AbstractAs the new-born brain grows and matures, the neurons within migrate, form synapses, and assemble into functional networks, all whilst changing in morphology and adapting to function. The first two weeks of life are a particularly dynamic period, with the adaptations in electrical properties and activity underpinned by the composition and biophysical properties of the neuronal membrane. The action potential (AP) is the foundation of neuronal communication, being the primary indicator of neuronal activity, the progenitor of neurotransmitter release and synaptogenesis, implicated in the formation of neuronal networks, and controlling gene expression. Changes to the AP waveform over maturation are therefore simultaneously indicative of the underlying biophysics and influential to neuronal structure and function. Even subtle changes to the AP waveform can contain a wealth of information, but thorough biophysical characterisation of the AP waveform with neuronal development has remained elusive. Such characterisation requires a combination of established experimental techniques, as well as carefully optimised models to elucidate the underlying membrane properties, ion channel populations and neuronal conductances that give the observed dynamics their particular development-dependent shape.
By fitting computational conductance-based Hodgkin-Huxley models to electrophysiological data, the underlying biophysics can be illuminated. But, with multiple and developmentally-dynamic parameters to consider, accurate and fast optimisation techniques are needed. Voltage-based fitting methods can produce complex parameter error landscapes with local or narrow minima, requiring computationally expensive algorithms to return the maximal conductances associated with various channels. By instead algebraically solving the expressions for the ion channel gating variables and computing the difference between passive and active neuronal currents, parameter optimisation can be reduced to a simple linear sum of currents. The minimisation of the residual of this sum can be presented graphically for the different maximal conductances of the Hodgkin-Huxley model in multidimensional yet simple parameter landscapes that allow for intuitive interpretation and fast model optimisation.
Via the combination of current-clamp whole-cell electrophysiology, dye-filling and confocal imaging, the biophysical characteristics of regular-spiking excitatory neurons in the somatosensory cortex of neonatal mice aged between 3 and 11 postnatal days were investigated. The experimental protocols used facilitated consideration of the development of ion channel populations, membrane thickness, cell morphology and gap junctions when painting a picture of neuronal biophysics.
This work demonstrates that postnatal neuronal development is correlated with large increases in the height and speed of APs. Passive membrane dynamics are observed to mature, with analysis via a two-compartment model of exponential decay revealing a developmentally dependent fast passive current sink in some cells. Further to this, these passive dynamics are manipulated to reveal surface area predictions that correlate with morphological observations. A surprising diversity of neuronal morphologies is found within the layer IV barrel cortex, along with evidence of gap junction coupling between neurons, with potential implications for development-dependent regulation of neural networks. Mathematical models, built on artificial data sets and optimised for multiple parameters via residual current minimisation, demonstrate better robustness to noise than models optimised via voltage-comparison methods. Resultant multiple-channel neuron models can be used to provide a probe of the biophysical heterogeneity of maturing neuronal populations.
The combination of electrophysiological, computational and imaging techniques allows us to make biophysically complex predictions of neuronal development, to produce a picture of the changing biophysical nature of these excitatory neurons as they approach maturity, and to elucidate the dynamics that drive maturation of neuronal excitability.
|Date of Award||23 Jan 2019|
|Supervisor||Michael C Ashby (Supervisor) & Nathan F Lepora (Supervisor)|