Characterising information correlation in a stochastic Izhikevich neuron

Citation

Yang Z, Gandhi VS, Karamanoglu M & Graham B (2015) Characterising information correlation in a stochastic Izhikevich neuron. In: Proceedings of the International Joint Conference on Neural Networks 2015. 2015. 2015 International Joint Conference on Neural Networks (IJCNN), Killarney, Ireland, 12.07.2015-17.07.2015. New York: IEEE. https://doi.org/10.1109/IJCNN.2015.7280534

Abstract
The Izhikevich spiking neuron model is a relatively new mathematical framework which is able to represent many observed spiking neuron behaviors, excitatory or inhibitory, by simply adjusting a set of four model parameters. This model is deterministic in nature and has achieved wide applications in analytical and numerical analysis of biological neurons due largely to its biological plausibility and computational efficiency. In this work we present a stochastic version of the Izhikevich neuron, and measure its performance in transmitting information in a range of biological frequencies. The work reveals that the deterministic Izhikevich model has a wide information transmission range and is generally better in transmitting information than its stochastic counterpart.

Keywords
Izhikevich neuron; Information content; Mutual information; Probability; Correlation

People

Professor Bruce Graham

Emeritus Professor, Computing Science