Clinical neurophysiology
Detecting epileptiform activity
In the MSU–Henry Ford Health collaboration, I preprocess and segment EEG and MEG recordings, train models from prior studies, and compare how well they detect interictal epileptiform discharges.
EEG · MEG · Neural networks
Cognitive neuroscience
Measuring attention with EEG
I built a Python pipeline that aligns visual stimuli with eight-channel EEG, reduces eye-movement artifacts, and measures P300 amplitude and latency. Diagnostic plots let me check each session’s signal quality.
Python · MNE · Event-related potentials
Computer vision
Recognizing manipulated faces
I compared features from pretrained vision models for deepfake detection. A two-layer classifier using InternVL3-78B features reached 89.9% accuracy on a balanced test set of 2,000 images.
Read the paper →