System Identification of a Multi-timescale Adaptive Threshold Neuronal Model

TL;DR

This paper presents a novel linear parameter estimation method for a multi-timescale adaptive threshold neuronal model, enabling accurate prediction of neuron firing times using both synthetic and experimental data.

Contribution

It introduces a linearly parametrized model and a prediction error framework for efficient parameter identification in neuronal models, outperforming existing methods.

Findings

Accurately fits the MAT model to synthetic and experimental data.

Superior performance compared to existing neuronal identification methods.

Effective in predicting precise neuron firing times.

Abstract

In this paper, the parameter estimation problem for a multi-timescale adaptive threshold (MAT) neuronal model is investigated. By manipulating the system dynamics, which comprise of a non-resetting leaky integrator coupled with an adaptive threshold, the threshold voltage can be obtained as a realizable model that is linear in the unknown parameters. This linearly parametrized realizable model is then utilized inside a prediction error based framework to identify the threshold parameters with the purpose of predicting single neuron precise firing times. The iterative linear least squares estimation scheme is evaluated using both synthetic data obtained from an exact model as well as experimental data obtained from in vitro rat somatosensory cortical neurons. Results show the ability of this approach to fit the MAT model to different types of fluctuating reference data. The performance of the proposed approach is seen to be superior when comparing with existing identification approaches used by the neuronal community.