Tinnitus is an auditory phantom sound perception in the absence of any external source which manifests as a ringing, whistling or buzzing in the ear or head.
It is usually related to hearing loss, either caused naturally (ageing) or accidentally (induced by noise or inner ear conditions).
With a steadily ageing population, work and leisure noise exposure, tinnitus becomes more and more relevant to many individuals of the society.
Between 12% to 30% experience some sort of tinnitus and about 1% to 3% of the general population suffers considerably under chronic tinnitus.
Tinnitus intensity but especially related distress vary greatly between persons.
To this day, the disease is neither curable nor properly understood by science.
Some neuroscientific brain imaging studies have investigated differences between tinnitus patients and healthy controls.
Results were rather inconsistent and unspecific so that no useful insights about mechanisms and treatments could be derived.
This can be explained by the small sample sizes used in the studies as well as shortcomings in study design and analysis.
With the use of extensive MRI data in combination with further health and biomarkers from the UKBB, we plan to surpass these shortcomings.
For the sake of analysis, current state of the art methods of brain imaging and machine learning will be used.
These methods ensure high quality and consistency of analyses and minimize human interaction or bias.
The aim of this data analysis is thus to predict tinnitus and its severity with neuroimaging, namely MRI structural and functional data including a set of control variables available from the UKBB.
More concretely, we want to examine how
– brain structure (grey and white matter)
– brain connectivity
are able to predict a) tinnitus incidence and b) tinnitus severity with machine learning classifiers.
In addition, results will be associated with genetic data to identify possible predispositions.
Data from several time points will be used to test the robustness of the found effects as well as to better understand the natural history of tinnitus.
So far, epidemiological research on tinnitus has never considered the inclusion of neuroscientific data and is traditionally bound to clinical data or surveys.
Thus, the method and scale of this project are unseen in the field of tinnitus and will inform future scientific studies.
Results may help to elucidate basic mechanisms, individual subtypes, and finally individual treatments for tinnitus.
