Real-time segmentation and tracking of brain metabolic state in ICU EEG recordings of burst suppression

M. B. Westover, S. Ching, M. M. Shafi, S. S. Cash, E. N. Brown

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review

1 Scopus citations


Introduction: Burst suppression – a discontinuous electroencephalographic (EEG) pattern in which flatline (suppression) and higher voltage (burst) periods alternate systematically but with variable burst and suppression durations (see Figure 14.1) – is a state of profound brain inactivation. Burst suppression is inducible by high doses of most anesthetics (Clark & Rosner 1973) or in profound hypothermia (e.g. used for cerebral protection in cardiac bypass surgeries) (Stecker et al. 2001); may occur pathologically in patients with coma after cardiac arrest or trauma as a manifestation of diffuse cortical hypoxicischemic injury (Young 2000), or in a form of early infantile encephalopathy (“Othahara syndrome”) (Ohtahara & Yamatogi 2006); and as a non-pathological finding in the EEGs of premature infants known as “trace alternant” or “trace discontinu.” The fact that these diverse etiologies produce similar brain activity have led to the current consensus view that (i) burst suppression reflects the operation of a low-order dynamic process which persists in the absence of higher-level brain activity, and (ii) there may be a common pathway to the state of brain inactivation. Four cardinal phenomenological features of burst suppression have been established through a variety of EEG and neurophysiological studies (Akrawi et al. 1996; Amzica 2009; Ching et al. 2012). First, burst onsets are generally spatially synchronous (i.e., bursts begin and end nearly simultaneously across the entire scalp), except in cases of large-scale cortical deafferentation (Niedermeyer 2009), in which cases regional differences in blood supply and autoregulation may prevent the uniformity typically associated with burst suppression. A caveat here is related to recent evidence that suggests that, on a local circuit level, the onset of bursts may exhibit significant heterogeneity (Lewis et al. 2013). Second, the fraction of time spent in suppression– classically quantified using the burst suppression ratio (BSR) – increases monotonically with the level of brain inactivation. For example, the BSR increases with increasing doses of anesthetic or hypothermia, eventually reaching 100% as the EEG becomes isoelectric (flatline).

Original languageEnglish
Title of host publicationAdvanced State Space Methods for Neural and Clinical Data
PublisherCambridge University Press
Number of pages15
ISBN (Electronic)9781139941433
ISBN (Print)9781107079199
StatePublished - Jan 1 2015


Dive into the research topics of 'Real-time segmentation and tracking of brain metabolic state in ICU EEG recordings of burst suppression'. Together they form a unique fingerprint.

Cite this