/seizures_recognition

Primary LanguageJupyter Notebook

Epileptic Seizure Recognition

Use various time series analysis techniques to detect epileptic seizures at electroencephalogram readings.

Data

We use a dataset from here.

To download it, just type:

python download.py

The script will save 5 .npy files to data/. These are matrices whose rows are single-channel EEG segments, each one having 23.6-second duration. The matrices correspond to 5 subsets of the dataset:

  • The sets are denoted as Z, O, N, F, and S.
  • Z - healthy volunteers, surface EEG recordings, eyes open.
  • O - healthy volunteers, surface EEG recordings, eyes closed.
  • F - patients, seizure-free intervals, epileptogenic zone.
  • N - patients, seizure-free intervals, hippocampal formation of the opposite hemisphere of the brain.
  • S - patients, seizure activity.