Computerized seizure detection of complex partial seizures

Anthony M Murro, Don W. King, Joseph R. Smith, Brian B. Gallagher, Herman F. Flanigin, Kimford Meador

Research output: Contribution to journalArticlepeer-review

83 Scopus citations


In this study, we describe a computerized method that uses 3 quantified EEG features and discriminant analysis to automatically detect seizure EEG. The quantified EEG features were relative amplitude, dominant frequency and rhythmicity. Using EEGs recorded from intracranial electrodes, the seizure detection method was applied to consecutive non-overlapping 2-channel EEG epochs. A seizure detection, sensitivity, ranging from 90% to 100%, was associated with a false positive detection rate of 1.5-2.5/h. The performance of the seizure detection method remained stable for EEG recorded over variable time periods.

Original languageEnglish (US)
Pages (from-to)330-333
Number of pages4
JournalElectroencephalography and Clinical Neurophysiology
Issue number4
StatePublished - Jan 1 1991


  • Discriminant analysis
  • Electroencephalography
  • Epilepsy
  • Fourier analysis
  • Seizure
  • Temporal lobe

ASJC Scopus subject areas

  • Clinical Neurology
  • Neuroscience(all)


Dive into the research topics of 'Computerized seizure detection of complex partial seizures'. Together they form a unique fingerprint.

Cite this