Study on the role of inhibition on granule cells of the dentate gyrus in the propagation of activity and computation in the hippocampus

Electrophysiological recordings in the hippocampus have revealed a tight control of inhibitory interneurons over the Granule Cells (GC), the principal excitatory neurons of the Dentate Gyrus (DG). This excitation/inhibition balance is crucial for information transmission and likely relies on inhibitory synaptic plasticity. Previous experiments have shown that Long-Term Potentiation (LTP) of the Perforant Pathway, not only potentiates glutamatergic synapses but also decreases feed-forward inhibition in the DG, facilitating activity propagation in the circuit. To investigate this phenomenon, we built a population computational model where neurons were described by Izhikevich's equations. The model contained the minimal elements required to reproduce the neuronal dynamics reported experimentally in the DG. The results obtained from the numerical integration of the model equations, before and after LTP induction, support the counterintuitive experimental observation of an LTP-induced synaptic depression of the feed-forward inhibitory connection. We find that LTP increases the efficiency of the glutamatergic input to recruit the inhibitory network of the hilar region, resulting in an average reduction of the basket cell population activity. The predictions of the model were supported electrophysiologically in mice, in an in vitro preparation of intracellular patch-clamp recordings after in vivo LTP induction. Furthermore, we found that LTP preferentially decreased perisomatic inhibitory inputs without affecting distal dendritic inputs. The potential functional implications of the found dendritic/perisomatic imbalance induced by LTP was further investigated using a detailed computational model of GCs. Overall, our findings suggest that LTP rebalance the inhibitory network in the DG resulting in a reduction in the tight control of basket cells over GCs firing, increasing burst firing and reliability in response to a constant input pattern. As a consequence, information transmission to CA3 is enhanced without affecting pattern separation. Overall, the result of this thesis support a network mechanism operated by synaptic plasticity and based on the regulation of perisomatic inhibition in GCs to control information transmission in the hippocampus.