“…Numerous methods have been developed to determine this and broadly fall into two categories. Conditioned behavior models (Rüttiger et al, 2003;Brozoski and Bauer, 2016;Pace et al, 2016) are often regarded as the more accurate, and require lengthy prior training of animals to perform or refrain from certain behaviors, such as licking, during the presence of an ongoing sound. Automatic response methods (Turner et al, 2006;Lobarinas et al, 2013) have the advantage of requiring no training, and exploit involuntary responses, such as the acoustic startle response, in conjunction with stimuli related to the possible tinnitus (e.g., a short gap in an ongoing pure tone) to modify this depending on tinnitus status, but are subject to caveats and controversies (Campolo et al, 2013;Lobarinas et al, 2013), and show inconsistent replicability in humans (Fournier and Hébert, 2013;Shadwick and Sun, 2014;Boyen et al, 2015).…”