Differences in electroencephalographic non-rapid-eye movement sleep slow-wave characteristics between young and old mice

Abstract

Changes in sleep pattern are typical for the normal aging process. However, aged mice show an increase in the amount of sleep, whereas humans show a decrease when aging. Mice are considered an important model in aging studies, and this divergence warrants further investigation. Recently, insights into the network dynamics of cortical activity during sleep were obtained by investigating characteristics of individual electroencephalogram (EEG) slow waves in young and elderly humans. In this study, we investigated, for the first time, the parameters of EEG slow waves, including their incidence, amplitude, duration and slopes, in young (6 months) and older (18-24 months) C57BL/6J mice during undisturbed 24 h, and after a 6-h sleep deprivation (SD). As expected, older mice slept more but, in contrast to humans, absolute NREM sleep EEG slow-wave activity (SWA, spectral power density between 0.5-4 Hz) was higher in the older mice, as compared to the young controls. Furthermore, slow waves in the older mice were characterized by increased amplitude, steeper slopes and fewer multipeak waves, indicating increased synchronization of cortical neurons in aging, opposite to what was found in humans. Our results suggest that older mice, in contrast to elderly humans, live under a high sleep pressure.

Figure 1. Time course of vigilance states and absolute slow-wave activity (SWA, 0.5–4 Hz) in non-rapid-eye movement (NREM) sleep, and episode frequency histograms of vigilance states.

(a) Time course of vigilance states and absolute SWA in NREM sleep, for 24-h baseline, 6-h sleep deprivation (SD, hatched bar) and 18-h recovery for the two age groups, young (black circles, n = 9) and old mice (gray circles, n = 24). Curves connect 2-h values (mean ± SEM) of Waking, NREM, REM sleep and absolute SWA. The black and white bars indicate the light-dark cycle. Horizontal gray lines represent significant differences between young and old across the 48-h period and black (young) and gray (old) asterisks significant differences between recovery and baseline day (unpaired and paired t-tests, p < 0.05 after significant 1-, 2- or 3-way ANOVA, interaction factors ‘age’, ‘age*time of day’ and ‘age*time of day*day’). (b) Episode frequency histograms of Waking (upper), NREM (middle) and REM sleep (lower panel) in the Light (L, left) and Dark (D, right) period during baseline for young (black bars, n = 9) and old mice (gray bars, n = 24) (mean values ± SEM). Episodes are partitioned into ten exponentially increased duration bins from 4 to >1024 s (x axis designates the upper limit of each bin for the following bins: 0–4, 5–8, 9–16, 17–32, 33–64, 65–128, 129–256, 257–512, 513–1024, >1024). Asterisks indicate significant differences between young and old animals (unpaired t-tests, p < 0.05 after significant 2- or 3-way ANOVA, interaction factors ‘age*duration’, ‘age*Light-dark’ and ‘age*Light-dark*duration’).

Figure 2. Spectral distribution of electroencephalogram (EEG) power density (mean values ± SEM) in the frequencies between 0.5 and 25 Hz for young (black, n = 9) and old mice (gray, n = 24) in Waking (top left), non-rapid-eye movement (NREM, top right) and REM sleep (bottom left) computed for pooled values of the 24-h baseline day.

Asterisks indicate significant differences across the frequency bins between the two groups (unpaired t-tests, p < 0.05 after significant 2-way ANOVA, interaction factors ‘age*frequency bins’).

Figure 3. Average electroencephalogram (EEG) slow wave and Sigma amplitude.

(a) Representative EEG trace of a slow wave followed by a spindle. Time point of 0 ms was arbitrarily defined to emphasize the example of the spindle-event that follows a random slow-wave. (b) Representative average EEG slow wave (μV, top graphs) and Sigma amplitude (%, 9–13 Hz, bottom graphs) of one young (left, black) and one old mouse (right, gray). (c) Average Sigma amplitude (%, mean values ± SEM) in the young (black, n = 9) and aged mice (gray, n = 9). The asterisk indicates significant difference between young and old mice for the period of the trough (unpaired t-test, p = 0.0286).

Figure 4. Slow-wave incidence (n/min) and number of slow waves with more than one peak (multipeak waves) during five 2-h time points across the light period of baseline and recovery for young (left, n = 9) and old mice (right, n = 9).

Slow waves were equally subdivided into five amplitude quintiles ranging from 0–100%. (a) The wave incidence (mean values ± SEM) changes significantly in the following time points in all amplitude quintiles in both groups: 0–2 h vs. 10–12 h in baseline, 6–8 h vs. 10–12 h in recovery, 6–8 h in baseline vs. 6–8 h in recovery (paired t-tests, p < 0.05 after significant 2-way ANOVA, interaction factors ‘time of day*amplitude levels’). (b) Multipeak waves are expressed as a percentage of the total number of slow waves (mean values ± SEM). They change significantly in the following time points in all amplitude quintiles in both groups: 0–2 h vs. 10–12 h in baseline, 6–8 h vs. 10–12 h in recovery, 6–8 h in baseline vs. 6–8 h in recovery. Significant differences were found between the groups, with old mice having overall less multipeak waves (paired and unpaired t-tests, p < 0.05 after significant 2-way ANOVA with interaction factors ‘age*time of day’ and ‘age*amplitude levels’).

Figure 5. Relationship between absolute non-rapid-eye movement (NREM) sleep slow-wave activity (SWA, 0.5–4 Hz) and multipeak waves (expressed as a percentage of the total number of slow waves) computed for the six 2-h intervals across the light period (time points 0–2, 2–4, 4–6, 6–8, 8–10, 10–12 h) of baseline for young (black, n = 9) and old mice (gray, n = 9) for the highest-amplitude slow waves (80–100%) (mean values ± SEM).

Dashed lines depict linear regression (R² are significantly different between young and old mice, unpaired t-test p = 0.013).

Figure 6. Slope of the first and second segment of the slow waves during five 2-h time points across the light period of baseline and recovery for young (left, n = 9) and old mice (right, n = 9).

Slow waves were equally subdivided into five amplitude quintiles ranging from 0–100% and slopes (mean values ± SEM) are expressed as % of the 12 h mean. (a) The slope of the first segment (Slope 1) changes significantly in the 0–2 h vs. 10–12 h in baseline, only in the old mice in all amplitude quintiles (paired t-tests, p < 0.05 after significant 2-way ANOVA with interaction factors ‘age*time of day’). (b) The slope of the second segment (Slope 2) changes significantly in the following conditions: 0–2 h vs. 10–12 h in baseline in young mice in the 80–100% quintile, 0–2 h vs. 10–12 h in baseline in old mice in all amplitude quintiles, 6–8 h vs. 10–12 h in recovery in the young mice in the 80–100% quintile, 6–8 h vs. 10–12 h in recovery in the old mice in all amplitude quintiles, 0–2 h in baseline vs. 6–8 h in recovery in the old in the 80–100% quintile and 6–8 h in baseline vs. 6–8 h in recovery in all amplitude quintiles in the young and old mice (paired t-tests, p < 0.05 after significant 2-way ANOVA with interaction factors ‘age*time of day’).