A proposed role for electrical coupling in the neocortical slow oscillation.

Abstract

We constructed a computational thalamocortical network model for study of the neocortical slow oscillation. It incorporated a number of neuronal types, both excitatory and inhibitory, each model neuron simulated as a multicompartment entity with numerous membrane conductances. As in previous experimental and modeling studies, simulated slow oscillations primarily depended on recurrently connected deep intrinsic bursting (IB) pyramidal cells, with NMDA receptors being critical as well as intrinsic membrane conductances (e.g. persistent Na⁺); and with repolarization to the Down state dependent on intrinsic (slow Ca²⁺-dependent K⁺) and synaptic (GABA_B receptor mediated) conductances. Furthermore, however, we now can account for additional features of the slow oscillation: the frequent occurrence of spikelets, the presence of very fast ripple-like oscillations, and the transition to so-called fast runs (10 to ∼20 Hz bursty oscillations). These latter phenomena depended in our model on electrical coupling via gap junctions between pyramidal neurons. The importance of gap junctions is supported by previous experimental data on the ripple-blocking effect of halothane, as well as by data from the in vitro hippocampus.