Simulation of gamma rhythms in networks of interneurons and pyramidal cells

Abstract

Networks of hippocampal interneurons, with pyramidal neurons pharmacologically disconnected, can generate gamma-frequency (20 Hz and above) oscillations. Experiments and models have shown how the network frequency depends on excitation of the interneurons, and on the parameters of GABA(A)-mediated IPSCs between the interneurons (conductance and time course). Here we use network simulations to investigate how pyramidal cells, connected to the interneurons and to each other through AMPA-type and/or NMDA-type glutamate receptors, might modify the interneuron network oscillation. With or without AMPA-receptor mediated excitation of the interneurons, the pyramidal cells and interneurons fired in phase during the gamma oscillation. Synaptic excitation of the interneurons by pyramidal cells caused them to fire spike doublets or short bursts at gamma frequencies, thereby slowing the population rhythm. Rhythmic synchronized IPSPs allowed the pyramidal cells to encode their mean excitation by their phase of firing relative to the population waves. Recurrent excitation between the pyramidal cells could modify the phase of firing relative to the population waves. Our model suggests that pools of synaptically interconnected inhibitory cells are sufficient to produce gamma frequency rhythms, but the network behavior can be modified by participation of pyramidal cells.