Identification of functional synaptic plasticity from spiking activities using nonlinear dynamical modeling

摘要: Abstract This paper presents a systems identification approach for studying the long-term synaptic plasticity using natural spiking activities. consists of three modeling steps. First, multi-input, single-output (MISO), nonlinear dynamical neuron model is formulated to estimate and represent strength in means functional connectivity between input output neurons. Second, this MISO extended nonstationary form track time-varying properties strength. Finally, Volterra method used extract learning rule, e.g., spike-timing-dependent plasticity, explanation input–output nonstationarity as consequence past patterns. framework developed study underlying mechanisms memory formation behaving animals, may serve computational basis building next-generation adaptive cortical prostheses.