Electrophysiological recordings have demonstrated a tight inhibitory control of hilar interneurons over Dentate Gyrus granule cells (DGgc) (Bragin et al. 1995; Permía-Andrade et al. 2014). This excitation/inhibition balance is crucial for information transmission (Bartos et al., 2001) and likely relies on inhibitory synaptic plasticity (Vogels et al., 2011). Our experiments show that LTP induction in the Perforant Pathway (PP), not only potentiates glutamatergic synapses but unexpectedly decreases feed-forward inhibition in the DG, facilitating activity propagation in the circuit and modifying the long-range connectivity in the brain. To investigate this phenomenon, we propose to study a circuit of populations of point neurons described by the Izhikevich model. The model contains entorhinal cortex (EC) neurons, DGgc, mossy cells, basket cells and hilar interneurons. The proportion of neurons per population and the connectivity of the neural network is based on anatomical published data and is fitted to achieve experimental electrophysiological in vivo recordings (Permía-Andrade et al. 2014). The study of the effect of LTP in the local circuit of the DG is performed in the model adapting synaptic weights in the EC projections. The results obtained from the model, before and after LTP induction, support the counterintuitive experimental observation of synaptic depression in the feed-forward inhibitory connection induced by LTP. We show that LTP increases the efficiency of the glutamatergic input to recruit the inhibitory network, resulting in a reciprocal cancellation of the basket cell population activity. We validate the result of the model by electrophysiological experiments inducing LTP in the PP of anaesthetized mice _in vivo_ and recording excitatory and inhibitory currents in vitro in the same animals. Overall, our findings suggest that LTP of the EC input increases the excitation/inhibition balance, and facilitates activity propagation to the next station in the circuit by recruiting an interneuron-interneuron network that inhibits the tight control of basket cells over DGgc firing.

Funded by Conselleria d’Innovació, Recerca I Turisme of the Government of the Balearic Islands and European Social Fund (FPI/1900/2016), María de Maeztu Program for Units of Excellence in R&D (MDM-2017-0711), Ministerio de Economía y Competitividad (MINECO) FEDER funds (# GC2018-101055-B-I00) and Severo Ochoa Program for Centres of Excellence in R&D (SEV- 2017-0723).

Trabajo presentado en la 30th Annual Computational Neuroscience Meeting (CNS*2021), celebrada online del 3 al 7 de julio de 2021.

Peer reviewed