• DocumentCode
    1812990
  • Title

    Auto-regressive analysis of EEG reveals brain´s response to injury

  • Author

    Goel, Vaibhava ; Brambrink, Ansgar M. ; Baykal, Ahmet ; Thakor, Nitish V.

  • Author_Institution
    Dept. of Biomed. Eng., Johns Hopkins Univ. Sch. of Med., Baltimore, MD, USA
  • fYear
    1994
  • fDate
    3-6 Nov 1994
  • Firstpage
    219
  • Abstract
    The authors hypothesize that an insight into brain injury during hypoxia-asphyxia and recovery can be gained by quantifying the changes in frequency characteristics of EEG. Neonatal piglets (1-2 weeks old) are hypoxiated and asphyxiated to simulate severe birth related insult. Continuous Auto-Regressive (AR) analysis of EEG records is done to obtain three dominant frequencies corresponding to three poles in the spectrum. The results show an increase in power in all three frequency components (upto 200% of baseline) during hypoxia, the rate of increase being maximum for the highest frequency component. The power goes down to approximately zero within 40-60 secs. of asphyxia. Recovery in power occurs in a bursting fashion, most clearly elicited in the lowest frequency. This result is consistent with burst-suppression during recovery reported in the literature
  • Keywords
    autoregressive processes; electroencephalography; medical signal processing; 40 to 60 s; EEG frequency characteristics; autoregressive EEG analysis; brain injury response; burst-suppression; continuous autoregressive analysis; dominant frequencies; hypoxia-asphyxia; neonatal piglets; power recovery; severe birth related insult simulation; Asphyxia; Biomedical engineering; Biomedical imaging; Brain injuries; Brain modeling; Electroencephalography; Equations; Frequency; Pediatrics; Sequences;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Engineering in Medicine and Biology Society, 1994. Engineering Advances: New Opportunities for Biomedical Engineers. Proceedings of the 16th Annual International Conference of the IEEE
  • Conference_Location
    Baltimore, MD
  • Print_ISBN
    0-7803-2050-6
  • Type

    conf

  • DOI
    10.1109/IEMBS.1994.411826
  • Filename
    411826