Post-synaptic facilitation and network dynamics underlying stimulus-specific combination sensitivity

Combination-sensitive neurons (CSNs) transform specific combinations of sensory features into selective neural responses, supporting the temporal organization of complex signals. We developed a biophysically realistic computational model, constrained by intrinsic and synaptic properties, to examine how temporally ordered stimulus pairs can be associated across delays of hundreds of milliseconds. The model shows that upstream inhibitory-delay circuits can transiently preserve information about the first stimulus while progressively shaping the excitability of an output neuron. Together with post-inhibitory rebound and facilitation, these dynamics create a delay-dependent primed state that funnels separated inputs into a narrow coincidence-detection window. This mechanism enables sharply timed, selective responses to stimulus pairs without requiring sustained spiking in the output neuron. Our results provide a circuit-level link between coincidence detection and longer-timescale temporal integration, offering a general framework for temporal coding in sensory systems.