Back
Phenomenological models for single-neuron energetics
{Despite representing only 2\textpercent of the body mass in humans, the brain requires up to 20\textpercent of the oxygen and 25\textpercent of the glucose supply to function, which is striking, yet remarkably little given the variety of tasks it has to tackle. Many models have been developed to account for neuronal behaviors but none of the standard implementations account for energy consumption. Despite the insights these models provided, they cannot faithfully characterize neuronal states involving high metabolic stress, for instance in epilepsy [1] or many neurodegenerative diseases [2]. This work aims at filling the gap by introducing new and augmented integrate-and-fire models which account for energy requirements associated to neuronal activity. These models reproduce qualitatively crucial behaviors such as depolarization blocks and bistability while remaining computationally efficient and analytically tractable, thus enabling the study of large neuronal ensembles. Therefore, they provide a bridge between the complexity of detailed conductance-based models [3], that can capture a large spectrum of the neuronal behaviours but become both intractable and computationally expensive, and simple models such as the leaky integrate-and-fire neuron, that are numerically efficient but cannot reproduce important dynamical features such as depolarization block. As the new models phenomenologically reproduce the influence of ATP/ADP levels and ion concentrations on the membrane potential [4], they enabled us to propose straightforward and falsifiable mechanisms for several important phenomena. The poster will detail some of these possible mechanisms, focusing on epilepsy, as well as up-and-down states [5]. Each case will be illustrated by both theoretical analysis and numerical simulations to show how the mechanisms captured by the models can provide an intuitive explanation to these behaviors.}
@misc{item_3162712, title = {{Phenomenological models for single-neuron energetics}}, booktitle = {{Bernstein Conference 2019}}, abstract = {{Despite representing only 2\textpercent of the body mass in humans, the brain requires up to 20\textpercent of the oxygen and 25\textpercent of the glucose supply to function, which is striking, yet remarkably little given the variety of tasks it has to tackle. Many models have been developed to account for neuronal behaviors but none of the standard implementations account for energy consumption. Despite the insights these models provided, they cannot faithfully characterize neuronal states involving high metabolic stress, for instance in epilepsy [1] or many neurodegenerative diseases [2]. This work aims at filling the gap by introducing new and augmented integrate-and-fire models which account for energy requirements associated to neuronal activity. These models reproduce qualitatively crucial behaviors such as depolarization blocks and bistability while remaining computationally efficient and analytically tractable, thus enabling the study of large neuronal ensembles. Therefore, they provide a bridge between the complexity of detailed conductance-based models [3], that can capture a large spectrum of the neuronal behaviours but become both intractable and computationally expensive, and simple models such as the leaky integrate-and-fire neuron, that are numerically efficient but cannot reproduce important dynamical features such as depolarization block. As the new models phenomenologically reproduce the influence of ATP/ADP levels and ion concentrations on the membrane potential [4], they enabled us to propose straightforward and falsifiable mechanisms for several important phenomena. The poster will detail some of these possible mechanisms, focusing on epilepsy, as well as up-and-down states [5]. Each case will be illustrated by both theoretical analysis and numerical simulations to show how the mechanisms captured by the models can provide an intuitive explanation to these behaviors.}}, pages = {286--287}, year = {2019}, slug = {item_3162712}, author = {Fardet, T and Levina, A} }