SEARCH

SEARCH BY CITATION

Keywords:

  • electroencephalogram;
  • memory consolidation;
  • sigma activity;
  • sleep;
  • sleep spindles;
  • slow-wave activity

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Conflicts of Interest
  8. References
  9. Supporting Information

Previous studies suggest that sleep-specific brain activity patterns such as sleep spindles and electroencephalographic slow-wave activity contribute to the consolidation of novel memories. The generation of both sleep spindles and slow-wave activity relies on synchronized oscillations in a thalamo-cortical network that might be implicated in synaptic strengthening (spindles) and downscaling (slow-wave activity) during sleep. This study further examined the association between electroencephalographic power during non-rapid eye movement sleep in the spindle (sigma, 12–16 Hz) and slow-wave frequency range (0.1–3.5 Hz) and overnight memory consolidation in 20 healthy subjects (10 men, 27.1 ± 4.6 years). We found that both electroencephalographic sigma power and slow-wave activity were positively correlated with the pre–post-sleep consolidation of declarative (word list) and procedural (mirror-tracing) memories. These results, although only correlative in nature, are consistent with the view that processes of synaptic strengthening (sleep spindles) and synaptic downscaling (slow-wave activity) might act in concert to promote synaptic plasticity and the consolidation of both declarative and procedural memories during sleep.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Conflicts of Interest
  8. References
  9. Supporting Information

Evidence from molecular to behavioural studies indicates that sleep fosters learning, memory and underlying neural plasticity (Diekelmann and Born, 2010). However, the contribution of the predominant sleep stages, rapid eye movement (REM) and non-REM (NREM) sleep, and of associated neural processes to the consolidation of newly encoded memories in the major memory systems, declarative and non-declarative memory, remains to be further elucidated.

Recent findings suggest that sleep spindles, brief but powerful bursts of synchronous neuronal firing during mammalian stage 2 sleep in the electroencephalogram (EEG) sigma range (12–16 Hz), and EEG slow-wave activity (SWA), highly-synchronized oscillations during NREM sleep (0.1–3.5 Hz), might represent oscillatory activity that promotes sleep-related synaptic plasticity (Rosanova and Ulrich, 2005; Tononi and Cirelli, 2006).

Animal studies indicate that sleep spindles might foster intracellular Ca2+-dependent mechanisms leading to the consolidation of new memories through synaptic strengthening (Steriade, 2003). In contrast, SWA and underlying ‘down-states’ of extensive hyperpolarization and neuronal silencing and subsequent ‘up-states’ of depolarization and intense neuronal firing (Steriade, 2006) might sharpen the information-to-noise ratio and novel memory representations through synaptic downscaling (Tononi and Cirelli, 2006).

Earlier work proposed that NREM sleep might preferentially promote declarative memory, whereas REM sleep might predominantly facilitate non-declarative learning (Plihal and Born, 1997). Consistently, NREM sleep-related sleep spindle density has been shown to be enhanced during nocturnal sleep following training on a declarative word-pair association task (Gais et al., 2002). Additionally, the increase in sleep spindles was correlated with the overnight retention on a declarative word-pair association task (Schabus et al., 2004). In contrast, Rasch et al. (2009) observed a positive correlation between performance gains on a procedural finger-tapping task and spindle activity in a night with pharmacologically suppressed REM sleep. However, it is not clear whether this correlation was driven by implicit or explicit components of this task. Furthermore, Fogel et al. (2007) showed an increase in EEG sigma power after training on a simple procedural pursuit task that primarily involves implicit motor skill learning. Thus, to date, it is not clear whether sleep spindles preferentially enhance declarative or non-declarative memories, or both (Fogel and Smith, 2011).

Other studies have focused on SWA. A longstanding line of research indicates that SWA represents a homeostatic marker that increases as a function of prior waking time and declines across subsequent NREM sleep (Borbely, 1982). More recently, Tononi and Cirelli (2006) proposed that slow-wave homeostasis reflects net synaptic strength. This hypothesis is supported by a study showing a correlation between the local increase in SWA over task-related cortical areas during NREM sleep after training and the post-sleep improvement on a rotation adaptation task (Huber et al., 2004). Complementary, Huber et al. (2006) demonstrated attenuation in SWA over sensorimotor cortical areas after a period of immobilization of the contralateral arm. Evidence for a causal function of SWA came from a seminal study showing that the induction of slow oscillations during NREM sleep by transcranial direct current stimulation enhanced the retention of newly encoded declarative memories (Marshall et al., 2006). To date, it is a subject of debate whether neural mechanisms related to SWA promote memory through an active strengthening (Diekelmann and Born, 2010) or synaptic downscaling (Tononi and Cirelli, 2006), and to what extent SWA contributes to which types of memories.

The aim of this study was to further elucidate the impact of EEG spindle band (sigma) activity and SWA on memory consolidation during sleep. Specifically, we tested the hypothesis that increased EEG sigma and SWA during NREM sleep would represent oscillatory patterns that provide favourable conditions for the consolidation of both newly acquired declarative and non-declarative memories in healthy humans.

Materials and Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Conflicts of Interest
  8. References
  9. Supporting Information

Subjects

Twenty healthy volunteers (10 men and 10 women) aged 21–37 years (27.1 ± 4.6 years), all right-handed, participated in the study. All participants underwent an extensive physical examination, including an electrocardiogram (ECG), EEG and a routine laboratory screening to rule out any somatic disorder. In addition, a urine drug screening demonstrated that all participants were free of any benzodiazepines, barbiturates, amphetamines or opiates. All subjects underwent a thorough psychiatric evaluation by an experienced psychiatrist, including a Structured Clinical Interview for Diagnosis (SCID) and an evaluation of the family history. Subjects with a personal or family history of psychiatric disorders or primary sleep disorders according to DSM-IV were excluded. All subjects were completely free of any medication, did not consume any alcohol or caffeine during the period of the study, and were non-smokers. A sleep diary ensured that subjects’ typical sleep schedules approximated the imposed sleep schedule in the laboratory. All participants were informed in detail about the study and provided their written informed consent prior to the onset. The study had been approved by the local ethics committee and has been carried out in accordance with the Declaration of Helsinki.

Experimental design

The present investigation was part of a study on the effects of cholinomimetics on REM sleep regulation (Nissen et al., 2006). For the present analysis, only the drug-free nights in the sleep laboratory were retained. Of the 20 subjects in the study, one subject had to be excluded from the main analysis because this subject was a statistical outlier on EEG spectral sigma power (> 3 SDs greater than the mean of the other subjects). To address the problem of reporting outlying values, we report the results with and without the outlier in the Results. The experimental night (23:00–07:00 hours) followed a first adaptation and screening night in the sleep laboratory. Two memory tasks were performed at 21:30 hours prior to the experimental night (encoding), and on the following morning between 07:30 and 08:00 hours (recall), that is, 30–60 min after ‘lights-on’ at 07:00 hours. The order of the tasks was counterbalanced across subjects to control for sequence effects.

Declarative word-list task

The word-list of 15 items was taken from the German version of the Rey Auditory Verbal Learning Test (RAVLT; Rey, 1964). In the evening, the test list of 15 words was read aloud to the subjects in five trials according to the standard procedures outlined earlier (Lezak, 1995). Each trial involved reading of the test list followed by free recall in any order. For each trial, the number of correctly retrieved words was noted. Not exactly reported or associated words were regarded as incorrect. On the following morning, subjects were asked to recall as many words as possible from the evening list without further learning in one single trial. In contrast to the standard procedure of the RAVLT, no interference list and no recognition condition were used. As outcome measures for acquisition in the evening, the number of correctly retrieved items in the last trial was assessed. To measure overnight memory consolidation, the retention rate (%) of correctly retrieved words in the morning trial referred to the last evening trial was calculated.

Mirror-tracing task

The mirror-tracing task used in the present investigation, including the line-drawn stimuli, was identical with the one described previously (Plihal and Born, 1997). In this task, subjects were required to trace different line-drawn stimuli using a stylus with an electronic light sensor that measures: (i) draw time; (ii) number of errors; and (iii) the mean error time collapsed across all six figures. Visual access to the stimuli was provided only indirectly via a mirror. During the learning condition in the evening, mirror-tracing a star was repeated until the participant reached the criterion of a maximum of six errors. Then, six line-drawn figures were presented consecutively. During the retrieval condition in the morning, subjects were asked to trace the star one single time to keep conditions comparable, followed by the six figures. Performance in the learning conditions was assessed by measuring: (i) the number of trials to reach the criterion for tracing the star; (ii) the mean draw time; (iii) the mean total error count; and (iv) the mean error time. During the retrieval period in the following morning, the same measures were assessed again. Overnight improvement was expressed as the performance change in: (i) mean draw time; (ii) mean error count; and (iii) mean error time from evening to morning as a percentage (%).

Sleep recordings

Polysomnography was recorded from 23:00 hours to 07:00 hours. Sleep recordings were scored by experienced raters based on standard criteria (Rechtschaffen and Kales, 1968). The setup included the EEG electrodes C3-A2 and C4-A1, submental electromyogram, vertical and horizontal electrooculogram and ECG. The following variables of sleep continuity and architecture were assessed: sleep latency (defined as the period between bedtime until sleep onset); sleep period time (defined as the period between sleep onset and final awakening); sleep efficiency (defined as the percentage of total sleep time referred to the time in bed); waking after sleep onset (defined as the time spent awake after sleep onset); as well as the time spent in sleep stages 1, 2, slow-wave sleep (combined stages 3 and 4) and REM sleep (as percentage of sleep period time).

An all-night spectral power analysis was performed on the C3-A2 derivation in 30-s epochs for which sleep stages had been determined. Data were recorded with a sampling rate of 200 Hz and a resolution of 16 bits. Signals were recorded with a time constant of 0.3 s and low-pass-filtered at 70 Hz. Spectral estimates for each epoch were obtained as the average of 22 overlapping FFT windows (512 data points, 2.56 s) covering a 30-s epoch to obtain the spectral power within that epoch, resulting in a spectral resolution of 0.39 Hz. A Welch taper was applied to each FFT window after demeaning and detrending the data in that window. The spectral power values were then log-transformed (base e) and continuously stored on disk. All subsequent steps including the statistical analysis were performed on these logarithmic values, which have more symmetrically distributed errors than raw spectral power. Artefact rejection used an automatic method discarding epochs due to abnormal total- or gamma-band power values relative to a 10-min moving window (Feige et al., 1999). The log spectra of the remaining epochs were averaged across all NREM sleep epochs. Spectral band power was calculated for the following theoretical frequency ranges: delta (0.1–3.5 Hz); theta (3.5–8 Hz); alpha (8–12 Hz); sigma (12–16 Hz); slow sigma (12–14 Hz); fast sigma (14–16 Hz); slow beta (16–20 Hz); fast beta (20–24 Hz); and gamma (24–50 Hz). Spectral band power between lower and higher boundaries was computed by adding all frequency bins with a nominal (centre) frequency f, for which lower ≤ f < higher. In most studies, analyses using absolute and relative spectral power show similar results (Krystal et al., 2002). In this study we used absolute spectral power for our primary analyses because we were interested in the broadest possible understanding of the association of EEG spectral power and memory consolidation. As a well-established measure for sleep spindle activity, we used EEG sigma activity in the frequency range of 12–16 Hz (Dijk et al., 1993). Additionally, we analysed 12–14 Hz slow sigma (spindle) and fast 14–16 Hz sigma (spindle) activity (Fogel and Smith, 2011). Because spindles are a hallmark of NREM sleep stage 2, we additionally calculated sigma spectral power in stage 2 separately.

Statistical analysis

Data were analysed using the Statistical Package for Social Sciences (SPSS) version 18.0 (SPSS, Chicago, IL, USA). Descriptive presentation of the data includes mean values and standard deviations (SDs). To test correlational hypotheses, Pearson’s correlation coefficients were calculated. The level of significance was set at < 0.05.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Conflicts of Interest
  8. References
  9. Supporting Information

Memory performance in the evening and morning

Parameters of memory performance in the evening (learning) and morning (retrieval) are listed in Table 1. Correlational analyses revealed no significant association between baseline measures before sleep in both memory tasks and EEG activity in the sigma or slow-wave frequency range (> 0.1).

Table 1.   Memory performance in the evening and morning
 Memory parameterMean value
  1. Memory parameters represent means ± SD.

Word-list task
 EveningCorrectly retrieved words13.8 ± 1.3
 MorningCorrectly retrieved words11.8 ± 2.6
Overnight retention rate (%)84.7 ± 13.9
Mirror-tracing task
 EveningDraw time (s)65.2 ± 25.2
Error count (n)7.4 ± 6.6
Error time (s)5.7 ± 4.8
 MorningDraw time (s)50.8 ± 17.5
Error count (n)3.7 ± 3.5
Error time (s)2.3 ± 2.4
Overnight improvement draw time (%)19.7 ± 12.6

Overnight memory consolidation

Our primary analysis on overnight memory consolidation and EEG sigma and SWA revealed significant positive correlations for sigma activity and SWA with both declarative and procedural memory consolidation (Table 2).

Table 2.   EEG slow-wave and sigma activity of the experimental night, and correlations with the overnight consolidation of procedural and declarative memories
 Correlation with overnight improvement in MT draw time (%)Correlation with overnight word-list retention rate (%)
EEG spectral power (log)  r P r P
  1. Sleep parameters represent means ± SD.

  2. EEG, electroencephalogram; MT, mirror-tracing; SWA, slow-wave activity.

  3. Significant results are given in bold.

SWA (delta)5.4 ± 0.5 0.462 0.023 0.448 0.027
Theta3.2 ± 0.4 0.439 0.030 0.3330.082
Alpha2.1 ± 0.50.3690.0600.3180.092
Sigma1.7 ± 0.4 0.448 0.027 0.397 0.046
 Stage 2 sigma1.8 ± 0.4 0.470 0.021 0.411 0.040
Slow sigma1.4 ± 0.4 0.509 0.013 0.445 0.028
 Stage 2 slow sigma1.5 ± 0.5 0.540 0.009 0.465 0.022
Fast sigma0.3 ± 0.40.0340.4450.1290.299
 Stage 2 fast sigma0.3 ± 0.40.0480.4220.1320.295
Slow beta0.1 ± 0.30.2190.1840.1470.274
Fast beta−0.9 ± 0.30.0070.4890.0390.438
Gamma−1.8 ± 0.3−0.2390.162−0.0620.400

Subsequent analyses revealed that the association between sigma activity and memory consolidation was exclusively driven by a significant correlation with slow sigma activity in the frequency range of 12–14 Hz, whereas no correlation with fast sigma activity in the frequency range of 14–16 Hz was observed (Table 2). The association between sigma power and overnight memory consolidation was slightly more pronounced for sigma power during stage 2 sleep.

Fig. 1 depicts the main findings of the significant correlations between EEG slow sigma activity and SWA and the overnight memory consolidation on the word-list and mirror-tracing task.1

image

Figure 1.  Relationship between slow electroencephalographic (EEG) sigma (a) and SWA (b) and overnight declarative memory consolidation expressed as correctly retrieved words in the morning trial referred to the last evening trial (%). Relationship between EEG slow sigma (c) and SWA (d) and overnight procedural memory consolidation expressed as the % improvement in mirror-tracing draw time from the evening to the morning session. EEG SWA values refer to NREM sleep, EEG sigma values to stage 2 sleep. As depicted, significant positive correlations between both EEG sigma and SWA and overnight consolidation of declarative and procedural memories were observed. Solid and dashed lines depict the linear regression result with 95% confidence intervals.

Download figure to PowerPoint

To further explore the origin of the effect between EEG slow sigma activity (stage 2) and SWA and memory consolidation, we calculated the same correlations for subsequent NREM sleep periods of each NREM–REM sleep cycle separately. For declarative memory consolidation, we observed a robust association with SWA in the first NREM sleep period (= 0.559, = 0.006), whereas the association with slow sigma activity tended to be weaker in the first period (= 0.405, = 0.043). In contrast, during the second NREM sleep period, the association with SWA tended to be weaker (= 0.418, = 0.042), whereas the association with slow sigma activity tended to be stronger (= 0.485, = 0.021) than in the first NREM sleep period. For procedural memory consolidation, the correlations with SWA appeared to be relatively stable across the first (= 0.439, = 0.016) and second NREM sleep period (= 0.496, = 0.018), while slow sigma activity showed a tendency towards a stronger association in the second versus the first NREM sleep period (= 0.679, = 0.001 versus = 0.511, = 0.013). Statistical comparisons between the correlation coefficients of NREM sleep period 1 and 2 did not yield statistically significant differences (> 0.2). Analyses with later NREM sleep periods did not show any significant correlations with measures of memory consolidation (> 0.1 for all analyses), presumably due to limited statistical power related to a decreasing number of subjects reaching > 2 NREM sleep periods.

Fig. 2 depicts the prominent associations between SWA in the first NREM sleep period and slow sigma activity in the second NREM sleep period with both declarative and procedural memory consolidation.

image

Figure 2.  Relationship between electroencephalographic (EEG) SWA in the first non-rapid eye movement (NREM) episode with (a) overnight declarative memory consolidation expressed as correctly retrieved words in the morning trial referred to the last evening trial (%) and (b) overnight procedural memory consolidation expressed as the % improvement in mirror-tracing draw time from the evening to the morning session. Relationship between EEG slow sigma activity for the second NREM episode with (c) declarative memory consolidation and (d) procedural memory consolidation. EEG slow sigma values refer to stage 2 sleep. As depicted, significant positive correlations between first NREM episode EEG SWA and second NREM episode slow sigma activity with overnight consolidation of declarative and procedural memories were observed. Solid and dashed lines depict the linear regression result with 95% confidence intervals.

Download figure to PowerPoint

To control for potential effects of age on our results, we first correlated all spectral parameters with age, yielding no significant association in the present sample (> 0.05). However, there was a non-significant negative correlation between SWA and age (= −0.332, = 0.082), which is in line with a developmental decline of SWA. Therefore, we repeated all analyses using partial correlations with age as a control variable. Importantly, the association between slow sigma and SWA and overnight memory consolidation remained significant when correcting for age (< 0.05). These findings indicate that the reported results are not sufficiently explained by the effects of age.

To investigate the relationship between EEG SWA and sigma activity, we tested the association between SWA and sigma activity in the slow (12–14 Hz) and fast (14–16 Hz) frequency range. Importantly, there was no significant correlation between the main spectral outcome parameters SWA and slow sigma activity – neither across the whole night (r = 0.332, = 0.165) nor for each NREM–REM sleep cycle separately (> 0.05). We found a positive correlation between SWA and fast sigma activity (r = 0.497, P = 0.030).

Additional exploratory analyses for spectral parameters in the other EEG frequency bands revealed no significant correlation with overnight memory consolidation, except for a positive correlation between theta power and overnight improvement in mirror-tracing (= 0.439, = 0.030; Table 1). With regard to polysomnography, selectively stage 1 sleep (% SPT) was negatively correlated with the overnight recall rate of the word-list task. Table 3 lists the polysomnographic parameters for the experimental night.

Table 3.   Polysomnographic parameters for the experimental night
PolysomnographyMeans ± SD
  1. Sleep parameters represent means ± SD.

  2. REM, rapid eye movement sleep; SWS, slow-wave sleep.

Sleep latency (min)17.7 ± 14.2
Sleep period time (SPT, min)457.0 ± 22.6
Sleep efficiency (%)90.1 ± 5.7
Waking after sleep onset (min)25.5 ± 15.6
Sleep stage, as percentage of sleep period time
 Waking (%)10.3 ± 7.6
 Stage 1 (%)8.7 ± 3.9
 Stage 2 (%)53.2 ± 5.5
 SWS (%)10.3 ± 7.6
 REM (%)22.1 ± 4.3

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Conflicts of Interest
  8. References
  9. Supporting Information

The results of this study are consistent with the initial hypothesis that sleep-specific brain activity patterns in the form of EEG sleep spindles and SWA during NREM sleep are associated with the overnight consolidation of newly encoded declarative and procedural memories. This has been demonstrated by spectral analysis of EEG sleep in healthy human subjects showing that both EEG sigma (as an index for sleep spindle activity) and SWA were positively correlated with the pre–post-sleep consolidation on a declarative word-list and a procedural mirror-tracing task.

Recently, there has been growing evidence that sleep spindles and SWA promote the consolidation of new memories and underlying synaptic refinements (for review, see Diekelmann and Born, 2010). Preclinical findings indicate that oscillations in the spindle frequency range can trigger intracellular mechanisms, such as the Ca2+-dependent activation of the calcium-calmodulin kinase II (Sejnowski and Destexhe, 2000), that ultimately facilitate synaptic long-term potentiation (LTP), thought to be the major molecular correlate of learning (Kandel, 2001). Given that these oscillations are widely distributed across the brain, it seems plausible that sleep spindles foster the strengthening of memory traces in different memory systems. In fact, there is evidence that spindle activity occurs preferentially at synapses previously potentiated by tetanizing stimulation (Werk et al., 2005), and that repeated spindle-associated spike discharges can induce LTP-like changes in neocortical synapses (Rosanova and Ulrich, 2005).

These observations are in line with the ‘active consolidation’ hypothesis (Diekelmann and Born, 2010) positing that novel and initially unstable memories are strengthened and consolidated during subsequent periods of sleep (McClelland et al., 1995). Particularly, the results of our study are consistent with the emerging notion that sleep spindles are not only implicated in declarative memory formation that has been shown to rely on a hippocampal-neocortical dialogue during NREM sleep (Sirota and Buzsaki, 2005), but also in procedural learning (Fogel and Smith, 2006). The present results confirm the association of EEG power in the sleep spindle (sigma) range, and both declarative memory consolidation (Schabus et al., 2004) and ‘off-line’ gains in motor performance. Notably, selectively EEG power in the slow sleep spindle frequency range (12–14 Hz) was associated with overnight memory consolidation in the present study.

Another line of research proposes that synapses that are strengthened during wakefulness undergo a process of generalized downscaling during high-amplitude oscillations in the EEG slow-wave frequency range during NREM sleep (‘synaptic homeostasis’ hypothesis; Tononi and Cirelli, 2006). Following this hypothesis, the beneficial effect of sleep on performance results from a refinement of the signal-to-noise-ratio in neural networks restoring the brain’s ability to acquire new information under conditions of limited energy and space. The association between the overnight consolidation of declarative and procedural memories and SWA observed in the present study is consistent with this hypothesis.

Taken together, our results are in line with both the ‘active consolidation’ and the ‘synaptic homeostasis’ hypothesis, and it seems plausible that these processes are not mutually exclusive. Rather they might act in concert, resulting in: (i) a sleep-dependent strengthening of relevant memory traces, for example supported by sleep spindles; and (ii) a sharpening of the information-to-noise ratio resulting from a generalized synaptic downscaling during SWA. Alternatively, EEG activity in the slow-frequency bands (including theta and alpha; please refer to Tables 2 and S1, online supplement) might reflect similar or even identical pathways for memory consolidation. However, based on studies examining the relationship of SWA and sigma activity (Dijk et al., 1993), and the current observation that SWA and slow sigma activity were not correlated, slow sleep spindle activity and SWA more likely represent, at least in part, distinct processes.

Our observation of a tendency towards a stronger correlation between overnight memory consolidation and SWA in the first NREM sleep period and sigma power in the second NREM sleep period fits to previous models of sleep regulation, as it is known that SWA, as a homeostatic marker of sleep intensity (Borbely, 1982), prevails in the beginning of the night, whereas spindle activity reaches its maximum later in the night (De Gennaro and Ferrara, 2003).

As a major limitation, it has to be noted that the findings of our study are only correlational and not causal in nature. Thus, two significant alternative hypotheses need to be addressed. First, EEG sigma and SWA might represent markers of general cognitive abilities (Bodizs et al., 2005), and subjects with high power values in these frequency ranges might show better memory consolidation independently of sleep. In the current study, we did not further assess general cognitive performance. However, we observed a specific association between EEG sigma and SWA and the pre–post-sleep consolidation, but no correlation between these EEG parameters and memory performance in the evening (learning). This observation is more consistent with the notion of sleep-specific effects (Gais et al., 2002). Second, EEG sigma and SWA might have no causal relationship with memory processes at all, but might represent epiphenomena of an underlying unknown process driving both memory and EEG rhythms. However, as outlined earlier, a number of animal and human studies have provided evidence for a cause-effect-chain, suggesting that we observed not a mere epiphenomenon but relevant sleep-specific processes.

In conclusion, our findings are consistent with the hypothesis that sleep provides particular conditions, such as neural mechanisms underlying sleep spindles and SWA, which promote the consolidation of newly acquired memory traces in both major memory systems, the declarative and non-declarative system.

Footnotes
  • 1

    The data including the subject defined as an outlier are provided in an additional Table S1 in the online supplement. It is to note that the inclusion of this subject did not change (but rather increased) the level of significance of the main findings (slow sigma activity and SWA correlations).

Disclosure of Conflicts of Interest

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Conflicts of Interest
  8. References
  9. Supporting Information

This is not an industry-supported study. All authors have declared no conflicts of interest.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Conflicts of Interest
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Conflicts of Interest
  8. References
  9. Supporting Information

Table S1. EEG activity of the experimental night and correlations with the overnight consolidation of procedural and declarative memories for all subjects including the statistical outlier in EEG-sigma activity.

FilenameFormatSizeDescription
jsr1017_sm_ts1.doc42KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.