Project Summary
This semester or MSc thesis project will investigate how large-scale brain networks reorganize during neurofeedback treatment for chronic tinnitus. [1] Using longitudinal fMRI data acquired over 3–4 months per subject, the student will analyse functional connectomes across multiple sessions to identify individual-specific connectivity patterns. A key objective is to apply brain fingerprinting approaches to track network stability and change over time. [2,3] These neural markers will be linked to clinical measures such as tinnitus distress and symptom improvement. The project contributes to precision neuroscience approaches for better understanding and treating chronic subjective tinnitus.
Background
Chronic tinnitus is characterized by the persistent perception of phantom sounds and can be associated with substantial distress, sleep disruption, anxiety, and reduced quality of life. While peripheral auditory damage often triggers tinnitus, increasing evidence indicates that central brain mechanisms and large-scale network alterations play a critical role in its persistence and clinical heterogeneity. Neuroimaging studies have identified abnormal activity and connectivity in auditory, limbic, salience, and default-mode networks. [4] Conventional fMRI analyses, however, are largely limited to regional activation or pairwise functional connectivity, which may fail to capture the complex and distributed nature of tinnitus-related brain dynamics. Connectomics provides a whole-brain framework to quantify how functional networks are organized, segregated, and integrated. Within this framework, brain fingerprinting has emerged as a powerful approach to characterize individual-specific connectivity profiles and their longitudinal stability. [5] Applying these methods to tinnitus fMRI neurofeedback data offers a unique opportunity to link network-level reorganization to clinical improvement and to identify potential biomarkers of treatment response.
Project Description
The project focuses on the analysis of longitudinal fMRI neurofeedback data acquired over repeated sessions in 21 patients with chronic tinnitus. The student will reconstruct functional connectomes using established brain parcellations and quantify their evolution across time. Brain fingerprinting and graph-theoretical metrics will be used to assess within-subject stability and between-subject specificity of connectivity patterns. These measures will be statistically associated with clinical outcomes, including tinnitus distress scores. This is a data-driven project requiring strong computational skills, with extensive use of MATLAB and python for neuroimaging and network analyses.
Requirements
- Background or strong interest in neuroscience, neuroimaging, or biomedical engineering
- Interest in tinnitus research and clinical neuroscience
- Solid command of MATLAB (required)
- Experience with python for data analysis (preferred)
- Interest in connectomics, network science, or graph-theoretical methods
- Good analytical skills and motivation to work with a large neuroimaging dataset and in an interdisciplinary team
This project is offered in the scope of a larger effort to characterize network-level changes in the tinnitus brain (supported by the American Tinnitus Association).
Contact
Please get in touch with nicolas.gninenko@epfl.ch with your CV & short statement of research interests and I’ll quickly get back to you. Project starting date: Jan/Feb 2026.
References
[1] Gninenko, N. et al. (2024). Functional MRI neurofeedback for reducing tinnitus distress. Radiology.
[2] Finn, E.S. et al. (2015). Functional connectome fingerprinting. Nature Neuroscience.