Changes in EEG permutation entropy in the evening and in the transition from wake to sleep

Abstract

Quantifying the complexity of the EEG signal during prolonged wakefulness and during sleep is gaining interest as an additional mean to characterize the mechanisms associated with sleep and wakefulness regulation. Here, we characterized how EEG complexity, as indexed by Multiscale Permutation Entropy (MSPE), changed progressively in the evening prior to light off and during the transition from wakefulness to sleep. We further explored whether MSPE was able to discriminate between wakefulness and sleep around sleep onset and whether MSPE changes were correlated with spectral measures of the EEG related to sleep need during concomitant wakefulness (theta power-Ptheta: 4-8 Hz). To address these questions, we took advantage of large datasets of several hundred of ambulatory EEG recordings of individual of both sexes aged 25-101 years. Results show that MSPE significantly decreases before light off (i.e. before sleep time) and in the transition from wakefulness to sleep onset. Furthermore, MSPE allows for an excellent discrimination between pre-sleep wakefulness and early sleep. Finally, we show that MSPE is correlated with concomitant Ptheta. Yet, the direction of the latter correlation changed from before light-off to the transition to sleep. Given the association between EEG complexity and consciousness, MSPE may track efficiently putative changes in consciousness preceding sleep onset. An MSPE stands as a comprehensive measure that is not limited to a given frequency band and reflects a progressive change brain state associated with sleep and wakefulness regulation. It may be an effective mean to detect when the brain is in a state close to sleep onset.

Keywords: complexity; electroencephalography; entropy; multiscale analysis; sleep need.

Figure 1.

(Color online) Illustration of the MSPE algorithm. (A) The coarse-graining procedure for scale factor of 3. Each black dot represents a data point in the original time series. (B) The ordinal patterns in MSPE calculation with embedding dimension of 4 and time lag of 1. The circle dots in (B) represent the data points in a time series, and the combination of four numbers under a rectangle or a horizontal line stands for an ordinal pattern of the segment in the rectangle or right above the line. The segments in the rectangles have the same pattern [1–4].

Figure 2.

Schematic diagram of the timeline in the analyses. (A) The timeline for the analysis on PhysioNet dataset. MSPE and P_theta were evaluated in three different periods, that is, 2 h pre-light-off, the sleep transition from light-off to sleep onset, and the first sleep cycle. (B) The timeline for the analysis on SHHS dataset. The included subjects must have a sleep latency more than 10 min. MSPE and P_theta were computed over each 30 s epoch within the 10 min immediately preceding and following sleep onset.

Figure 3.

(color online) Variations in MSPE and P_theta values before light off. (A) Average value of PE for all the participants in PhysioNet datasets using different scale factors. For the calculation of PE, the embedding dimension m was set as 3. (B) Average MSPE at m = 3, 4, or 5 at each time bin; shade areas represent the standard errors of the mean. (C) Average P_theta at each time bin; shade areas represent the standard errors of the mean. (D) p-values of the Spearman correlation between MSPE (with m set as 3, 4, or 5) and concomitant P_theta over each time bin.

Figure 4.

The values of MSPE (A) and P_theta (B) during pre-light-off wakefulness, pre-sleep wakefulness and first sleep NREM-REM cycle. Each dot represents the median value of MSPE or P_theta for a participant during the corresponding period. The box-plots illustrate the distribution of these median values for all the participants in PhysioNet dataset. The symbol ‘*’ represents for a significant difference of median values between groups (post hoc tests of ANOVA, p < 0.05).

Figure 5.

(Color online) Variations in MSPE and P_theta values within 10 min immediately before sleep-onset. (A) Average value of PE for all the participants in PhysioNet datasets using different scale factors. For the calculation of PE, the embedding dimension m was set as 3. (B) Average MSPE at m = 3, 4, or 5 at each time bin; shade areas represent the standard errors of the mean. (C) Average P_theta at each time bin; shade areas represent the standard errors of the mean. (D) p-Values of the Spearman correlation between MSPE (with m set as 3, 4, or 5) and concomitant P_theta over each time bin.

Figure 6.

(Color online) The Spearman’s rho and its p-value between MSPE and TBR during (A) the 2 h pre-light-off with PhysioNet dataset and (B) the 10 min before sleep-onset with SHHS dataset. Here, m was set as 3 in the calculation of MSPE and TBR represents the ratio of EEG power in theta band and beta band.

Figure 7.

(Color online) The AUC values to differentiate states before and after sleep-onset. (A) AUC values of PE obtained with different scale factors and different embedding dimensions (3, 4, or 5). AUC values above the dashed line correspond to an excellent ability of the test. (B) ROC curves of P_theta and MSPE obtained with m = 3, 4, or 5.