Abstract
Animal models have significantly contributed to understanding the pathophysiology of chronic subjective tinnitus. They are useful because they control etiology, which in humans is heterogeneous; employ random group assignment; and often use methods not permissible in human studies. Animal models can be broadly categorized as either operant or reflexive, based on methodology. Operant methods use variants of established psychophysical procedures to reveal what an animal hears. Reflexive methods do the same using elicited behavior, for example, the acoustic startle reflex. All methods contrast the absence of sound and presence of sound, because tinnitus cannot by definition be perceived as silence.