The erroneous and widely held assumption that subjective tinnitus is generated in the inner ear, was for a long time an obstacle to an adequate and effective treatment of this symptom. Its continued persistence after section of the auditory nerve, and the observation that in many patients tinnitus is still perceived, even after restoration of their hearing, support the view that the associated "phantom-like" perceptions must originate in the brain; as we know today, these perceptions have their origin in neuroplastic, reorganization processes occurring in the auditory cortex. Such processes might be triggered by damages to the synapses between the inner hair cells (IHC) and the dendrites of the peripheral axons of the spiral ganglion neurons, leading to frequency-specific deafferentiations in the auditory pathway. The highly spezialized sensory cell synapses of the auditory pathway are extremely vulnerable to ionic disturbances occurring in the synaptic cleft, particularly to fluctuations of the Mg concentrations. Physiologically, the high discharge rate related to their sensory function is modulated in the afferent neurons through Mg-blockade of the NMDA receptors channels. The removal of the Mg-blockade brought about by a reduction of Mg ion concentrations in the synaptic cleft, results in an increased Ca-conductance through NMDA receptor channels, leading to a permanent depolarization of the postsynaptic membrane and consequently to deafferentation of auditory pathways for certain frequency ranges due predominantly to neurotoxic effects of external glutamate. If care is taken that Mg ions concentrations is restored to physiologic levels in the vicinity of the postsynaptic side of IHC or that the deafferentiated acoustic fibers are blocked in the initial stage by local anesthesia, the tinnitus would not arise. These new findings have led to the development of several promising methods (such as the repetitive transcranial magnetic stimulation and the combination of vagus nerve stimulation with selected, multiple sound frequencies), that are based on a modulation of cortical neuroplasticity aiming to a "reversal" of the acoustic "phantom perception" at the cortical level.