This thesis contributes the synaptic update rule finder (SURF), a framework for automatically inferring the update rule and naive synaptic weights controlling the circuit underlying a given plastic behaviour, from a time series of neural activity observations from that circuit over the course of learning in a real organism.

The objective of this work is to characterize the neural activity within the medial prefrontal cortex (mPFC) and basolateral amygdala (BLA) of rats, during a non-stationary two-armed bandit-style task. We performed simultaneous multi-unit extracellular recording in the mPFC and BLA.