Limits...
Assessment of human sleep depth is being de-standardized by recently advised EEG electrode locations.

Kemp B, van Someren P, Roessen M, van Dijk JG - PLoS ONE (2013)

Bottom Line: These might well differ in slow wave amplitudes which would bias the scorings.The resulting de-standardization most likely affects any patient group because the amplitude differences were consistent over diagnoses, genders, and age.Derivation-dependent amplitude thresholds were proposed to reduce the scoring bias.

View Article: PubMed Central - PubMed

Affiliation: Department of Neurology, Leiden University Medical Centre, Leiden, The Netherlands. bk@hsr.nl

ABSTRACT
Human sleep depth was traditionally assessed by scoring electro-encephalographic slow-wave amplitudes at the globally standardized C4-M1 electrode derivation. Since 2007, the American Association of Sleep Medicine (AASM) has accepted three additional derivations for the same purpose. These might well differ in slow wave amplitudes which would bias the scorings. Some derivations might also introduce large inter-individual variability. We compared mean and variability of slow wave amplitudes between six derivations including the four AASM ones. Slow wave amplitudes in those derivations were simultaneously measured using automated analysis in 29 patients. Each amplitude was divided by the average from the six derivations, thus removing shared factors such as age, gender and sleep depth while retaining factors that differ between the derivations such as caused by local skull characteristics, electrode distance and neuronal dipole orientation. The remaining inter-individual variability differed significantly and up to a factor of two between the AASM derivations. The amplitudes differed significantly and up to 60% between the AASM derivations, causing substantial scoring bias between centres using different derivations. The resulting de-standardization most likely affects any patient group because the amplitude differences were consistent over diagnoses, genders, and age. Derivation-dependent amplitude thresholds were proposed to reduce the scoring bias. However, it would be better to settle on just one derivation, for instance Cz-Oz or Fpz-Cz because these have lowest variability while matching the traditional C4-M1 amplitudes.

Show MeSH

Related in: MedlinePlus

Compared EEG derivations.Locations of electrodes involved in the studied derivations (left) with 30 seconds of EEG recorded in +/−100 µV panes (right). Top: top view of the head (nose pointing downwards) with Cz-Oz, Fz-Cz and Fpz-Cz EEG derivations. Bottom: front view of the head with C4-M1, F4-M1 and Fpz-E1 EEG derivations. Note the difference in average amplitude between the derivations. Note also that the large waves at 7 and 23 seconds peak at different locations.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC3733975&req=5

pone-0071234-g001: Compared EEG derivations.Locations of electrodes involved in the studied derivations (left) with 30 seconds of EEG recorded in +/−100 µV panes (right). Top: top view of the head (nose pointing downwards) with Cz-Oz, Fz-Cz and Fpz-Cz EEG derivations. Bottom: front view of the head with C4-M1, F4-M1 and Fpz-E1 EEG derivations. Note the difference in average amplitude between the derivations. Note also that the large waves at 7 and 23 seconds peak at different locations.

Mentions: An Embla Titanium recorder simultaneously recorded EEG's from F4, C4, Fpz, Fz, Cz, Oz, E1 and M1 electrodes, each one referred to the Pz electrode. From these eight EEG's, the F4-M1, Fz-Cz, Fpz-E1, C4-M1, Cz-Oz and Fpz-Cz EEG derivations (Fig. 1) were constructed off-line by subtraction. The frequency bandwidth was 0.15–120 Hz and the EEG was digitized at 256 Hz. Sleep stages were scored according to version 1 of the AASM manual [1]. All recordings were included that showed at least two periods of deep NonREM sleep (stage N3, based on C4-M1), each period lasting at least five minutes and being free of artifacts or intrusion from other stages. In order to possibly capture some variations caused by sleep physiology and also to insure that no overlapping periods could be selected, the two periods had to be separated by REM sleep.


Assessment of human sleep depth is being de-standardized by recently advised EEG electrode locations.

Kemp B, van Someren P, Roessen M, van Dijk JG - PLoS ONE (2013)

Compared EEG derivations.Locations of electrodes involved in the studied derivations (left) with 30 seconds of EEG recorded in +/−100 µV panes (right). Top: top view of the head (nose pointing downwards) with Cz-Oz, Fz-Cz and Fpz-Cz EEG derivations. Bottom: front view of the head with C4-M1, F4-M1 and Fpz-E1 EEG derivations. Note the difference in average amplitude between the derivations. Note also that the large waves at 7 and 23 seconds peak at different locations.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC3733975&req=5

pone-0071234-g001: Compared EEG derivations.Locations of electrodes involved in the studied derivations (left) with 30 seconds of EEG recorded in +/−100 µV panes (right). Top: top view of the head (nose pointing downwards) with Cz-Oz, Fz-Cz and Fpz-Cz EEG derivations. Bottom: front view of the head with C4-M1, F4-M1 and Fpz-E1 EEG derivations. Note the difference in average amplitude between the derivations. Note also that the large waves at 7 and 23 seconds peak at different locations.
Mentions: An Embla Titanium recorder simultaneously recorded EEG's from F4, C4, Fpz, Fz, Cz, Oz, E1 and M1 electrodes, each one referred to the Pz electrode. From these eight EEG's, the F4-M1, Fz-Cz, Fpz-E1, C4-M1, Cz-Oz and Fpz-Cz EEG derivations (Fig. 1) were constructed off-line by subtraction. The frequency bandwidth was 0.15–120 Hz and the EEG was digitized at 256 Hz. Sleep stages were scored according to version 1 of the AASM manual [1]. All recordings were included that showed at least two periods of deep NonREM sleep (stage N3, based on C4-M1), each period lasting at least five minutes and being free of artifacts or intrusion from other stages. In order to possibly capture some variations caused by sleep physiology and also to insure that no overlapping periods could be selected, the two periods had to be separated by REM sleep.

Bottom Line: These might well differ in slow wave amplitudes which would bias the scorings.The resulting de-standardization most likely affects any patient group because the amplitude differences were consistent over diagnoses, genders, and age.Derivation-dependent amplitude thresholds were proposed to reduce the scoring bias.

View Article: PubMed Central - PubMed

Affiliation: Department of Neurology, Leiden University Medical Centre, Leiden, The Netherlands. bk@hsr.nl

ABSTRACT
Human sleep depth was traditionally assessed by scoring electro-encephalographic slow-wave amplitudes at the globally standardized C4-M1 electrode derivation. Since 2007, the American Association of Sleep Medicine (AASM) has accepted three additional derivations for the same purpose. These might well differ in slow wave amplitudes which would bias the scorings. Some derivations might also introduce large inter-individual variability. We compared mean and variability of slow wave amplitudes between six derivations including the four AASM ones. Slow wave amplitudes in those derivations were simultaneously measured using automated analysis in 29 patients. Each amplitude was divided by the average from the six derivations, thus removing shared factors such as age, gender and sleep depth while retaining factors that differ between the derivations such as caused by local skull characteristics, electrode distance and neuronal dipole orientation. The remaining inter-individual variability differed significantly and up to a factor of two between the AASM derivations. The amplitudes differed significantly and up to 60% between the AASM derivations, causing substantial scoring bias between centres using different derivations. The resulting de-standardization most likely affects any patient group because the amplitude differences were consistent over diagnoses, genders, and age. Derivation-dependent amplitude thresholds were proposed to reduce the scoring bias. However, it would be better to settle on just one derivation, for instance Cz-Oz or Fpz-Cz because these have lowest variability while matching the traditional C4-M1 amplitudes.

Show MeSH
Related in: MedlinePlus