Skip to main content
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2017 Aug 22.
Published in final edited form as: Curr Biol. 2016 Jul 28;26(16):2127–2136. doi: 10.1016/j.cub.2016.06.044

Feedback-controlled transcranial alternating current stimulation reveals functional role of sleep spindles in motor memory consolidation

Caroline Lustenberger 1, Michael R Boyle 1,2, Sankaraleengam Alagapan 1, Juliann M Mellin 1, Bradley V Vaughn 3, Flavio Fröhlich 1,2,3,4,5
PMCID: PMC4996680  NIHMSID: NIHMS798575  PMID: 27476602

Summary

Transient episodes of brain oscillations are a common feature of both the waking and sleeping brain. Sleep spindles represent a prominent example of a poorly understood transient brain oscillation that is impaired in disorders such as Alzheimer’s disease and schizophrenia. Yet, the causal role of these bouts of thalamo-cortical oscillations remains unknown. Demonstrating a functional role of sleep spindles in cognitive processes has so far been hindered by the lack of a tool to target transient brain oscillations in real-time. Here, we show for the first time selective enhancement of sleep spindles with non-invasive brain stimulation in humans. We developed a system that detects sleep spindles in real-time and applies oscillatory stimulation. Our stimulation selectively enhanced spindle activity as determined by increased sigma activity after tACS application. This targeted modulation caused significant enhancement of motor memory consolidation that correlated with the stimulation-induced change in fast spindle activity. Strikingly, we found a similar correlation between motor memory and spindle characteristics during the sham night for the same spindle frequencies and electrode locations. Therefore, our results directly demonstrate a functional relationship between oscillatory spindle activity and cognition.

Introduction

Oscillatory patterns are fundamental to the organization of thalamo-cortical activity and are conserved across species [1, 2]. The presence of oscillations at different frequencies is dynamically regulated as a function of overall behavioral state and moment-to-moment fluctuations in cognitive demands [14]. The transient occurrence of pronounced rhythmic activity is commonly observed in recordings of cortical network dynamics. However, the causal role of the dynamic occurrence of brain oscillations remains poorly understood. Most prominently, sleep spindles are transient EEG oscillations between 11 and 16 Hz [5, 6]. The functional role of sleep spindles in cognitive processes has been hypothesized but not yet been directly demonstrated [7, 8]. Besides the issue that the majority of previous studies on the role of sleep spindles are based on correlations between sleep spindles and memory consolidation, the few studies that manipulated sleep using tones, electrical stimulation or pharmacology increased sleep spindles indirectly by enhancing slow oscillations/slow wave sleep [913]. This fundamental gap in our understanding of these thalamo-cortical oscillations is the result of the lack of a tool to monitor and selectively enhance transient epochs of oscillatory activity in real-time in humans. Transcranial alternating current stimulation (tACS) applies a weak electrical current to the scalp and recent evidence demonstrates that tACS is capable of inducing frequency-specific effects on brain dynamics [1419] and can be used to identify the functional role of brain oscillations in cognition [1922]. Yet no approach to selectively target transient oscillations has been described. Animal studies and computational models showed that the effectiveness of transcranial electrical stimulation (tES) relies on the internal network dynamics; therefore stimulation paradigms that resemble the temporal structure of endogenous activity patterns are the most effective paradigms [16, 2327]. Based on these findings, we hypothesized that real-time detection of transient oscillations that triggers short epochs of tACS resembling the targeted endogenous oscillation provides a means to boost transient oscillations. Sleep spindles represent the ideal target oscillation to apply this approach for several reasons: (1) Sleep spindles are clearly defined and dominant distinct oscillations during non-rapid eye movement (NREM) sleep that can be targeted in real-time (2) So far, no approach was described that enhanced sleep spindle activity without increasing other sleep oscillations or the time spent in specific sleep stages [1012] (3) Their proposed role in cognitive processes such as memory consolidation still needs to be demonstrated, and (4) Several psychiatric and neurologic disorders are hallmarked by sleep spindle deficits, such as Alzheimer’s disease [28], autism [29] and schizophrenia [3034]. We used an EEG feedback-controlled approach that restricts the application of tACS (FB-tACS) in the spindle frequency range to when a sleep spindle during NREM sleep is detected and therefore only enhances neuronal networks when spindle activity is prevailing. We further performed two learning paradigms (a declarative word-pair and procedural motor sequence tapping task) that have typically been used to demonstrate sleep-dependent memory consolidation [8, 10]. This approach enabled us to ask the question if sleep spindles play a causal role in memory consolidation. This is a question of significant translational relevance given the number of neurological and psychiatric conditions associated with memory impairment [33, 34]. We found that spindle FB-tACS caused an enhancement of cortical synchronization in the spindle frequency range that intensified the spindling process and improved memory consolidation.

Results

16 male participants underwent a screening night and thereafter completed two study nights (randomized, counter-balanced crossover design), one with spindle FB-tACS (verum) and one without stimulation (sham). During both study nights participants performed an associative word-pair (declarative) and motor sequence tapping task (procedural) in the evening and were retested in the morning to assess sleep-dependent memory consolidation. All-night polysomnographic recordings (8 h, EEG, EOG, and EMG) were collected. Participant-adapted thresholds based on spectral power values and spindle characteristics obtained during the screening night EEG (Fz-CPz) were used to simultaneously evaluate in real-time if (1) the participant was in NREM sleep and (2) spindle activity reached an individually defined threshold (Figure 1A, Figure S1, and Supplemental Experimental Procedures) during the study nights. If (1) and (2) were met, short epochs of alternating currents with a spindle-like waveform were applied bi-frontally during the verum condition (1 mA 12 Hz sine wave, 1-s duration at maximum amplitude, 0.25-s linear ramp up and 0.25-s linear ramp down, Figure 1B). Each stimulation was followed by a 6.5 s timeout (no stimulation even if spindles are present). Our electrode montage resulted in a stimulation that encompassed broadly frontal and centro-parietal regions (Figure 2). Participants were successfully blinded to stimulation condition, as the 2 participants that reported sensation of electrical stimulation did so during the sham night. One subject was excluded from stimulation-related EEG analysis due to bad signal quality (see Supplemental Experimental Procedures).

Figure 1. Feedback-controlled spindle tACS.

Figure 1

(A) Graphical representation of real-time spindle detection and feedback-controlled transcranial current stimulation. (B) Schematic of tACS current source and stimulation electrode configuration; stimulation electrode placement according to International 10–20 locations. See also Figure S1.

Figure 2. Electric field modeling of used electrode montage.

Figure 2

Electrodes were mounted over F3, F4 and Cz (return electrode). Field modeling of 1mA tACS was performed using HDExplore v2.3 (Soterix Medical Inc.). This electrode montage led to a broad field distribution over frontal to parietal regions with greatest magnitude of the electric field localized to areas underneath and between the electrodes. Left: Axial view (MNI position of white circle {-6,-3, 50}), Right: Sagittal view (MNI position of white circle {-7,-10, 32}). L: Left, R: Right, A: Anterior, P: Posterior.

TACS was restricted to NREM episodes with prevailing sleep spindle activity

Our spindle detection algorithm led to tACS application solely when sleep spindle activity was prevailing as illustrated in Figure 3. In all participants spindle activity was significantly higher at and around the algorithm spindle detection time-point (“stimulation onset”) compared to the rest of the epoch as verified by the Hilbert amplitude between 11–16 Hz during the sham nights (Figure 3C). Furthermore, combining the NREM and sigma threshold detection allowed for a successful identification of prevailing spindle activity during NREM sleep with a negligibly low number of stimulations during REM or wakefulness (Figure S1E and Table S1).

Figure 3. Spindle FB-tACS only applies tACS when spindle activity (11–16 Hz) is prevailing in the EEG.

Figure 3

(A) Single EEG trace of a representative participant with a detected spindle (pink dashed line) using our online spindle detection algorithm. The algorithm we used detected a spindle using two criteria. The first one was the spindle activity threshold and the second one was the number of peaks above this threshold. When the algorithm detected 5 peaks above the threshold (refers to time point 0 ms, pink dashed line), stimulation started for 1.5 s in the verum condition. Online spindle detection was used to control the stimulation start ensuring cortical stimulation exclusively during NREM spindles. Trace was obtained from a sham night (Fz-CPz, only triggering, no stimulation).

(B) Spectrogram of Fz-CPz of a representative participant during sham night shows that tACS triggers were present during sleep spindles as indicated by increased spindle activity (10–16 Hz) around 0 (represents onset of tACS for verum condition).

(C) Spindle (11–16 Hz) Hilbert amplitude averaged spindle triggers of Fz-CPz during sham night. Each line represents a participant (n = 16). Lower panel illustrates within-subject statistics. An unpaired one-sided t-test (right-tailed) was performed for the spindle Hilbert amplitude at each time-point of the illustrated epoch to the overall mean of the epoch (-2.5 to 7.5 s around trigger) for all correct NREM spindle triggers. Grey-black colored bars illustrate the number of participants showing significant increased spindle amplitude at the respective time-point compared to the mean of the whole epoch. Around 0 ms (“stimulation onset”) all participant showed significantly increased, prevailing spindle activity compared to the rest of the epoch.

See also Tables S1 and S4.

Spindle FB-tACS improved motor memory consolidation

We found superior motor memory consolidation (absolute overnight difference, Figure 4A) assessed by speed for correct trials (reduction in response time, a measure for the tapping time between key presses) after spindle FB-tACS (−21.01 ± 5.72 ms) compared to sham (−10.97 ± 7.69 ms; robust linear mixed model factor condition: F(1,11.8) = 5.7, p = 0.035). 12 of 16 participants (responders) showed this beneficial effect of spindle FB-tACS on motor memory consolidation (Figure 4B). This effect was not driven by baseline performance differences since the response time in the evening was not different between sham and verum conditions (factor condition: F(1,11.8) = 0.0, p = 0.97). Furthermore, the reported motor sequence speed gains cannot simply be explained by an improvement in attentional reaction time, as performance in a psychomotor vigilance task was not significantly affected by stimulation (q.v. Supplemental Experimental Procedures, Table S2). Number of errors and number of correctly tapped sequences were not affected by stimulation (Figure S2; all p for factor condition > 0.1). Number of correctly tapped sequences has previously been used as a measure for speed [10, 35, 36]. However, this measure likely assesses both accuracy and speed, because it is dependent on number of errors (accuracy) and response time (speed). Indeed, we found that overnight changes in correctly tapped sequences was negatively correlated with number of errors (pooled data for both conditions, r(30) = −0.59, p<0.001). In addition, we found that decreased response time (increase in speed) across the sleep period was related to an increase in the number of correctly tapped sequences (pooled data for both conditions, r(30) = −0.52, p<0.005). Of note, speed and accuracy were not significantly correlated and therefore represent two independent components (pooled data for both conditions, r(30) = −0.25, p>0.1). Hence, it is important to separate those two components of motor learning because they might be differentially affected by stimulation. Thus, stimulation effects might be masked if combination measures (e.g. number of correct sequences) are used. To confirm that the speed aspect of the number of correctly tapped sequences was also significantly affected by stimulation condition, we further controlled the robust linear mixed model (dependent variable: number of correctly tapped sequences) for accuracy by including number of errors as a covariate. This corrected model indeed revealed a significant effect of stimulation condition on number of correctly tapped sequences (F(1,10.9) = 5.17, p = 0.04) further confirming that specifically speed was significantly modulated by FB-tACS.

Figure 4. Spindle FB-tACS increases motor sequence tapping speed (response time).

Figure 4

(A) Spindle FB-tACS caused superior speed improvement (reduction in response time) compared to a night with sham condition as verified with a robust linear mixed model analysis (* n = 16, F(1,11.8) = 5.7, p = 0.035). Bars illustrate mean + s.e.m.

B) Difference of overnight speed gain (verum – sham) for each individual. Black bars illustrate participants with superior overnight speed gain during verum compared to sham (responders, n = 12) and grey bars indicate participants with inferior overnight speed gain during verum compared to sham (non-responders, n = 4).

See also Figure S2, Table S2.

Spindle FB-tACS had no effect on declarative memory (difference in number of recalled word-pairs: sham: 8.00 ± 1.23 words; verum: 7.94 ± 1.07 words; F(1,11.8) = 0.00, p = 0.97). Collectively, spindle FB-tACS improved sleep-related gains in motor sequence tapping speed but had no influence on motor sequence accuracy or declarative memory.

Spindle FB-tACS had no effect on sleep architecture but increased post-stimulation spindle activity

Given this beneficial effect of FB-tACS on motor memory consolidation, we next investigated whether FB-tACS enhanced sleep spindle activity. We hypothesized that a selective enhancement of spindle activity by stimulation was the underlying mechanism of this memory improvement. We first excluded the possibility that overall effects on the macroscopic structure of sleep could account for the effect on memory. None of the time spent in individual sleep stages or total sleep time were significantly different between the sham and verum conditions (Table 1, all p of factor condition > 0.1). Furthermore, also no significant effect of condition on sleep architecture was found if only motor memory responders (n=12) were included (all p of factor condition > 0.1). Due to the pronounced stimulation artifact (within ~2s around tACS start and around 4.3 – 6s (caused by internal source switching in the stimulator), Figure S3, see Supplemental Experimental Procedures for details) analysis was only possible in a stimulation free interval. Thus, we then examined how short epochs of 12Hz-tACS affected the NREM sleep EEG in a short stimulation-free interval after the tACS artefact (2 – 4.3 s after stimulation onset, see Supplemental Experimental Procedures for details) compared to sham condition (only spindle detection trigger, no tACS applied). For this analysis, 15 out of 16 participants were included due to unusable EEG for one participant (see Supplemental Experimental Procedures). We performed the analysis separately for NREM sleep stage 2 (N2) and 3 (N3) to account for number of included trials, light (N2) and deep sleep (N3), and different thalamic hyperpolarization levels (see Supplemental Experimental Procedures for details). Spindle FB-tACS led to a broad increase in spindle activity around 11–16 Hz only in N2 averaged over all electrodes, with motor memory responders (n = 11) showing an increase in very fast spindle frequencies (15 – 16 Hz) compared to non-responders (n = 4, show decrease; Figure 5). Besides a selective increase in spindle activity, our stimulation also significantly reduced power in the delta and theta range in N2 (Figure 5) and N3 (Figure S4). Since sigma activity overlaps with alpha activity during wakefulness one might argue that the stimulation leads to arousal that could explain an increase in sigma activity. However, our results clearly show that this is not the case: (1) wakefulness alpha is between 8–12 Hz whereas our increase in spindle activity is between 12–16 Hz (Figure S5) (2) The spectrogram after the stimulation has a similar profile for sham and verum epochs, looking clearly different from a typical wakefulness (eyes closed) period, and (3) number of wakefulness periods and perceived sleep depth were not significantly different between conditions (Table 1 and Table S2).

Table 1.

Sleep architecture comparison between sham and verum condition (n = 16)

Sham Verum Statistics
Mean (SEM) Mean (SEM) Factor Condition (p) Interaction
Condition × Session (p)
Factor
Session (p)
Total sleep time (min) 447.0 (4.5) 447.6 (4.1) >0.1 >0.1 >0.1
Sleep efficiency (%) 93.1 (0.9) 93.3 (0.9) >0.1 >0.1 >0.1
Sleep latency (min) 10.6 (2.0) 12.8 (3.2) >0.1 >0.1 >0.1
WASO (%) 5.3 (0.8) 4.6 (0.6) >0.1 >0.1 >0.1
Stage 1 (%) 3.4 (0.6) 3.1 (0.3) >0.1 >0.1 >0.1
Stage 2 (%) 50.2 (1.7) 49.9 (1.9) >0.1 >0.1 >0.1
Stage 3 (%) 19.4 (1.7) 18.7 (1.7) >0.1 >0.1 >0.1
NREM sleep (%) 69.6 (1.3) 68.6 (1.2) >0.1 >0.1 >0.1
REM sleep (%) 20.1 (0.9) 21.6 (0.9) >0.1 >0.1 >0.1

Figure 5. FB-tACS increases spindle activity during NREM stage 2 sleep that is related to stimulation-induced motor sequence tapping speed gains.

Figure 5

(A) Difference of spectrograms (verum – sham) averaged for all channels for longest artefact free interval during NREM stage 2 (N2, 2 – 4.3 s) and corresponding p-values of a paired t-test between sham and verum condition (p values > 0.1 are black, pink rectangles highlight window with increased spindle activity).

(B) Detailed analysis of increased spindle activity window during N2 (11 – 16 Hz, pink window in A). Spectrogram values were averaged over time for the selected time window and plotted for each frequency bin and channel. This analysis was done for responders (n = 11, superior speed gain in motor sequence task for verum condition compared to sham) and non-responders separately (n = 4).

(C) Topographical representation of Pearson correlation coefficients between the spindle activity difference (n = 15; pooled for responders and non-responders) for 15–16 Hz (black rectangle in B) with the difference (verum – sham) in overnight speed gain (Figure 4B). Superior speed gain in verum condition compared to sham is reflected in a negative number as superior speed means reduced response time. Thus, negative correlation coefficients show that more spindle activity increase is related to a more pronounced sleep-dependent response time decrease (speed increase) in the verum condition compared to sham. Electrodes (black dots) that showed a significant correlation (Pearson) are marked with grey dots (p < 0.05) and electrodes that showed a trend-level with white dots (p ≥ 0.05 and p < 0.1). The size of the cluster (4 neighboring electrodes with grey and white dots) was significant after performing a supra-threshold cluster analysis (see Supplemental Experimental Procedures).

See also Figures S3-S6 and Table S3.

FB-tACS induced enhancement of spindle activity predicted improvement in motor memory consolidation

In order for sleep spindle activity to promote motor memory speed gains, the stimulation-induced increase in spindle activity should be related to the improvement in motor memory consolidation. Given that non-responders and responders mainly differed in spindle activity increase for very fast frequencies (15–16 Hz) we restricted our correlation analysis to this frequency window. Indeed, we found a significant negative correlation between the verum-related change in response time and spindle activity for the very fast spindle frequency range indicating that the increase in fast sleep spindle activity predicted reduction in tapping time (increase in speed) due to verum stimulation (Figures 5C). This negative correlation was found globally but only reached trend-level or significance for mainly parietal and occipital electrodes (Pearson correlation of merged parieto-occipital cluster (4 electrodes): r(13) = −0.65, p = 0.009, cluster survives supra-threshold cluster analysis, see Supplemental Experimental Procedures). No significant correlation was found for delta-theta activity reduction in tapping time due to verum stimulation (Figure S6). The number of applied stimulations during the verum night (Table S3) was not related to the FB-tACS related motor memory improvement (r (14) = −0.09, p = 0.75). Considering that we only encountered a spindle increase in sleep stage 2, we further performed the same analysis including only the number of stimulations during stage 2. Again, no significant correlation was found (r(14) = −0.05, p = 0.86).

Spindle characteristics and sleep-dependent motor memory consolidation are similarly correlated during the sham night

To further confirm the role of fast sleep spindles in motor memory consolidation, we finally examined whether a similar relationship exists between motor memory consolidation and different NREM sleep spindle characteristics (e.g. density) in the absence of stimulation (sham). Overnight change in response time was negatively correlated with spindle density and duration, again for the same frequency bins (15–16 Hz for density and 14.5–16 Hz for duration) and posterior electrodes (Figure 6). This finding convincingly confirms that characteristics of fast spindles, specifically density and duration, are important for sleep-dependent motor memory consolidation.

Figure 6. Relationship between sleep-dependent motor memory consolidation and spindle characteristics in absence of stimulation.

Figure 6

Two dimensional hierarchical cluster trees (dendrogramm) and heat plots of the r-values of the correlation between sleep – dependent reduction in response time (speed gain) during the sham night and (A) spindle density, and (B) spindle duration. Colored branches illustrate clusters with an Euclidean distance below 1.3. Negative correlation coefficients show that more pronounced appearance of the respective spindle characteristic was reflected in sleepdependent response time decrease (speed increase). Right column illustrates corresponding correlation coefficient (r) topographical plots of clustered frequency bands (based on clustering in dendrogram). Electrodes (black dots) that showed significant correlations (Pearson correlation) are marked with grey dots (p < 0.05) and electrodes that showed a trend-level with white dots (n = 16, p ≥ 0.05 and p < 0.1). The size of the cluster in A (6 neighboring electrodes with grey and white dots) was trend-level after performing a suprathreshold cluster analysis (see Supplemental Experimental Procedures), the cluster in B (8 neighboring electrodes) was significant.

Discussion

We established a successful framework to investigate the functional role of specific transient brain oscillations in cognitive processes by applying targeted, individualized and feedback-controlled weak electrical brain stimulation. We found that spindle FB-tACS can enhance sleep spindle activity in a broad frequency range during NREM stage 2 sleep without increasing other sleep rhythms or time spent in individual sleep stages. Furthermore, spindle FB-tACS enhanced motor sequence consolidation by means of increased speed, and fast sleep spindle activity played a functional role in this gain. We therefore provide the first direct demonstration of the functional role of sleep spindle activity in motor memory consolidation.

Sleep spindles have previously been hypothesized to benefit memory formation [8]. For instance, sleep-dependent improvements in declarative and procedural learning paradigms correlated with sleep spindle characteristics [35, 37, 38]. Furthermore, spindles were increased during sleep following the training of these learning paradigms compared to a control condition [3942]. In further support of a central role of sleep spindles in memory processes, patients with schizophrenia show a pronounced reduction in sleep spindles that correlates with deficits in sleep-dependent motor memory consolidation [33, 34]. However, these studies were restricted to correlations leaving it unclear whether learning-associated changes in sleep spindle dynamics are an epiphenomenon or indeed play a functional role in memory consolidation. Previous attempts in manipulating sleep in humans (e.g. auditory stimulation, pharmacology or slow-oscillatory direct current stimulation) were only successful in enhancing sleep spindles as a side effect of enhancing slow oscillations [9, 10, 12, 13] or the time spent in sleep stages, such as slow wave sleep [11]. In addition, tES approaches so far only enhanced declarative memory but failed to improve procedural tasks [43] even though one of the studies reported increases in sleep spindle measures along with enhanced slow oscillations/slow wave sleep [10]. A possible explanation for this missing effect on procedural memory is that the reported significant increase in sleep spindles were only found for slow-frequency spindles but not for the fast spindles [10]. In addition all studies using tES to modulate NREM sleep and enhance memory consolidation applied either slow-oscillatory tDCS/ACS (0.75 Hz) or tDCS [43] and were therefore not optimized to selectively target sleep spindles. We are the first here to selectively enhance sleep spindle activity along with motor memory consolidation using FB-tACS throughout nocturnal sleep and therefore provide a functional role of these oscillations in cognitive processes.

Spindle FB-tACS specifically enhanced sleep-dependent speed gains and not accuracy in a motor sequence tapping paradigm reflected in a significant decrease of response time, but not error rate. This is in accordance with previous studies that mainly found a robust effect of sleep on speed measures, e.g. [10, 11, 36, 44, 45]. However, most of these studies used number of correct sequences per trial as a measure for speed. Our results revealed that the number of correct trials is not independent of the error rate and therefore relates to the accuracy of the performance. In addition, some studies also indicate a beneficial effect of sleep on the error rate (accuracy) [46]. In other words, changes/variations in error rate might be reflected in the number of correct sequences and could therefore mask/confound sleep and intervention condition effects on speed measures. By including error rate as a covariate in our model, stimulation condition had a significant effect on number of correctly tapped sequences, showing that spindle FB-tACS selectively enhanced sleep-dependent speed benefits but not accuracy. Collectively, our findings argue for the use of more “pure” measures of speed in motor sequence tapping tasks, e.g. by focusing on the response time of correctly tapped sequences or controlling for the error rate in future models. Of note, future studies will be needed to investigate more complex “real life” motor tasks that benefit from sleep and to relate those findings to sleep spindles/FB-tACS. To elucidate the specific changes sleep spindles have on motor memory the use of simple motor tasks can be of advantage. In our case, the task design enabled us to differentiate speed from accuracy. Nevertheless, from a translational point of view, it remains to be investigated whether “real-life” memory impairments could benefit from our stimulation approach.

It has recently been shown that tACS effects on cognitive performance are dependent on basal cognitive performance [47]. Indeed, we also see a significant negative correlation (r (14) =−0.86, p = 0.004) for sham motor memory consolidation and FB-tACS related increases, indicating that the more participants already benefit from sleep in motor sequence memory during the sham night the less they further improve due to FB-tACS stimulation. However, as Santarnecchi et al. [47] address in their publication, we cannot rule out that the correlation we see is confounded by regression to the mean. Future studies are needed to support this specific finding.

Analysis of spindle activity during tACS was not possible due to the stimulation artifact. However, spindle activity was enhanced in a specific window after tACS. We hypothesize that our stimulation increases properties of the next spindle that followed the initially detected/and stimulated one (either probability of having a spindle, duration or amplitude of the spindle). Intriguingly, the increase starts around 3.5 s after the previous spindle started. According to literature, sleep spindles seem to have a prominent reoccurrence rate of around 3–5 s [4850]. This also points to the idea that the spindle following the stimulation might have been affected or more likely to be started. Since we have clear analysis limitations due to strong, multiple artifacts in our verum night (and therefore not a continuous recording that can be investigated), we further investigated correlations of motor memory benefits and spindle characteristics during the sham night. Our findings show that mainly spindle density and duration are correlated with the sleep benefit on motor sequence tapping speed, which gives a first hint that changes of sleep spindle duration and/or density might have given rise of our verum related motor memory improvement. Further studies are needed to determine the effect of our stimulation on the stimulated spindle itself and specific spindle characteristics in a dataset that can be continuously analyzed. This might be achieved by performing animal experiments with intracranial electrodes where spiking activity can be obtained during stimulation (e.g. [23]) or by developing and applying highly sophisticated tACS artifact removal algorithms that can perfectly separate neuronal activity from artifact in the same frequency range. Long-lasting effects of tACS have been reported, but previous reports did not find any long-lasting effects of tACS during wakefulness for very short applications of tACS (e.g. 1 s [51, 52]). However, they focused on alpha activity during wakefulness and a different mechanism might hold true for sleep. A recent report from our group [53] showed stimulation effects that outlast stimulation depended on the specific state in wakefulness. Moreover, using a computational model, the study suggested that recurrent connections in thalamo-cortical and cortico-cortical loops led to outlasting effects of stimulation. A similar mechanism is plausibly in play in our current study. During sleep, short stimulation bursts may activate the thalamo-cortical loop which leads to a more pronounced subsequent spindle or increase the likelihood of occurrence of the next spindle after stimulation. Along this line we could speculate that we only see a spindle activity increase in N2 but not N3 because the stimulation was not strong enough to modulate thalamo-cortical system properties during N3 since this state is characterized by more delta/slow waves and less sleep spindles than N2 [5, 54]. Future studies in animal models will be needed to explore spindle changes after tACS and to establish the mechanism of this tACS induced after effect in spindle activity using cortical and subcortical invasive recordings. The stimulation-induced overnight gains in motor sequence learning were mediated by the fast sleep spindle activity which is in line with previous literature showing a correlation of motor memory exclusively with fast spindle characteristics [39]. We replicate this correlation in the sham night with different spindle characteristics and found the same frequency bins and electrodes of spindle density and duration significantly correlating with motor memory consolidation. Several studies hypothesized that slow and fast frequency spindles might serve different functions [5, 39, 55]. Each spindle type shows a different topography with slower spindle frequencies (around 12 Hz) being preferentially visible over frontal areas whereas fast sleep spindles (around 14 Hz) are more pronounced over centro-parietal regions [5, 54, 56, 57]. Considering that our correlations with motor memory consolidation were restricted to the fast spindle frequency range, our results underline the assumption that slow and fast sleep spindles might serve different functions. Therefore, our results highlight the importance of separating slow and fast frequencies for future analyses of sleep spindles. The correlation between spindle activity and memory consolidation is seen mainly over posterior regions. Of note, the negative correlations are global but only reach significance in the posterior regions. Thus it is important to take our spatial findings with caution. Fronto-parietal regions (e.g. SMA, M1, precuneus), cerebellum, and basal ganglia are implicated in motor sequence learning [58, 59]. Interestingly, frontal regions seem to be more activated during early stages of learning while parietal regions seem to get more involved during later stages [58, 60]. In addition, Shadmehr and Holcomb showed that hours after performing a visuomotor task (consolidation), there is a shift in the neural representation of the internal model by means of activation in posterior parietal regions instead of frontal regions [61]. Thus, our correlation which is strongest at parietal regions might point to the main parietal involvement in motor sequence consolidation. Yet, it is still unclear which specific cortical regions might be involved in sleep-dependent memory consolidation. However, based on our findings future studies should target specifically posterior brain regions using faster frequencies (e.g. 15 Hz tACS) to optimally benefit motor memory consolidation.

In addition, the question arises whether synchronization of frontal oscillatory activity might play a role for the efficacy of our applied stimulation on behavior. It still remains to be investigated whether spindle synchronized across cortical regions are essential for memory consolidation to occur or whether only spindles localized to brain regions involved in performing the task are necessary. Future studies are needed to investigate this idea by e.g. using only left or right hemispheric stimulation, or out of phase stimulation.

Besides sleep spindles, slow waves have been proposed to play an important role in memory consolidation [8]. However, we found a superior sleep-dependent speed gain for verum condition despite a spindle FB-tACS induced decrease in delta and theta power, pointing to a limited role of slow waves in this specific process. Along this line, some studies have suggested a role of spindles in motor memory consolidation in dissociation from effects mediated by the slow waves. Using tones to reduce slow waves and REM sleep without changing sleep spindles, Genzel et al. [62] were able to preserve the consolidation of procedural and declarative memory. Enhancing slow wave activity (SWA, 0.5 – 4 Hz) but decreasing spindle activity using the GABA reuptake inhibitor Tiagabine led to diminished memory consolidation in a motor sequence tapping task [63]. Finally, patients with schizophrenia who show reduced motor sequence consolidation also exhibit a pronounced decrease in sleep spindles with negligible changes in slow wave activity [3234, 64]. Here we show that selective spindle enhancement had no effect on declarative memory consolidation despite this hypothesis from previous studies [8]. A possible explanation for this missing effect could be the reduction in delta activity because sleep spindles might only be beneficial for this memory type in combination with slow waves [65]. However, further studies are needed to delineate the importance of coalescence of slow waves and spindles for declarative memory, e.g. by applying spindle tACS time-locked to slow wave up-states.

Our results further suggest that sleep spindles and slow waves cannot be independently modulated. Along this line, previous studies have shown that specific sleep spindle characteristics and slow waves are inversely related [5, 54, 6669]. For instance, spindle density and spindle frequency are reduced in early NREM sleep, in the middle of NREM cycles and N3 when SWA is maximal [54]. Further studies underlining this notion found less spindle activity in the recovery night after sleep deprivation that is marked by increased SWA [5, 68] or reported negative correlations between spindle measures (e.g. sigma activity) and SWA during NREM sleep [69, 70]. Our results further support and extend the notion that SWA and sleep spindles share a reciprocal relationship.

Collectively, spindle FB-tACS revealed the functional relationship between fast sleep spindles and motor memory consolidation. Thus, our findings serve as an important starting point to develop neuro-therapeutics for treating motor memory impairments afflicting patients with psychiatric and neurological disorders [33, 34, 64] and older individuals [71]. Future studies, however, are needed to further find optimized stimulation parameters by means of ideal stimulation location (centro-parietal instead of frontal) and (spindle) frequency applied (e.g. 15 Hz instead of 12 Hz). In addition, future research is needed to evaluate whether other frequencies temporally far from spindle activity worsen the performance. In a broader context, our results provide convincing evidence that targeted and individualized stimulation approaches are fundamental for selectively boosting transient brain oscillations. Furthermore, our study provides a model paradigm for establishing the functional role of transient brain oscillations in human behavior. Our FB-tACS design is a radical departure from the former stimulation approach because it takes individual, endogenous network activity into account. Stimulation success likely depends on the underlying network activity as has been convincingly shown in in-vivo, in-vitro and computational studies [16, 2326]. This is why feedback-controlled approaches provide a promising starting point for individualized treatment paradigms that successfully target pathological network dynamics with non-invasive brain stimulation.

Supplementary Material

Highlights.

  • Feedback-controlled tACS (FB-tACS, 12 Hz) boosted subsequent sleep spindle activity

  • FB-tACS enhanced sleep-dependent motor but not declarative memory consolidation

  • Stimulation-induced fast spindle activity changes predicted motor memory benefits

  • Correlation of spindles and motor memory in sham session agree with FB-tACS results

In Brief.

Lustenberger et al. engineered a novel feedback-controlled spindle stimulation approach that selectively targeted and modulated sleep spindles in real-time. This approach revealed, for the first time, that fast sleep spindles play a functional role in motor memory consolidation.

Acknowledgments

We thank Trisha Burrello and Steve Ferrin for helping in data acquisition, and Kyle Kalkowski for his helpful contributions to the study. We are grateful to Dr. John Gilmore, Dr. Franz Hamilton and Kristin Sellers for critical reading of the manuscript. Research reported in this publication was partially supported by the National Institute of Mental Health under Award Number R01MH101547 (to F.F.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. This work was also partially supported by UNC Psychiatry, UNC School of Medicine (to F.F.), and the Swiss National Science Foundation (to C.L., grant P2EZP3-152214). UNC has filed provisional patents on tACS-related technology with Flavio Frohlich as the lead inventor. No licensing has occurred. Flavio Frohlich is the founder and majority shareholder of Pulvinar Neuro LLC.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Author Contributions

C.L., B.V. and F.F. designed the study. C.L, M.B and F.F. designed online spindle detection algorithm and feed-back controlled stimulation. C.L., M.B. and J.M. conducted data collection. C.L., M.B. and S.A performed data analysis. All authors contributed to writing the manuscript.

References

  • 1.Buzsaki G, Logothetis N, Singer W. Scaling brain size, keeping timing: evolutionary preservation of brain rhythms. Neuron. 2013;80:751–764. doi: 10.1016/j.neuron.2013.10.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Buzsaki G, Draguhn A. Neuronal oscillations in cortical networks. Science. 2004;304:1926–1929. doi: 10.1126/science.1099745. [DOI] [PubMed] [Google Scholar]
  • 3.Harris KD, Thiele A. Cortical state and attention. Nat Rev Neurosci. 2011;12:509–523. doi: 10.1038/nrn3084. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Lee SH, Dan Y. Neuromodulation of brain states. Neuron. 2012;76:209–222. doi: 10.1016/j.neuron.2012.09.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.De Gennaro L, Ferrara M. Sleep spindles: An overview. Sleep Med Rev. 2003;7:423–440. doi: 10.1053/smrv.2002.0252. [DOI] [PubMed] [Google Scholar]
  • 6.Warby SC, Wendt SL, Welinder P, Munk EG, Carrillo O, Sorensen HB, Jennum P, Peppard PE, Perona P, Mignot E. Sleep-spindle detection: crowdsourcing and evaluating performance of experts, non-experts and automated methods. Nat Methods. 2014;11:385–392. doi: 10.1038/nmeth.2855. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Fogel SM, Smith CT. The function of the sleep spindle: a physiological index of intelligence and a mechanism for sleep-dependent memory consolidation. Neurosci Biobehav Rev. 2011;35:1154–1165. doi: 10.1016/j.neubiorev.2010.12.003. [DOI] [PubMed] [Google Scholar]
  • 8.Rasch B, Born J. About Sleep's Role in Memory. Physiol Rev. 2013;93:681–766. doi: 10.1152/physrev.00032.2012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Del Felice A, Magalini A, Masiero S. Slow-oscillatory transcranial direct current stimulation modulates memory in temporal lobe epilepsy by altering sleep spindle generators: A possible rehabilitation tool. Brain stimulation. 2015;8:567–573. doi: 10.1016/j.brs.2015.01.410. [DOI] [PubMed] [Google Scholar]
  • 10.Marshall L, Helgadottir H, Molle M, Born J. Boosting slow oscillations during sleep potentiates memory. Nature. 2006;444:610–613. doi: 10.1038/nature05278. [DOI] [PubMed] [Google Scholar]
  • 11.Mednick SC, McDevitt EA, Walsh JK, Wamsley E, Paulus M, Kanady JC, Drummond SP. The Critical Role of Sleep Spindles in Hippocampal-Dependent Memory: A Pharmacology Study. J Neurosci. 2013;33:4494–4504. doi: 10.1523/JNEUROSCI.3127-12.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Ngo HV, Martinetz T, Born J, Molle M. Auditory Closed-Loop Stimulation of the Sleep Slow Oscillation Enhances Memory. Neuron. 2013;78:545–553. doi: 10.1016/j.neuron.2013.03.006. [DOI] [PubMed] [Google Scholar]
  • 13.Westerberg CE, Florczak SM, Weintraub S, Mesulam MM, Marshall L, Zee PC, Paller KA. Memory improvement via slow-oscillatory stimulation during sleep in older adults. Neurobiology of Aging. 2015;36:2577–2586. doi: 10.1016/j.neurobiolaging.2015.05.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Boyle MR, Frohlich F. EEG feedback-controlled transcranial alternating current stimulation; Neural Engineering (NER), 2013 6th International IEEE/EMBS Conference on. (IEEE); 2013. pp. 140–143. [Google Scholar]
  • 15.Helfrich RF, Schneider TR, Rach S, Trautmann-Lengsfeld SA, Engel AK, Herrmann CS. Entrainment of brain oscillations by transcranial alternating current stimulation. Curr Biol. 2014;24:333–339. doi: 10.1016/j.cub.2013.12.041. [DOI] [PubMed] [Google Scholar]
  • 16.Schmidt SL, Iyengar AK, Foulser AA, Boyle MR, Frohlich F. Endogenous cortical oscillations constrain neuromodulation by weak electric fields. Brain Stimul. 2014;7:878–889. doi: 10.1016/j.brs.2014.07.033. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Vossen A, Gross J, Thut G. Alpha Power Increase After Transcranial Alternating Current Stimulation at Alpha Frequency (alpha-tACS) Reflects Plastic Changes Rather Than Entrainment. Brain Stimul. 2014 doi: 10.1016/j.brs.2014.12.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Herrmann CS, Struber D, Helfrich RF, Engel AK. EEG oscillations: From correlation to causality. Int J Psychophysiol. 2015 doi: 10.1016/j.ijpsycho.2015.02.003. [DOI] [PubMed] [Google Scholar]
  • 19.Herrmann CS, Rach S, Neuling T, Struber D. Transcranial alternating current stimulation: a review of the underlying mechanisms and modulation of cognitive processes. Front Hum Neurosci. 2013;7:279. doi: 10.3389/fnhum.2013.00279. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Fröhlich F. Endogenous and exogenous electric fields as modifiers of brain activity: rational design of noninvasive brain stimulation with transcranial alternating current stimulation. Dialogues Clin Neurosci. 2014;16:93–102. doi: 10.31887/DCNS.2014.16.1/ffroehlich. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Lustenberger C, Boyle MR, Foulser AA, Mellin JM, Frohlich F. Functional role of frontal alpha oscillations in creativity. Cortex. 2015;67:74–82. doi: 10.1016/j.cortex.2015.03.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Santarnecchi E, Polizzotto, Nicola R, Godone M, Giovannelli F, Feurra M, Matzen L, Rossi A, Rossi S. Frequency-Dependent Enhancement of Fluid Intelligence Induced by Transcranial Oscillatory Potentials. Current Biology. 2013;23:1449–1453. doi: 10.1016/j.cub.2013.06.022. [DOI] [PubMed] [Google Scholar]
  • 23.Ali MM, Sellers KK, Frohlich F. Transcranial alternating current stimulation modulates large-scale cortical network activity by network resonance. J Neurosci. 2013;33:11262–11275. doi: 10.1523/JNEUROSCI.5867-12.2013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Fröhlich F, McCormick DA. Endogenous electric fields may guide neocortical network activity. Neuron. 2010;67:129–143. doi: 10.1016/j.neuron.2010.06.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Ozen S, Sirota A, Belluscio MA, Anastassiou CA, Stark E, Koch C, Buzsaki G. Transcranial electric stimulation entrains cortical neuronal populations in rats. J Neurosci. 2010;30:11476–11485. doi: 10.1523/JNEUROSCI.5252-09.2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Reato D, Gasca F, Datta A, Bikson M, Marshall L, Parra LC. Transcranial electrical stimulation accelerates human sleep homeostasis. PLoS Comput Biol. 2013;9:e1002898. doi: 10.1371/journal.pcbi.1002898. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Brittain J-S, Probert-Smith P, Aziz TZ, Brown P. Tremor suppression by rhythmic transcranial current stimulation. Current Biology. 2013;23:436–440. doi: 10.1016/j.cub.2013.01.068. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Rauchs G, Schabus M, Parapatics S, Bertran F, Clochon P, Hot P, Denise P, Desgranges B, Eustache F, Gruber G. Is there a link between sleep changes and memory in Alzheimer's disease? Neuroreport. 2008;19:1159. doi: 10.1097/WNR.0b013e32830867c4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Limoges É, Mottron L, Bolduc C, Berthiaume C, Godbout R. Atypical sleep architecture and the autism phenotype. Brain. 2005;128:1049–1061. doi: 10.1093/brain/awh425. [DOI] [PubMed] [Google Scholar]
  • 30.Ferrarelli F. Sleep in Patients With Schizophrenia. Current sleep medicine reports. 2015;1:150–156. doi: 10.1007/s40675-015-0010-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Ferrarelli F, Huber R, Peterson MJ, Massimini M, Murphy M, Riedner BA, Watson A, Bria P, Tononi G. Reduced sleep spindle activity in schizophrenia patients. Am J Psychiatry. 2007;164:483–492. doi: 10.1176/ajp.2007.164.3.483. [DOI] [PubMed] [Google Scholar]
  • 32.Ferrarelli F, Peterson MJ, Sarasso S, Riedner BA, Murphy MJ, Benca RM, Bria P, Kalin NH, Tononi G. Thalamic dysfunction in schizophrenia suggested by whole-night deficits in slow and fast spindles. Am J Psychiatry. 2010;167:1339–1348. doi: 10.1176/appi.ajp.2010.09121731. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Manoach DS, Thakkar KN, Stroynowski E, Ely A, McKinley SK, Wamsley E, Djonlagic I, Vangel MG, Goff DC, Stickgold R. Reduced overnight consolidation of procedural learning in chronic medicated schizophrenia is related to specific sleep stages. J Psychiatr Res. 2010;44:112–120. doi: 10.1016/j.jpsychires.2009.06.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Wamsley EJ, Tucker MA, Shinn AK, Ono KE, McKinley SK, Ely AV, Goff DC, Stickgold R, Manoach DS. Reduced sleep spindles and spindle coherence in schizophrenia: mechanisms of impaired memory consolidation? Biol Psychiatry. 2012;71:154–161. doi: 10.1016/j.biopsych.2011.08.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Rasch B, Pommer J, Diekelmann S, Born J. Pharmacological REM sleep suppression paradoxically improves rather than impairs skill memory. Nat Neurosci. 2009;12:396–397. doi: 10.1038/nn.2206. [DOI] [PubMed] [Google Scholar]
  • 36.Walker MP, Brakefield T, Morgan A, Hobson JA, Stickgold R. Practice with sleep makes perfect: sleep-dependent motor skill learning. Neuron. 2002;35:205–211. doi: 10.1016/s0896-6273(02)00746-8. [DOI] [PubMed] [Google Scholar]
  • 37.Clemens Z, Fabo D, Halasz P. Overnight verbal memory retention correlates with the number of sleep spindles. Neuroscience. 2005;132:529–535. doi: 10.1016/j.neuroscience.2005.01.011. [DOI] [PubMed] [Google Scholar]
  • 38.Holz J, Piosczyk H, Feige B, Spiegelhalder K, Baglioni C, Riemann D, Nissen C. EEG sigma and slow-wave activity during NREM sleep correlate with overnight declarative and procedural memory consolidation. J Sleep Res. 2012;21:612–619. doi: 10.1111/j.1365-2869.2012.01017.x. [DOI] [PubMed] [Google Scholar]
  • 39.Barakat M, Doyon J, Debas K, Vandewalle G, Morin A, Poirier G, Martin N, Lafortune M, Karni A, Ungerleider LG, et al. Fast and slow spindle involvement in the consolidation of a new motor sequence. Behav Brain Res. 2011;217:117–121. doi: 10.1016/j.bbr.2010.10.019. [DOI] [PubMed] [Google Scholar]
  • 40.Gais S, Molle M, Helms K, Born J. Learning-dependent increases in sleep spindle density. J Neurosci. 2002;22:6830–6834. doi: 10.1523/JNEUROSCI.22-15-06830.2002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Johnson LA, Blakely T, Hermes D, Hakimian S, Ramsey NF, Ojemann JG. Sleep spindles are locally modulated by training on a brain-computer interface. Proc Natl Acad Sci U S A. 2012;109:18583–18588. doi: 10.1073/pnas.1207532109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Schmidt C, Peigneux P, Muto V, Schenkel M, Knoblauch V, Munch M, de Quervain DJ, Wirz-Justice A, Cajochen C. Encoding difficulty promotes postlearning changes in sleep spindle activity during napping. J Neurosci. 2006;26:8976–8982. doi: 10.1523/JNEUROSCI.2464-06.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Barham MP, Enticott PG, Conduit R, Lum JA. Transcranial electrical stimulation during sleep enhances declarative (but not procedural) memory consolidation: Evidence from a meta-analysis. Neuroscience & Biobehavioral Reviews. 2016;63:65–77. doi: 10.1016/j.neubiorev.2016.01.009. [DOI] [PubMed] [Google Scholar]
  • 44.Nishida M, Walker MP. Daytime naps, motor memory consolidation and regionally specific sleep spindles. PLoS One. 2007;2:e341. doi: 10.1371/journal.pone.0000341. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Brawn TP, Fenn KM, Nusbaum HC, Margoliash D. Consolidating the effects of waking and sleep on motor-sequence learning. J Neurosci. 2010;30:13977–13982. doi: 10.1523/JNEUROSCI.3295-10.2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Walker MP, Brakefield T, Seidman J, Morgan A, Hobson JA, Stickgold R. Sleep and the time course of motor skill learning. Learn Mem. 2003;10:275–284. doi: 10.1101/lm.58503. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Santarnecchi E, Muller T, Rossi S, Sarkar A, Polizzotto N, Rossi A, Kadosh RC. Individual differences and specificity of prefrontal gamma frequency-tACS on fluid intelligence capabilities. Cortex. 2016;75:33–43. doi: 10.1016/j.cortex.2015.11.003. [DOI] [PubMed] [Google Scholar]
  • 48.Achermann P, Borbely AA. Low-frequency (< 1 Hz) oscillations in the human sleep electroencephalogram. Neuroscience. 1997;81:213–222. doi: 10.1016/s0306-4522(97)00186-3. [DOI] [PubMed] [Google Scholar]
  • 49.Evans B, Richardson N. Demonstration pf a 3–5s periodicity between the spindle bursts in NREM sleep in man. Journal of Sleep Research. 1995;4:196–197. [Google Scholar]
  • 50.Olbrich E, Achermann P. Analysis of the temporal organization of sleep spindles in the human sleep EEG using a phenomenological modeling approach. Journal of biological physics. 2008;34:241–249. doi: 10.1007/s10867-008-9078-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Strüber D, Rach S, Neuling T, Herrmann CS. On the possible role of stimulation duration for after-effects of transcranial alternating current stimulation. Frontiers in cellular neuroscience. 2015;9 doi: 10.3389/fncel.2015.00311. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Vossen A, Gross J, Thut G. Alpha Power Increase After Transcranial Alternating Current Stimulation at Alpha Frequency (alpha-tACS) Reflects Plastic Changes Rather Than Entrainment. Brain Stimul. 2015;8:499–508. doi: 10.1016/j.brs.2014.12.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Alagapan S, Schmidt SL, Lefebvre J, Hadar E, Shin HW, Frohlich F. Modulation of Cortical Oscillations by Low-Frequency Direct Cortical Stimulation is State-Dependent. PLoS biology. 2016 doi: 10.1371/journal.pbio.1002424. In press. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Andrillon T, Nir Y, Staba RJ, Ferrarelli F, Cirelli C, Tononi G, Fried I. Sleep spindles in humans: insights from intracranial EEG and unit recordings. J Neurosci. 2011;31:17821–17834. doi: 10.1523/JNEUROSCI.2604-11.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Lustenberger C, Wehrle F, Tushaus L, Achermann P, Huber R. The Multidimensional Aspects of Sleep Spindles and Their Relationship to Word-Pair Memory Consolidation. Sleep. 2015 doi: 10.5665/sleep.4820. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.De Gennaro L, Ferrara M, Bertini M. Topographical distribution of spindles: variations between and within nrem sleep cycles. Sleep Res Online. 2000;3:155–160. [PubMed] [Google Scholar]
  • 57.Jobert M, Poiseau E, Jahnig P, Schulz H, Kubicki S. Topographical analysis of sleep spindle activity. Neuropsychobiology. 1992;26:210–217. doi: 10.1159/000118923. [DOI] [PubMed] [Google Scholar]
  • 58.Hikosaka O, Nakamura K, Sakai K, Nakahara H. Central mechanisms of motor skill learning. Curr Opin Neurobiol. 2002;12:217–222. doi: 10.1016/s0959-4388(02)00307-0. [DOI] [PubMed] [Google Scholar]
  • 59.Honda M, Deiber M-P, Ibáñez V, Pascual-Leone A, Zhuang P, Hallett M. Dynamic cortical involvement in implicit and explicit motor sequence learning. A PET study. Brain. 1998;121:2159–2173. doi: 10.1093/brain/121.11.2159. [DOI] [PubMed] [Google Scholar]
  • 60.Toni I, Krams M, Turner R, Passingham RE. The time course of changes during motor sequence learning: a whole-brain fMRI study. Neuroimage. 1998;8:50–61. doi: 10.1006/nimg.1998.0349. [DOI] [PubMed] [Google Scholar]
  • 61.Shadmehr R, Holcomb HH. Neural correlates of motor memory consolidation. Science. 1997;277:821–825. doi: 10.1126/science.277.5327.821. [DOI] [PubMed] [Google Scholar]
  • 62.Genzel L, Dresler M, Wehrle R, Grozinger M, Steiger A. Slow wave sleep and REM sleep awakenings do not affect sleep dependent memory consolidation. Sleep. 2009;32:302–310. doi: 10.1093/sleep/32.3.302. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 63.Feld GB, Wilhelm I, Ma Y, Groch S, Binkofski F, Molle M, Born J. Slow wave sleep induced by GABA agonist tiagabine fails to benefit memory consolidation. Sleep. 2013;36:1317–1326. doi: 10.5665/sleep.2954. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 64.Seeck-Hirschner M, Baier PC, Sever S, Buschbacher A, Aldenhoff JB, Goder R. Effects of daytime naps on procedural and declarative memory in patients with schizophrenia. J Psychiatr Res. 2010;44:42–47. doi: 10.1016/j.jpsychires.2009.05.008. [DOI] [PubMed] [Google Scholar]
  • 65.Molle M, Born J. Slow oscillations orchestrating fast oscillations and memory consolidation. Prog Brain Res. 2011;193:93–110. doi: 10.1016/B978-0-444-53839-0.00007-7. [DOI] [PubMed] [Google Scholar]
  • 66.Himanen SL, Virkkala J, Huhtala H, Hasan J. Spindle frequencies in sleep EEG show U-shape within first four NREM sleep episodes. J Sleep Res. 2002;11:35–42. doi: 10.1046/j.1365-2869.2002.00273.x. [DOI] [PubMed] [Google Scholar]
  • 67.Steriade M, Amzica F. Coalescence of sleep rhythms and their chronology in corticothalamic networks. Sleep Res Online. 1998;1:1–10. [PubMed] [Google Scholar]
  • 68.Dijk DJ, Hayes B, Czeisler CA. Dynamics of electroencephalographic sleep spindles and slow wave activity in men: effect of sleep deprivation. Brain Res. 1993;626:190–199. doi: 10.1016/0006-8993(93)90579-c. [DOI] [PubMed] [Google Scholar]
  • 69.Uchida S, Maloney T, March J, Azari R, Feinberg I. Sigma (12–15 Hz) and delta (0.3–3 Hz) EEG oscillate reciprocally within NREM sleep. Brain research bulletin. 1991;27:93–96. doi: 10.1016/0361-9230(91)90286-s. [DOI] [PubMed] [Google Scholar]
  • 70.Aeschbach D, Borbely AA. All-night dynamics of the human sleep EEG. J Sleep Res. 1993;2:70–81. doi: 10.1111/j.1365-2869.1993.tb00065.x. [DOI] [PubMed] [Google Scholar]
  • 71.Fogel SM, Albouy G, Vien C, Popovicci R, King BR, Hoge R, Jbabdi S, Benali H, Karni A, Maquet P, et al. fMRI and sleep correlates of the agerelated impairment in motor memory consolidation. Hum Brain Mapp. 2014;35:3625–3645. doi: 10.1002/hbm.22426. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

RESOURCES