Computational analyses of learning networks

As summarized in the present volume, the mechanisms that underlie learning and memory in invertebrates are the subject of intense study. Behavioral studies encompass many learning protocols, ranging from simple forms of nonassociative and associative learning such as habituation, sensitization, and classical and operant conditioning to higher forms of learning, such as observational learning. Moreover, mechanistic analyses span levels of biological organization ranging from gene regulatory networks to signal transduction cascades, single cells, and neural networks, as well as spanning temporal domains ranging from milliseconds to days. The vast amounts of data, their diversity, and their inherent complexity make it difficult to arrive at an intuitive and comprehensive understand of relationships among molecular, cellular, and network properties and cognitive processes such as learning and memory. p0030 Mathematical models and computer simulations provide quantitative and modifiable frameworks for representing, integrating, and manipulating complex data sets. The development and analyses of computational models of learning and memory mechanisms are helping to elucidate biological principles that are common in diverse species, common to more than one form of learning, and common to learning-induced changes in different types of behaviors. This chapter provides a brief overview of several well-characterized computational models of the biological processes that underlie learning and memory in invertebrates. It focuses on analyses of nonassociative and associative learning