Biophysical cortical column model for optical signal analysis

Voltage-sensitive dye imaging (VSDI) is a powerful modern neuroimaging technique whose application is expanding worldwide because it offers the possibility to monitor the neuronal activation of a large population with high spatial and temporal resolution. In this thesis, we investigate the biological sources of the voltage-sensitive dye signal (VSD signal), since this question remains unresolved in the literature. What does the voltage-sensitive dye imaging signal measures? This question is difficult to resolve at the physiological level as the signal is multi-component: The dye reflects the dynamics of the membrane potential of all membranes in the neuronal tissue, including all layers of the circuitry, all cell types (excitatory, inhibitory, glial) and all neuronal compartments (somas, axons, dendrites). To answer this question, we propose to use a biophysical cortical column model, at a mesoscopic scale, taking into account biological and electrical neural parameters of the laminar cortical structure. The model is based on a cortical microcircuit, whose synaptic connections are made between six specific populations of neurons, excitatory and inhibitory neurons in three main layers. Each neuron is represented by a reduced compartmental description with conductance-based Hodgkin-Huxley neuron model. The model is fed by a thalamic input with increasing activity, background activity and lateral connections. Isolated neurons and network behavior have been adjusted to fit data published in the literature. The so-calibrated model offers the possibility to compute the VSD signal with a linear formula. We validated the model by comparing the simulated and the measured VSD signal. Thanks to the compartmental construction of this model, we confirm and quantify the fact that the VSD signal is the result of an average from multiple components, with excitatory dendritic activity of superficial layers as the main contribution. It also suggests that inhibitory cells, spiking activity and deep layers are contributing differentially to the signal dependently on time and response strength. We conclude that the VSD signal has a dynamic multi-component origin and propose a new framework for interpreting VSD data.