CAJAL: A Simulator for Neurobiological Networks

Abstract

At the outset, it is necessary to draw a sharp dichotomy between two classes of software that are used to perform computer simulations of neural networks. In one class are the many simulators for artificial neural networks and connectionist applications. In the other, are the relatively few simulators for real biological neural networks and neurobiology applications. Although the two classes share certain surface characteristics (e.g. methods of representing network architectures), they are fundamentally different in concept, design, implementation and user applications. This article is devoted to neurobiological network simulators and describes one of these, called CAJAL, which we have developed and used rather intensively over the last three years. Some previous articles [KB92], [BK92] describe earlier versions of CAJAL. Here, drawing freely on [BK92], we elaborate and update these earlier accounts of CAJAL to give a description of the current version, called CAJAL-3.0. Having called attention to the dichotomy in the universe of neural network simulators, it is appropriate to draw a distinction within the class of neurobiological simulators, namely, between simulators for single biological neurons and simulators for arbitrary networks of biological neurons. CAJAL is intended primarily as a network simulator. However, it can also be used effectively to study and simulate a single neuron in great detail.KeywordsGraphical User InterfaceOrdinary Differential EquationDendritic TreeBiological Neural NetworkNeural Network SimulatorThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.