Recent information analyses suggest that latency to the first evoked spike of a cortical neuron following stimulus presentation may carry a relatively large proportion of transmitted information about the stimulus. We have previously investigated the use of ensemble first-spike latencies as a code for sound source direction at the level of primary auditory cortex (Al) in the cat. Several decoding issues exist in terms of how to properly frame the ensemble neural responses as probability distributions, and subseuently as likelihood functions or posterior probability distributions. In auditory cortex, first-spike latency is dependent not only on the direction of the sound source, but also on the intrinsic intensity of the sound source. This ambiguity, among others, must be addressed if first-spike latency is to be considered as a viable code for sound idrection. We approach this analysis as a parameter estimation problem where several parameters are considered as nuisance parameters and sound direction is considered to be the parameter of interest. Integrated likelihood and profile likelihood methods will be discussed for eliminating nuisance parameters to achieve a likelihood function for sound source direction.