Neural Counting and Photon Counting in the Presence of Dead Time

The usual stimulus-based neural counting model for audition is found to be mathematically identical to the well-know semiclassical formalism for photon counting. In particular, we explicitly demonstrate the equivalence of McGill's noncentral negative binomial distribution and Pe {r} {r}{}ina's multimode confluent hypergeometric distribution for a coherent signal imbedded in chaotic noise. Dead-time corrections, important both in neural counting and in photon counting, are incorporated in a generalized form of this distribution. Some specific implications of these results are discussed.