In the present study, we used an innovative music-rest interleaved fMRI paradigm to investigate the neural correlates of tinnitus distress. Tinnitus is a poorly-understood hearing disorder where individuals perceive sounds, in the absence of an external source. Although the great majority of individuals habituate to chronic tinnitus and report few symptoms, a minority report debilitating distress and annoyance. Prior research suggests that a diverse set of brain regions, including the attention, the salience, and the limbic networks, play key roles in mediating both the perception of tinnitus and its impact on the individual; however, evidence of the degree and extent of their involvement has been inconsistent. Here, we minimally modified the conventional resting state fMRI by interleaving it with segments of jazz music. We found that the functional connectivity between a set of brain regions–including cerebellum, precuneus, superior/middle frontal gyrus, and primary visual cortex–and seeds in the dorsal attention network, the salience network, and the amygdala, were effective in fractionating the tinnitus patients into two subgroups, characterized by the severity of tinnitus-related distress. Further, our findings revealed cross-modal modulation of the attention and salience networks by the visual modality during the music segments. On average, the more bothersome the reported tinnitus, the stronger was the exhibited inter-network functional connectivity. This study substantiates the essential role of the attention, salience, and limbic networks in tinnitus habituation, and suggests modulation of the attention and salience networks across the auditory and visual modalities as a possible compensatory mechanism for bothersome tinnitus.