Modeling active dendritic processes in pyramidal neurons

The role of active ion channels in dendritic function is among the most interesting and complex aspects of information processing in single neurons While the behavior of isolated channels or the passive electrical properties of dendrites can be studied in isolation, the interaction of multiple nonlinear ionic currents within a geometrically complex structure is described by equations that cannot be solved analytically and often resist intuition. Detailed computer models thus provide a crucially needed framework within which hypotheses about active dendritic mechanisms can be expressed and tested. Realistic neuronal models help determine whether known biophysical properties can account for experimental observations and provide insight into computational functions of these mechanisms. The chief obstacle in constructing such models is obtaining constraints for their parameters. Until recently, not enough experimental data were available on the physiological properties of cortical neurons to be able to create accurate models. In early models of hippocampal and neocortical neurons (e. properties of the channels were mixed and matched from cortical and noncortical neurons, including motoneurons in the spinal cord, sympathetic ganglion cells, and even the squid giant axon. Still scantier data were available regarding the spatial distribution of channels present in dendrites, axons, and synaptic terminals. Fortunately, the application of dendritic patch recordings, ca2+ imaging, and immunocytochem-istry and other molecular techniques is now yielding an abundance of precise data on the properties and localization of ion channels in the dendritic membrane. Furthermore , dendritic recordings and imaging of ca2+-or voltage-sensitive dyes provide windows into dendritic behavior not available with traditional somatic recordings. Thus the time is ripe for detailed models of active dendritic computation. The basics of compartmental models are presented in chapter 3, this volume, and numerical methods for solving them in chapter 14. In the present chapter, we show how compartmental simulations can be used to model neurons with active dendritic ion channels; our focus is chiefly on modeling pyramidal neurons such as those of the neocortex and hippocampus, but much of the material can be applied more widely. In section 5.2, we present definitions of channel types, their densities, and their localization, and we discuss the steps involved in combining experimental data into a