones. We adjust the anatomic and electrical parameters of the model to agree with available intracellular data from stellate cells in the AVCN of the mouse and the cat and compare the response of the model to injected current with responses recorded in vitro. The model shows approximately linear current-voltage characteristics for small hyperpolarizing currents. The model's input resistance and the time course of its response to hyperpolarizing current applied at the soma are comparable with those measured from stellate cells in vitro. In response to sustained depolarizing current, the model fires repetitively with nearly perfect regularity, a property also observed in vitro. Auditory nerve inputs to the cell are modeled as deadtime-modified Poisson processes with a multiexponential adaptation in the Poisson rate. We are able to adjust the number, rate, and location of excitatory and inhibitory inputs to the model and succeed in simulating chopper response patterns seen in vivo. Chopper units exhibit a variety of regularity and adaptation patterns in response to tone stimuli. Physiological data from brain slice experiments and experiments in vivo imply that this heterogeneity is primarily due to differences in input configurations. By systematically varying the number and position of excitatory and inhibitory inputs, we can simulate a range of chopper response patterns. We quantify the regularity of the model's response using the coefficient of variation (CV) of the interspike interval. We find that the CV decreases, i.e., the regularity increases, as the number of converging inputs or their distance from the soma increases. The regularity of the output is more sensitive to the number of converging inputs than to their location on the dendritic tree. The statistics of the first spike latency (FSL) are also sensitive to the configuration of excitatory inputs. The mean and minimum FSL are more sensitive to the electrotonic distance of the inputs from the soma than to the number of inputs, whereas the standard deviation of the FSL is highly dependent on the number of converging inputs and is nearly independent of their location. We show that the model fires less regularly when inhibitory inputs are activated and that certain inhibitory input configurations can result in rapid, transient adaptation. The simulations with inhibitory inputs support the hypothesis that a major difference between transient and sustained choppers is the presence of inhibitory inputs.