Functional near-infrared spectroscopy (fNIRS) biomarkers of tinnitus severity within tinnitus subtypes

Tinnitus, the perception of sounds not present externally, can significantly impair quality of life. Currently there are no clinically available objective measures to enable monitoring of tinnitus-related changes in brain activity. Our previous work demonstrated sensitivity of a non-invasive imaging technique called functional near-infrared spectroscopy (fNIRS) to changes in brain activity associated with tinnitus presence and severity. In this study we aimed to assess whether fNIRS features could quantify tinnitus severity on a continuous scale, a requirement for effective monitoring of treatments. We collected fNIRS data at rest and in response to auditory and visual stimuli from a group of 103 individuals with tinnitus and 50 controls. We identified fNIRS-derived measures of connectivity which showed significant group differences and used these features in a linear regression model to predict tinnitus-related distress as rated using the Tinnitus Handicap Inventory (THI), and tinnitus loudness and annoyance scores. We assessed correlations between predicted and actual severity in the large tinnitus dataset as well as in two cohorts with shared clinical features- those with severe tinnitus (THI≥58), and cochlear implant users who experienced tinnitus. Findings showed predicted THI ratings correlated significantly with actual THI scores in our large group of individuals with tinnitus although predictive precision was limited, likely due to heterogeneity within this group. Predictions improved when considering both the severe cohort and cochlear implant users with tinnitus. A biomarker of tinnitus severity could provide a tool to monitor treatment effectiveness, particularly within tinnitus subtypes with shared clinical and neural characteristics.