Synthetic EEG with GANs
Undergraduate Researcher · Marmara University · October 2025 – present
EEG data from stroke patients is scarce and valuable. In this work I try to augment it with deep learning.
What I do
- GAN-based data augmentation to generate synthetic EEG signals from stroke-patient recordings (PyTorch).
- Built the preprocessing pipeline in MNE-Python: raw EEG filtering, epoch segmentation and preparation for model training.
Tools
PyTorch · GAN · MNE-Python · NumPy · SciPy