Database Restricted Access
Real-Time Segmentation of Burst Suppression Patterns in Critical Care EEG Monitoring
Published: Dec. 7, 2023. Version: 1.0
When using this resource, please cite:
(show more options)
Shafi, M., Moura Junior, V., Gupta, A., Ghanta, M., & Westover, M. B. (2023). Real-Time Segmentation of Burst Suppression Patterns in Critical Care EEG Monitoring (version 1.0). Brain Data Science Platform. https://doi.org/10.60508/yvna-zd74.
Brandon Westover M, Shafi MM, Ching S, Chemali JJ, Purdon PL, Cash SS, Brown EN. Real-time segmentation of burst suppression patterns in critical care EEG monitoring. J Neurosci Methods. 2013 Sep 30;219(1):131-41. doi: 10.1016/j.jneumeth.2013.07.003. Epub 2013 Jul 23. PMID: 23891828; PMCID: PMC3939433.
This dataset is intended to allow reproducibility of the results reported in
the following manuscript: "Real-time segmentation of burst suppression patterns in critical care EEG monitoring. J Neurosci Methods. 2013 Sep 30;219(1):131-41".
Brief abstract: Automatic computational analysis of burst suppression patterns in neurological critically ill patients can provide important clinical information regarding brain state, for example during recovery from hypoxic ischemic encephalopathy and for guiding medical treatment of refractory status epilepticus with pharmacologically induced coma. A robust method is presented for automated segmentation of adult burst suppression ICU EEG patterns, and is validated against 20 manual EEG segmentations encompassing a wide range of different burst suppression depths and patterns. Expert-level automated real-time segmentation of burst suppression EEG data is demonstrated.
Here is a summarized version of the background section with references removed and length reduced by about 20%:
Burst suppression is an EEG pattern of profound brain inactivation characterized by periods of suppressed brain activity alternating with bursts of activity. It occurs spontaneously in neonatal development and in conditions like anoxic brain injury. Burst suppression can also be induced by anesthetics to achieve surgical anesthesia or treat status epilepticus, or by hypothermia during cardiac surgery.
In pharmacologically-induced burst suppression, higher anesthetic concentrations lead to longer suppressions and eventually continuous suppression. In pathological states, longer spontaneous suppressions indicate worse prognosis. The burst suppression ratio (BSR), measuring the percentage of suppressed time, is an important indicator of suppression intensity that medical personnel aim to control.
Despite recommendations for BSR use, current practice involves visual EEG examination and qualitative burst suppression assessment. Continuous EEG monitoring is labor intensive, so an objective, reproducible, validated method to quantify burst suppression depth would be beneficial. The first step is EEG segmentation into suppressions and bursts.
Previous publications describe various methods for automated segmentation, but have mostly focused on neonates and surgical anesthesia rather than adult ICU patients. Validation against human expert segmentation has been limited.
In this work we developed an algorithm that automatically segments adult ICU burst suppression in real-time. Algorithm performance is validated against manual segmentation by two electroencephalographers. Automated segmentation agreement with experts is comparable to inter-expert agreement.
A representative sample of burst-suppression EEG recordings from critically ill neurological patients was identified by retrospective review of clinical EEG reports from all ICU patients who underwent continuous EEG monitoring at the Massachusetts General Hospital between August 2010 and March 2012. From these, we selected the first 20 consecutive EEGs reported to show burst suppression patterns. Recordings were only included for analysis if patients were being treated within an intensive care unit (ICU) at the time of the recording; EEG recordings from surgical procedures were excluded. We restricted selection to recordings no longer that 90 minutes. Given the retrospective nature of the study, we did not control the length of these EEG recordings. Recording durations ranged from 21 to 76 minutes (see Table 1). Each record was manually segmented in its entirety as described below.
All EEGs were recorded using 19 silver/silver chloride electrodes, affixed to the scalp according to the international 10-20 system. Data were recorded at 512 or 256 Hz, using XLTEK clinical EEG equipment (Natus Medical Inc, Oakville, Canada), subsequently downsampled to 200Hz.
Clinical information was gathered from review of written inpatient medical notes, imaging studies and reports, EEG reports, and discharge summaries. Baseline demographic data (age, gender), primary admission diagnosis, and the identity of anesthetic agents administered at the time of EEG recording were noted (Table 1 of the published manuscript). Review of clinical and EEG data was carried out with the approval of the local institutional review board. The study was approved by the Massachusetts General Hospital Human Research Committee.
Clinical expert classification of bursts and suppressions
A common approach to quantitatively distinguishing bursts and suppressions is to define suppressions as EEG segments of less than a threshold voltage value, typically set at between 0.5-20 μV, which last at least 0.5 seconds. However, prescriptive definitions such as these are problematic, because burst suppression in different pathological conditions or induced by different anesthetics vary widely in amplitude and spectral characteristics (see Fig. 1 of the published manuscript). Moreover, such definitions are not used in clinical practice which relies instead on clinician visual pattern recognition. We therefore adopted an empirical approach to defining bursts and suppressions by having two experienced clinical electroencephalographers independently segment all 20 EEG records manually, and defined “definite” burst or suppression epochs as those portions of the EEG so classified by both reviewers. The algorithms developed below are optimized with respect to these consensus EEG epochs.
Expert data segmentation
Two experienced clinical electroencephalographers (MBW, MMS) independently reviewed and each manually performed an exhaustive segmentation of all twenty burst-suppression EEGs using custom viewing and annotation software written in-house, implemented in Matlab (Natus, MA). The reviewers were instructed to mark the beginning and end of all instances of “suppressions”; all remaining EEG segments were classified as non-suppressions, referred to hereafter as “bursts”. As explained above, to simulate actual clinical practice, reviewers were instructed not to apply explicit technical definitions of “bursts” and “suppressions”; rather, reviewers were to segment the each record according to clinical judgment based on experience. Data review and segmentation was conducted while viewing data from the complete array of recording channels displayed in average referential montage.
Assessment Inter-rater agreement
Inter-rater agreement between manual segmentations of the two expert reviewers was examined in three ways. First, for each of the twenty records, percent agreement was calculated as the fraction of each EEG to which the two reviewers assigned the binary burst vs suppression classification; a global percent agreement was also calculated for all 20 records considered as a whole.
Second, we assessed inter-rater agreement using Cohen’s Kappa (κ) values separately for each record and for all records taken as a whole (Carletta 1996). Kappa values were calculated as: κ = (a – c)/(1-c), where 'a' is the fraction of EEG data points classified identically by both reviewers (observed inter-rater agreement), and 'c' is the hypothetical probability of chance agreement obtained by assuming that each rater randomly assigned categories to EEG data points using the observed relative frequencies of each category. (κ is a measure of the degree of inter-rater agreement beyond that theoretically achievable by chance; see Supplemental Material.)
Finally, a histogram was computed for the lengths of EEG segments, measured in seconds, on which the two raters disagreed. Summary statistics were computed for this “discrepancy histogram” including the median, mode, and standard deviation.
Automatic segmentation of burst suppression data
Recursive variance estimation
Our objective was to develop a causal burst suppression classifier, i.e. a classifier that declares each new incoming data point to be part of a burst or suppression based only on past data. We evaluated a simple classification scheme employing a thresholding operation applied to a recursive estimate of the local signal variance. See the published manuscript for details, and the code, provided here: Burst-Suppression-Segmentation
Validation of automated segmentation: inter-rater agreement analysis
Statistical performance of automatic burst suppression segmentations achieved via the adaptive variance thresholding equations with optimized parameters was evaluated by computing global and record-by-record inter-rater agreement and Cohen’s Kappa statistics for the agreement between the automated method and each of the two expert reviewers.
Quantification of Burst Suppression Depth
The burst suppression probability (BSP) is a recently proposed measure of burst suppression depth which, compared with the more 'traditional burst suppression ratio' (BSR), provides a superior tradeoff between signal smoothness and responsiveness to changes in the EEG; (see Supplementary Material of the published manuscript). The BSP represents an estimate of the instantaneous probability that the EEG is in the suppressed state , and can be computed recursively on a sample-by-sample basis in real-time. Examples of the result of passing the binary data output from the segmentation algorithm through the BSP algorithm with parameters optimized for adult ICU EEG data are presented in Figure 5.
Validation of automated segmentation: root mean squared error of BSP
We further validated our segmentation algorithm by comparing the BSP traces computed from the output of our automated segmentation algorithm with those derived from manual segmentations. Signal comparisons were carried out using a root mean-squared error criterion. Performance of the algorithm was assessed by comparing the discrepancy between the two expert’s BSP signals with respect to each other, with the discrepancies between the automated algorithm’s BSP signal and those of each human experts.
Data is provided in MATLAB format (.mat).
- EEG files:
- p1_data.mat, ..., p20_data.mat: 20 EEG files. Each contains:
- Channels 1x19 2056 cell -- names of EEG channels
- Fs 1x1 8 double -- sampling rate (200Hz for all files)
- data 19xNt Nt double (size of EEG data; 19 x timesteps)
- p1_data.mat, ..., p20_data.mat: 20 EEG files. Each contains:
- annotation files from two EEG experts, marking the locations of suppression periods:
- INDS_mo_1_data.mat,...,INDS_mo_20_data.mat, and INDS1_data.mat,...,INDS20_data.mat
- these each contain:
- Channels 1x19 2056 cell - names of EEG channels
- Fs 1x1 8 double - sampling rate (200Hz)
- INDS 1xNt double - 0s and 1s (annotated suppressions)
- data 19xNt double - EEG data
- single channel of EEG and binary annotations from algorithm:
- these each contain:
- s 1xNt double - 1 EEG channel
- z 1xNt logical - 0s and 1s (detected suppressions)
Code for reading the data and reproducing results in the published manuscript are provided at:
Review of clinical and EEG data was carried out with the approval of the local institutional review board. The study was approved by the Massachusetts General Hospital Human Research Committee. All data provided is de-identified.
Conflicts of Interest
- Westover MB, Ching S, Shafi MM, Cash SS, Brown EN. Real-time segmentation and tracking of brain metabolic state in ICU EEG recordings of burst suppression. Annu Int Conf IEEE Eng Med Biol Soc. 2013;2013:7108-11. doi: 10.1109/EMBC.2013.6611196. PMID: 24111383; PMCID: PMC3939432.
- sign the data use agreement for the project