MSc Thesis: Brain age predicition using resting-state electroencephography (EEG)

The electroencephalogram (EEG) is a tool that allows non-invasive monitoring of the brain activity. Thin electrodes placed on the scalp detect tiny electrical charges that result from the activity of brain cells. Given the simple and low-cost setup, in clinical medicine the EEG is used by neurologists to evaluate brain disorders (e.g. epilepsy), to diagnose diseases (e.g. Alzheimer’s disease, narcolepsy), or simply to monitor blood flow in the brain. [1]

Within the context of this thesis, a dataset containing the EEGs of chronic pain patients (chronic backpain, chronic widespread pain, joint pain, and neuropathic pain) and an age-matching healthy control group is provided by the University Hospital rechts der Isar. Previous work on EEG data of chronic pain patients has shown an increased brain connectivity at specific frequencies in frontal brain areas when compared to the healthy control group. [2] Furthermore, longitudinal studies indicate a global brain network change in chronic pain patients. [3] In this thesis we will investigate how brain age is affected by chronic diseases. Self-supervised learning and transfer learning techniques will be leveraged to understand the high-level context of an EEG. The learned EEG representation is then used to predict the brain age. We will analyse the differences in brain age between chronic pain patients and the healthy control group.

Your qualifications:

  • Background in computer science, engineering, mathematics, or similar studies
  • Basic knowledge of signal processing
  • Advanced knowledge of machine learning
  • Advanced programming skills in Python and a common DL framework (PyTorch, Tensorflow, JAX)
  • Independent working style with strong interest in teamwork and methodic research

What we offer:

  • Ability to perform cutting edge research in the field of neuroscience and deep learning
  • Close supervision and access to high-end computer hardware
  • Closely working and collabroting in an inter-disciplinary team of experts in signal processing, deep learning, neuroscience and medicine
  • This project is targeting publication at leading neuroscience journals

How to apply:

Please send us a short e-mail with your CV and grade report to oezguen.turgut@tum.de.

References

[1] Johns Hopkins Medicine. “Electroencephalogram (EEG)". https://www.hopkinsmedicine.org/health/treatment-tests-and-therapies/electroencephalogram-eeg
[2] Dinh et al. “Brain dysfunction in chronic pain patients assessed by resting-state electroencephalography.” Pain 160.12 (2019): 2751. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7195856/
[3] Heitmann et al. “Longitudinal resting-state electroencephalography in patients with chronic pain undergoing interdisciplinary multimodal pain therapy.” Pain 163.9 (2022): e997-e1005. https://pubmed.ncbi.nlm.nih.gov/35050961/

Özgün Turgut
Özgün Turgut
PhD Student

My research interests focus on signal processing using deep learning methods.