Phase Lag Index Histogram Features for Identifying Epileptic Seizure EEG Signals

We compute the phase lag index between signals from different brain regions and represent its distribution as a histogram. Using short segments taken from longer EEG recordings, we distinguish seizure from non-seizure signals.

Brain activity can be recorded as electroencephalogram signals, or EEG, using electrodes placed on the head. We study how a computer can distinguish periods with epileptic seizures from periods without seizures in long EEG recordings from people with epilepsy.

Instead of examining only the waveform at one location, we look at timing relationships between signals recorded at different locations. One measure of these relationships is called the phase lag index. Our method randomly selects short sections from a long recording and calculates this measure between signals from different brain regions. It then creates a histogram, which records how often different values occur. This summary becomes a clue for distinguishing seizure and non-seizure signals.

Experiments using the publicly available CHB-MIT Scalp EEG database demonstrated that the method could distinguish the two types of signal. The results show the usefulness of summarizing relationships between EEG signals recorded at different locations.

• Masayoshi Oguri, Tetsuya Okazaki, Tohru Okanishi, Masashi Nishiyama, Sotaro Kanai, Hiroyuki Yamada, Kaoru Ogo, Takashi Himoto, Yoshihiro Maegaki, Ayataka Fujimoto,
  Phase Lag Analysis Scalp Electroencephalography May Predict Seizure Frequencies in Patients with Childhood Epilepsy with Centrotemporal Spikes,
  Yonago Acta Medica, Vol. 66, No. 1, pp. 48 - 55, February 2023.
• Masashi Nishiyama, Yoshio Iwai, Masayoshi Oguri, and Yoshihiro Maegaki
  Phase Lag Index Histogram Features for Identifying Epileptic Seizure EEG Signals, Proceedings of IEEE 10th Global Conference on Consumer Electronics (GCCE), OS-HPL, pp. 167 - 170, October 2021.