The Neuroid revisited: A heuristic approach to model neural spike trains

Abstract

Introduction: Since it was introduced in 2012, the Neuroid has been used to aid in understanding how functionally different neural populations contribute to sensory information processing. However, insights about whether this neuron-model could perform better than others or about when its utilization should be considered have not been provided yet. Methods: In an attempt to address this issue, a comparison between the Neuroid and the leaky-integrate-and-fire (LIF) model in terms of accuracy and computational cost was performed. Both models were tested for different stimulation amplitudes and stimulation periods, with time step sizes ranging from 10 -4 to 1 ms. Results: It was found that, although the Neuroid was able to produce more accurate results than its original version, its accuracy was lower than the achieved with the LIF model solved by the forward Euler method. On the other hand, the Neuroid performed its calculations in an amount of time significantly lower (Mulfactorial ANOVA test, p < 0.05) than that required by the LIF model when it was solved by using the forward Euler method. Moreover, it was possible to use Neuroid-based networks to replicate biologically relevant firing patterns produced by low-scale networks composed of more detailed neuron-models. Conclusion: Results suggest that the Neuroid could be an interesting choice when computational resources are limited, although its use might be restricted to a narrow band of applications.Keywords Neuroid, Spiking neuron-model, Frequency-intensity curve, Accuracy, Computational cost, Heuristic. This is an Open Access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.How to cite this article: Prada EJA, Arteaga IAB, Martínez AJD. The Neuroid revisited: a heuristic approach to model neural spike trains.