Determining sleep quality in children with sleep disordered breathing: EEG spectral analysis compared with conventional polysomnography

Abstract

Study objectives: To identify the extent of sleep disruption in children with various severities of sleep disordered breathing (SDB) using both conventional visually scored assessment of sleep stages and arousal indices together with EEG power spectral analysis.

Design: Sleep stages and power spectral analysis of the sleep EEG in children with varying severities of SDB with matched control subjects with no history of snoring were compared across the whole night, across sequential hours from sleep onset, and across sleep stages.

Measurements: Overnight polysomnography was performed on 90 children (49M/41F) aged 7-12 y with SDB and 30 age-matched healthy controls (13M/17F). Sleep stages were visually scored and the EEG spectra were analyzed in 5-s epochs.

Results: Conventional visual scoring indicated that, although sleep duration was reduced in severely affected children, sleep quality during the essential stages of SWS and REM was preserved, as evidenced by the lack of any significant decrease in their duration in SDB severity groups. This finding was supported by the lack of substantial differences in EEG spectral power between the groups over the whole night, within specific hours, and in individual sleep stages.

Conclusions: Both conventional scoring and EEG spectral analysis indicated only minor disruptions to sleep quality in children with SDB when assessed across the night, in any specific hour of the night, or in any specific sleep stage. These results suggest that reduced daytime functioning previously reported in children with SDB may not be due to sleep disruption. We speculate that in children, in contrast to adults, a stronger sleep drive may preserve sleep quality even in severe SDB.

Figure 1A-E

Proportional spectral power distribution across the first 6 hours from clinically defined sleep onset, with power in each hour being expressed as a percentage of the sum of the 6 hours within each frequency band. (A) delta, (B) theta, (C) alpha, (D) sigma, (E) beta. □ Controls PS Mild OSA ▪ Mod/Sev OSA; *P < 0.05 between groups. (Statistical differences between Hours are not shown on graph)

Figure 2A-E

Proportional spectral power distribution in each of the five sleep stages as well as in WASO periods as a percentage of (A) delta, (B) theta, (C) alpha, (D) sigma, (E) beta. □ Controls PS Mild OSA ▪ Mod/Sev OSA; *P < 0.05 between groups. (Statistical differences between Stages are not shown on graph)