Network burst activity in hippocampal neuronal cultures: the role of synaptic and intrinsic currents

The goal of this work was to define the contributions of intrinsic and synaptic mechanisms toward spontaneous network-wide bursting activity, observed in dissociated rat hippocampal cell cultures. This network behavior is typically characterized by short-duration bursts, separated by order of magnitude longer interburst intervals. We hypothesize that while short-timescale synaptic processes modulate spectro-temporal intraburst properties and network-wide burst propagation, much longer time-scales of intrinsic membrane properties such as persistent sodium (Na-p) currents govern burst onset during interburst intervals. To test this, we used synaptic receptor antagonists picrotoxin, 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), and 3-(2-carboxypiperazine-4-yl) propyl-1-phosphonate (CPP) to selectively block GABA(A), AMPA, and NMDA receptors and riluzole to selectively block Na-p channels. We systematically compared intracellular activity (recorded with patch clamp) and network activity (recorded with multielectrode arrays) in eight different synaptic connectivity conditions: GABA(A) + NMDA + AMPA, NMDA + AMPA, GABA(A) + AMPA, GABA(A) + NMDA, AMPA, NMDA, GABA(A), and all receptors blocked. Furthermore, we used mixed-effects modeling to quantify the aforementioned independent and interactive synaptic receptor contributions toward spectro-temporal burst properties including intraburst spike rate, burst activity index, burst duration, power in the local field potential, network connectivity, and transmission delays. We found that blocking intrinsic Na-p currents completely abolished bursting activity, demonstrating their critical role in burst onset within the network. On the other hand, blocking different combinations of synaptic receptors revealed that spectro-temporal burst properties are uniquely associated with synaptic functionality and that excitatory connectivity is necessary for the presence of network-wide bursting. In addition to confirming the critical contribution of direct excitatory effects, mixed-effects modeling also revealed distinct combined (non-linear) contributions of excitatory and inhibitory synaptic activity to network bursting properties.