We report on development and use of dynamic-clamp technology to understand how synchronous neuronal activity is generated in the hippocampus and entorhinal cortex. We find that “hard” real-time dynamic-clamp systems, characterized by very small maximal errors in timing of feedback, are necessary for cases in which fast voltage-gated channels are being mimicked in experiments. Using a hard real-time system to study cellular oscillations in entorhinal cortex, we demonstrate that the stochastic gating of persistent Na+ channels is necessary for cellular oscillations, and that cellular oscillations lead to dynamic changes in gain for conductance-based synaptic inputs. At the network level, we review experiments demonstrating that oscillating entorhinal stellate cells synchronize best via mutually excitatory interactions. Next, we show that cellular oscillations are volatile in the hypothesized “high-conductance” state, thought to occur in vivo, and suggest alternate means by which coherent activity can be generated in the absence of strong cellular oscillations. We close by discussing future developments that will increase the utility and widespread use of the dynamic-clamp method.