Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Dec 1.
Published in final edited form as: Neurobiol Learn Mem. 2019 Oct 14;166:107100. doi: 10.1016/j.nlm.2019.107100

Dynamics of Sleep Spindles and Coupling to Slow Oscillations Following Motor Learning in Adult Mice

Korey Kam a, Ward D Pettibone a, Kaitlyn Shim a, Rebecca K Chen a, Andrew W Varga a
PMCID: PMC6910722  NIHMSID: NIHMS1545113  PMID: 31622665

Abstract

Sleep spindles have been implicated in motor learning in human subjects, but their occurrence, timing in relation to cortical slow oscillations, and relationship to offline gains in motor learning have not been examined in animal models. In this study, we recorded EEG over bilateral primary motor cortex in conjunction with EMG for 24 hours following a period of either baseline handling or following rotarod motor learning to monitor sleep. We measured several biophysical properties of sleep spindles and their temporal coupling with cortical slow oscillations (SO, <1 Hz) and cortical delta waves (1–4 Hz). Following motor learning, we found an increase in spindles during an early period of NREM sleep (1–4 hours) without changes to biophysical properties such as spindle power, peak frequency and coherence. In this same period of early NREM sleep, both SO and delta power increased after motor learning. Notably, a vast majority of spindles were associated with minimal SO power, but in the subset that were associated with significant SO power (>1 z-score above the population mean), spindle-associated SO power was greater in spindles following motor learning compared to baseline sleep. Also, we did not observe a group-level preferred phase in spindle-SO or spindle-delta coupling. While SO power alone was not predictive of motor performance in early NREM sleep, both spindle density and the difference in the magnitude of the mean resultant vector length of the phase angle for SO-associated spindles, a measure of its coupling precision, were positively correlated with offline change in motor performance. These findings support a role for sleep spindles and their coupling to slow oscillations in motor learning and establish a model in which spindle timing and the brain circuits that support offline plasticity can be mechanistically explored.

Keywords: rotarod, motor learning, spindles, slow oscillation, delta waves, coupling

INTRODUCTION

A role for sleep in the processing of motor skill in humans has been observed for several motor tasks, most commonly in the finger-tapping motor sequence task and serial reaction time test (Fischer, Hallschmid et al. 2002, Walker, Brakefield et al. 2002, King, Hoedlmoser et al. 2017), but also in continuous tracking/pursuit rotor tasks (Smith and MacNeill 1994, Fogel and Smith 2006) and forms of motor adaptation such as mirror-tracing (Plihal and Born 1997, Mantua, Baran et al. 2016) or joystick/computer mouse adaptation tasks (Debas, Carrier et al. 2010, Albouy, Vandewalle et al. 2013). The role of sleep in augmenting offline gains in motor performance appears to be greater with increasing motor task complexity (Kuriyama, Stickgold et al. 2004), an observation recapitulated in animal models (Nagai, de Vivo et al. 2016).

Early work on human motor sequence learning task identified that motor performance improvement correlated with non-REM (NREM) stage 2 sleep (Walker, Brakefield et al. 2002), a stage characterized by the presence of cortical spindles. Several subsequent studies have identified an increase in spindle density following a motor learning task (Fogel and Smith 2006, Fogel, Smith et al. 2007, Nishida and Walker 2007) and the existence of specific correlations between overnight performance improvements on motor tasks and spindle density (Walker, Brakefield et al. 2002, Rasch, Pommer et al. 2009, Ackermann and Rasch 2014), including over somatotopically relevant motor cortical areas (Johnson, Blakely et al. 2012). Post-sleep motor performance correlated with increased sleep spindle density following administration of a selective serotonin reuptake inhibitor (SSRI), which enhanced NREM stage 2 sleep at the expense of REM (Rasch, Pommer et al. 2009). Selective spindle augmentation with transcranial alternating current stimulation (tACS) was also shown to enhance motor memory consolidation (Lustenberger, Boyle et al. 2016).

Animal models of motor learning would be useful for not only helping to develop a causal role for spindles, but also for identifying the neuroanatomical areas that spindles might functionally couple, thereby improving motor performance. Normal sleep is critical to this offline gain in performance, as sleep disruption has been shown to impair motor learning in rodents on a standard rotarod task (Yang, Lai et al. 2014), a rotarod gist learning task (Pettibone, Kam et al. 2019), complex wheel running task (Nagai, de Vivo et al. 2016), and skilled reaching task (Varga, Kang et al. 2014). Additionally, replay of skilled reach task-related ensembles in the motor cortex linked to bursts of spindle activity, supporting the reactivation during sleep hypothesis of memory consolidation (Ramanathan, Gulati et al. 2015). However, whether sleep spindle properties change following motor learning in adult mice, recapitulating the observation in human subjects, has not been investigated. In particular, the precise temporal coupling of spindles to cortical slow oscillations has been important for hippocampus-dependent fear memories in rodents (Latchoumane, Ngo et al. 2017) and for motor sequence task learning in human subjects (Demanuele, Bartsch et al. 2017). In this study, we examined changes to sleep spindle features, power in the slow oscillation (SO, 0.5–1 Hz) and delta oscillation (1–4 Hz) bands, and phase coupling of spindles to these slower oscillations in adult mice during sleep following rotarod motor learning in comparison to baseline sleep, and whether such changes predict offline changes in rotarod performance.

METHODS

Animals and Behavioral Paradigm

Fifteen C57/BL6 mice (8 female and 7 male) were obtained from the Jackson Laboratory (Bar Harbor, Maine) and used between ages 5 to 7 months old (6.1±0.6 months old). Animals were housed on a 12 hr/12 hr light/dark schedule with lights on at 9:00am (Zeitgeber time (ZT0)) and provided ad libitum access to food and water. Prior to the experiment, surgically implanted mice (at least 7 days from surgical date) were handled daily before rotarod exposure and had one full day of acclimation to tethered EEG/EMG recording system. The experimental timeline consisted of a Baseline session at ZT0 (Handling session control on the static rod of the rotarod apparatus) followed by EEG/EMG recording until ZT0 next day. Then motor learning began at ZT0 (day 2) and consisted of 10 trials on the rotarod, with each trial capped at 5 minutes with a 3 minute inter-trial interval, followed by EEG/EMG recording until the next day. Finally, motor testing began at ZT0 (day 3) with the same rotarod trial parameters as motor learning. All experiments were approved by the Institution of Animal Care and Use Committee of the Icahn School of Medicine at Mount Sinai and were carried out in accordance with the National Institutes of Health guidelines.

Surgery

Implantation of EEG/EMG electrodes were performed as previously described (Kam, Duffy et al. 2016). Mice were anesthetized with inhaled isoflurane and placed in a stereotaxic apparatus (David Kopf). After exposing the skull, 5 electrodes were positioned. Two subdural electrodes (2.5 mm diameter screws with tapered tips, Pinnacle Technologies), symmetrically placed over left and right primary motor cortices (1.5 mm anterior to Bregma, ± 2.0 mm lateral to the midline) served as EEG electrodes. Two epidural screw electrodes were placed above the cerebellum to serve as reference and ground. A bipolar, twisted stainless steel electrode (California Fine Wire Co.) inserted into the nuchal muscles served as an electromyogram (EMG) site. After implantation, a 6-pin connector (Millmax) was centered over the skull with dental cement (Dentsply) and the animal was placed in its home cage on top of a heating pad set to 37°C (Harvard Apparatus) until fully ambulatory. All animals were supplied with subcutaneous hydration and pain control (buprenorphine) following surgery.

Rotarod

Mice were habituated to the room containing the rotarod for at least 7 days before initiating behavioral training. Mice were placed on an accelerating rotarod (Ugo Basile) in which the rod accelerated from 4 to 40 RPM over 300 seconds. Mice completed 10 consecutive rotarod trails with an inter-trial interval of 3 minutes. Each trial terminated when the mouse fell off the rod, when the animal clung to the rod for a full 360° rotation, or when the maximum time of 300 seconds was reached. All rotarod training took place between the hours ZT0 and ZT2.

Video-EEG/EMG Recording

As previously described (Kam, Duffy et al. 2016), recordings were performed in the home cage (12”x7”) coupled with a suspended multichannel commutator (Pinnacle Technologies) to allow for freely behaving/sleeping behavior. All of these recordings were performed in the same room containing the rotarod. This setup allows for access to food pellets, water and nesting material over the entire ~23-hour continuous recording session. Signals were acquired at 1000 Hz sampling rate and bandpass filtered from 0.5–100 Hz (Pinnacle). Simultaneous video was recorded continuously at 10 frames per second (synchronized with the EEG record) during both light and dark periods using an infrared LED camera with Sirenia Video Acquisition software (Pinnacle).

Data Analysis

Data analysis was performed using MATLAB (MathWorks) with the FieldTrip (Oostenveld, Fries et al. 2011), CircStat (Berens 2009, Kempter, Leibold et al. 2012), and FMA (Hazan, Zugaro et al. 2006) toolboxes.

Sleep-wake analysis

Sleep/wake scoring was performed as previously described in a semi-automated way(Kam, Duffy et al. 2016). Video-EEG/EMG was analyzed continuously to characterize behavioral states in one second epochs. Behavioral states (wakefulness, NREM and REM sleep) were classified in an epoch-free approach based on these criteria:

  • Time-varying ratio of theta over delta power (θ, 5–10 Hz; δ, 1–4 Hz) using both the right and left primary motor cortex lead.

  • Presence of slow waves (delta power, 1–4 Hz) defined as segments with greater than 1 zscore analyzed from both the right and left primary motor cortex lead across the entire recording.

  • Movement, detected by EMG and confirmed by simultaneous manual review of video.

REM sleep was defined by a high ratio of theta/delta power (ratio > 2.5), and little or no movement of the body (based on EMG < 1 zscore). In addition, a criterion for REM sleep was that the prior behavioral state was NREM sleep (which is the normal pattern for sleep in rodents). REM sleep segments separated by less than 3 seconds were merged because these were periods when small twitches or slight posture changes appeared to interrupt an otherwise continuous period of REM sleep. If movement was < 1 zscore, but other criteria for REM were not met, the behavioral state was classified as NREM sleep or quiet wakefulness. NREM was discriminated from quiet wakefulness based on power in the delta band and presence of putative spindles. Thus, NREM sleep showed a lower ratio of theta/delta power (<2.5) than quiet wakefulness. Sleep episodes were confirmed manually by reviewing video and finding that mice had a curled or resting body position. Periods with relatively low delta power (< 1 zscore) and minimal movement (< 1 zscore) were designated as quiet wakefulness. Periods with movement for >3 seconds were classified as active wakefulness and included exploration/walking, grooming, sniffing, consummatory behavior (eating/drinking), and spontaneous arousals from. All spectral thresholds were verified manually for each recording.

Spindle detection and definitions of spindle biophysical properties

Sleep spindles were detected by a normalized squared signal algorithm that has been previously described(Khodagholy, Gelinas et al. 2017). In brief, the raw EEG from each cortical lead was bandpass filtered in the 10–16 Hz range using a zero-phase 4th order butterworth filter. Spindle onset/offset were defined as events in NREM sleep having an amplitude of the filtered signal greater than 3 zscores above the baseline (for a minimum of 0.5 sec. and a maximum of 3 sec. in duration) and having a peak spindle power of at least 5 zscores above the baseline while residing between the spindle onset/offset (to satisfy “waxing and waning” appearance). All subsequent spindle measures (count, density, duration, frequency, power) were the average of the right and left primary motor cortex with the exception of spindle coherence. Spindle duration was defined as the sample time from offset to onset for each spindle. Detections were visually inspected for accuracy for each recording session.

Spindle power was computed using a Morlet wavelets approach. EEG data were convolved with a complex Morlet wavelet (5-cycle density per frequency). Raw wavelet power spectra were then normalized by computing the wavelet power spectrum on a 5 second baseline segment that preceded each spindle onset. Spindle power was normalized to account for relative eventrelated baseline shifts (such as learning-induced increases to broadband power) throughout a recording and limited to NREM sleep.

Normalized spindle power was defined as:

wavelet powerduring spindle−wavelet powerin 5  second prior to spindle onsetwavelet powerin 5  second prior to spindle onset ×100%  change

within the spindle frequency band (10–16 Hz). Spindle peak frequency was defined as the frequency at maximal wavelet power within the spindle frequency band (10–16 Hz). Spindle coherence was computed using a multitaper FFT approach between right and left primary motor cortices (Schoffelen, Poort et al. 2011).

Spindle coupling measures

Spindle coupling analyses were performed on the right M1 channel only, as there were no differences in spindle counts, coherence, or power between the right and left channel. To determine SO-associated spindles, FFT power spectra was computed from 2 seconds before to 2 seconds after the onset of each detected spindle (a 4 second time window which includes the duration of each spindle) using the multitaper frequency transformation with slepian tapers (Mitra and Pesaran 1999). A spindle was defined as an “SO-associated spindle” within each condition per animal if the relative power in the SO band (0.5–1 Hz) normalized to broadband power (0.5–100 Hz) during this time window (±2 seconds from spindle onset) was above 1 zscore for all detected spindles per recording session. Relative SO power was done to compare across animals. Spindle-SO phase coupling was computed using the Hilbert transform for instantaneous phase of the SO cycle surrounding identified spindles. The mean resultant vector length and preferred angle were computed on spindle-SO phase distributions for SO-associated spindles in each animal and each recording. Because our EEG recordings are subdural, the SO trough (negative peak) was used as the −180° reference point. Thereafter, an SO cycle peak was defined at 0° (successive positive wave peak) while the subsequent SO cycle trough was defined at +180° (negative wave polarity down) (Rasch, Buchel et al. 2007, Molle, Eschenko et al. 2009, Niethard, Ngo et al. 2018). The same procedure was applied for spindle-delta phase coupling using the 1–4 Hz frequency band to define delta-associated spindles.

The deviance of each spindle’s peak power relative to the SO peak phase (0°) was defined as the mean absolute value phase for all spindle-SO events in the 1–4 hours of NREM sleep. Circular distributions having a significantly preferred spindle-SO phase relationship were defined as having a kappa (κ concentration) concentration parameter greater than 0.1 (the concentration parameter of a von Mises distribution that is inversely related to circular variance) (Khodagholy, Gelinas et al. 2017).

Statistical Analysis

All data are reported as mean ± SEM with the p criterion set to 0.05. For comparison of two means, either paired t-tests or two sample t-tests were used. The Kolmogorov-Smirnov test was used to infer if motor learning or baseline spindle event SO or delta power came from the same distribution. Circular distributions of spindle-SO or spindle-delta coupling for each recording were summarized by the kappa concentration parameter (κ). Watson-Williams (WW) test of circular inference was used to assess differences in circular distributions of spindle phase to either SO or delta between baseline and motor learning conditions. Spearman’s correlation was used to assess bivariate relationships between motor performance metrics and spindle measures. A circular-linear correlation was performed to assess preferred phase (spindle-SO or spindle-delta) to power associations (SO or delta absolute power). The Chi-squared test was used to compare proportions of groups with preferred phase between behavioral conditions.

RESULTS

An early increase in spindles after motor learning

To identify sleep spindles, we recorded EEG/EMG from bilateral M1 of adult mice after both a baseline handling control on day 1 beginning at ZT0 and after 10 trials of rotarod on day 2 also beginning at ZT0 (Figure 1A). Sleep/wake states were scored and sleep spindles were detected from NREM sleep bouts only (Figure 1B–C). Detection of cortical sleep spindles across all baseline NREM sleep resulted in a mean spindle density of 2.2 ± 0.7 per minute of NREM sleep, consistent with spindle density observed in several other studies where spindle density has ranged from 1.67/min to 3/min in NREM sleep in mice (Kim, Latchoumane et al. 2012, Rovo, Matyas et al. 2014, Kim, Hwang et al. 2015, Uygun, Katsuki et al. 2019).

Figure 1: Identification of sleep spindles after motor learning.

Figure 1:

Open in a new tab

A. Animals naïve to the rotarod were handled and placed on the rotarod (without rotation) for ten 5-minute trials with 3-minute inter-trial interval (Baseline, B). A recording period lasting ~23 hrs in the home cage (Day 1) is followed by motor learning, ten trials on the rotarod with 3-minute inter-trial interval (motor learning, ML). A second recording period lasting ~23 hrs in the home cage (Day 2) is then followed by motor testing (MT). See the Methods section for additional parameters.

B. An example hypnogram of sleep/wake states obtained from the recording period following motor learning (ML). Gray indicates Wakefulness, Blue indicates NREM sleep and Orange indicates REM sleep. For scoring of sleep/wake states in mice, refer to the Methods section.

C. Continuous wavelet transform of a detected spindle recorded from the right primary motor cortex following ML.

D. Raw EEG from the right primary motor cortex, spindle bandpass filtered (10–16 Hz) and slow wave bandpass filtered (0.5–4 Hz, SO+delta band) channels below.

To evaluate the dynamics of sleep spindles following motor learning, we compared their occurrence across all spontaneous NREM sleep bouts to their occurrence following baseline handling until ZT0 the following day. Example spindle time course scatter plots are shown in Figure 2A (following baseline) and Figure 2B (following motor learning). While there were no differences in spindle density between conditions across the full recording period of ~23 hours (Figure 2C, paired t-test, p = 0.654), the spindle time course scatter plots suggested an early transient increase in spindle occurrence. Quantification of spindles during this transient period in the first 1–4 hours of NREM sleep (termed 1–4 hours of NREM sleep) revealed a significant increase in spindle density during this time following motor learning in comparison to baseline handling (Figure 2D, paired t-test, p = 0.020).

Figure 2: An early transient increase in spindles following motor learning.

Figure 2:

Open in a new tab

A. An example time course scatter plot of detected spindles in R-M1 by relative change in spindle band power in the recording following handling baseline. Time begins at the onset of NREM sleep for each recording. Red indicates maximal spindle-power 2D density, while blue indicates minimal spindle-power 2D density (color-coded for visual purposes only).

B. An example time course scatter plot of detected spindles in R-M1 by relative change in spindle band power in the recording following motor learning in the same animal from A. Note an intense density of spindles in the first 1–4 hours of NREM sleep following motor learning.

C. Mean spindle density (average of Right and Left-M1 channels) across all NREM sleep between the end of the experimental condition and ZT 0 the following day (~23 hours) in Baseline versus Motor learning conditions was not significantly different (paired t-test, p = 0.654).

D. Mean spindle density (average of Right and Left-M1 channels) in 1–4 hours of NREM sleep was greater following Motor learning versus Baseline conditions (paired t-test, p = 0.020). Gray bar with error represents mean ± SEM throughout.

Spindle features and sleep architecture are not different between motor learning and baseline sleep

To characterize whether other biophysical properties of sleep spindles change following motor learning during this transient stage, we performed spectral analyses on spindles that occurred in post-motor learning sleep compared to baseline sleep. We found no differences in spindle power, coherence, duration, or the peak spindle frequency (Supp. Figure 1A–J). Because sleep spindles can also occur during the transition to REM sleep in mice (Vyazovskiy, Achermann et al. 2004), we quantified the number of NREM to REM transitions, total number of REM bouts, NREM and REM duration in the first 1–4 hours since NREM sleep onset. We did not find differences in any of these variables (Supp. Figure 2A–D). Therefore, unlike increases in spindle density following motor learning, both the spectral features of spindles and sleep architecture examined here did not change with motor learning.

SO and delta power increase during the initial 1–4hr of NREM sleep after motor learning

Having identified an event-driven, spindle density increase in the first 1–4hrs of NREM sleep following motor learning, we performed power analyses across this early NREM sleep period to determine if broad power changes had also occurred (Figure 3A–B). The change in power spectra between motor learning and baseline behavioral conditions indicated that most animals had increases in power below 10 Hz (Figure 3C). Band-specific quantification during this 1–4hr period of NREM sleep showed that while spindle power did not change within animals (Figure 3D, paired t-test, p = 0.387), both slow oscillation (SO, 0.5–1 Hz; Figure 3E, paired t-test, p = 0.014) and delta wave (1–4 Hz; Figure 3F, paired t-test, p = 0.022) power increased following motor learning.

Figure 3: An early increase in SO and delta power following motor learning.

Figure 3:

Open in a new tab

A. Individual (gray) and group-level (orange) power spectra during Baseline NREM sleep in the 1–4 hours since NREM onset. Orange shading indicates 95% CI.

B. Individual (gray) and group-level (yellow) power spectra during Motor learning NREM sleep in the 1–4 hours since NREM onset. Yellow shading indicates 95% CI.

C. Individual (gray) and group-level (purple) difference in power spectra between Motor learning and Baseline NREM sleep in the 1–4 hours since NREM onset. Purple shading indicates 95% CI.

D. Absolute spindle power was not significantly different between motor learning and baseline sleep (paired t-test, p = 0.387).

E. Absolute SO power was significantly different between motor learning and baseline sleep (paired t-test, p = 0.014).

F. Absolute delta power was significantly different between motor learning and baseline sleep (paired t-test, p = 0.022).

Spindle coupling to the phase of the slow oscillation or delta wave after motor learning

To explore the coupling between spindles and SOs, we first ranked all spindles detected based on the relative SO power around each detected spindle (±2 second window from each spindle onset). We then defined SO-associated spindles as those spindles having SO power within this window above 1 zscore threshold per each individual animal (Figure 4A–B). Analyses on measures of coupling were subsequently restricted to only those SO-associated spindle events. Defined in this manner, the average proportion of SO-associated spindles was 15.7% ± 0.3% following motor learning and 15.5% ± 0.2% following baseline handling (p=0.450, paired t-test), and Supplemental Table 1 lists the total number of identified spindles and those identified as SO-associated for each animal per condition. There was a significant rightward shift in the pooled distribution of SO power surrounding spindles following motor learning compared to baseline handling (Figure 4B, KS test, p<0.0001). Next, we computed the phase of the filtered SO (0.5–1Hz) at which peak spindle power occurred for each SO-associated spindle and observed no phase preference in the pooled distribution of SO phase angles between motor learning (κ concentration = 0.05, Figure 4C) and handling conditions (κ concentration = 0.03, Figure 4C). The distribution of spindle-SO phase angles as a function of SO power across behavioral groups (Figure 4D) and the mean spindle-SO phase angle and mean resultant vector length for each animal (Figure 4E) did not indicate a phase preference following either baseline handling or motor learning. Furthermore, there was no difference in the proportion of animals with a significantly preferred spindle-SO phase distribution (κ concentration > 0.1 in 12/15 at baseline, 14/15 following motor learning, χ2 = 0.288, p = 0.591) between motor learning (mean κ concentration parameter = 0.34±0.17) and baseline sleep (mean κ concentration parameter = 0.31±0.18). When comparing between behavioral conditions, we did not find a significantly preferred phase of spindle-SO coupling (as defined by mean κ concentration parameter).

Figure 4: Phase coupling of spindles and slow oscillation (0.5–1 Hz) power after motor learning.

Figure 4:

Open in a new tab

A. Example histogram of relative SO power (0.5–1 Hz) surrounding spindles restricted to 1–4 hours of NREM sleep in motor learning and baseline. Dashed lines indicate 1 zscore cutoffs per condition.

B. Pooled histogram shows a significant difference in relative SO power around spindles in sleep after motor learning compared to baseline (n=15 mice, KS test, p<0.0001). Colored lines indicate distribution fits.

C. Pooled polar histogram shows the lack of preferred phase in spindle coupling to SO phase restricted to spindle in the 104 hours of NREM sleep following motor learning (κ concentration = 0.05) and baseline (κ concentration = 0.03).

D. Polar scatter plots of the distribution of SO (0.5–1 Hz) phases at which spindles occur with corresponding normalized SO power on the radial axis during 1–4 hours of NREM sleep following baseline handling (left, blue dots) or motor learning (right, orange dots) (n=15 mice). SO phases span ±180° with 0° at the peak and ±180° at the trough of a SO cycle. Neither baseline sleep (blue, circular-linear coeff. = 0.032, p = 0.708) nor Motor learning sleep (orange, circular-linear coeff. = 0.060, p = 0.285) exhibited a circular-linear association.

E. Group-level circular plot of the mean spindle-SO phase angle and resultant vector length (radial axis) per animal in sleep following baseline control (left) or following motor learning (right). Line length is the magnitude of the mean resultant vector length (RVL) for each animal. Each colored line represents the same animal ID throughout this and all subsequent figures.

Prior work has suggested that spindles occurring at the peaks of slow oscillations (i.e. 0 degrees) have the highest likelihood of inducing behavioral changes in the context of fear memory (Latchoumane, Ngo et al. 2017). We hypothesized that spindle-SO coupling could converge toward an SO peak (0 degrees) following motor learning without significant differences in the overall mean phase angle of SO-associated spindles versus baseline. In order to evaluate this possibility, we calculated the mean absolute value of the phase angle of SO-associated spindles per animal during the transient 1–4 hours of NREM sleep. This measure was not significantly different between motor learning and baseline handling (in sleep following baseline = 140°±27° and in sleep after motor learning = 116°±50°, WW test, p = 0.158). We also determined the proportion of SO-associated spindles per animal per condition that occurred near the peak (±10 degree-bin around 0 degrees, paired t-test, p = 0.738) or trough (±10 degree-bin around 180 degrees, paired t-test, p = 0.528). Both of these measures were not different between motor learning and baseline.

The mean resultant vector length (RVL) is a measure of the variance of phase angles such that a completely uniform distribution of angles between 0 and ±180 degrees results in a RVL of 0, whereas a pure concentration of angles at a single value results in a RVL of 1. The mean RVL of SO-associated spindles during the transient 1–4 hours of NREM sleep was also not different between behavioral conditions (paired t-test, p = 0.382)

Delta waves (1–4 Hz) are another oscillation characteristic of NREM sleep distinct from the SO, for which there is evidence for coupling with spindles, primarily from human work (Yordanova, Kirov et al., Demanuele, Bartsch et al. 2017). Like the subset of SO-associated spindles, delta-associated spindles were also a minor fraction (Supp. Figure 3A–B), but notably, the pooled distribution of delta power surrounding spindles was left shifted (i.e. lower power) following motor learning versus baseline handling (Supp. Figure 3B, KS test, p<0.0001). Within the fraction of delta-associated spindles, we found no differences in preferred delta phase (Supp. Figure 3C), the relationship between spindle-delta phase angles and delta power (Supp. Figure 3D), or an individual animal’s spindle-delta mean resultant vector length (Supp. Figure 3E) between baseline and motor learning conditions. We computed the same spindle coupling measures with delta waves, and while we found a significant difference in the mean absolute phase angle (in sleep following baseline = 67°±53° and in sleep after motor learning = 129°±51°, WW test, p = 0.013), we did not observe significant differences in the proportion of delta-associated spindles near the peak (paired t-test, p = 0.422) or the trough (paired t-test, p = 0.5928), or the magnitude of the mean RVL between baseline and motor learning conditions (paired t-test, p = 0.978). As we observed with spindle-SO phase distributions, the proportion of mice showing preferred spindle-delta phase distributions was also not significant between motor learning and baseline sleep (κ concentration > 0.1 in 13/15 at baseline, 13/15 following motor learning, χ2 = 0.289, p = 0.591). These results suggest that motor learning did not enhance spindle to delta peak coupling.

Spindle measures during initial sleep correlate with motor performance

Because spindles have been shown to associate with sleep-dependent motor learning, we next examined associations between offline change in rotarod motor performance across successive days (Figure 1A) and spindle features and spindle-SO coupling in the early 1–4 hours of intervening NREM sleep. Rotarod performance curves on day 1 and day 2 are shown in Figure 5A and 5B respectively. Offline changes were next measured as changes in the mean latency to fall for the first 3 trials of day 2 in comparison to the last 3 trials of day 1 (abbreviated as F3D2/L3D1). This inter-session performance change has been used previously (Nagai, de Vivo et al. 2016) to assess offline change in both rodent rotarod and complex wheel running, minimizes dependence on intra-session learning during wakeful practice, and also has precedence as a measure of offline change in human motor sequence task learning (Walker, Brakefield et al. 2003, Djonlagic, Saboisky et al. 2012). Mean offline change by this measure was 97% ± 6% and F3D2/L3D1 ≥ 100% observed in 7/15 mice. Spindle density during 1–4 hours of NREM sleep was significantly positively correlated with offline change in motor consolidation measured as F3D2/L3D1 (Figure 5C, Spearman correlation, rho = 0.628, p = 0.014). In addition, the difference in mean resultant vector length of spindle-SO phase coupling between sleep following motor learning and sleep following baseline handling during the initial 1–4 hours of NREM sleep was also significantly positively associated with offline change in motor performance measured as F3D2/L3D1 (Figure 5D, Spearman correlation, rho = 0.521, p = 0.048). Furthermore, although SO and delta power both increased following motor learning during the initial 1–4 hours of NREM sleep, there was no significant correlation between event-defined SO power surrounding spindles (Figure 5E, Spearman correlation, rho = 0.375, p = 0.169) or delta power surrounding spindles (Figure 5F, Spearman correlation, rho = 0.132, p = 0.639) and offline change as expressed by F3D2/L3D1.

Figure 5: Early spindle measures correlate with motor performance.

Figure 5:

Open in a new tab

A. Day 1 trial-by-trial rotarod performance (n=15 mice, solid blue line is mean with blue shade indicating 95% CI).

B. Day 2 trial-by-trial rotarod performance for the same mice given ad libitum sleep immediately after Day 1 testing.

C. A significant positive association was observed between offline motor performance measured as the mean of the first 3 trials on day 2 relative to the last 3 trials of day 1 (F3D2/L3D1) and spindle density in 1–4 hours of NREM sleep (rho = 0.628, p = 0.014, Spearman correlation).

D. A significant positive association was observed between offline motor performance measured as F3D2/L3D1 and the difference in spindle-SO mean mean resultant vector length (MRVL SO) from motor learning to baseline (rho = 0.521, p = 0.048).

E. No correlation between F3D2/L3D1 and the median SO power surrounding spindles in 1–4 hours of NREM sleep (Spearman correlation, rho = −0.304, p = 0.271).

F. No correlation between F3D2/L3D1 and the median delta power surrounding spindles in 1–4 hours of NREM sleep (Spearman correlation, rho = 0.142, p = 0.611).

DISCUSSION

In this study, we examined the dynamics of sleep spindles after motor learning in adult mice. We found a specific early yet transient increase in spindles during the first 1 to 4 hour period of NREM sleep following motor learning without changes in spectral or other biophysical properties of spindles. Moreover, spindle count and density in this early window of NREM sleep following initial experience with the rotarod motor learning task predicted subsequent offline change in performance observed 24 hours later. These observations are largely consistent with related phenomena in human subjects, where increases in sleep spindles have been observed following finger-tapping motor sequence task learning (Fogel and Smith 2006, Fogel, Smith et al. 2007, Nishida and Walker 2007), and spindle density correlated with subsequent offline improvements in motor performance (Walker, Brakefield et al. 2002, Rasch, Pommer et al. 2009, Ackermann and Rasch 2014) including over somatotopically relevant motor cortical areas (Johnson, Blakely et al. 2012). One potential difference across species is that the spindle increase in mice appears to occur in the first several hours of NREM sleep, whereas the same dynamic has not clearly been observed in human subjects across a full night of sleep. That said, when human sleep spindles are divided into slow (<13 Hz) and fast (≥13 Hz) subtypes, there has been evidence for the specific role of fast spindles in motor learning (Morin, Doyon et al. 2008, Tamaki, Matsuoka et al. 2009, Barakat, Doyon et al. 2011, Tamaki, Huang et al. 2013), and a large epidemiological study of sleep spindles demonstrated that fast spindle density peaked in the first NREM cycle, whereas slow spindle density remained relatively stable across the night in young adults under age 40 (Purcell, Manoach et al. 2017). Changes in spindle density in rodents have also been observed following odor-reward pairing learning (Eschenko, Molle et al. 2006, Molle, Eschenko et al. 2009). Our current observations were present in a group of mice that was evenly split between sexes. We did not monitor estrous phase in the female mice, a factor which may influence the findings, although no effect of estrous phase was found on rotarod performance in C57Bl/6 mice (Meziane, Ouagazzal et al. 2007).

In addition to examining the presence and biophysical properties of spindles as well as spindle-rich NREM-REM sleep transitions (Vyazovskiy, Achermann et al. 2004), we were interested in the timing of spindles in relation to slower cortical rhythms in the SO (0.5–1 Hz) and delta (1–4 Hz) bands. This interest was based on work from both animals (Steriade, Nunez et al. 1993, Steriade 2006, Latchoumane, Ngo et al. 2017) and human subjects (Molle, Marshall et al. 2002, Molle, Bergmann et al. 2011, Molle and Born 2011, Staresina, Bergmann et al. 2015, Muehlroth, Sander et al. 2019) that has suggested spindles (especially fast spindles) are modulated by the phase of the cortical slow oscillation, particularly at the peak up-state which follows a SO trough down-state (Latchoumane, Ngo et al. 2017). In human subjects, coupling of spindles to slow oscillations was increased following word pair learning (Molle, Eschenko et al. 2009), and coupling to delta oscillations was increased following motor learning in both schizophrenics (Demanuele, Bartsch et al. 2017) and healthy controls (Yordanova, Kirov et al.). Here, we observed that both SO and delta power were increased following motor learning, however, these measures alone were not predictive of offline performance change. This observation is consistent with the SO and delta power increases observed following motor learning in rodents (Hanlon, Faraguna et al. 2009, Binder, Baier et al. 2012, Oyanedel, Binder et al. 2014) and humans (Huber, Ghilardi et al. 2004, Molle, Eschenko et al. 2009). Within the fraction of spindles defined as either SO-associated or delta-associated, we did not observe a strong phase preference for sleep spindles with either the slow oscillation or delta wave at both baseline and following motor learning. This observation is similar to the lack of slow oscillation phase preference observed for spindles in rats following odor-reward learning (Molle, Eschenko et al. 2009). While we additionally observed no significant difference in the variance of the phase coupling with the SO (expressed by the magnitude of the mean resultant vector) of spindles between baseline and post-learning conditions, we did observe that the magnitude of the change in the mean resultant vector length significantly predicted the magnitude of the mean change in offline motor performance. We interpret this to mean that while the preferred phase of the slow oscillation in which spindles occur may be individual to each animal, increased fidelity of spindle timing to that particular phase can be a marker of increased motor learning.

How spindles contribute to motor learning is still being elucidated, but there are some emerging clues in both animals and human subjects. Single unit recordings in rat motor cortex during a skilled reaching task and subsequent sleep showed that task-related ensembles were replayed during sleep, and notably, replay was linked to the coincidence of slow wave events (Gulati, Ramanathan et al. 2014) and bursts of spindle activity (Ramanathan, Gulati et al. 2015).

Additionally, an evoked spindle stimulation pattern in rat somatosensory cortex led to an NMDA receptor-dependent short-term potentiation and an L-type Ca2+ channel-dependent long-term potentiation (Rosanova and Ulrich 2005). This potentiation may be mediated by increased cortical dendritic calcium activity and synchronization observed during spindles (Seibt, Richard et al. 2017). Calcium activity in cortical pyramidal cells was particularly increased when spindles co-occurred with slow oscillation up-states compared to calcium activity during isolated spindles or slow oscillations (Niethard, Ngo et al. 2018). When human subjects exposed to an odor during motor sequence task learning were re-exposed to the odor during NREM stage 2 sleep, spindles displayed a shift to increased mean frequency and increased amplitude, and subsequent motor performance was increased in comparison to performance following sleep without such re-exposure (Laventure, Fogel et al. 2016). Interestingly, in a very similar paradigm, re-exposure to an odor experienced during motor sequence task learning exclusively in NREM stage 3 sleep did not impart subsequent benefit on motor performance (Rasch, Buchel et al. 2007), suggesting this effect may be sleep stage-specific.

Our findings here support the theory of sleep as a permissive state for potentiation of motor circuit reactivation for motor-type learning (Genzel, Kroes et al. 2014). We hypothesize that potentiation of cortical proprioceptive inputs from somatosensory cortex onto motor cortical neurons may be particularly important for this in rotarod motor learning. Related coordination of NREM oscillations such as sharp wave-ripples in the hippocampus, during which neuronal replay of spatial experience is thought to occur, with neocortical spindles are thought to be important for non-motor forms of memory consolidation (Wilson and McNaughton 1994, Nadasdy, Hirase et al. 1999) (Siapas and Wilson 1998), and ripple disruption in sleep impairs forms of spatial memory consolidation (Girardeau, Benchenane et al. 2009, Ego-Stengel and Wilson 2010).

While the EEG recordings in this work were done over bilateral primary motor cortex, one limitation to bear in mind is that we cannot ascertain that changes to spindles are specific to motor cortex. Spindles occur simultaneously in mice over broad swaths of cortex bilaterally (Kim, Hwang et al. 2015). There is some suggestion that in addition to spindles that are truly global, there are some spindles with a slight anterior predominance and some with a slight posterior predominance. We focused on M1 in part due to observations regarding the promotion of dendritic spines during sleep in layer 5 M1 pyramidal neurons (Yang, Lai et al. 2014). That said, the rotarod task has a strong proprioceptive component, likely mediated by the cerebellum (Ito 2002, Nolan, Malleret et al. 2003, Sausbier, Hu et al. 2004) and primary somatosensory cortex (S1) (Matyas, Sreenivasan et al. 2010) and a timing/rhythm generation component from the striatum (Costa, Cohen et al. 2004, Dang, Yokoi et al. 2006). In human subjects, sleep spindles have been found to be associated with binding motor task-relevant cortical and subcortical regions (Boutin, Pinsard et al. 2018), and that spindle-associated striatal reactivation correlated with offline improvement in a motor sequence learning task (Fogel, Albouy et al. 2017). Additional work will be needed to unravel how sleep in general and spindles in particular impact plasticity in the various circuits that support motor learning.

Overall our current findings contribute to the further understanding of the role of sleep spindles in motor learning, and importantly, establish a model in which future studies evaluating spindle timing and neuroanatomical location can be completed in a relevant motor learning tasks with graded difficulty in order to better understand the mechanisms underlying sleep-associated brain plasticity.

Supplementary Material

1

Acknowledgements:

We thank Drs. Ricardo Osorio, Indu Ayappa, David Rapoport, Anna Mullins and Ankit Parekh for reading the manuscript and providing useful insights. This work was supported by the philanthropy of the James Kuhn Friends of Sleep Medicine, the Leon Levy Foundation Neuroscience Fellowship (A.W.V.), the American Sleep Medicine Foundation Bridge to Success Award 153-BS-16 (A.W.V.), a Friedman Brain Institute Scholars Award (A.W.V), Alzheimer’s Association grant 2018-AARG-589632, Merck Investigator Studies Program, and by NIA grants AG056682 and AG059179 (A.W.V.),

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 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.

Conflict of Interest: The authors declare no competing financial interests.

REFERENCES

  1. Ackermann S and Rasch B (2014). “Differential effects of non-REM and REM sleep on memory consolidation?” Curr Neurol Neurosci Rep 14(2): 430. [DOI] [PubMed] [Google Scholar]
  2. Albouy G, Vandewalle G, Sterpenich V, Rauchs G, Desseilles M, Balteau E, Degueldre C, Phillips C, Luxen A and Maquet P (2013). “Sleep stabilizes visuomotor adaptation memory: a functional magnetic resonance imaging study.” J Sleep Res 22(2): 144–154. [DOI] [PubMed] [Google Scholar]
  3. Barakat M, Doyon J, Debas K, Vandewalle G, Morin A, Poirier G, Martin N, Lafortune M, Karni A, Ungerleider LG, Benali H and Carrier J (2011). “Fast and slow spindle involvement in the consolidation of a new motor sequence.” Behav Brain Res 217(1): 117–121. [DOI] [PubMed] [Google Scholar]
  4. Berens P (2009). “CircStat: A MATLAB Toolbox for Circular Statistics.” Journal of Statistical Software 31(10): 1–21. [Google Scholar]
  5. Binder S, Baier PC, Molle M, Inostroza M, Born J and Marshall L (2012). “Sleep enhances memory consolidation in the hippocampus-dependent object-place recognition task in rats.” Neurobiol Learn Mem 97(2): 213–219. [DOI] [PubMed] [Google Scholar]
  6. Boutin A, Pinsard B, Bore A, Carrier J, Fogel SM and Doyon J (2018). “Transient synchronization of hippocampo-striato-thalamo-cortical networks during sleep spindle oscillations induces motor memory consolidation.” Neuroimage 169: 419–430. [DOI] [PubMed] [Google Scholar]
  7. Costa RM, Cohen D and Nicolelis MAJCB (2004). “Differential corticostriatal plasticity during fast and slow motor skill learning in mice.” 14(13): 1124–1134. [DOI] [PubMed] [Google Scholar]
  8. Dang MT, Yokoi F, Yin HH, Lovinger DM, Wang Y and Li Y (2006). “Disrupted motor learning and long-term synaptic plasticity in mice lacking NMDAR1 in the striatum.” Proceedings of the National Academy of Sciences 103(41): 15254–15259. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Debas K, Carrier J, Orban P, Barakat M, Lungu O, Vandewalle G, Hadj Tahar A, Bellec P, Karni A, Ungerleider LG, Benali H and Doyon J (2010). “Brain plasticity related to the consolidation of motor sequence learning and motor adaptation.” Proc Natl Acad Sci U S A 107(41): 17839–17844. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Demanuele C, Bartsch U, Baran B, Khan S, Vangel MG, Cox R, Hamalainen M, Jones MW, Stickgold R and Manoach DS (2017). “Coordination of Slow Waves With Sleep Spindles Predicts Sleep-Dependent Memory Consolidation in Schizophrenia.” Sleep 40(1). [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Djonlagic I, Saboisky J, Carusona A, Stickgold R and Malhotra A (2012). “Increased sleep fragmentation leads to impaired off-line consolidation of motor memories in humans.” PLoS One 7(3): e34106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Ego-Stengel V and Wilson MA (2010). “Disruption of ripple-associated hippocampal activity during rest impairs spatial learning in the rat.” Hippocampus 20(1): 1–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Eschenko O, Molle M, Born J and Sara SJ (2006). “Elevated Sleep Spindle Density after Learning or after Retrieval in Rats.” Journal of Neuroscience 26(50): 12914–12920. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Fischer S, Hallschmid M, Elsner AL and Born J (2002). “Sleep forms memory for finger skills.” Proc Natl Acad Sci U S A 99(18): 11987–11991. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Fogel S, Albouy G, King BR, Lungu O, Vien C, Bore A, Pinsard B, Benali H, Carrier J and Doyon J (2017). “Reactivation or transformation? Motor memory consolidation associated with cerebral activation time-locked to sleep spindles.” PLoS One 12(4): e0174755. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Fogel SM and Smith CT (2006). “Learning-dependent changes in sleep spindles and Stage 2 sleep.” J Sleep Res 15(3): 250–255. [DOI] [PubMed] [Google Scholar]
  17. Fogel SM, Smith CT and Cote KA (2007). “Dissociable learning-dependent changes in REM and non-REM sleep in declarative and procedural memory systems.” Behav Brain Res 180(1): 48–61. [DOI] [PubMed] [Google Scholar]
  18. Genzel L, Kroes MC, Dresler M and Battaglia FP (2014). “Light sleep versus slow wave sleep in memory consolidation: a question of global versus local processes?” Trends Neurosci 37(1): 10–19. [DOI] [PubMed] [Google Scholar]
  19. Girardeau G, Benchenane K, Wiener SI, Buzsaki G and Zugaro MB (2009). “Selective suppression of hippocampal ripples impairs spatial memory.” Nat Neurosci 12(10): 1222–1223. [DOI] [PubMed] [Google Scholar]
  20. Gulati T, Ramanathan DS, Wong CC and Ganguly K (2014). “Reactivation of emergent task-related ensembles during slow-wave sleep after neuroprosthetic learning.” Nat Neurosci 17(8): 1107–1113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Hanlon EC, Faraguna U, Vyazovskiy VV, Tononi G and Cirelli C (2009). “Effects of skilled training on sleep slow wave activity and cortical gene expression in the rat.” Sleep 32(6): 719–729. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Hazan L, Zugaro M and Buzsaki G (2006). “Klusters, NeuroScope, NDManager: a free software suite for neurophysiological data processing and visualization.” J Neurosci Methods 155(2): 207–216. [DOI] [PubMed] [Google Scholar]
  23. Huber R, Ghilardi MF, Massimini M and Tononi G (2004). “Local sleep and learning.” Nature 430(6995): 78–81. [DOI] [PubMed] [Google Scholar]
  24. Ito M (2002). “Historical Review of the Significance of the Cerebellum and the Role of Purkinje Cells in Motor Learning.” 978(1): 273–288. [DOI] [PubMed] [Google Scholar]
  25. Johnson LA, Blakely T, Hermes D, Hakimian S, Ramsey NF and Ojemann JG (2012). “Sleep spindles are locally modulated by training on a brain-computer interface.” Proceedings of the National Academy of Sciences 109(45): 18583–18588. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Johnson LA, Blakely T, Hermes D, Hakimian S, Ramsey NF and Ojemann JG (2012). “Sleep spindles are locally modulated by training on a brain-computer interface.” Proc Natl Acad Sci U S A 109(45): 18583–18588. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kam K, Duffy AM, Moretto J, LaFrancois JJ and Scharfman HE (2016). “Interictal spikes during sleep are an early defect in the Tg2576 mouse model of beta-amyloid neuropathology.” Sci Rep 6: 20119. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Kempter R, Leibold C, Buzsaki G, Diba K and Schmidt R (2012). “Quantifying circular-linear associations: hippocampal phase precession.” J Neurosci Methods 207(1): 113–124. [DOI] [PubMed] [Google Scholar]
  29. Khodagholy D, Gelinas JN and Buzsaki G (2017). “Learning-enhanced coupling between ripple oscillations in association cortices and hippocampus.” Science 358(6361): 369–372. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Kim A, Latchoumane C, Lee S, Kim GB, Cheong E, Augustine GJ and Shin HS (2012). “Optogenetically induced sleep spindle rhythms alter sleep architectures in mice.” Proc Natl Acad Sci U S A 109(50): 20673–20678. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Kim D, Hwang E, Lee M, Sung H and Choi JH (2015). “Characterization of topographically specific sleep spindles in mice.” Sleep 38(1): 85–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. King BR, Hoedlmoser K, Hirschauer F, Dolfen N and Albouy G (2017). “Sleeping on the motor engram: The multifaceted nature of sleep-related motor memory consolidation.” Neurosci Biobehav Rev 80: 1–22. [DOI] [PubMed] [Google Scholar]
  33. Kuriyama K, Stickgold R and Walker MP (2004). “Sleep-dependent learning and motor-skill complexity.” Learn Mem 11(6): 705–713. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Latchoumane CV, Ngo HV, Born J and Shin HS (2017). “Thalamic Spindles Promote Memory Formation during Sleep through Triple Phase-Locking of Cortical, Thalamic, and Hippocampal Rhythms.” Neuron 95(2): 424–435 e426. [DOI] [PubMed] [Google Scholar]
  35. Laventure S, Fogel S, Lungu O, Albouy G, Sevigny-Dupont P, Vien C, Sayour C, Carrier J, Benali H and Doyon J (2016). “NREM2 and Sleep Spindles Are Instrumental to the Consolidation of Motor Sequence Memories.” PLoS Biol 14(3): e1002429. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Lustenberger C, Boyle MR, Alagapan S, Mellin JM, Vaughn BV and Frohlich F (2016). “Feedback-Controlled Transcranial Alternating Current Stimulation Reveals a Functional Role of Sleep Spindles in Motor Memory Consolidation.” Curr Biol 26(16): 2127–2136. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Mantua J, Baran B and Spencer RM (2016). “Sleep benefits consolidation of visuo-motor adaptation learning in older adults.” Exp Brain Res 234(2): 587–595. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Matyas F, Sreenivasan V, Marbach F, Wacongne C, Barsy B, Mateo C, Aronoff R and Petersen CCH (2010). “Motor Control by Sensory Cortex.” 330(6008): 1240–1243. [DOI] [PubMed] [Google Scholar]
  39. Meziane H, Ouagazzal A-M, Aubert L, Wietrzych M and Krezel W (2007). “Estrous cycle effects on behavior of C57BL/6J and BALB/cByJ female mice: implications for phenotyping strategies.” 6(2): 192–200. [DOI] [PubMed] [Google Scholar]
  40. Mitra PP and Pesaran B (1999). “Analysis of dynamic brain imaging data.” Biophys J 76(2): 691–708. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Molle M, Bergmann TO, Marshall L and Born J (2011). “Fast and slow spindles during the sleep slow oscillation: disparate coalescence and engagement in memory processing.” Sleep 34(10): 1411–1421. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Molle M and Born J (2011). “Slow oscillations orchestrating fast oscillations and memory consolidation.” Prog Brain Res 193: 93–110. [DOI] [PubMed] [Google Scholar]
  43. Molle M, Eschenko O, Gais S, Sara SJ and Born J (2009). “The influence of learning on sleep slow oscillations and associated spindles and ripples in humans and rats.” Eur J Neurosci 29(5): 1071–1081. [DOI] [PubMed] [Google Scholar]
  44. Molle M, Marshall L, Gais S and Born J (2002). “Grouping of spindle activity during slow oscillations in human non-rapid eye movement sleep.” J Neurosci 22(24): 10941–10947. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Morin A, Doyon J, Dostie V, Barakat M, Hadj Tahar A, Korman M, Benali H, Karni A, Ungerleider LG and Carrier J (2008). “Motor sequence learning increases sleep spindles and fast frequencies in post-training sleep.” Sleep 31(8): 1149–1156. [PMC free article] [PubMed] [Google Scholar]
  46. Muehlroth BE, Sander MC, Fandakova Y, Grandy TH, Rasch B, Shing YL and WerkleBergner M (2019). “Precise Slow Oscillation-Spindle Coupling Promotes Memory Consolidation in Younger and Older Adults.” Sci Rep 9(1): 1940. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Nadasdy Z, Hirase H, Czurko A, Csicsvari J and Buzsaki G (1999). “Replay and time compression of recurring spike sequences in the hippocampus.” J Neurosci 19(21): 9497–9507. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Nagai H, de Vivo L, Bellesi M, Ghilardi MF, Tononi G and Cirelli C (2016). “Sleep Consolidates Motor Learning of Complex Movement Sequences in Mice.” Sleep. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Niethard N, Ngo HV, Ehrlich I and Born J (2018). “Cortical circuit activity underlying sleep slow oscillations and spindles.” Proc Natl Acad Sci U S A 115(39): E9220–E9229. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Nishida M and Walker MP (2007). “Daytime Naps, Motor Memory Consolidation and Regionally Specific Sleep Spindles.” PLoS ONE 2(4): e341. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Nolan MF, Malleret G, Lee KH, Gibbs E, Dudman JT, Santoro B, Yin D, Thompson RF, Siegelbaum SA and Kandel ERJC (2003). “The hyperpolarization-activated HCN1 channel is important for motor learning and neuronal integration by cerebellar Purkinje cells.” 115(5): 551–564. [DOI] [PubMed] [Google Scholar]
  52. Oostenveld R, Fries P, Maris E and Schoffelen JM (2011). “FieldTrip: Open source software for advanced analysis of MEG, EEG, and invasive electrophysiological data.” Comput Intell Neurosci 2011: 156869. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Oyanedel CN, Binder S, Kelemen E, Petersen K, Born J and Inostroza M (2014). “Role of slow oscillatory activity and slow wave sleep in consolidation of episodic-like memory in rats.” Behav Brain Res 275: 126–130. [DOI] [PubMed] [Google Scholar]
  54. Pettibone WD, Kam K, Chen RK and Varga AW (2019). “Necessity of Sleep for Motor Gist Learning in Mice.” Frontiers in Neuroscience 13(293). [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Plihal W and Born J (1997). “Effects of early and late nocturnal sleep on declarative and procedural memory.” J Cogn Neurosci 9(4): 534–547. [DOI] [PubMed] [Google Scholar]
  56. Purcell SM, Manoach DS, Demanuele C, Cade BE, Mariani S, Cox R, Panagiotaropoulou G, Saxena R, Pan JQ, Smoller JW, Redline S and Stickgold R (2017). “Characterizing sleep spindles in 11,630 individuals from the National Sleep Research Resource.” Nature Communications 8: 15930. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Ramanathan DS, Gulati T and Ganguly K (2015). “Sleep-Dependent Reactivation of Ensembles in Motor Cortex Promotes Skill Consolidation.” PLoS Biol 13(9): e1002263. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Rasch B, Buchel C, Gais S and Born J (2007). “Odor cues during slow-wave sleep prompt declarative memory consolidation.” Science 315(5817): 1426–1429. [DOI] [PubMed] [Google Scholar]
  59. Rasch B, Pommer J, Diekelmann S and Born J (2009). “Pharmacological REM sleep suppression paradoxically improves rather than impairs skill memory.” Nat Neurosci 12(4): 396–397. [DOI] [PubMed] [Google Scholar]
  60. Rosanova M and Ulrich D (2005). “Pattern-specific associative long-term potentiation induced by a sleep spindle-related spike train.” J Neurosci 25(41): 9398–9405. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Rovo Z, Matyas F, Bartho P, Slezia A, Lecci S, Pellegrini C, Astori S, David C, Hangya B, Luthi A and Acsady L (2014). “Phasic, nonsynaptic GABA-A receptor-mediated inhibition entrains thalamocortical oscillations.” J Neurosci 34(21): 7137–7147. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Sausbier M, Hu H, Arntz C, Feil S, Kamm S, Adelsberger H, Sausbier U, Sailer C, Feil R and Hofmann F. J. P. o. t. N. A. o. S. (2004). “Cerebellar ataxia and Purkinje cell dysfunction caused by Ca2+-activated K+ channel deficiency.” 101(25): 9474–9478. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Schoffelen JM, Poort J, Oostenveld R and Fries P (2011). “Selective movement preparation is subserved by selective increases in corticomuscular gamma-band coherence.” J Neurosci 31(18): 6750–6758. [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Seibt J, Richard CJ, Sigl-Glockner J, Takahashi N, Kaplan DI, Doron G, de Limoges D, Bocklisch C and Larkum ME (2017). “Cortical dendritic activity correlates with spindle-rich oscillations during sleep in rodents.” Nat Commun 8(1): 684. [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Siapas AG and Wilson MA (1998). “Coordinated Interactions between Hippocampal Ripples and Cortical Spindles during Slow-Wave Sleep.” Neuron 21(5): 1123–1128. [DOI] [PubMed] [Google Scholar]
  66. Smith C and MacNeill C (1994). “Impaired motor memory for a pursuit rotor task following Stage 2 sleep loss in college students.” J Sleep Res 3(4): 206–213. [DOI] [PubMed] [Google Scholar]
  67. Staresina BP, Bergmann TO, Bonnefond M, van der Meij R, Jensen O, Deuker L, Elger CE, Axmacher N and Fell J (2015). “Hierarchical nesting of slow oscillations, spindles and ripples in the human hippocampus during sleep.” Nat Neurosci 18(11): 1679–1686. [DOI] [PMC free article] [PubMed] [Google Scholar]
  68. Steriade M (2006). “Grouping of brain rhythms in corticothalamic systems.” Neuroscience 137(4): 1087–1106. [DOI] [PubMed] [Google Scholar]
  69. Steriade M, Nunez A and Amzica F (1993). “A novel slow (< 1 Hz) oscillation of neocortical neurons in vivo: depolarizing and hyperpolarizing components.” J Neurosci 13(8): 3252–3265. [DOI] [PMC free article] [PubMed] [Google Scholar]
  70. Tamaki M, Huang TR, Yotsumoto Y, Hamalainen M, Lin FH, Nanez JE Sr., Watanabe T and Sasaki Y (2013). “Enhanced spontaneous oscillations in the supplementary motor area are associated with sleep-dependent offline learning of finger-tapping motor-sequence task.” J Neurosci 33(34): 13894–13902. [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Tamaki M, Matsuoka T, Nittono H and Hori T (2009). “Activation of fast sleep spindles at the premotor cortex and parietal areas contributes to motor learning: a study using sLORETA.” Clin Neurophysiol 120(5): 878–886. [DOI] [PubMed] [Google Scholar]
  72. Uygun DS, Katsuki F, Bolortuya Y, Aguilar DD, McKenna JT, Thankachan S, McCarley RW, Basheer R, Brown RE, Strecker RE and McNally JM (2019). “Validation of an automated sleep spindle detection method for mouse electroencephalography.” Sleep 42(2). [DOI] [PMC free article] [PubMed] [Google Scholar]
  73. Varga AW, Kang M, Ramesh PV and Klann E (2014). “Effects of acute sleep deprivation on motor and reversal learning in mice.” Neurobiol Learn Mem 114: 217–222. [DOI] [PMC free article] [PubMed] [Google Scholar]
  74. Vyazovskiy VV, Achermann P, Borbely AA and Tobler I (2004). “The dynamics of spindles and EEG slow-wave activity in NREM sleep in mice.” Arch Ital Biol 142(4): 511–523. [PubMed] [Google Scholar]
  75. Walker MP, Brakefield T, Morgan A, Hobson JA and Stickgold R (2002). “Practice with sleep makes perfect: sleep-dependent motor skill learning.” Neuron 35(1): 205–211. [DOI] [PubMed] [Google Scholar]
  76. Walker MP, Brakefield T, Seidman J, Morgan A, Hobson JA and Stickgold R (2003). “Sleep and the time course of motor skill learning.” Learn Mem 10(4): 275–284. [DOI] [PMC free article] [PubMed] [Google Scholar]
  77. Wilson MA and McNaughton BL (1994). “Reactivation of hippocampal ensemble memories during sleep.” Science 265(5172): 676–679. [DOI] [PubMed] [Google Scholar]
  78. Yang G, Lai CS, Cichon J, Ma L, Li W and Gan WB (2014). “Sleep promotes branch-specific formation of dendritic spines after learning.” Science 344(6188): 1173–1178. [DOI] [PMC free article] [PubMed] [Google Scholar]
  79. Yordanova J, Kirov R, Verleger R and Kolev V “Dynamic coupling between slow waves and sleep spindles during slow wave sleep in humans is modulated by functional pre-sleep activation.” (2045–2322 (Electronic)). [DOI] [PMC free article] [PubMed]

Associated Data

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

Supplementary Materials

1
NIHMS1545113-supplement-1.docx (1.8MB, docx)

RESOURCES