Post-synaptic facilitation and network dynamics underlying stimulus-specific combination sensitivity
Combination-sensitive neurons (CSNs) integrate multiple stimulus features to generate behaviorally meaningful responses. While such neurons are well studied in fast timescale systems such as bat echolocation, the mechanisms enabling extended temporal integration in species like songbirds remain poorly understood. Using a computational model, we show that syllable-specific neurons in the songbird auditory system function as coincidence detectors whose selectivity depends on both post-inhibitory facilitation and persistent network activity. This persistence serves as a "memory trace", allowing precise association of sequential syllables across hundreds of milliseconds. We propose that this dynamic interplay between intrinsic neuronal properties and convergent synaptic inputs gives rise to a higher-order "meta-combination sensitivity" enabling the auditory system to transform discrete acoustic events into temporally extended percepts. Our findings provide a mechanistic framework that bridges theories of coincidence detection with longer-timescale working memory, highlighting the importance of distributed network mechanisms for auditory temporal coding.