Mapping the electrophysiological marker of sleep depth reveals skill maturation in children and adolescents

Abstract

Electroencephalographically (EEG) recorded slow wave activity (SWA, 1-4.5Hz), reflecting the depth of sleep, is suggested to play a crucial role in synaptic plasticity. Mapping of SWA by means of high-density EEG reveals that cortical regions showing signs of maturational changes (structural and behavioral) during childhood and adolescence exhibit more SWA. Moreover, the maturation of specific skills is predicted by the topographical distribution of SWA. Thus, SWA topography may serve as a promising neuroimaging tool with prognostic potential. Finally, our data suggest that deep sleep SWA in humans is involved in cortical development that optimizes performance.

Fig. 1

Electrode localization and regions of interest (ROIs). (A) Co-registration of electrodes with T1-weighted magnetic resonance images. Standard electrodes are labeled on the right side (RM = right mastoid). (B) Layout of high density EEG electrode net in top view (adapted from the Electrical Geodesics, Inc.). Based on Talairach coordinates (Lancaster et al., 1997; Lancaster et al., 2000), electrodes were assigned to ROIs, which were defined according to Brodmann areas (BA) associated with a skill. The table contains the electrode numbers and BAs for all ROIs (color coded). Remaining electrodes were attributed to closest ROIs (colored open circles, i.e. 5, 12, 38, 121 to ‘cognitive control’; 6, 13, 29, 111 to ‘complex motor’, and 52 to ‘simple motor’; this was the case in n = 11).

Fig. 2

Maturation of sleep slow-wave activity (SWA). (A) SWA distribution for age groups: Top 25% of SWA values are mapped. (B). Regions of interest (ROIs) based on anatomical electrode localization using Brodmann areas (see Fig. 1). (C). SWA topography maturation reflected by the ROI with maximal SWA (SWA maxima located in vision in n = 5 subjects, visuomotor n = 3, simple motor n = 3, complex motor n = 12, language n = 6, cognitive control n = 34). The age of subjects with maximal SWA in the same ROI was averaged (thin vertical lines) and variability is presented as mean ± 2* SE.

Fig. 3

Relationship between SWA maturation index (SWAMI) and skills in ‘complex motor’. SWAMI identified which ROI contained the single electrode with maximal SWA in each subject (see Figs. 1 and 2). SWAMI thus represents the maturational stage of a subject's topographical SWA map. SWAMI predicted complex motor skills (Spearman, R = 0.31, p = 0.02, n = 56, 4.4–25.9 y). Points are color coded for age groups (black = youngest, light blue = oldest, colors refer to same age groups as in Fig. 2). The Spearman correlations between SWAMI and skills for the other ROIs revealed R_visuomotor = 0.12, p_visuomotor = 0.46; R_{simple motor} = 0.27, p_{simple motor} = 0.09; R_language = 0.33, p_language = 0.14; R_{cognitive control} = 0.21, p_{cognitive control} = 0.35, see Fig. S2.

Fig. 4

Maturation of SWA maturation index (SWAMI) and skills in ‘complex motor’. Data were fitted with a double exponential function [f = a*(1 − exp(−b* x))+ c *(1 − exp(−d* x))]. The results show that SWAMI (solid, black) increases before skills (dashed, blue, R_SWAMI = 0.59, p_SWAMI<0.0001, R_skill = 0.82, p_skill<0.0001, both n = 56). For example, when y is fixed at −1, skills are delayed relative to SWAMI by 3.7 years. The dotted line represents the extrapolation of the fit to younger ages. Gray dots represent SWAMI of subjects that were not included in the curve fit (n = 7). Both variables were z-scored across subjects.

Fig. 5

Gray matter (gray, dashed line) and skills (blue, solid line) fitted by a double exponential function [f = a*(1 − exp(−b*x))+c *(1 − exp(−d* x)), both n = 44, R_Skills = 0.65, p_Skills<0.0001, R_{Gray matter} = 0.45, p_{Gray matter} = 0.03]. SWAMI of an extended sample was added (n = 63, black, dotted line and open circles (same fit, R_SWAMI = 0.77, p_SWAMI<0.0001)). All variables were z-scored across subjects, and gray matter values were inverted.