Elsevier

Neurocomputing

Volumes 58–60, June 2004, Pages 41-46
Neurocomputing

Networks of neurons that emit and recognize signatures

https://doi.org/10.1016/j.neucom.2004.01.020Get rights and content

Abstract

Recent experiments have revealed the presence of neural fingerprints in the activity of several neurons of the pyloric central pattern generator of crustacean. These signatures consist of specific spike timings in the bursting activity of the neurons. The existence of cellular mechanisms to identify the origin of individual neural signals, and the study of information processing based on this identification have been neglected in the context of theoretical approaches to the nervous system. In this paper, we present a simple model to study the ability of a neural network to process information based on the emission and recognition of neural signatures.

Introduction

Neurons of the stomatogastric central pattern generator (CPG) of the lobster [5] have specific neural signatures in the form of characteristic interspike intervals at the beginning of each burst [6]. These neuron-specific firing patterns depend on the synaptic organization of the network [1], [4], [6] and can be reconfigured by neuromodulatory action. These facts raise several intriguing questions: Does the nervous system have mechanisms to generate neuron signatures and the ability to process information using these fingerprints? Do neuron signatures enhance the capacity of a network to perform a given task?

As a first step to answer this last question we have developed a simple network of neurons that are able to emit and recognize specific neural fingerprints. We have studied the self-organizing properties within this network. Fast transitions of the collective activity emerge as a function of a stimulus introduced in a few neurons within the network. Information processing based on the identification of specific neural signatures can be a general and powerful strategy for neural systems.

Section snippets

Neuron and network models

As a first approach to study information processing with neuron signatures, we have built a network of binary neurons. Each neuron in the network responds to the recognition of two signatures (A and B). When A or B signatures are recognized, the neuron emits the same signature to its neighbors with a probability equal to 0.8. If no signature is recognized, the neuron emits another signature, C, with a probability equal to 0.05. A schematic representation of this behavior can be seen in Fig. 1

Results

Initially, the neurons are silent and no stimulus is given to the network. At time step 5000, signature A is introduced in eight neurons chosen randomly within the network. This stimulus is kept for 10000 time steps. The network evolves freely for 5000 time steps. Then signature B is introduced as the external input to the network. These sequence is repeated once more with both signatures.

Fig. 2 shows the number of neurons emitting a particular signature at each time step (panels corresponding

Discussion

Characteristic interspike interval signatures have been found in several cells of the pyloric central pattern generator of crustacean. These neuron fingerprints may play an important role for fast and fine tuning of CPG rhythms. Recent neurophysiological experiments show that modulatory inputs can modify CPG neuron signatures [6]. In model experiments we have seen that the characteristic CPG triphasic rhythm evolves to other types of rhythms when signatures are changed [2]. Thus, dynamic

Acknowledgements

This work was supported by Fundación BBVA.

References (6)

  • R. Latorre et al.

    Characterization of triphasic rhythms in central pattern generators (I): interspike interval analysis

    Lect. Notes Comput. Sci.

    (2002)
  • R. Latorre, F.B. Rodrı́guez, P. Varona, Effect of individual spiking activity on rhythm generation of Central Pattern...
  • L.G. Morris et al.

    Muscles express motor patterns of non-innervating neural networks by filtering broad-band input

    Nature Neurosci.

    (2000)
There are more references available in the full text version of this article.
View full text