Interictal Spike detection from EEG with stationary waveform classification using Support Vector Machine (SVM)

This was an original Research aimed at devising a deep learning algorithm to learn and diagnose epileptogenic patterns from an EEG which indicate that a person has had an epileptic seizure.

The research employed analyzing the nonlinear energy operator (NEO) of the stationary waves and then thresholding with a semi constant calculated from the energy of the waves followed by a supervised classification of a 70ms window of the wave.

We obtained around 97% accuracy on testing with real-world data (opensource epileptogenic EEG).

This research has been presented in an IEEE conference (ICEECCOT 2018).

Check out my code on GitHub.