Point-process modeling of excitatory/inhibitory interactions in LSO neurons
Johnson, Don H.
Doctor of Philosophy thesis
The thesis of this work is that the representation of sensory information in the discharges of single neurons can be characterized by point process models. The work focuses on the encoding of binaural information in the discharges of neurons in the cat lateral superior olivary (LSO). LSO neurons are postulated to be involved in encoding interaural level difference (ILD) and interaural time-of-arrival difference (ITD)--binaural cues for localizing and detecting high frequency signals. These cues are extracted from the binaural inputs by the interaction of the excitatory input (from the ipsilateral ear) with the inhibitory input (from the contralateral ear) and encoded in the discharges of LSO neurons. I have developed a unifying point process model that describes the temporal and statistical characteristics of the discharges of all three LSO unit types. The model is based on an intrinsic recovery function that is independent of the stimulus conditions and the timing of previous discharges. The effects of the stimulus conditions and the duration of the previous interval are described by operators (e.g. scaling and shifting) that are applied to the intrinsic recovery function. The resulting model successfully predicts the complex effects of the stimulus conditions on the statistical and temporal characteristics of LSO neuron discharges. Excitatory and inhibitory effects are clearly distinguished and it is concluded that the sustained discharges of LSO neurons convey information about both the average sound level and the ILD. The ILD affects also the relative latency of the excitatory and inhibitory inputs and modulates the tightly distributed timing of the initial discharges to a monaural stimulus. Thus, the model quantifies the representation of ILD in the timing of LSO neurons; pressing the question of the utilization of information represented in the detailed timing of neural activity as opposed to the general consensus that only the discharge rate is relevant for neural information processing. Specific implications about the properties of LSO cells membranes are made based on the correlation between the operators used in the model and the effects they describe and can direct the construction of a compartmental model of LSO single-neuron.