On the Biological Plausibility of Artificial Metaplasticity - Foundations on Natural and Artificial Computation - page 119

On the Biological Plausibility of Artificial
Metaplasticity

Diego Andina and Javier Ropero-Peláez

Group for Automation in Signal and Communications, Technical University of
Madrid, Spain and
Federal University of A.B.C, Brazil

Abstract. The training algorithm studied in this paper is inspired by
the biological metaplasticity property of neurons. Tested on different
multidisciplinary applications, it achieves a more ecient training and
improves Artificial Neural Network Performance. The algorithm has been
recently proposed for Artificial Neural Networks in general, although for
the purpose of discussing its biological plausibility, a Multilayer Percep-
tron has been used. During the training phase, the artificial metaplastic-
ity multilayer perceptron could be considered a new probabilistic version
of the presynaptic rule, as during the training phase the algorithm assigns
higher values for updating the weights in the less probable activations
than in the ones with higher probability.

1 Introduction

Artificial Metaplasticity (AMP) term was first introduced by Andina [1] for

an Artificial Neural Network (ANN) of the Multilayer Perceptron type (MLP),
referred as AMMLP. During the AMMLP training phase, the matrix weight

et al

W
that models the synaptic strength of its artificial neurons is updated according to
the probability of the input patterns and therefore of the corresponding synap-
tic activations. The concept of biological metaplasticity was defined in 1996 by
W.C. Abraham [2] and now is widely applied in the fields of biology, neuro-
science, physiology, neurology and others [2, 3, 4]. The prefix “meta” comes from
Greek and means “beyond” or “ above”. In neuroscience and other fields “meta-
plasticity” indicates a higher level of plasticity, expressed as a change or trans-
formation in the way synaptic ecacy is modified. Metaplasticity is defined as
the induction of synaptic changes, that depends on prior synaptic activity [3, 5] .

Metaplasticity is due, at least in part, to variations in the level of postsynaptic
depolarization that induce synaptic changes. These variations facilitate synap-
tic potentiation and inhibit synaptic depression in depressed synapses (and vice
versa in potentiated synapses). The direction and the degree of the synaptic
alteration are functions of postsynaptic depolarization during synaptic activa-
tion. Upregulation -incrementing, reinforcement of synaptic ecacy- is termed

This research has been supported by Group for Automation in Signal and Commu-
nications, GASC/UPM.