Using Two Series-Connected Reservoirs in the Prediction Task of Field Excitatory Postsynaptic Potentials in the Rodent Hippocampus
摘要
Abstract
We propose and test an approach to predict local electrical potential signals in the rodent hippocampal trisynaptic circuit using reservoir computing on the basis of experimentally obtained biological data. The results of prediction are compared for two predictor architectures: one with one reservoir and with two reservoirs connected in series. The results obtained can be used in tasks of restoring hippocampal activity using neurohybrid chips.