Lecture Notes in Computer Science 1240 Edited by G. Goos, J. Hartmanis and J. van Leeuwen Advisory Board: W. Brauer D. Gries J. Stoer Jos6 Mira Roberto Moreno-Diaz Joan Cabestany (Eds.) Biological dna Artificial Computation: From Neurosciene to Technology International Work-Conference on Artificial and Natural Neural Networks, IWANN' 97 Lanzarote, Canary Islands, Spain June 4-6, 1997 Proceedings regnirpS Series Editors Gerhard Goos, Karlsruhe University, Germany Juris Hartmanis, Cornell University, NY, USA Jan van Leeuwen, Utrecht University, The Netherlands Volume Editors Jos6 Mira Departamento de InteligenciaArtificial Universidad Nacional de Educaci6n a Distancia Senda del Rey s/n, E-28040 Madrid, Spain E-mail: [email protected] Roberto Moreno-Diaz Centro Inter. de Investigacion en Ciencias de la Computacion Universidad de las Palmas de Gran Canaria Campus de Tafira, E-35017, Canary Islands, Spain E-mail: roberto @grumpy.dis.ulpgc.es Joan Cabestany Departament d'Enginyeria Electronica, Universitat Polit6cnica de Catalunya C/Gran Capita s/n, E-08034 Barcelona, Spain E-mail: [email protected] Cataloging-in-Publication data applied for Die Deutsche Bibliothek - CIP-Einheitsaufnahme Biological and artificial computation : from neuroscience to neurotechnology ; proceedings / International Work Conference on Artificial Neural Networks, IWANN '97, Lanzarole, Canary Islands, Spain, June 4 - 6, 1997. Jos~ Mira ... (ed.). - Berlin ; Heidelberg ; New York ; Barcelona, Budapest ; tlong Kong ; London ; Milan ; Paris ; Santa Clara ; Singapore ; Tokyo : Springer, 1997 (Lecture notes ni computer science ; 1240) Vol. ISBN 3-540-63047-3 CR Subject Classification (1991): 1.2,F.1.1, C.1.3, C.2.1, G.1.6, 2.5.1, B.7.1, J.1, J.2 ISSN 0302-9743 ISBN 3-540-63047-3 Springer-Verlag Berlin Heidelberg New York This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, re-use of illustrations, recitation, broadcasting, reproduction on microfilms or in any other way, and storage in data banks. Duplication of this publication orp arts thereof is permitted only under the provisions of the German CopyrightL aw of September 9, 1965, in its current version, and permission for use musta lways be obtained from Springer -Verlag. Violations are liable for prosecution under the German Copyright Law. (cid:14)9 Springer-Verlag Berlin Heidelberg 1997 Printed in Germany Typesetting: Camera-ready by author SPIN 10548822 06/3142 - 5 4 3 2 1 0 Printed on acid-free paper Preface Neural computation is considered here in the dual perspective of analysis (as reverse engineering) and synthesis (as direct engineering). As a science of analysis, neural computation seeks to help neurology, brain theory, and cognitive psychology in the understanding of the functioning of the suow~en system by means of computational models of neurons, neural nets, and subcellular processes, with the possibility of using electronics and computers as a "laboratory" in which cognitive processes can be simulated and hypothesperso ven without having to act directly upon living beings. As a direct engineering (how can we build sub-symbolic intelligent machines?), neural computation seeks to complement the symbolic perspective of artificial intelligence (AI), using the biologically inspired models of distributed self- programming and self-organizing networks, to solve those non-algorithmic problems of function approximation and pattern classification having to do with changing and only partially known environments. Fault tolerance and dynamic reconfiguration are other basic advantages of neural nets. In the sea of meetings, congresses, and workshops on ANNs, IWANN'97, the fourth International Work-Conference on Artificial Neural Networks, that took place in Lanzarote, Canary Islands (Spain), 4 - 6 June, 1997, focused on the three subjects that most worry us: (1) The search for biologically inspired new models of local computation architectures and learning along with the organizational principles behind the complexity of intelligent behavior. (2) The search for some methodological contributions in the analysis and design of knowledge-based ANNs, instead of "blind nets", and in the reduction of the knowledge levetlo the sub-symbolic implementation level. (3) The cooperation with symbolic AI, with the integration of connectionist and symbolic processing in hybrid and multi-strategy approaches for perception, decision, and control tasks, as well as for case-based reasoning, concept formation, and learning. To contribute to the formulation and partial solution of these global topics, IWANN'97 offered a brain-storming interdisciplinary forum in advanced neural computation for scientists and engineers/~om biology, neuroanatomy, computational neurophysiology, molecular biology, biophysics, mathematics, computer science, artificial intelligence, parallel computing, electronics, cognitive sciences, and all the concerned applied domains (sensory systems and Signal processing, monitoring, diagnosis, classification, and decision making, intelligent control and supervision, perceptual robotics and communication systems). The papers presented here correspond to talks delivered at IWANN'97, organized by the Universidad National de Educaci6n a Distancia (LrNED), Madrid, Universidad de Las Palmas de Gran Canaria, and Universidad Polit6cnica de IV Catalunya, in cooperation with the Asociaci6n Espafiola de Redes Neuronales (AERN), IFIP Working Group in Neural Computer Systems, WG10.6, Spanish RIG IEEE Neural Networks Council, tahned UK&RI Communication Chapter of IEEE. Sponsorship has been obtained from the Spanish CICYT and DGICYT (MEC) atnhde organizing universities ,DENIL( Las Palmas, and Catalunya). After the evaluation process, 241 papers were accepted for oral presentation or poster, according to the recommendations of reviewers and the author's preferences. The three extended papers corresponding to the invited speakers (DeFelipe, Eckhom, and Ierme) have been included as introductions to the corresponding topics of neuroscience, neurmaold eling perception, in and implementation, We would like to thank all the authors as well as all the members of the international program committee for their labor in the production and evaluation of the papers. Only by proceeding with this severe averaging of the external experts' reviews, could web e sure to maximize the originality, technical quality, and scientific relevance of this event. We also would like to mention the effort of the authors of rejected papers, mainly because they were immature proposals or topics not covered by IWANN. Last but not least, the editors would like to thank Springer-Verlag, in particular Alfred Hofrnann, for the continuous and excellent cooperative collaboration fromt he first IWANN in Granada (1991, LNCS 540), the successive meetings in Sitges (1993, LNCS 686) and Torremolinos (1995, LNCS 930), annodw Lanzarote. in The papers published in this volume present the current situation in natural and artificial neural nets, with a significant increase in the contributions related to the biological foundations of neural computation and the computational perspective of neuroscience. We have organized the papers in the following sections: o~o Biological Foundations of Neural Computation ~ Formal Tools and Computational Models of Neurons and Neural Nets Architectures o~o Plasticity Phenomena (Maturing, Learning and Memory) o~o Complex Systems Dynamics o~o Cognitive Science and IA ~:o Neural Nets Simulation, Emulation and Implementation o~o Methodology for Data Analysis, Selection Task and Nets Design .~o Neural Neneorks for Communications, Control and Robotics This book endeavors to summarize the state of the art in neural computation with a focus on biologically inspired models of the natural nervous system. The complexity of the nervous system is now accepted, and a significant part of the scientific community has returned to anatomy and physiology, rejecting the temptation to use models which are clearly insufficient to cope with this complexity. At the same time there is an increasing interest in the use of computational models of neural networks to improve our understanding of the functional organization of the brain. Finally, there is iiv also evidence of a lack of formal tools enabling the hybridization of the symbolic and connectionistic perspectives of artificial intelligence in the common goal of making computational the knowledge of human experts in technical domains related with perception, communication, and control. All these developments, as reported in these proceeding, are needed in order to bring neuroscience and computation closer together. To recognize the disparity that exists between the richness and fineness of the nervous system and the crudeness we use in handling it is a good step forward. Madrid, March 1997 J. Mira Mira R. Moreno-Diaz J. Cabestany Moncusi Contents I. Biological Foundations of Neural Computation Microcircuits in the Brain J. De Felipe Some Reflections on the Relationships Between Neuroscience and Computation 51 J.Mira, A.E. Delgado Different Types of Temporal Correlations Obtained in Pairs of Thalamic Visual Neurons Suggest Different Functional Patterns of Connectivity 27 .C Rivadulla, J. Cudeiro Development of On-Off and Off-On Receptive Fields Using a Semistochastic Model 53 E. M. Muro, P. Isasi, M. A. Andrade, F. Mordn The Classification of Spatial, Chromatic, and Intensity Features of Simple Visual Stimuli by Network a of Retinal Ganglion Cells 44 .S Shoham, R. Osan, J. Ammermuller, A. Branner, E. Ferndndez, R. A. Normann Geometric Model of Orientation Tuning Dynamics in Neurons Striate 45 t.A. Shevelev , K.A. Saltykov, G.A. Sharaev Neuronal Circuitry in the Medial Cerebral Cortex of Lizards 16 J.A. Luis de la lglesia, .C L6pez-Garcia Interactions Between Environmental and Hormonal Oscillations Induce Plastic Changes in a Neuroendocrine Simple Transducer 27 R. Alonso, .I L@ez-Coviella, F. Herndndez-Diaz, P. Abreu, E. Salido, L. Tabares Current Source Density Analysis as a Tool to Constrain the Parameter Space Hippocampal in 1AC Neuron Models 28 P. Varona, ..J M. lbarz, J. A. Sigiienza, .O Herreras Spontaneous Activity of Hippocampal Cells in Various Physiological States 19 N. Stollenwerk, .L Men~ndez de la Prida, J. .V Sdnchez-Andrds Neural Network Model of Striatal Complex 301 .B Aleksandrovsky, .F Bracher, .G Lynch, .R Granger Symmetry and Self-Organization of the Oculo-Motor Neural Integrator 611 .T .J Anastasio Quantal Neural Mechanisms Underlying Movement Execution and Motor Learning 421 J.M. Delgado-Garcia, A. J.A. Gruart, Domingo, J.A. Trigo A Model of Cerebellar Saccadic Motor Learning Using Qualitative Reasoning 331 J.L. Krichmar, G.A. Ascolg .L Hunter, J.L. Olds Balance Between Intercellular Coupling and Input Resistence as a Necessary Requirement for Oscillatory Electrical Activity in Pancreatic slleC-31 146 E. Andreu, .R omares, P .B Soria, J.V. Sdnchez-Andr& Mechanisms of Synchronization in the Hippocampus and Its Role 154 Along Development L. Mendndez de K Prida la J. S6nchez-Andr~s , Analysis of Synfire Chains Above Saturation 162 R,M. Reyes, C.J. P&ez Vicente Allometry in the Justo Gonzalo's Model of the Sensorial Cortex 169 L Gonzalo 2. Formal Tools and Computational Models of Neurons and Neural Net Architectures 178 Systems Models of Retinal CeUs: A Classical Example .R Moreno-Diaz A Generic Formulation of Neural Nets as a Model of Parallel and 591 Self-Programming Computation .J Mira, J.C. Herrero, A.E. Delgado Using an Artificial Neural Network for Studying the Interneuronal 207 Layer of a Leech Neuronal Circuit J.M. Santos, .L Szczupak IX Capacity and Parasitic Fixed Points Control in a Recursive Neural Network 217 .V Gimdnez , M. Pdrez-Castellanos, J. Rios Carrion, F. de Mingo The Use of Prior Knowledge in Neural Network Configuration and Training 227 M. Hilario, A. Rida A Model for Heterogeneous Neurons and Its Use in Configuring Neural Networks for Classification Problems 237 J.J. Valdds, R. Garcia A Computation Theory for Orientation-Selective Simple Cells Based on the MAP Estimation Principle and Markov Random Fields 247 M. N. Shirazi, .Y Nishikawa Competition Between Feed-Forward and Lateral Information Processing in Layered Neural Networks 257 A. .C .C Coolen, L. Viana Computing Ftmctions with Neurons Spiking in Temporal Coding 562 B. Ruf An Introduction to Fuzzy State Automata 372 L. Reyneri Statistical Analysis of Regularization Constant - From Bayes, MDL and NIC Points of View 284 S.-i. AmarL N.Murata Building Digital Libraries from Paper Documents, Using ART Based Neuro-Fuzzy Systems 294 R. Sam Guadarrama, Y.A. Dimitriadis , G.1. Sainz Palmero, J,M. Cano lzquierdo, J. L6pez Coronado Parallelization of Connectionist Models Based on a Symbolic Formalism 304 J. Santos, M. Cabarcos, R.P. Otero , J. Mira Generic Neural Network Model and Simulation Toolkit 313 M. Garcla del Valle, .C Garcia-Orellana, F.J. Ldpez-Aligud, I. Acevedo-Sotoca IIX A Neural-Fuzzy Technique for Interpolating Spatial Data via the Use of Learning Curve 323 P.M. Wong, ..W.K Wong, C.C. Fung, TD.Gedeon Task Decomposition Based on Class Relations: A Modular Neural Network Architecture for Pattern Classification 330 B.-L. ,uL .M Ito Lower Bounds of Computational Power of a Synaptic Calculus 340 JP. Neto, J.F. Costa, .H Coelho Feed Forward Neural Network Entities 349 A. Hadjiprocopis, .P Smith 3. Plasticity Phenomena (Maturing, Learning and Memory) Astrocytes and Slow Learning in the Formation of Distal Cortical Associations 360 J.G. Wallace, .K Bluff Adaptation and Other DynamicE ffects on Neural Signal Transfer 370 .L Orz6, .E Ldbos Hebbian Learning in Networks of Spiking Neurons Using Temporal 380 Coding .B Ruf .M Schmitt An Associative Learning Model for Coupled Neural Oscillators 390 .J Nishii Random Perturbations to Hebbian Synapses of Associative Memory 398 Using a Genetic Algorithm A. lmada, .K Araki 408 Phase Memory in Oscillatory Networks M.G. Kuzmina, LL Surina Strategies for Autonomous Adaptation and Learning in Dynamical 417 Networks . N.H Farhat, E. Del Moral Hernandez, Lee G.-H. Modeling the Parallel Development of Multiple Featuremaps and 427 Topography in Visual Cortex .A.W Fellenz