These articles question the basis on which tinnitus is measured in animal models. And hence the articles question whether the pre-clinical studies demonstrating efficacy of compound x,y, or z live up to their promise in the "human test lab" so to speak...
A (concerned) passerby
It would seem that ABR testing is a far more robust method for monitoring tinnitus-related alterations in the auditory system in both animals and humans. The startle response is just a cheap and easy way to measure this. It has its limitations and researchers often make conclusions without considering them.
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4335042/