Abstract: Phantom sound perception (tinnitus) arises from aberrant central auditory pathways. Subjective tinnitus is typically caused by hearing loss. Auditory cortex (AC) in tinnitus animal models exhibits increased spontaneous neural firing and synchrony. Another subset of tinnitus exists whereby patients (80%) can manipulate pitch, tone and/or volume of tinnitus with head and neck/jaw movements. Termed somatic or somatosensory tinnitus (ST), this modulable form of phantom sound perception is associated with cortical & subcortical plasticity. Somatosensory/trigeminal neural inputs to auditory brainstem and cortex likely underlie ST etiology & maintenance. Thus, two forms of subjective tinnitus exist; 1. non-somatic tinnitus (non-ST; hearing loss etiology) 2. somatic (somatosensory /trigeminal etiology). Our group was the first to report equivalent objective neural correlates in human, non-ST, using non-invasive brain imaging with functional near-infrared spectroscopy (fNIRS). In our completed R21 we successfully validated an adapted fNIRS probe for the external auditory canal (EAC). Our unique design that was coupled with traditional scalp fNIRS probes, passes NIR light into ventral AC to measure hemodynamic responses (HRs), resting state functional neural connectivity (RSFC). We replicated predictable/published increases in AC HRs during silence in non-ST participants that is suppressed with sound stimulation. These objective correlates of non-ST in humans have not been evaluated in ST. Knowledge gaps in both non-ST & ST include: 1. Do objective changes in AC HRs & RSFC in non-ST tinnitus relate to tinnitus severity? 2. Does ST both at rest & during a somatic maneuver /somatosensory stimulation yield objective changes in AC HRs and RSFC? 3. Can objective neural correlates from non-ST & ST be used to construct predictive modeling/machine learning to distinguish these two forms of subjective tinnitus? Our central hypothesis is that human non-ST & ST will reveal objective measurable patterns of AC activity (HRs) & connectivity (RSFC) that is tinnitus severity specific that can be discerned with machine learning algorithms: We will investigate HRs and RSFC in subjective (AIM 1) and somatic (AIM 2) tinnitus variants. The data from these two AIMS will be considered with several tinnitus indices/questionnaires to determine how objective brain responses in tinnitus potentially correlate with tinnitus severity. Data generated from AIMS 1 and 2 will then provide the necessary information to build machine learning algorithms that can then be used to predict HR and RSFC in human subjects going forward as a clinical and research tool. Based on subjective severity and modulation properties, the goal of this AIM is to establish predictive neural maps that may begin to individualize underlying properties of pathology that may be variant from participant to participant. In the age of personalized medicine, this will be an important contribution to identify dis...