Artificial NeuroGlial Networks

Artificial NeuroGlial Networks

Ana Belén Porto Pazos (University of A Coruña, Spain), Alberto Alvarellos González (University of A Coruña, Spain) and Félix Montañés Pazos (University of A Coruña, Spain)
Copyright: © 2009 |Pages: 5
DOI: 10.4018/978-1-59904-849-9.ch026
OnDemand PDF Download:


More than 50 years ago connectionist systems (CSs) were created with the purpose to process information in the computers like the human brain (McCulloch & Pitts, 1943). Since that time these systems have advanced considerably and nowadays they allow us to resolve complex problems in many disciplines (classification, clustering, regression, etc.). But this advance is not enough. There are still a lot of limitations when these systems are used (Dorado, 1999). Mostly the improvements were obtained following two different ways. Many researchers have preferred the construction of artificial neural networks (ANNs) based in mathematic models with diverse equations which lead its functioning (Cortes & Vapnik, 1995; Haykin, 1999). Otherwise other researchers have pretended the most possibly to make alike these systems to human brain (Rabuñal, 1999; Porto, 2004). The systems included in this article have emerged following the second way of investigation. CSs which pretend to imitate the neuroglial nets of the brain are introduced. These systems are named Artificial NeuroGlial Networks (ANGNs) (Porto, 2004). These CSs are not only made of neuron, but also from elements which imitate glial neurons named astrocytes (Araque, 1999). These systems, which have hybrid training, have demonstrated efficacy when resolving classification problems with totally connected feed-forward multilayer networks, without backpropagation and lateral connections.
Chapter Preview


The ANNs or CSs emulate the biological neural networks in that they do not require the programming of tasks but generalise and learn from experience. Current ANNs are composed by a set of very simple processing elements (PEs) that emulate the biological neurons and by a certain number of connections between them.

Until now, researchers that pretend to emulate the brain, have tried to represent in ANNs the importance the neurons have in the Nervous System (NS). However, during the last decades research has advanced remarkably in the Neuroscience field, and increasingly complex neural circuits, as well as the Glial System (GS), are being observed closely. The importance of the functions of the GS leads researchers to think that their participation in the processing of information in the NS is much more relevant than previously assumed. In that case, it may be useful to integrate into the artificial models other elements that are not neurons.

Since the late 80s, the application of innovative and carefully developed cellular and physiological techniques (such as patch-clamp, fluorescent ion-sensible images, confocal microscopy and molecular biology) to glial studies has defied the classic idea that astrocytes merely provide a structural and trophic support to neurons and suggests that these elements play more active roles in the physiology of the Central Nervous System.

New discoveries are now unveiling that the glia is intimately linked to the active control of neural activity and takes part in the regulation of synaptic neurotransmission (Perea & Araque, 2007). Abundant evidence has suggested the existence of bidirectional communication between astrocytes and neurons, and the important active role of the astrocytes in the NS’s physiology (Araque et al., 2001; Perea & Araque, 2005). This evidence has led to the proposal of a new concept in synaptic physiology, the tripartite synapse, which consists of three functional elements: the presynaptic and postsynaptic elements and the surrounding astrocytes (Araque et al., 1999). The communication between these three elements has highly complex characteristics, which seem to reflect more reliably the complexity of the information processing between the elements of the NS (Martin & Araque, 2005).

Key Terms in this Chapter

Evolutionary Computation: Solution approach guided by biological evolution, which begins with potential solution models, then iteratively applies algorithms to find the fittest models from the set to serve as inputs to the next iteration, ultimately leading to a model that best represents the data.

Glial Sytem: Commonly called glia (greek for “glue”), are non-neuronal cells that provide support and nutrition, maintain homeostasis, form myelin, and participate in signal transmission in the nervous system. In the human brain, glia cells are estimated to outnumber neurons by about 10 to 1.

Astrocytes: Astrocytes are a sub-type of the glial cells in the brain. They perform many functions, including the formation of the blood-brain barrier, the provision of nutrients to the nervous tissue, and play a principal role in the repair and scarring process in the brain. They modulate the synaptic transmission and recently their crucial role in the information processing was discovered.

Genetic Algorithms: Genetic algorithms (GAs) are adaptive heuristic search algorithm premised on the evolutionary ideas of natural selection and genetic. The basic concept of GAs is designed to simulate processes in natural system necessary for evolution, specifically those that follow the principles first laid down by Charles Darwin of survival of the fittest. As such they represent an intelligent exploitation of a random search within a defined search space to solve a problem.

Hybrid Training: earning method that combines the supervised and unsupervised training of Connectionist Systems.

Backpropagation algorithm: A supervised learning technique used for training ANNs, based on minimising the error obtained from the comparison between the outputs that the network gives after the application of a set of network inputs and the outputs it should give (the desired outputs).

Synapse: Specialized junctions through which the cells of the nervous system signal to each other and to non-neuronal cells such as those in muscles or glands.

Artificial Neural Network: A network of many simple processors (“units” or “neurons”) that imitates a biological neural network. The units are connected by unidirectional communication channels, which carry numeric data. Neural networks can be trained to find nonlinear relationships in data, and are used in applications such as robotics, speech recognition, signal processing or medical diagnosis.

Complete Chapter List

Search this Book: