Brain stimulation-induced neuroplasticity underlying therapeutic response in phantom sounds.

Noninvasive brain stimulation can modify phantom sounds for longer periods by modulating neural activity and putatively inducing regional neuroplastic changes. However, treatment response is limited and there are no good demographic or clinical predictors for treatment outcome. We used state-of-the-art voxel-based morphometry (VBM) to investigate whether transcranial magnetic stimulation-induced neuroplasticity determines therapeutic outcome. Sixty subjects chronically experiencing phantom sounds (i.e., tinnitus) received repetitive transcranial magnetic stimulation (rTMS) of left dorsolateral prefrontal and temporal cortex according to a protocol that has been shown to yield a significantly higher number of treatment responders than sham stimulation and previous stimulation protocols. Structural magnetic resonance imaging was performed before and after rTMS. In VBM whole-brain analyses (P < 0.05, FWE corrected), we assessed longitudinal gray matter changes as well as structural connectivity between the ensuing regions. We observed longitudinal mesoscopic gray matter changes of left dorsolateral prefontal (DLPFC), left operculo-insular, and right inferior temporal cortex (ITC) in responders (N = 22) but not nonresponders (N = 38), as indicated by a group × time interaction and post-hoc tests. These results were neither influenced by age, sex, hearing loss nor by tinnitus laterality, duration, and severity at baseline. Furthermore, we found robust DLPFC-insula and insula-ITC connectivity in responders, while only relatively weak DLPFC-insula connectivity and no insula-ITC connectivity could be demonstrated in nonresponders. Our results reinforce the implication of nonauditory brain regions in phantom sounds and suggest the dependence of therapeutic response on their neuroplastic capabilities. The latter in turn may depend on (differences in) their individual structural connectivity. Hum Brain Mapp 39:554-562, 2018.