Leader neurons in leaky integrate and fire neural network simulations

TL;DR

This study identifies properties of leader neurons in simulated 2D neural networks using leaky integrate-and-fire models, revealing how certain neurons initiate bursts and can be predicted based on their connectivity and response patterns.

Contribution

It characterizes leader neuron properties in simulations and develops a predictive formula based on linear analysis of network parameters.

Findings

Leader neurons are excitatory with low firing thresholds.

They send signals to many excitatory and few inhibitory neurons.

A predictive formula can identify potential leader neurons.

Abstract

Several experimental studies show the existence of leader neurons in population bursts of 2D living neural networks. A leader neuron is, basically, a neuron which fires at the beginning of a burst (respectively network spike) more often that we expect by looking at its whole mean neural activity. This means that leader neurons have some burst triggering power beyond a simple statistical effect. In this study, we characterize these leader neuron properties. This naturally leads us to simulate neural 2D networks. To build our simulations, we choose the leaky integrate and fire (lIF) neuron model. Our lIF model has got stable leader neurons in the burst population that we simulate. These leader neurons are excitatory neurons and have a low membrane potential firing threshold. Except for these two first properties, the conditions required for a neuron to be a leader neuron are difficult to identify and seem to depend on several parameters involved in the simulations themself. However, a detailed linear analysis shows a trend of the properties required for a neuron to be a leader neuron. Our main finding is: A leader neuron sends a signal to many excitatory neurons as well as to a few inhibitory neurons and a leader neuron receives only a few signals from other excitatory neurons. Our linear analysis exhibits five essential properties for leader neurons with relative importance. This means that considering a given neural network with a fixed mean number of connections per neuron, our analysis gives us a way of predicting which neuron can be a good leader neuron and which cannot. Our prediction formula gives us a good statistical prediction even if, considering a single given neuron, the success rate does not reach hundred percent.