Abstract

Sleep promotes the consolidation of newly acquired associative memories. Here we used neuronal oscillations in the human EEG to investigate sleep-dependent changes in the cortical memory trace. The retrieval activity for object–color associations was assessed immediately after encoding and after 3 hr of sleep or wakefulness. Sleep had beneficial effects on memory performance and led to reduced event-related theta and gamma power during the retrieval of associative memories. Furthermore, event-related alpha suppression was attenuated in the wake group for memorized and novel stimuli. There were no sleep-dependent changes in retrieval activity for missed items or items retrieved without color. Thus, the sleep-dependent reduction in theta and gamma oscillations was specific for the retrieval of associative memories. In line with theoretical accounts on sleep-dependent memory consolidation, decreased theta may indicate reduced mediotemporal activity because of a transfer of information into neocortical networks during sleep, whereas reduced parietal gamma may reflect effects of synaptic downscaling. Changes in alpha suppression in the wake group possibly index reduced attentional resources that may also contribute to a lower memory performance in this group. These findings indicate that the consolidation of associative memories during sleep is associated with profound changes in the cortical memory trace and relies on multiple neuronal processes working in concert.

INTRODUCTION

Sleep plays a pivotal role in the consolidation of newly acquired memories, transforming initially labile memories into more stable and lasting memory representations (Rasch & Born, 2013; Diekelmann & Born, 2010; Stickgold, 2005). In particular, it has been demonstrated that the consolidation of explicitly encoded associative memories is promoted by slow wave sleep (Rasch, Buchel, Gais, & Born, 2007; Marshall, Helgadóttir, Mölle, & Born, 2006). Although theoretical models presume profound sleep-dependent changes in the cortical engram (e.g., Born & Wilhelm, 2012; Marshall & Born, 2007; Tononi & Cirelli, 2006), it has rarely been investigated how episodes of sleep effect memory representations in the human cerebral cortex.

According to an active system consolidation account, sleep promotes the transfer of novel memories from neuronal networks in the mediotemporal lobe into lasting memory representations in neocortical networks (Born & Wilhelm, 2012; Rasch & Born, 2007). It is assumed that the transfer of information relies on an active replay of newly encoded information during sleep, as suggested by studies in rodents (Ji & Wilson, 2007; Wilson & McNaughton, 1994) and recent evidence in humans (Deuker et al., 2013). Second, according to the synaptic homeostasis hypothesis, wakefulness is assumed to lead to a general increase in neuronal activity, whereas sleep promotes a global downscaling of synaptic strength (Tononi & Cirelli, 2006). By a proportional reduction in synaptic strength, sleep may lead to a relative enhancement of recently activated synapses. So far, this theoretical idea was investigated in animal models (Liu, Faraguna, Cirelli, Tononi, & Gao, 2010; Vyazovskiy, Cirelli, Pfister-Genskow, Faraguna, & Tononi, 2008) and recently also in the human brain in wake EEG recordings (Kuhn et al., 2016; Landsness et al., 2011). For example, Kuhn and colleagues (2016) found decreased levels of occipital theta and gamma oscillatory power after sleep, compared with sleep-deprived control group, although possible relations to memory performance were not directly tested in this study. Importantly, these two theoretical ideas outlined here are not mutually exclusive, and the consolidation of novel memories may rely on both processes working in concert (Diekelmann & Born, 2010).

A first study on sleep-dependent changes in the cortical memory representation by Gais and colleagues (2007) compared the retrieval activity before (pre) and after (post) an interval of sleep or sleep deprivation. Using fMRI the authors found a redistribution of retrieval activity from mediotemporal to neocortical neuronal networks, which is in support of an active system consolidation account.

Neuronal oscillations play a pivotal role in memory processes (Fell & Axmacher, 2011). In the human MEG and EEG, a particular focus lies on oscillations in the theta (4–8 Hz), alpha (8–14 Hz). and gamma (30–100 Hz) ranges. Increases in theta oscillations are associated with episodic encoding (Friese et al., 2013; Staudigl & Hanslmayr, 2013; Osipova et al., 2006; Sederberg, Kahana, Howard, Donner, & Madsen, 2003) and retrieval (Gruber, Tsivilis, Giabbiconi, & Müller, 2008; Osipova et al., 2006). Functionally, there is good evidence that theta oscillations index the communication between mediotemporal and neocortical networks, for example, during spatial learning (Jones & Wilson, 2005) and associative memory processes (Backus, Schoffelen, Szebényi, Hanslmayr, & Doeller, 2016; Brincat & Miller, 2015; Kaplan et al., 2014; Staudigl & Hanslmayr, 2013). A suppression in alpha activity during episodic encoding (Friese et al., 2013; Hanslmayr, Spitzer, & Bäuml, 2009) and retrieval (Klimesch, 1997) is assumed to facilitate cortical processing. More specifically, alpha oscillations index the gating of incoming information by inhibiting task-irrelevant regions and routing it to task-relevant regions (Jensen & Mazaheri, 2010). Furthermore, modulations in the alpha frequency mark states of wakefulness and alertness (Klimesch, 1999). Finally, gamma oscillations (30–100 Hz) in the visual cortex reflect object representation processes (Hassler, Trujillo-Barreto, & Gruber, 2011; Fries, 2009; Bertrand & Tallon-Baudry, 2000), accompanying successful episodic encoding (Friese et al., 2013; Osipova et al., 2006) and retrieval (Gruber et al., 2008; Osipova et al., 2006).

In the present encephalographic (EEG) study, we further investigate sleep-dependent changes in the cortical engram. Neuronal oscillations were analyzed during the retrieval of associative memories pre and post 3 hr of nocturnal sleep, compared with a wake control group (see Figure 1). The main focus of the analysis was on the sleep-dependent changes in associative memory circuits during memory retrieval. We thus investigated the changes in neuronal oscillatory activity in the theta, alpha, and gamma frequencies from a pre to a post sleep retrieval phase, compared with a wake group. To see if sleep-dependent changes would be specific to the retrieval of associative memories, sleep-dependent changes for the processing of novel stimuli, forgotten stimuli, and items that were recognized without the associated color were included in the analysis. Here, a specific focus was on correctly identified novel items, as an indicator for sleep-dependent changes in non-mnemonic processing of visual stimuli.

Figure 1. 

The experimental procedure and the stimuli used for the associative memory task. (A) The experimental schedule of the sessions of the pre–post control group design. The sessions started with the encoding task around 6:00 p.m., after the setup of the EEG apparatus. Participants then performed two retrieval phases, pre and post a sleep (experimental group) or a wake (control group) interval. Participants of the sleep group were equipped with a polysomnography apparatus and went to bed for 3 hr between 8:00 p.m. and 0:00 a.m. The d2 test of attention was applied in both groups, before the second encoding phase, to test for differences in attention and concentration endurance. (B) Associative pictorial memory task. During encoding, participants had to form associations between the items and the background color (yellow or green). During retrieval, participants were asked to recognize objects and to recall the associated color among distractor items. Stimulus presentation was identical for encoding and retrieval: Each trial started with a blank screen (1 sec), followed by a fixation point (variable duration of 0.5–1.0 sec) and the presentation of a target stimulus (2 sec). The trial terminated with a question mark that remained until a response was given.

Figure 1. 

The experimental procedure and the stimuli used for the associative memory task. (A) The experimental schedule of the sessions of the pre–post control group design. The sessions started with the encoding task around 6:00 p.m., after the setup of the EEG apparatus. Participants then performed two retrieval phases, pre and post a sleep (experimental group) or a wake (control group) interval. Participants of the sleep group were equipped with a polysomnography apparatus and went to bed for 3 hr between 8:00 p.m. and 0:00 a.m. The d2 test of attention was applied in both groups, before the second encoding phase, to test for differences in attention and concentration endurance. (B) Associative pictorial memory task. During encoding, participants had to form associations between the items and the background color (yellow or green). During retrieval, participants were asked to recognize objects and to recall the associated color among distractor items. Stimulus presentation was identical for encoding and retrieval: Each trial started with a blank screen (1 sec), followed by a fixation point (variable duration of 0.5–1.0 sec) and the presentation of a target stimulus (2 sec). The trial terminated with a question mark that remained until a response was given.

METHODS

Participants

Twenty-six university students voluntarily participated in the experiment, 13 in a sleep group (9 women, Mage = 19.9 years, SDage = 2.4 years) and 13 in a wake control group (9 women, Mage = 20.1 years, SDage = 1.6 years). None of the participants reported any neurological or psychiatric record or sleep disorders in an anamnesis session. The experimental procedure was conducted in accordance with the World Medical Association's Declaration of Helsinki (59th WMA general assembly, Seoul, 2008), and informed written consent was obtained from each participant. Three additional participants were not included in the analysis because of technical problems (n = 1), insufficient trials for the EEG analysis (fewer than 10 artifact-free trials for item–color hits or correctly identified novel items, n = 1, or a lack of missed items, n = 1), or incomplete data assessments (headache, n = 1).

Participants got out of bed at seven in the morning on the day of the experiment and refrained from caffeine or alcohol consumption on this day (after having one coffee or tea for breakfast). On average, participants reported 7.6 hr of sleep (SD = .9) during the night before the experiment.

Stimuli and Procedure

The stimulus set consisted of 400 grayscaled pictures of objects (e.g., plants, animals, clothes, tools), taken from a standard picture library (Hemera Photo Objects), presented at a visual angle of 6.2 × 6.2°.

The experimental procedure is depicted in Figure 1A. The sessions started around 6:00 p.m., after setting up the EEG apparatus. During encoding, participants saw 300 pictures in front of a yellow or a green square (see Figure 1B) and were instructed to form object–color associations. They were told the example of “a flower on green grass” for a flower that is presented on a green background. To assess the sleep-dependent changes in neuronal activity during retrieval, we employed a pre–post control group design: During encoding, participants learned 300 object–color associations. The first retrieval session (pre) followed after a 5-min break. One hundred fifty pictures from encoding were intermixed with 50 new stimuli and were presented on a gray background. For each picture, participants had to indicate whether they retrieved the picture with the associated color (old, green or old, yellow), retrieved the picture without a color (old, no color), or did not see this picture before (new). Here, the response option “old, no color” was included as a response option to avoid extensive color guesses and thereby increase the signal to noise ratio for correctly retrieved item–color associations. A second, identical retrieval session (post) followed around 0:00 a.m., after a 180-min sleep interval (experimental group) or an equally long awake interval (control group). In the post sleep retrieval session, the second half of the initially encoded stimuli was presented (150 stimuli) along with 50 additional distracters.

Stimulus presentation was identical for encoding and retrieval: Each trial started with a blank screen (1 sec), followed by a fixation point (variable duration of 0.5–1.0 sec) and the presentation of a target stimulus (2 sec). The trial terminated with a question mark. Response keys were pressed with different fingers of the right hand. The procedure was demonstrated in 10 training trials prior to each phase of the actual experiment. The stimuli were counterbalanced between sessions (encoding and retrieval, as well as pre- and postretrieval) equally in both groups (sleep and wake). Object color allocations were randomized for each participant.

Between pre- and postretrieval sessions, participants were freed from the EEG equipment and had a meal. Participants in the sleep group changed their clothes, were equipped with a polysomnography recording, and went to bed for 3 hr between 8:00 p.m. and 0:00 a.m. The polysomnography apparatus was monitored online, and participants were only awakened from the Sleep Stages 1 and 2 to avoid inertia. Participants in the control group saw animated movies to avoid interference with the real objects used as stimulus material (cf. Alger, Lau, & Fishbein, 2010).

In the sleep group, participants’ sleep was assessed with an Embla Titanium polysomnography system (TNI medical, Würzburg, Germany). The recording montage included six EEG electrodes (F3, F4, C3, C4, P3, P4) and references (M1, M2), such as an EOG and an electromyogram at submental muscles. Sleep recordings were classified according to the criteria of the American Academy of Sleep Medicine (Iber, Ancoli-Israel, Chesson, & Quan, 2007). The interrater reliability between two coders was assessed for 28 random frames of 30 sec (Cohen's κ = .75).

The d2 test of attention (Brickenkamp, 2002) was applied before the second encoding phase to test for differences in attention and concentration endurance after the sleep and the wake interval. In this test, participants are required to cross out target letters among distractors (every letter d with 2 lines as targets within a series of the letters d, p, or q with 1, 2, 3, or 4 lines) in a limited amount of time. Scores of the test were calculated and compared between groups: total (number of all processed letters minus number of all errors), accuracy (number of errors), and concentration endurance (number of correct responses minus confusion errors, i.e., mistakenly marked letters).

Sleep-dependent memory consolidation was tested for participants overall memory performance and their associative memory performance by the means of corrected recognition scores (Pr, i.e., hits minus false alarms; Snodgrass & Corwin, 1988). For the overall memory performance score, all remembered items were counted as hits and misclassified new items were counted as false alarms, irrespective of the color judgment. For participants' associative memory performance, the proportion of remembered items with correct color judgments (counted as color hits) was corrected by the proportion of remembered items with wrong color judgments (counted as color false alarms, i.e., an estimate for color guesses given that the object was correctly remembered). Both Pr scores were subjected to separate mixed-model ANOVAs with the factors Group (sleep, wake) and Retrieval session (pre, post). For all significant main effects and interactions, Greenhouse–Geisser corrected p values are reported along with the effect sizes, quantified by partial eta squared ηp2.

EEG Recording and Analyses

The EEG was recorded from 128 active electrodes using a BioSemi Active-Two amplifier system (BioSemi, Amsterdam, Netherlands) at a sampling rate of 512 Hz in a shielded room. A horizontal and vertical EOG was applied to monitor eye movements and blinks. Two additional electrodes (CMS [common mode sense] and DRL [driven right leg]; cf. www.biosemi.com/faq/cms&drl.htm) served as reference and ground.

Before the analysis, continuous EEG data were high-pass filtered at 0.5 Hz and eyeblinks and muscle artifacts were detected using an independent component procedure and removed after visual inspection (Chaumon, Bishop, & Busch, 2015). EEG data were then segmented into epochs from −1000 msec to 3000 msec with regard to the stimulus onset. Further artifacts and noisy trials were removed by the means of statistical correction of artifacts in dense array studies (SCADS; Junghöfer, Elbert, Tucker, & Rockstroh, 2000) used in several former studies (e.g., Gruber et al., 2008; Gruber & Müller, 2006). Furthermore, we applied a correction of saccade-related transient potentials (COSTRAP; Hassler et al., 2011) used in several previous publications (Köster, Friese, Schöne, Trujillo-Barreto, & Gruber, 2014; Friese et al., 2013; Hassler, Friese, Martens, Trujillo-Barreto, & Gruber, 2013; Hassler et al., 2011) to remove miniature eye movement artifacts (Yuval-Greenberg, Tomer, Keren, Nelken, & Deouell, 2008). Because of artifact correction procedures, approximately 10% of the original trials were removed. Throughout further analyses, an average reference was used. To obtain the spectral power over time, the trial data were convoluted using Morlet's wavelets with seven cycles (Bertrand & Pantev, 1994) at a resolution of 1 Hz. To account for the variability in frequency bands across individuals (Klimesch, 1997), we identified the peak theta, alpha, and gamma frequencies, individually for each participant, based on the mean spectral activity across the most relevant conditions (item–color associations and correct rejections; cf. Friese et al., 2013). This resulted in (mean ± SD) 4.22 Hz ± 1.53 Hz theta frequencies, 10.92 Hz ± 2.07 Hz alpha frequencies, and 60.37 Hz ± 8.43 Hz gamma frequencies. Event-related spectral changes for each condition and frequency band were then calculated as the relative signal change of the post stimulus spectral activity, relative to a 500-msec to 100-msec prestimulus baseline, in percent. Throughout all further analyses, the relative signal change at individual peak frequencies was used. For all topographical analyses, the relative signal change values at individual peak frequencies were averaged over the full time period of stimulus presentation (0–2000 msec) for all comparisons. The main focus of this study was sleep-dependent changes in associative memory processes. Furthermore, to test whether or not sleep-dependent changes in EEG activity would be specific to the retrieval of associative memories, we also included the trials of correctly identified novel items, hits without color judgment, and missed items.

Grand Mean Oscillatory Activity

In a first step, we analyzed the grand mean spectral changes relative to the baseline in the first retrieval session, namely prior to the group-specific experimental manipulation (sleep or wake). Values at each electrode were averaged over all trials of the conditions entered in the analyses. The signal changes for individual theta, alpha, and gamma frequencies were tested against baseline using a cluster mass permutation t test (two-sided, dependent samples; Maris & Oostenveld, 2007). To further characterize the event-related signal changes over the stimulus presentation time, we calculated the grand mean activity at each time point, at frequency-specific electrode clusters, that is, at the most significant clusters for theta, alpha, and gamma.

Sleep-dependent Changes in Neuronal Oscillatory Activity during Memory Retrieval

The sleep-dependent changes in retrieval activity were analyzed, following the rationale of the pre–post control group design: The pre–post differences in the three frequency bands (theta, alpha, gamma) were compared between the wake and the control group to obtain the sleep-specific changes in neuronal oscillatory activity. Specifically, we applied a two-level statistical analysis (cf. Obleser, Wöstmann, Hellbernd, Wilsch, & Maess, 2012). At the first or subject level, we obtained t values for the trial data of individual pre and post differences using an unpaired t test (testing for significant differences, two-sided, independent samples). This was done separately for the four conditions included in the analysis (item–color hits, novel items, hits without color, and misses). Attaining first-level statistics rather than using only average power changes compared with baseline has the advantage that the pre–post differences that are subsequently entered into group statistics are effectively standardized for across-trial variances and differences in the number of trials between participants and conditions. At the second or group level, the individual t maps of all four conditions were used as the input of a combined cluster mass permutation t test (testing for significant differences, two-sided, independent samples; adapted from Maris & Oostenveld 2007; implemented in fieldtrip, Oostenveld, Fries, Maris, & Schoffelen, 2011). This was to check for significant differences in changes from pre to post (first level) between the sleep and wake groups (second level). Specifically, the cluster statistic was calculated as the sum over the t values of neighboring electrodes with a cluster inclusion criterion of p < .05, separately for each condition included in the analysis. The significance of the cluster statistics was then calculated from a combined permutation distribution obtained from 1000 Monte Carlo iterations with randomly assigned sleep and wake groups, including the clusters of all four conditions. The results are cluster-wise p values, which are not affected by inflated false-positive rates otherwise arising from multiple comparisons. The rationale for including all four conditions into a combined permutation was to consider the whole variation for pre–post differences when identifying clusters that exceed chance level.

RESULTS

Behavioral Results

Participants in the sleep group showed normal sleep profiles for an early night sleep (mean ± SD): The total sleep time was 137.9 ± 20.0 min with pronounced episodes of slow wave sleep (Stage 1 sleep [S1]: 14.3 ± 12.2 min; Stage 2 sleep [S2]: 51.4 ± 19.9 min; slow wave sleep [S3]: 66.5 ± 31.9 min; rapid eye movement sleep [REM]: 9.2 ± 11.3 min; wake after sleep onset: 25.8 ± 39.7 min).

The results of the d2 attention and concentration endurance scores are reported in Table 1. There were no differences in the total score, the accuracy, or the concentration endurance score between the sleep and wake groups.

Table 1. 

Attention and Concentration Endurance

d2 ScoreSleepWaketp
Total 459.1 (75.6) 496.8 (58.3) −1.43 >.15 
Accuracy 15.1 (7.6) 23.3 (19.3) −1.44 >.15 
Concentration endurance 184.5 (28.3) 187.6 (57.5) −.18 >.25 
d2 ScoreSleepWaketp
Total 459.1 (75.6) 496.8 (58.3) −1.43 >.15 
Accuracy 15.1 (7.6) 23.3 (19.3) −1.44 >.15 
Concentration endurance 184.5 (28.3) 187.6 (57.5) −.18 >.25 

Means ± SD are displayed. p Values denote differences between the sleep and wake groups (results of pairwise comparisons, two-sided). Total = number of processed letters minus number of errors; Accuracy = number of errors; Concentration endurance = number of correct responses minus confusion errors.

Over both groups (sleep and wake) and conditions (pre and post), participants remembered (mean ± SD) 77.6 ± 8.7% of the target pictures (hits) while judging 21.5 ± 8.7% as new (misses). Of the distracters, 78.5 ± 15.2% were correctly recognized as new (correct rejections) and, accordingly, 22.4 ± 15.2% were classified old (false alarms). Regarding the retrieval of object–color associations, 46.9 ± 12.8% of all target pictures were recollected with the correct background color (hit, color), 16.5 ± 7.3% were retrieved with false color judgments (false alarm, color), and 13.7 ± 5.4% were retrieved without color judgment.

For the memory performance of the sleep and wake groups in the pre and post phases, see Table 2. Overall memory performance showed a decay over time, F(1, 24) = 19.88, p < .001, ηp2 = .453, and a sleep-dependent consolidation effect. This is indicated by a significant Phase × Group interaction, F(1, 24) = 5.88, p = .023, ηp2 = .197, as well as a significant decay in memory performance in the wake, but not in the sleep group (see Table 2). Associative memory performance also decreased from pre to post retrieval, F(1, 24) = 16.05, p < .001, ηp2 = .401. Although the Phase × Group interaction was not significant, a consolidation effect is indicated by a reduction in the wake, but not in the sleep group (see Table 2). Although baseline differences between the sleep and wake groups contribute to these effects, baseline performances for overall and the associative memory did not differ significantly between groups, t = −1.03, p = .32, and, t = −.77, p = .45.

Table 2. 

Overall and Associative Memory Performance

PrePosttp
Memory Performance (Pr) 
Sleep .59 (.23) .53 (.16) 1.31 >.20 
Wake .66 (.12) .47 (.14) 5.44 <.001 
 
Associative Memory Performance (Pr) 
Sleep .30 (.24) .21 (.13) 1.77 >.10 
Wake .36 (.10) .21 (.08) 4.99 <.001 
PrePosttp
Memory Performance (Pr) 
Sleep .59 (.23) .53 (.16) 1.31 >.20 
Wake .66 (.12) .47 (.14) 5.44 <.001 
 
Associative Memory Performance (Pr) 
Sleep .30 (.24) .21 (.13) 1.77 >.10 
Wake .36 (.10) .21 (.08) 4.99 <.001 

Means ± SD are displayed. p Values denote significant changes between retrieval phases (results of pairwise, two-sided comparisons).

EEG Results

Grand Mean Oscillatory Activity during Retrieval

The signal changes in oscillatory power in the first retrieval phase revealed significant increases in theta power at frontal (cluster test against baseline: p < .001) and posterior (p = .016) electrodes. At frontal electrodes, theta activity reached a peak around 600 msec after stimulus onset. Overall alpha power was suppressed across all electrodes (p < .001). Relative to the theta increase, the onset and the peak of the alpha suppression were slightly delayed (∼100 msec). The gamma power showed an increase in the occipital region (p = .050), with a peak around 300 msec and a decrease to baseline level after stimulus offset. We furthermore found unexpected event-related decreases in temporal gamma activity (both ps < .021). Grand mean event-related changes in the pre retrieval phase are displayed in Figure 2. Notably, a comparison of the pre retrieval oscillatory activity between the sleep and wake groups confirmed that there were no differences in retrieval activity between both groups. This is, there were no significant clusters for any frequency band in any of the conditions included in the analyses (all ps > .116).

Figure 2. 

Grand mean spectral power in the theta, alpha, and gamma frequencies during retrieval. Topographical maps of the relative signal changes (RSCs) in grand mean activity for item–color hits, correctly identified novel items, hits without color judgments and misses. Topographies show the averaged power for the whole duration of the stimulus presentation (0 to 2000 msec) at individual theta, alpha, and gamma frequencies. Only significant differences (p < .05) are displayed. White discs indicate the electrodes of statistically significant clusters (p < .05), compared with baseline. Line plots show the temporal evolution of the event-related signal changes per time point for the frontal (theta), overall (alpha), or occipital (gamma) clusters. Shaded areas are 95% confidence intervals.

Figure 2. 

Grand mean spectral power in the theta, alpha, and gamma frequencies during retrieval. Topographical maps of the relative signal changes (RSCs) in grand mean activity for item–color hits, correctly identified novel items, hits without color judgments and misses. Topographies show the averaged power for the whole duration of the stimulus presentation (0 to 2000 msec) at individual theta, alpha, and gamma frequencies. Only significant differences (p < .05) are displayed. White discs indicate the electrodes of statistically significant clusters (p < .05), compared with baseline. Line plots show the temporal evolution of the event-related signal changes per time point for the frontal (theta), overall (alpha), or occipital (gamma) clusters. Shaded areas are 95% confidence intervals.

Sleep-dependent Changes in Neuronal Oscillatory Activity during Memory Retrieval

To test the sleep-dependent changes in the retrieval activity, we compared the time-dependent changes (post–pre) between the sleep and wake groups (Figure 3B) for all conditions included in the analysis. For correctly retrieved item–color associations, we found reduced event-related theta activity after the sleep interval and an increase in event-related theta activity after the wake interval (Figure 3A, left column). Group differences were found in a significant cluster in the posterior region (p = .024) and in a cluster over the left temporal-posterior region (p = .015), see Figure 3B and C, left column. The event-related alpha suppression was reduced in the wake group, whereas the sleep group maintained the same level of alpha suppression. This Group effect in the alpha frequency is significant in frontal (p = .011) and parietal (p = .041) clusters. Sleep-dependent changes in the gamma frequency were found at posterior electrodes, namely the electrodes with the highest gamma power in the grand mean (see Figure 2). The sleep group showed a decrease in gamma power, whereas the wake group showed increased gamma (Figure 3AC, right panel). The parietal cluster was not significant in the cluster mass permutation test (p = .111), possibly because of the very focalized spatial distribution of parietal gamma power that does not survive the cluster permutation test. However, in a data-driven approach that evaluates the occipital electrode cluster or the peak gamma electrode of the grand mean, the analysis resulted in significant differences between the sleep and wake groups (cluster: t = 2.29, p = .041; peak electrode: t = 2.51, p = .027). Note that sleep-dependent changes in gamma were not found in the further conditions included in the analyses (cluster: all ps > .201; electrode: all ps > .267).

Figure 3. 

Spectral power for the retrieval of item–color associations. (A) Topographical maps of the time-dependent changes in retrieval activity from pre to post (post–pre) for the sleep and awake groups. Topographies depict relative signal changes (RSCs) for the whole time of the stimulus presentation (0–2000 msec). (B) Topographical maps show the sleep-dependent changes for retrieval activity for item–color associations, following the rationale of the pre–post control group design: The pre–post differences, as depicted in A, were compared between the wake and control groups (sleep–wake). Only significant differences (p < .05) are displayed. White discs indicate the electrodes of statistically significant clusters (p < .05), identified by a cluster mass permutation test. Note that sleep-dependent changes in gamma power were significant at the electrodes of the grand mean parietal gamma cluster (p = .041). (C) Time series show the averaged post–pre differences of the sleep group (blue) and the wake group (red), averaged over the electrodes in the clusters of B. Shaded areas are 95% confidence intervals.

Figure 3. 

Spectral power for the retrieval of item–color associations. (A) Topographical maps of the time-dependent changes in retrieval activity from pre to post (post–pre) for the sleep and awake groups. Topographies depict relative signal changes (RSCs) for the whole time of the stimulus presentation (0–2000 msec). (B) Topographical maps show the sleep-dependent changes for retrieval activity for item–color associations, following the rationale of the pre–post control group design: The pre–post differences, as depicted in A, were compared between the wake and control groups (sleep–wake). Only significant differences (p < .05) are displayed. White discs indicate the electrodes of statistically significant clusters (p < .05), identified by a cluster mass permutation test. Note that sleep-dependent changes in gamma power were significant at the electrodes of the grand mean parietal gamma cluster (p = .041). (C) Time series show the averaged post–pre differences of the sleep group (blue) and the wake group (red), averaged over the electrodes in the clusters of B. Shaded areas are 95% confidence intervals.

We further focused on the processing of novel stimuli to see whether sleep would likewise modulate the processing of novel stimuli. Sleep-dependent changes in the processing of novel stimuli did not reveal event-related changes in theta and gamma activity, but in alpha power (Figure 4A). Namely, there was a significant reduction in event-related alpha suppression from pre to post in the wake group but not in the sleep group (frontal cluster: p = .004; parietal cluster: p = .034; Figure 4B and C). For the further conditions, hits without color judgment and misses, there were no significant differences between the sleep and wake groups, but a marginally significant reduction in frontal alpha suppression (p = .069) after the wake compared with the sleep interval (Figure 5).

Figure 4. 

Spectral power for correctly identified new items. (A) In parallel to the analysis of the retrieval of item–color associations (see Figure 1), topographical maps depict time-dependent changes (post–pre) for correctly identified novel items in the sleep and awake groups. Topographies depict the relative signal changes (RSCs) between 0 and 2000 msec. (B) Sleep-dependent changes: The pre–post differences, as depicted in A, were compared between the wake and control groups (sleep–wake). Only significant differences (p < .05) are displayed. White discs indicate the electrodes of statistically significant clusters (p < .05). (C) Time series show the averaged post–pre differences of the sleep group (blue) and the wake group (red), averaged over the electrodes in the parietal cluster of B. Shaded areas are 95% confidence intervals.

Figure 4. 

Spectral power for correctly identified new items. (A) In parallel to the analysis of the retrieval of item–color associations (see Figure 1), topographical maps depict time-dependent changes (post–pre) for correctly identified novel items in the sleep and awake groups. Topographies depict the relative signal changes (RSCs) between 0 and 2000 msec. (B) Sleep-dependent changes: The pre–post differences, as depicted in A, were compared between the wake and control groups (sleep–wake). Only significant differences (p < .05) are displayed. White discs indicate the electrodes of statistically significant clusters (p < .05). (C) Time series show the averaged post–pre differences of the sleep group (blue) and the wake group (red), averaged over the electrodes in the parietal cluster of B. Shaded areas are 95% confidence intervals.

Figure 5. 

Spectral power for items retrieved without color and misses. Topographical maps depict sleep-dependent changes for (A) items retrieved without color and (B) items that were missed during retrieval. The pre–post differences (post–pre) were compared between the wake and control groups (sleep–wake), compare B of Figures 3 and 4. Topographies depict the relative signal changes (RSCs) between 0 and 2000 msec. Only significant differences (p < .05) are displayed. Black discs indicate the electrodes of the only cluster that was marginally significant in these two conditions (p = .069).

Figure 5. 

Spectral power for items retrieved without color and misses. Topographical maps depict sleep-dependent changes for (A) items retrieved without color and (B) items that were missed during retrieval. The pre–post differences (post–pre) were compared between the wake and control groups (sleep–wake), compare B of Figures 3 and 4. Topographies depict the relative signal changes (RSCs) between 0 and 2000 msec. Only significant differences (p < .05) are displayed. Black discs indicate the electrodes of the only cluster that was marginally significant in these two conditions (p = .069).

DISCUSSION

We analyzed neuronal oscillations in the human EEG during the retrieval of associative memories before and after a sleep, compared with a wake control group, to investigate sleep-dependent changes in the cortical engram. Sleep-dependent memory consolidation was indicated by the behavioral data and specifically modulated associative retrieval processes reflected in the theta and gamma frequency ranges: Event-related theta power was reduced after a sleep interval, in contrast to a wake interval, peaking at centroparietal electrodes. Likewise, occipital gamma activity was reduced after a sleep period but increased after a wake period. In contrast, we did not find sleep-dependent changes in the theta and gamma frequency for the processing of novel stimuli, hits without color judgments, or misses. Event-related alpha suppression for the retrieval of associative memories was reduced from pre to post in the wake group for the processing of novel stimuli and was also marginally reduced for items that were recognized without color, but not for missed items. Thus, the retrieval of associative memories after sleep was specifically associated with a reduction in theta and gamma activity, whereas a reduced alpha suppression after sleep seems to be a more general effect of sleep on the processing of visually presented stimuli.

The theta rhythm is a prominent marker of associative memory processes in the mediotemporal lobe (Lisman & Jensen, 2013; Fell & Axmacher, 2011). Furthermore, theta oscillations in neocortical neuronal networks were found to synchronize with theta oscillations in the mediotemporal lobe during associative retrieval processes in the monkey (Brincat & Miller, 2015) and human brain (Backus et al., 2016; Kaplan et al., 2014). The theta rhythm is assumed to index the interplay between mediotemporal and neocortical neuronal networks (Brincat & Miller, 2015; Kaplan et al., 2014; Köster et al., 2014; Nyhus & Curran, 2010). Here, we found reduced theta activity during memory retrieval after a sleep interval (Tononi & Cirelli, 2006). Reduced theta activity may possibly index a reduced and more efficient interplay between the mediotemporal and the cortical memory system, which would be in line with the active system account of memory consolidation (Born & Wilhelm, 2012). In particular, a reduced or more effective interplay between the mediotemporal and the cortical memory system may result from a partial transfer of information from mediotemporal into neocortical networks during sleep. Recent findings from fMRI research suggests that the MTL and neocortical engram is reactivated during sleep (Deuker et al., 2013) and facilitates the communication between MTL and neocortical networks in the human brain (Gais et al., 2007). Occipital gamma activity may reflect neuronal synchronization processes accompanying visual perception of objects (e.g., Osipova et al., 2006; for a review, see Tallon-Baudry & Bertrand, 1999), which have been shown to be sensitive to memory retrieval processes (Köster et al., 2014; Osipova et al., 2006). The reduction in occipital gamma activity has formerly been described for repeatedly presented objects (Hassler et al., 2011, 2013; Busch, Herrmann, Müller, Lenz, & Gruber, 2006) and is interpreted as a reduction in neocortical activity due to sharpened or globally reduced synaptic activity for the processing of stimulus material (Wiggs & Martin, 1998). Here we found a sleep-dependent reduction in occipital gamma during the retrieval of associative memories, as indicated by significant differences in the parietal cluster from the grand mean, but not in the cluster test. Both findings, namely, reduced theta and gamma activity during post sleep retrieval processes thus conform with a synaptic homeostasis view on sleep-dependent memory consolidation, namely that sleep promotes a global downscaling of synaptic strength, leading to a more effective processing in recently activated memory traces (Tononi & Cirelli, 2006). Noteworthy, a recent study by Kuhn et al. (2016) also found decreases in theta and gamma activity after sleep, which is also interpreted as an indicator of more general downscaling processes in the human cerebral cortex after sleep.

Critically, the reduction in theta and gamma activity was selectively observed for associative memories, but not for items retrieved without color. This substantiates the idea that theta and gamma neuronal oscillations index specific neuronal mechanisms that underpin the initially formation of lasting associative memories within the MTL and neocortical memory system (Lisman & Jensen, 2013), which is also assumed to play a key role in the consolidation of associative memories during sleep (Rasch & Born, 2013; Diekelmann & Born, 2010).

Modulations in alpha activity have been shown to reflect diurnal variation in alertness and wakefulness (Klimesch, 1999). Event-related alpha suppression is a marker of attentional processes (Berger, 1932) and is thought to gate task-relevant processes in neocortical networks (e.g., Jensen & Mazaheri, 2010; Klimesch, Sauseng, & Hanslmayr, 2007; Klimesch, 1999). Research suggests that event-related alpha suppression during mnemonic processing is rather associated with semantic processes accompanying visual perception processes (Hanslmayr et al., 2009). In this study, modulations in alpha activity were found in the wake group (where alpha suppression was reduced after sleep) and were not specific to memory retrieval but also found for novel stimuli and hits without color judgments. Thus, reduced alpha suppression may reflect decreases in attentional resources and the semantic processing of the stimuli due to the 3-hr period of wakefulness. These findings suggest that changes in the alpha band do not index processes that are specific to associative memory retrieval but may support the idea that sleep also promotes mnemonic functioning by a recovery of attentional resources. These may also contribute to a lower decay of memory performance found in the sleep group. This interpretation leaves open the question on why there were no sleep-dependent changes in alpha oscillations for missed items. Furthermore, the d2 test of attention and concentration endurance did not reveal any differences in attentional resources after the sleep or the wake interval, indicating that there was no severe effect of wakefulness on participants' task performance. To summarize, our findings substantiate the idea that sleep-dependent memory consolidation is a multifaceted phenomenon, relying on multiple neuronal mechanisms and processes working in concert.

To conclude, to our best knowledge, this is the first study that analyzed neuronal oscillatory activity during memory retrieval to investigate sleep-dependent changes in the cortical memory trace. Most importantly, we found a sleep-dependent reduction in theta activity, indicating a reduced or sharpened interplay between MTL and neocortical networks and a reduction in occipital gamma, indicating reduced synaptic activity in visual cortex. These findings highlight the notion that sleep leads to a marked transformation of cortical memory representations, presumably driven by active system consolidation and synaptic homeostasis processes working in concert. However, the results of this study should be substantiated in future studies and cross-validated with other methodologies, such as MEG or intracranial recordings. Methodologically, this study further illustrates that pre–post sleep comparisons are a promising approach to further elucidate the transformations that memory representations undergo during periods of sleep.

Reprint requests should be sent to Moritz Köster, Institute of Psychology, University of Münster, Fliednerstraße 21, 48149 Münster, Germany, or via e-mail: moritz.koester@uni-muenster.de.

REFERENCES

Alger
,
S. E.
,
Lau
,
H.
, &
Fishbein
,
W.
(
2010
).
Delayed onset of a daytime nap facilitates retention of declarative memory
.
PloS One
,
5
,
e12131
.
Backus
,
A. R.
,
Schoffelen
,
J. M.
,
Szebényi
,
S.
,
Hanslmayr
,
S.
, &
Doeller
,
C. F.
(
2016
).
Hippocampal-prefrontal theta oscillations support memory integration
.
Current Biology
,
26
,
450
457
.
Berger
,
H.
(
1932
).
über das Elektrenkephalogramm des Menschen
.
European Archives of Psychiatry and Clinical Neuroscience
,
97
,
6
26
.
Bertrand
,
O.
, &
Pantev
,
C.
(
1994
).
Stimulus frequency dependence of the transient oscillatory auditory evoked responses (40 Hz) studied by electric and magnetic recordings in human
. In
C.
Pantev
,
T.
Elbert
, &
B.
Lütkenhöner
(Eds.),
Oscillatory event-related brain dynamics
(pp.
231
242
).
New York
:
Plenum
.
Bertrand
,
O.
, &
Tallon-Baudry
,
C.
(
2000
).
Oscillatory gamma activity in humans: A possible role for object representation
.
International Journal of Psychophysiology
,
38
,
211
223
.
Born
,
J.
, &
Wilhelm
,
I.
(
2012
).
System consolidation of memory during sleep
.
Psychological Research
,
76
,
192
203
.
Brickenkamp
,
R.
(
2002
).
d2 Aufmerksamkeits-Belastungs-Test [Test of Attention]
.
Göttingen, Germany
:
Hogrefe
.
Brincat
,
S. L.
, &
Miller
,
E. K.
(
2015
).
Frequency-specific hippocampal-prefrontal interactions during associative learning
.
Nature Neuroscience
,
18
,
576
581
.
Busch
,
N. A.
,
Herrmann
,
C. S.
,
Müller
,
M. M.
,
Lenz
,
D.
, &
Gruber
,
T.
(
2006
).
A cross-laboratory study of event-related gamma activity in a standard object recognition paradigm
.
Neuroimage
,
33
,
1169
1177
.
Chaumon
,
M.
,
Bishop
,
D. V.
, &
Busch
,
N. A.
(
2015
).
A practical guide to the selection of independent components of the electroencephalogram for artifact correction
.
Journal of Neuroscience Methods
,
250
,
47
63
.
Deuker
,
L.
,
Olligs
,
J.
,
Fell
,
J.
,
Kranz
,
T. A.
,
Mormann
,
F.
,
Montag
,
C.
, et al
(
2013
).
Memory consolidation by replay of stimulus-specific neural activity
.
Journal of Neuroscience
,
33
,
19373
19383
.
Diekelmann
,
S.
, &
Born
,
J.
(
2010
).
The memory function of sleep
.
Nature Reviews Neuroscience
,
11
,
114
126
.
Fell
,
J.
, &
Axmacher
,
N.
(
2011
).
The role of phase synchronization in memory processes
.
Nature Reviews Neuroscience
,
12
,
105
118
.
Fries
,
P.
(
2009
).
Neuronal gamma-band synchronization as a fundamental process in cortical computation
.
Annual Review Neuroscience
,
32
,
209
224
.
Friese
,
U.
,
Köster
,
M.
,
Hassler
,
U.
,
Martens
,
U.
,
Barreto
,
N. T.
, &
Gruber
,
T.
(
2013
).
Successful memory encoding is associated with increased cross-frequency coupling between frontal theta and posterior gamma oscillations in human scalp-recorded EEG
.
Neuroimage
,
66
,
642
647
.
Gais
,
S.
,
Albouy
,
G.
,
Boly
,
M.
,
Dang-Vu
,
T. T.
,
Darsaud
,
A.
,
Desseilles
,
M.
, et al
(
2007
).
Sleep transforms the cerebral trace of declarative memories
.
Proceedings of the National Academy of Sciences, U.S.A.
,
104
,
18778
18783
.
Gruber
,
T.
, &
Müller
,
M. M.
(
2006
).
Oscillatory brain activity in the human EEG during indirect and direct memory tasks
.
Brain Research
,
1097
,
194
204
.
Gruber
,
T.
,
Tsivilis
,
D.
,
Giabbiconi
,
C. M.
, &
Müller
,
M. M.
(
2008
).
Induced electroencephalogram oscillations during source memory: Familiarity is reflected in the gamma band, recollection in the theta band
.
Journal of Cognitive Neuroscience
,
20
,
1043
1053
.
Hanslmayr
,
S.
,
Spitzer
,
B.
, &
Bäuml
,
K. H.
(
2009
).
Brain oscillations dissociate between semantic and nonsemantic encoding of episodic memories
.
Cerebral Cortex
,
19
,
1631
1640
.
Hassler
,
U.
,
Friese
,
U.
,
Martens
,
U.
,
Trujillo-Barreto
,
N.
, &
Gruber
,
T.
(
2013
).
Repetition priming effects dissociate between miniature eye movements and induced gamma‐band responses in the human electroencephalogram
.
European Journal of Neuroscience
,
38
,
2425
2433
.
Hassler
,
U.
,
Trujillo-Barreto
,
N.
, &
Gruber
,
T.
(
2011
).
Induced gamma band responses in human EEG after the control of miniature saccadic artifacts
.
Neuroimage
,
57
,
1411
1421
.
Iber
,
C.
,
Ancoli-Israel
,
S.
,
Chesson
,
A.
, &
Quan
,
S. F.
(
2007
).
The AASM manual for the scoring of sleep and associated events: Rules, terminology, and technical specification
(1st ed.).
Westchester, IL
:
American Academy of Sleep Medicine
.
Jensen
,
O.
, &
Mazaheri
,
A.
(
2010
).
Shaping functional architecture by oscillatory alpha activity: Gating by inhibition
.
Frontiers in Human Neuroscience
,
4
,
186
.
Ji
,
D.
, &
Wilson
,
M. A.
(
2007
).
Coordinated memory replay in the visual cortex and hippocampus during sleep
.
Nature Neuroscience
,
10
,
100
107
.
Jones
,
M. W.
, &
Wilson
,
M. A.
(
2005
).
Theta rhythms coordinate hippocampal-prefrontal interactions in a spatial memory task
.
PLoS Biology
,
3
,
2187
.
Junghöfer
,
M.
,
Elbert
,
T.
,
Tucker
,
D. M.
, &
Rockstroh
,
B.
(
2000
).
Statistical control of artifacts in dense array EEG/MEG studies
.
Psychophysiology
,
37
,
523
532
.
Kaplan
,
R.
,
Bush
,
D.
,
Bonnefond
,
M.
,
Bandettini
,
P. A.
,
Barnes
,
G. R.
,
Doeller
,
C. F.
, et al
(
2014
).
Medial prefrontal theta phase coupling during spatial memory retrieval
.
Hippocampus
,
24
,
656
665
.
Klimesch
,
W.
(
1997
).
EEG-alpha rhythms and memory processes
.
International Journal of Psychophysiology
,
26
,
319
340
.
Klimesch
,
W.
(
1999
).
EEG alpha and theta oscillations reflect cognitive and memory performance: A review and analysis
.
Brain Research Review
,
29
,
169
195
.
Klimesch
,
W.
,
Sauseng
,
P.
, &
Hanslmayr
,
S.
(
2007
).
EEG alpha oscillations: The inhibition–timing hypothesis
.
Brain Research Review
,
53
,
63
88
.
Köster
,
M.
,
Friese
,
U.
,
Schöne
,
B.
,
Trujillo-Barreto
,
N.
, &
Gruber
,
T.
(
2014
).
Theta-gamma coupling during episodic retrieval in the human EEG
.
Brain Research
,
1577
,
57
68
.
Kuhn
,
M.
,
Wolf
,
E.
,
Maier
,
J. G.
,
Mainberger
,
F.
,
Feige
,
B.
,
Schmid
,
H.
, et al
(
2016
).
Sleep recalibrates homeostatic and associative synaptic plasticity in the human cortex
.
Nature Communications
,
7
,
Article 12455
.
Landsness
,
E. C.
,
Ferrarelli
,
F.
,
Sarasso
,
S.
,
Goldstein
,
M. R.
,
Riedner
,
B. A.
,
Cirelli
,
C.
, et al
(
2011
).
Electrophysiological traces of visuomotor learning and their renormalization after sleep
.
Clinical Neurophysiology
,
122
,
2418
2425
.
Lisman
,
J. E.
, &
Jensen
,
O.
(
2013
).
The theta-gamma neural code
.
Neuron
,
77
,
1002
1016
.
Liu
,
Z. W.
,
Faraguna
,
U.
,
Cirelli
,
C.
,
Tononi
,
G.
, &
Gao
,
X. B.
(
2010
).
Direct evidence for wake-related increases and sleep-related decreases in synaptic strength in rodent cortex
.
Journal of Neuroscience
,
30
,
8671
8675
.
Maris
,
E.
, &
Oostenveld
,
R.
(
2007
).
Nonparametric statistical testing of EEG- and MEG-data
.
Journal of Neuroscience Methods
,
164
,
177
190
.
Marshall
,
L.
, &
Born
,
J.
(
2007
).
The contribution of sleep to hippocampus-dependent memory consolidation
.
Trends in Cognitive Science
,
11
,
442
450
.
Marshall
,
L.
,
Helgadóttir
,
H.
,
Mölle
,
M.
, &
Born
,
J.
(
2006
).
Boosting slow oscillations during sleep potentiates memory
.
Nature
,
444
,
610
613
.
Nyhus
,
E.
, &
Curran
,
T.
(
2010
).
Functional role of gamma and theta oscillations in episodic memory
.
Neuroscience and Biobehavioral Reviews
,
34
,
1023
1035
.
Obleser
,
J.
,
Wöstmann
,
M.
,
Hellbernd
,
N.
,
Wilsch
,
A.
, &
Maess
,
B.
(
2012
).
Adverse listening conditions and memory load drive a common alpha oscillatory network
.
Journal of Neuroscience
,
32
,
12376
12383
.
Oostenveld
,
R.
,
Fries
,
P.
,
Maris
,
E.
, &
Schoffelen
,
J. M.
(
2011
).
FieldTrip: Open source software for advanced analysis of MEG, EEG, and invasive electrophysiological data
.
Computational Intelligence and Neuroscience
,
2011
,
156869
.
Osipova
,
D.
,
Takashima
,
A.
,
Oostenveld
,
R.
,
Fernandez
,
G.
,
Maris
,
E.
, &
Jensen
,
O.
(
2006
).
Theta and gamma oscillations predict encoding and retrieval of declarative memory
.
Journal of Neuroscience
,
26
,
7523
7531
.
Rasch
,
B.
, &
Born
,
J.
(
2007
).
Maintaining memories by reactivation
.
Current Opinion in Neurobiology
,
17
,
698
703
.
Rasch
,
B.
, &
Born
,
J.
(
2013
).
About sleep's role in memory
.
Physiological Reviews
,
93
,
681
766
.
Rasch
,
B.
,
Buchel
,
C.
,
Gais
,
S.
, &
Born
,
J.
(
2007
).
Odor cues during slow wave sleep prompt declarative memory consolidation
.
Science
,
315
,
1426
1429
.
Sederberg
,
P. B.
,
Kahana
,
M. J.
,
Howard
,
M. W.
,
Donner
,
E. J.
, &
Madsen
,
J. R.
(
2003
).
Theta and gamma oscillations during encoding predict subsequent recall
.
Journal of Neuroscience
,
23
,
10809
10814
.
Snodgrass
,
J. G.
, &
Corwin
,
J.
(
1988
).
Pragmatics of measuring recognition memory: Applications to dementia and amnesia
.
Journal of Experimental Psychology: General
,
117
,
34
50
.
Staudigl
,
T.
, &
Hanslmayr
,
S.
(
2013
).
Theta oscillations at encoding mediate the context-dependent nature of human episodic memory
.
Current Biology
,
23
,
1101
1106
.
Stickgold
,
R.
(
2005
).
Sleep-dependent memory consolidation
.
Nature
,
437
,
1272
1278
.
Tallon-Baudry
,
C.
, &
Bertrand
,
O.
(
1999
).
Oscillatory gamma activity in humans and its role in object representation
.
Trends in Cognitive Science
,
3
,
151
162
.
Tononi
,
G.
, &
Cirelli
,
C.
(
2006
).
Sleep function and synaptic homeostasis
.
Sleep Medical Reviews
,
10
,
49
62
.
Vyazovskiy
,
V. V.
,
Cirelli
,
C.
,
Pfister-Genskow
,
M.
,
Faraguna
,
U.
, &
Tononi
,
G.
(
2008
).
Molecular and electrophysiological evidence for net synaptic potentiation in wake and depression in sleep
.
Nature Neuroscience
,
11
,
200
208
.
Wiggs
,
C. L.
, &
Martin
,
A.
(
1998
).
Properties and mechanisms of perceptual priming
.
Current Opinion in Neurobiology
,
8
,
227
233
.
Wilson
,
M. A.
, &
McNaughton
,
B. L.
(
1994
).
Reactivation of hippocampal ensemble memories during sleep
.
Science
,
265
,
676
679
.
Yuval-Greenberg
,
S.
,
Tomer
,
O.
,
Keren
,
A. S.
,
Nelken
,
I.
, &
Deouell
,
L. Y.
(
2008
).
Transient induced gamma-band response in EEG as a manifestation of miniature saccades
.
Neuron
,
58
,
429
441
.