Subtyping tinnitus based on fMRI dynamic functional connectivity - Project summary Subjective tinnitus is a sensation of ringing or other noise perceived without external sound. People with tinnitus often experience comorbidities, such as hearing, mental, and sleep problems. Thus, tinnitus can substantially decrease quality of life and is a major socioeconomic burden. There is currently no objective method for diagnosing tinnitus or predicting an optimal treatment for an individual. The large interindividual variability of tinnitus perceptual characteristics and response to tinnitus treatments suggests that many different forms of tinnitus exist. Interactions between brain regions or “functional connectivity” measured by neuroimaging has suggested the premise as a biomarker for tinnitus, as it has been shown to correlate with tinnitus characteristics, change after successful tinnitus treatment, and even predict an outcome of tinnitus treatment. However, relatively small sample sizes have resulted in inconsistent findings from neuroimaging studies of the heterogeneous tinnitus population. The limited spatial resolution of the conventional 3T functional magnetic resonance imaging (fMRI) has also complicated the investigation of the role of subcortical structures in tinnitus. Furthermore, most studies have ignored the potentially meaningful fluctuations of brain functional connectivity over time. This K99/R00 project investigates if tinnitus can be subtyped based on the functional connectivity of brain networks. In the mentored K99 phase, the candidate aims to subtype the UK Biobank fMRI dataset of 7,868 people with tinnitus using state-of-the-art machine learning and individual-level analyses as well as dynamic functional connectivity that also captures information about functional connectivity fluctuations over time. To achieve this goal, the candidate will get familiarized with tinnitus research and clinical auditory neuroscience. Additionally, she will extend her strong background in biomedical engineering with big data, machine learning, and individual-level analyses. Building on this in the independent R00 phase, the candidate will apply the methods used in the K99 phase on an ultra-high resolution 7T fMRI dataset that she will measure from 120 participants with tinnitus and 120 controls. The high resolution allows investigation of between-subtype differences in the small auditory pathway nuclei. Furthermore, to associate the subtyping with more detailed tinnitus characteristics than available in the UK Biobank, the candidate will also collect a comprehensive set of questionnaires about tinnitus and its comorbidities and measure psychoacoustic and hearing tests from the same participants. The study could increase knowledge of the neural basis of tinnitus, and subtyping can lead to more accurate diagnosis and targeted treatment. The mentored phase is conducted at Athinoula A. Martinos Center for Biomedical Imaging and Massachusetts Eye & Ear Tinnitus Clinic, which together provide an ideal environment for the candidate to complement her skillset for launching an independent research program focusing on the development of neuroimaging biomarkers for tinnitus and other problems related to auditory and speech processing.
