UC San Diego

Vocalizations are an essential medium for sexual signaling in zebra finches. Male zebra finches produce highly stereotyped complex sequences of vocal gestures during courtship. The acquisition and production of their songs occur through the song system which consists of a series of forebrain nuclei. The telencephalic sensorimotor nucleus known as HVC plays a central role in this complex system by controlling the temporal structure of birdsong and the order of syllables. We combine physical and mathematical approaches to explore how HVC neurons form activities that contribute to the downstream song production pathway, and leverage electrophysiology and imaging data from this nucleus to develop single neuron and neuron network models and verify their performance.

In Chapter 2, we explore a data assimilation method to transfer electrophysiology observations to HVC neuron models. We employ conductance-based models of single HVC neurons describing experimentally-verified ion currents, and estimate the time-invariant parameter values and dynamics of model variables with data assimilation. The effectiveness of the method is tested by accuracy of estimations of unknown variables and prediction of future neuron voltage under a different stimulus during numerical simulations.

Synaptical interactions among various types of HVC neurons are essential for this nucleus to function normally, so Chapter 3 proposes a neural network model containing two major types of neurons in the HVC. The model is based on recordings of individual neurons’ membrane potentials and experimentally observed synaptical currents. Model predictions are supported by multiple observations of behavior of the two types of neurons, especially their firing patterns during birdsong.

This dissertation explores dynamics and functions of neurons in HVC via modeling andinference approaches rooted in physics principles.