Subjects/Materials and Methods
EEG samples
EEG samples (n = 37) were collected prospectively between January 1, 2012, and
September 1, 2012. Samples consisted of 12 examples of lateralized periodic discharges
(LPDs), 6 of generalized periodic discharges (GPDs), 3 of bilateral independent periodic
discharges (BIPDs), 4 of lateralized rhythmic delta activity (LRDA), 4 of generalized
rhythmic delta activity (GRDA), 2 of generalized rhythmic spike-and-wave complexes
(GSWs), 1 of lateralized rhythmic spike-and-wave complexes (LSWs), and 5 of seizures.
Samples were obtained from unselected consecutive cases where the pattern of interest was
present. A greater emphasis was placed on periodic discharges, as in our experience they
exhibit the largest amount of variability of morphology and allow most components of the
terminology to be used.
Ten second to 60 s examples were presented as consecutive 10 s pages in a longitudinal
bipolar montage, with the low-pass filter at 70 Hz and the high-pass filter at 1 Hz. None of
the samples contained significant amounts of electrical noise, and thus the notch filter was
left off. Sensitivity was individually set for each example to optimize visualization. Raters
could not change these settings. The duration of each example was chosen to present enough
information for all terms to be applied, although it was expected that for some cases, the
determination of evolution and fluctuation could not be captured in samples of this brief
duration. The durations of nonictal (median 30 s and interquartile range [IQR] 30–40 s) and
ictal samples (median 40 s and IQR 30–50 s) were not significantly different (p = 0.29;
Mann-Whitney
U
-test).
Completing the IRA assessment was part of a certification test that each rater had to take to
participate in studies of the Critical Care EEG Monitoring Research Consortium
(CCEMRC). Before participating, the raters were invited to review a set of training slides
presenting the most recent version of the Critical Care EEG Terminology.
15
These slides can
be obtained on the website of the ACNS (a link is provided at the end of the article). The test
was designed with a Web-based survey engine (SurveyMonkey, Palo Alto, CA, U.S.A.).
Raters were free to complete the test in multiple sittings, and were allowed to refer to any
reference text while taking the test, including the recent version of the terminology and the
training slides. For each sample, they were asked to first decide whether the pattern
represented an unequivocal electrographic seizure (defined in the ACNS terminology as any
pattern composed of generalized spike-and-wave discharges at >3 Hz, or any pattern of
evolving discharges reaching a frequency of >4 Hz). In cases where raters opted to classify
the pattern as a definite seizure, the rest of the questions were skipped and the raters were
brought to the next sample. If they classified the pattern as not a definite seizure, then raters
were asked to describe it using 11 descriptors defined in the terminology (Table 1). We
derived four additional concepts about the plus (+) modifiers, namely “any +”, and the
presence or absence of individual plus modifiers (+F, superimposed fast activity; +R,
superimposed rhythmic delta activity; and +S, superimposed spikes/sharp waves or sharply
contoured delta activity), yielding a total of 15 concepts.
Gaspard et al.
Page 3
Epilepsia
. Author manuscript; available in PMC 2016 May 25.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript