Animal models of tinnitus require a behavioral correlate thereof. Various conditioned response methods and gap-startle reflex methods are in use and the outcomes generally correspond with putative electrophysiological substrates of tinnitus. However, for salicylate-induced tinnitus there is discordance between the behavioral and electrophysiological test results. As a result, it is not clear what the various tests are reflecting. A review of the, mostly sub-cortical, neural circuits that underlie the behavioral responses suggests that cortical electrophysiological correlates do not necessarily have to correspond with behavioral ones. Human objective correlates of tinnitus point heavily into cortical network, but not just auditory cortex, correlates of tinnitus. Furthermore, the synaptic mechanisms underlying spontaneous firing rate changes may be different from those involved in driven neural activity.