Skip to main content
NCBI home page
UKPMC Funders Author Manuscripts logoLink to UKPMC Funders Author Manuscripts
. Author manuscript; available in PMC: 2023 Apr 11.
Published in final edited form as: Neuropsychologia. 2010 Oct 23;49(1):115–123. doi: 10.1016/j.neuropsychologia.2010.10.025

Keeping time in your sleep: Overnight consolidation of temporal rhythm

Penelope A Lewis 1,2,✉, Tom J Couch 1, Matthew P Walker 3
PMCID: PMC7614423  EMSID: EMS172343  PMID: 20974158

Abstract

Temporal processing forms the basis of a vast number of human behaviours, from simple perception and action to tasks like locomotion, playing a musical instrument, and understanding language. Growing evidence suggests that these procedural skills are consolidated during sleep, however investigation of such learning has focused upon the order in which movements are made rather than their temporal dynamics. Here, we use psychophysics and neuroimaging to explore the possibility that temporal aspects of such skills are also enhanced over a period of sleep. Behaviourally, our examinations of motor (tapping a finger in time with a temporal rhythm) and perceptual (monitoring a temporal rhythm for deviants) tasks reveal post-sleep improvements in both domains. Functionally, we show that responses in striatum, supplementary motor area, and lateral cerebellum are stronger during motor, but not during perceptual timing post-sleep. Conversely, responses in posterior hippocampus are stronger during perceptual, but not motor timing post-sleep. Our data support the proposal that these two forms of timing draw on different brain mechanisms, with motor timing using a more automatic system while perceptual timing of the same rhythm requires a more cognitive form of processing.

Keywords: sleep, memory, time perception, striatum, hippocampus

Introduction

Humans spend a large proportion of their waking lives performing skilled tasks which require little attention. These include movement based activities like walking or riding a bicycle and perceptual activities like reading, understanding speech, or judging the depth of a visual scene. Careful analysis has demonstrated that many of these tasks can be broken down into independently learned temporal and ordinal elements (Ullen & Bengtsson, 2003). Furthermore, many such procedural skills have been shown to improve across periods of sleep (Walker & Stickgold, 2006; Walker & Stickgold, 2004; Born et al., 2006; Maquet et al., 2000; Laureys et al., 2002; Plihal & Born, 1997; Fischer et al., 2006; Fischer et al., 2002; Maquet et al., 2003; Fenn et al., 2003; Karni et al., 1994; Stickgold et al., 2000; Rauchs et al., 2005; Dang-Vu et al., 2006). Most examinations of such off-line enhancement have focussed on the order of responses (Walker et al., 2002; Cohen et al., 2005; Robertson et al., 2004; Spencer et al., 2006; Walker et al., 2005). Consolidation of the temporal aspects of these tasks therefore remains largely unexplored, though one study (Maquet et al., 2003) demonstrated that the emergent timing in visuomotor tracking is strengthened over sleep. In the current report, we aim to build on this work by determining whether the event timing in skilled rhythm processing (Zelaznik et al., 2000; Spencer et al., 2003; Zelaznik et al., 2002) also undergoes overnight consolidation, and to examine the neuroplasticity associated with such changes.

Studies of procedural learning have shown that the brain areas in which responses are enhanced as learning progresses (Doyon et al., 2009) commonly exhibit greater activity when the task is performed after sleep. This is true for both motor (Walker et al., 2005; Albouy et al., 2008) and perceptual (Walker et al., 2005) forms of learning. A wide range of evidence suggests that different temporal processing mechanisms are recruited in different forms of timing task (Rammsayer, 1999; Lewis & Miall, 2003; Zelaznik et al., 2002; Wiener et al., 2010), with the distinct, but potentially overlapping systems used in motor and perceptual timing (Clarke et al., 1996; Bueti et al., 2008; Wiener et al., 2010; Lewis & Miall, 2003) as a prime example. Here, we explore this distinction by studying post-sleep enhancements in brain activity during both forms of timing.

To allow examination of both motor and perceptual timing, we elected to use auditory rhythms akin to those found in music as our stimuli. In our motor paradigm, ‘Tapping’ (Wing & Kristofferson, 1973), participants synchronised button presses with a rhythm of auditory beeps which occurred in a repeating temporal pattern, then continued to tap the same temporal rhythm without external cues (Figure 1A). In our perceptual paradigm, ‘Monitoring’, participants listened to an equivalent auditory rhythm and monitored it for rare temporal deviants (Figure 1B), pressing a button when these were detected. We performed two experiments using these paradigms: one which examined the interaction between sleep and behavioural performance in the Tapping task, and a second which used functional magnetic resonance imaging (fMRI) to monitor sleep-related alterations in the brain responses associated with both Tapping and Monitoring tasks.

Figure 1. Tasks and paradigms.

Figure 1

Open in a new tab

A) The tapping task: in each trial, participants tapped their right index finger in time with an auditory rhythm, then continued to tap the same temporal rhythm once the auditory cues had ceased. Each oblong box represents a repetition of the full rhythm sequence (or ‘bar). B) The Monitoring task: participants listened to a repeating auditory rhythm and pressed a key when they detected a temporal deviant. Deviants were rare and consisted of an auditory beep which occurred 200 milliseconds to early or too late, disrupting the timing of the two adjacent intervals in the rhythm. C) The testing schedule for AM/PM/AM and PM/AM/PM groups in Experiment 1. D) The testing schedule for Sleep and Wake groups in Experiment 2.

Based upon a prior analysis of the systems used in automatic and cognitively controlled timing (Lewis & Miall, 2003), we expected post-sleep enhancements within the movement control system, particularly the supplementary motor area (SMA), cerebellum, and striatum, during motoric rhythm tapping (Wiener et al., 2010), and in higher cognitive areas, potentially including prefrontal, and parietal cortices, during perceptual rhythm Monitoring (Lewis and Miall 2006).

Materials and Methods

Participants

All participants were consenting, healthy, right handed, and had no history of psychiatric illness. 14 of these (7 male and 7 female, mean age 27) participated in Experiment 1, and 24 (12 male and 12 female, mean age 25, +/- SEM 1 year) in Experiment 2. All participants were instructed to abstain from alcohol, caffeine, and other drugs during, and for twenty four hours prior to, the experiment. Experiments were approved by the Liverpool research ethics committee.

Behavioural tasks

Experiment 1 – behavioural testing

Participants performed a motor synchronisation continuation task (Wing & Kristofferson, 1973) (Figure 1A, Tapping) in three sessions. The rhythm they learned was based on those in (Lewis et al., 2004) and consisted of eight temporal intervals: 107, 429, 214, 1065, 536, 643, 321, and 857 milliseconds, in that order, with each bar (repeating sequence of intervals) lasting 4172 milliseconds. Each trial was initiated by a press to the spacebar, and contained a synchronisation phase immediately followed by a continuation phase. During synchronisation, the rhythm was presented via auditory beeps (250 Hz for 25 ms) and was repeated 6 times (25 seconds total), during continuation auditory presentation stopped and participants were exposed to 47 seconds of silence, terminated by a high pitched beep. Participants were instructed to synchronise right index finger button presses with the beeps during synchronisation (6 bars), then continue to press the button in the same temporal sequence during continuation (11 bars). There were 8 synchronisation/continuation trials in every session.

Fourteen participants were randomly divided into two groups, determined by the time of day at which they were trained and tested: an AM/PM/AM group (seven participants) and a PM/AM/PM group (seven participants). In the AM/PM/AM group, Session 1 was performed in the morning of day 1, Session 2 that evening, and Session 3 in the morning of day 2 (Figure 1C). In the PM/AM/PM group, the order was reversed: Session 1 was performed on the evening of Day 1, Session 2 the next morning (Day 2), and Session 3 that evening. For these two groups, all experimental sessions began between 8:00 and 11:00 AM or 20:00 and 23:00 PM, and delays between sessions were always 12 (+/- 1) hours.

Experiment 2 – functional imaging

Participants in Experiment 2 were divided into Sleep and Wake groups, each of these comprising 6 males and 6 females (12 participants total per group, 24 total in the experiment). Sleep participants were trained in the evening and scanned next morning, Wake participants were trained in the morning and scanned that evening (Figure 1D). Each group was instructed to go about their normal routine (e.g. going to work, class, or sleep as usual). The Wake group was instructed to abstain from daytime napping.

The Tapping task in Experiment 2 was similar to that used in Experiment 1, but implemented two different rhythm sequences (A and B). Intervals in these rhythms were: 640, 160, 560, 960, 320, 400, 240, and 720 milliseconds in A, and 320, 1040, 800, 160, 240, 400, 480, and 560 milliseconds in B, and were presented in those orders using auditory beeps (again 250 Hz and 25 milliseconds duration) with a repeating bar of 4000 milliseconds. Participants attended a Training session and a Scanning session that were separated by 12 (+/- 1) hours. All experimental sessions began between 8:00 and 11:00 AM or 20:00 and 23:00 PM.

The first session, Training, was performed outside the MRI scanner. Participants learned either rhythm A or B (Learned rhythm, counterbalanced across participants) via a series of trials in which they listened to an auditory presentation of the rhythm and synchronised right index finger button presses to it for 6 bars (synchronization, 24 seconds), then continued to press the button in the same temporal rhythm in silence for a further 12 bars (continuation, 48 seconds). Continuation was terminated by a high-pitched beep. Each training session contained 8 trials of synchronisation followed by continuation, and each trial was initiated by a press to the spacebar. The second session, Scanning, was performed inside the MRI scanner. Learned and Unlearned rhythms were interleaved in 8 alternating trials, with each trial comprising 8 bars (repetitions) of synchronisation and 5 bars of continuation. These rhythm trials were randomly interleaved with 20 second epochs of a fixation baseline. Visual cues specifying ‘rest’, ‘synchronise’, or ‘continue’, were presented as appropriate.

In addition to the motor synchronisation/continuation task (Tapping, Figure 1A), participants performed a perceptual monitoring task (Monitoring, Figure 1B) during the scanning session of Experiment 2. In Monitoring, participants listened to rhythms A and B (Learned and Unlearned) in 10 alternating trials of 32 seconds, and responded by pressing a button with the right index finger when a beep was misplaced in time by 200 milliseconds (deviation). A maximum of 3 deviations occurred in any trial, and these could not occur within the first two bars of the rhythm. Participants practiced this task for two trials before the start of scanning to become familiar with the Unlearned rhythm. As with Tapping, trials of Monitoring were randomly interleaved with 20 second epochs of fixation baseline and visual cues specifying ‘rest’ or ‘monitor’ were presented as appropriate.

The same Learned and Unlearned rhythms were used for Monitoring as Tapping, but Monitoring was performed only at Scanning (not in training). This meant that monitored rhythms were learned via the synchronisation/continuation Tapping paradigm in the Training session. Participants performed both Monitoring and Tapping tasks in the fMRI scanner and the order in which these were performed was counterbalanced across participants, with data for each task collected in a separate scanning session (run).

Behavioural analysis

Performance on Tapping was assessed during the continuation phase. Each sequence of button presses was aligned with the presented rhythm using the longest and shortest intervals according to the method in (Lewis et al., 2004), and those estimates differing from the target interval by >95% were excluded as outliers. For each trial, the coefficient of variation (CV) for each target interval was calculated by dividing the mean of estimates by the standard deviation of estimates. CV’s from all 8 target intervals were then averaged to give a single measure for the trial. In the Monitoring task, behavioural performance was measured as Performance = (Hits – False Alarms) / total number of actual deviants, where false alarms included button presses >1 second after the deviant.

Equipment and fMRI parameters

The behavioural paradigms for both experiments were written in Cogent on a Matlab 6.5 base. We used a Domino 2 system from Micromint to log responses with accuracy ~1 millisecond. During fMRI scanning, we used an MR compatible audio setup from MR Confon to present auditory stimuli.

Functional imaging was performed on a 3-T Trio MR scanner (Siemens Vision, Erlangen, Germany) with an 8 channel head coil. We used echo-planar imaging to obtain image volumes with 31 contiguous oblique transverse slices every 2 sec (voxel size 3.5 3.5 2 mm, 80% gap, TE 30 milliseconds) covering the whole brain. Data for Tapping and Monitoring were collected in separate sessions (runs).

Functional analysis

Functional MRI images were analysed using the statistical parametric mapping (SPM2) software package (Wellcome Department of Cognitive Neurology, London, UK, http://www.fil.ion.ucl.ac.uk/spm). The functional volumes from each session were corrected for head motion by realigning with the first image, spatially normalised to an EPI template corresponding to the Montreal Neurological Institute (MNI) space, and smoothed using a Gaussian Kernel size of 5 mm full-width at half-maximum.

Localised analysis

To characterise functional responses, the data were examined using a 2-level random-effects analysis. First level (within participant) analyses were performed separately for Tapping and Monitoring since these data were acquired in separate sessions. In this analysis, the responses of individual participants were examined using an individual first-level design matrix. For Tapping, this model included four main regressors: Learned synchronisation, Learned continuation, Unlearned synchronisation, and Unlearned continuation. To ensure that results were not influenced by performance related brain responses, mean CV (calculated using the 20 seconds of continuation data in each trial) was included as a parametric regressor of no interest for each continuation block. For Monitoring, the first-level design matrix included four regressors: Learned and Unlearned rhythm trials, deviant presentation times, and button-press response times. To control for motion artefacts, six ridged body movement parameters were included as regressors of no interest in each design matrix. Parameter estimates reflecting the height of the hemodynamic response function for each regressor were calculated at each voxel. Contrast images providing a direct comparison of responses in Learned and Unlearned conditions were then calculated for both Tapping and Monitoring. The resulting images were used in a second-level random effects analyses that combined data across participants.

In order to isolate changes in activation which developed across sleep, and to compare and contrast these across tasks, a series of second-level contrasts were performed. To independently isolate activations within each task while controlling for session-specific effects, the interaction Sleep[Learned>Unlearned] < >Wake[Learned>Unlearned] was calculated. Comparison of Learned and Unlearned rhythms at the subject level served two functions. First, it controlled for circadian factors by removing activities associated with retrieval at a specific time of day, following the method of (Walker et al., 2005). Second, it allowed isolation of learning-related responses. The contrast images resulting from this subtraction were used to form a second-level one-way ANOVA which compared Sleep and Wake conditions. This analysis was performed separately for Tapping (contrast 1) and Monitoring (contrast 2).

To determine how overnight changes differed across the two tasks, the interaction terms generated by contrasts 1 and 2 were compared directly. This was achieved by forming a second level ANOVA with the two level factors: Tapping (Sleep and Wake), and Monitoring (Sleep and Wake). An SPM conjunction (Friston et al., 2005) was used to test for regions where activation was enhanced after sleep in both tasks (contrast 3). The interaction between task and group (contrast 4) was used to identify regions where overnight alterations in activation differed significantly between tasks, calculated as: Tapping [Sleep(Learned>Unlearned)>Wake(Learned>Unlearned)]<>Monitoring [Sleep(Learned>Unlearned)>Wake(Learned>Unlearned)].

Responses in contrasts 1-4 were considered significant at p < 0.001 uncorrected and a cluster size of k=5 voxels. To test for correlations between consolidation-related improvements in performance and neural activity, we extracted the parameter estimates associated with the contrast Learned – Unlearned for the group peak voxels in hippocampus, striatum, and dorsal cerebellum for each participant, and regressed these against improvement in performance (Train CV – Test CV) for Sleep and Wake groups.

Results

Experiment 1- behavioural testing

In Experiment 1, performance was assessed using a 2x2 mixed analysis of variance (ANOVA) with delay type (sleep/wake) as the within subject factor, group (AM/PM/AM or PM/AM/PM as determined by the time of day of the testing sessions) as the between subject factor, and the difference between CVs across offline delay periods (Delay1: [session1–session2], and Delay2: [session2–session3]) as the dependent variable. This identified a main effect of delay type (ANOVA F(1,12)=9.94, p<0.008), but no effect of group, and no interaction. Post-hoc two-tailed t-tests between adjacent sessions revealed improvement over delays containing sleep, but not over equivalent delays containing only wakefulness. Specifically, in the AM/PM/AM group, there was an overnight improvement (M=27% +/- 12% SEM) from the PM to the subsequent AM session (t(12)=2.44, p=0.031), and in the PM/AM/PM group there was an overnight improvement (M=32% +/- SEM 7%) from the AM to the subsequent PM session (t(12)=2.57, p=0.029), see figure 2A&B. Since performance improved only after periods of sleep, and considering that these improvements were maintained in subsequent sessions, these results indicate that learned representations of temporal rhythm were modulated offline, across a night of sleep.

Figure 2. Behavioural performance.

Figure 2

Open in a new tab

A) Behavioural results from Experiment 1 pooled across all trials in a session. From L to R results are shown for the AM/PM/AM, PM/AM/PM, and 24 hour control groups. Comparison of performance over the two 12 hour consolidation periods showed a significant decrease in CoV after each epoch of sleep, but not after equivalent epochs of wakefulness (t-test p=0.03 for both groups). B) To further illustrate these results, data from the AM/PM/AM and PM/AM/PM groups are shown in a trial-by-trial basis (each point represents a group mean). C) Behavioural results from the Monitoring task in Experiment 2. The figure demonstrates a marked improvement in Monitoring performance after sleep but not after a similar period of wakefulness. A 2-tailed t-test confirms that the difference in Monitoring percent correct [Learned – Unlearned] is greater in SLEEP than WAKE (p=0.006). Error bars show one SEM.

To determine whether there was an interaction between time of day and tapping performance, a paired two-tailed t-test was used to compare mean CV for session 1 across the two groups (AM/PM/AM and PM/AM/PM). This revealed no significant difference between performance on initial training in the AM and PM sessions (t(12)=0.53, p=0.481). As a second control for circadian influences upon performance, CVs in the first and last session (24 hours later) were compared using a 2x2 ANOVA with the factors group and session (first/last). This revealed a main effect of session, (F(1,12)=7.63, p=0.017), demonstrating that performance by the same participant was improved at retest 24 hours after the first session irrespective of whether this occurred in the morning or evening, Figure 2A. There was no interaction between group and session (first/last) (F(1,12)=0.21, p=0.656), showing that the time of day at which participants were trained or tested did not impact upon performance 24 hours later.

Overall, the behavioural data from Experiment 1 show marked improvement in tapping performance after a retention interval containing sleep, but not after an equivalent interval containing only wakefulness. Furthermore, these enhancements in performance were shown to be independent of diurnal test time.

Experiment 2 – functional imaging

Behavioural results

In Experiment 2, offline changes in Monitoring behaviour were assessed using a 2x2 mixed ANOVA with the factors sequence (Learned/Unlearned), and group (Sleep/Wake), and with Performance (see methods) as the dependent variable. This revealed a performance benefit for Learned sequences (main effect of learning, F(1,22)=22.76, p<0.001), with a specific facilitation of this benefit in those who had slept (interaction between sequence and group, F(1,22)=9.37, p=0.006). A post-hoc independent samples t-test comparing the differences in performance on Learned and Unlearned rhythms between groups reinforced this finding by revealing a greater advantage of learning in Sleep than Wake (two-tailed t(22)= -3.06, p=0.006, see Figure 2C). This overnight improvement suggests that the consolidation of timing behaviour observed in Experiment 1 is not merely associated with the motor components of tapping a rhythm, but also generalises to the perceptual domain.

Offline changes in Tapping performance on the learned sequences in Experiment 2 were assessed using a 2x2 mixed ANOVA with the factors session (session 1 / session 2) and group (Sleep/Wake), and with CV as the dependent variable. This showed no significant results (F(1,22)=0.93; p=0.346). As an alternate way of examining the impact of consolidation, a second 2x2 mixed ANOVA tested for differences in performance of Learned and Unlearned sequences during fMRI scanning. The factors were sequence (Learned/Unlearned) and group (Sleep/Wake). This also failed to reveal significant results (main effect of sleep F(1,22)=0.78; p=0.387).

Overall, the behavioural results from Experiment 2 provide evidence that a centralised (nonmotor) representation of the learned rhythm is strengthened across a night of sleep. Interestingly, this strengthened representation was indexed by enhanced performance on Monitoring, but not Tapping. The possibility that this difference between tasks may be due to difficulties that the noisy environment and constrained physical position associated with the fMRI scanner posed for skilled performance of the Tapping task is considered in the discussion.

fMRI Results

Within the functional data from Tapping, the interaction analysis (contrast 1) revealed increased activation throughout the motor system, with peak responses falling in the globus pallidus of the left striatum, the SMA, and the dorso-lateral cerebellum (p<0.001 uncorrected, Figure 3 & Table 1A). To test for a predictive relationship between overnight improvement in behavioural performance and the overnight alterations in brain responses while performing the task, parameter estimates for the peak voxels in each response were regressed against a measure of how much performance improved after consolidation (Training CV–Scanning CV) for the Learned sequence. In the Sleep group this revealed significant correlations for both striatum (p<0.05, R(12)=0.6) and cerebellum (p<0.05, R(12)=0.57), but not for the SMA. These correlations were not apparent in the Wake group (p= 0.91, R(12) = 0.04 for striatum, and p=0.85, R(12)=.03, for cerebellum) see Figure 4.

Figure 3. Enhanced fMRI response to motor and perceptual timing after sleep.

Figure 3

Open in a new tab

These results were calculated using the contrast SLEEP[Learned–Unlearned]–WAKE[Learned–Unlearned] for motor (Tapping) and perceptual (Monitoring) tasks. Peak parameter estimates are shown to the right (L=Learned, UL=Unlearned). Data are rendered on the SPM canonical brain at a visualisation threshold of p<0.005.

Table 1. functional responses.
Voxel count Z Coordinates Anatomical region
A) Tapping: SLEEP [ L - U ] - WAKE [ L - U ] (increases, contrast 1)
7 3.6 20 -30 -28 cerebellar culmen
5 3.5 -24 -22 0 globuspallidus
6 3.5 10 -16 58 medial superior frontal
gyrus (SMA)
6 3.4 18 64 2 frontal pole
B) Tapping: SLEEP [ L - U ] - WAKE [ L - U ] (decreases, contrast 1)
18 3.7 -22 -6 72 PMC
C) Monitor: SLEEP [ L - U ] - WAKE [ L – U ] (increases, contrast 2)
15 3.4 -32 -36 -2 posterior hippocampus
Open in a new tab

MNI coordinates for the peak voxels of clusters surviving at p=0.001 uncorrected and k=5.

Abbreviations: SMA = supplementary motor area, PMC = premotor cortex.

Figure 4. Correlation between post-sleep fMRI response and performance.

Figure 4

Open in a new tab

Responses in striatum and dorsal cerebellum correlate significantly with improvement in performance CV (Train-Test) after a delay containing sleep, but not after an equivalent delay containing wakefulness alone.

In Monitoring, the interaction analysis (contrast 2) revealed greater responses in left posterior hippocampus post-sleep, Figure 3, see Table 1C. The peak parameter estimate for this response correlated neither with performance of the Learned rhythm (R(12)=0.49, p= 0.11) nor with differences between performance on Learned and Unlearned rhythms (R(12)= -0.2, p=0.53) in the Sleep group.

To test for a dissociation between the brain structures showing enhanced post-sleep responses for Motor or Perceptual timing, results of the interaction analyses (contrasts 1 & 2 above) were compared for Tapping and Monitoring (contrast 1 < > contrast 2). This revealed that activation in striatum was significantly more enhanced in Tapping than in Monitoring, supporting the suggestion that motor timing engages striatum-dependent memory systems more strongly after sleep than perceptual timing. Finally, an SPM conjunction was performed across data from Tapping and Monitoring (contrasts 1 & 2 above) and revealed no region of common activity.

Overall our functional results show that the brain regions associated with rhythm processing are altered after a night of sleep. Furthermore, they support a dissociation between the brain regions involved in motor and perceptual timing, with components of the motor system showing enhanced responses during the former (Tapping), while the hippocampus shows enhanced responses during the latter (Monitoring) post-sleep.

Discussion

In this report we demonstrate that knowledge of a temporal rhythm can consolidate overnight. We also show that motor timing places a greater demand upon the SMA, lateral cerebellum, and striatum after sleep, while perceptual timing places a greater demand on posterior hippocampus under the same circumstances. This pattern of response supports a dissociation between the brain systems involved in automatic and cognitively controlled timing, with the former drawing more heavily on the motor system, while the latter draws more heavily upon higher cognitive areas (Lewis and Miall, 2003).

Overnight improvements in performance

Experiment 1 demonstrated that participants can tap a complex temporal rhythm with greater consistency after a night of sleep. Good performance on this task requires both a strong representation of the temporal durations in the rhythm and a highly controlled motor output. Because our behavioural analysis did not discriminate between these two factors, the improvement we observed could have been due to enhancements in either domain. Experiment 2 clarified this by revealing post-sleep improvements in perceptual monitoring of the rhythm (Figure 2C), thus demonstrating that such consolidation is not limited to motor control but also extends to a more abstract representation of temporal sequences.

Overnight enhancements in brain response

The additional neural responses elicited by both rhythm tasks when performed after a night of consolidation suggest that the way a temporal rhythm is represented in the brain alters across sleep. Importantly, the location of the observed increases differed across tasks, with cerebellum, striatum, and SMA more active during Tapping, while posterior hippocampus was more active during Monitoring post-sleep. The areas where responses were enhanced during Tapping are all strongly associated with time measurement (Macar et al., 1999; Buhusi & Meck, 2005; Ivry, 1997). Furthermore the SMA and cerebellum have been specifically linked to motor timing (Lewis & Miall, 2003; Wiener et al., 2010), and the lateral cerebellum may be particularly important for predicting the temporal dynamics of sequential movements (Miall et al., 1993; Miall et al., 1987; Ivry et al., 1988; Sakai et al., 2002).

Although there is as yet no strong consensus regarding the precise roles of these structures in a clock mechanism (Wiener et al., 2010) for a recent review, most models agree that a clock must contain both a measurement device which marks time in a reliable way, (e.g. along the lines of a pizo-electric crystal (Gibbon, 1977) or predictably decaying function (Staddon & Higa, 1999)), and some form of memory store. Having both of these components is important because it allows for comparison of currently measured intervals against remembered targets (Gibbon, 1977; Buhusi & Meck, 2005; Staddon & Higa, 1999; Bugman, 1998; Mauk & Buonomano, 2004). As the time measurement component is by definition a stable process, it may be unreasonable to expect the brain regions performing this type of function to show an altered pattern of activity after consolidation. Instead, post-consolidation increases in brain activity are likely associated with a strengthened mnemonic representation of the learned temporal durations, and may therefore relate to the memory component of the clock system.

The increases in brain response which we observed during post-sleep Tapping are easy to reconcile with a memory based interpretation since both the striatum (Cohen, 1984; Poldrack et al., 2001; Poldrack & Foerde, 2008; Poldrack & Packard, 2003) and cerebellum (Nixon & Passingham, 2001; Nixon & Passingham, 2000; Spencer & Ivry, 2009) are critical for motor learning, although the latter may be more important at earlier stages of consolidation (Doyon et al., 2009; Doyon et al., 2003). The increase in hippocampal activity which we observed during Monitoring post-sleep is also consistent with a memory based enhancement, since the hippocampus plays a well established role in declarative memory (Eichenbaum, 2006; Squire & Zola-Morgan, 1991). Furthermore, hippocampal damage has been shown to alter memory for the duration of learned target intervals in rats, leading to systematic errors in a temporal reproduction task (Meck et al., 1984). These data from the basis of a well-developed theory about the involvement of hippocampus in memory for temporal durations (Meck, 2005). Finally, because the hippocampus is closely interlinked with the prefrontal and parietal cortices (Eichenbaum 2000) it is plausible to assume that this structure acts as the memory store component of a timing system which draws upon these structures fore the time measurement component. The pattern of post-sleep enhancements we observed here is therefore compatible with the suggestion that motor timing draws more strongly on the motor control system while nonmotor timing draws on structures associated with higher cognitive processing (Lewis & Miall, 2003) even though we observed no enhancement in prefrontal or parietal cortices during post-sleep Monitoring.

Absence of overnight Tapping improvement in Experiment 2

A number of fMRI studies have reported evidence for sleep-dependent changes in memory representations that were not associated with concomitant behavioural improvements, see (Sterpenich et al., 2007; Sterpenich et al., 2009; Walker et al., 2005) for examples. Such findings support the widely held assumption that measurements of neural activity are more sensitive to the effects of consolidation than behavioural measures. The absence of an overnight improvement in Tapping performance in Experiment 2 raises the question of whether it is reasonable to make this same assumption in the current report. In considering this, we note that the marked overnight improvement by these same participants on the Monitoring task (Figure 2C) demonstrates that they had access to a strengthened representation of the rhythm post-sleep. Furthermore, although we did not find an overnight improvement in Tapping performance among these participants at the group level, there was a significant correlation between the extent to which individuals improved and the functional response in both striatum and cerebellum (Figure 4). In the context of the post-sleep Tapping enhancement observed in Experiment 1 (performed outside the fMRI scanner Figure 2A & B), these findings suggest that the atypical environment of the scanner (e.g. the cramped supine position and loud background noise) may have impaired some participants’ ability to express consolidation based memory enhancements via the Tapping task, thus precluding observation of a behavioural effect at the group level in that task.

Control for circadian effects and interference from daytime activity

Could the time of day at which testing sessions occurred provide alternative explanation for the differences we observed between test sessions in terms of both behavioural performance and brain responses? Sessions occurring directly after retention across wake and retention across sleep differed not only in the brain state experienced during the retention interval (asleep or awake), but also in the time of day at which they were administered (9 AM or 9 PM). It is possible that non-specific processes, such as attentional abilities are superior in the morning compared to the evening. With respect to Experiment 1, a circadian-based explanation of our findings appears unlikely due to the absence of significant inter-group differences in performance at the first session. Specifically, although the AM/PM/AM group executed their first session in the morning, while the PM/AM/PM group executed their first session in the evening, performance did not differ significantly across groups in this session (p=0.48). Additionally, comparison of performance from the first and last sessions, (AM to AM and PM to PM respectively for the two groups) revealed improvements across a 24 hour delay (p<0.02), indicating that such enhancements were not linked to circadian test time. In Experiment 2, the impact of circadian factors upon brain response was controlled by the use of two-tiered contrasts in which Learned and Unlearned rhythms collected in interleaved blocks within the same fMRI session were compared prior to contrasts between Sleep and Wake groups. This arrangement meant that responses associated with performing the task at a specific time of day were disambiguated from those relating to the difference between consolidation across sleep and wake. In summary, both the characteristics of our results in Experiment 1 and the two-tiered design in Experiment 2 help to minimize concern that differences in performance and brain response observed after consolidation across wake and sleep could have been due to circadian effects. Instead, it is more parsimonious to conclude that brain state during retention may have modulated these effects.

Summary

In sum, our data indicate that knowledge of a temporal rhythm is strengthened across a period of sleep and suggest that the striatum and cerebellum serve a memory function in rhythm tapping, while the posterior hippocampus plays a similar role in rhythm monitoring. These findings concur with the suggestion that structures within the movement control system are recruited during motor timing, while structures associated with higher cognitive function are recruited during more cognitively controlled perceptual timing. The current report extends prior work in this area by showing such responses are enhanced post-sleep.

Acknowledgements

This work was funded by a Wellcome Trust VIP award and a Biotechnology and Biological Sciences Research Council (BBSRC) New investigator award [BB/F003048/1] to PL. We thank Uta Noppenoy for helpful design suggestions and the staff at MARIARC, Liverpool University’s MRI centre, for technical assistance. We are grateful to Katharina Von Kriegstein, Simon Durrant, Atsuko Takashima, and Patti Adank for critical reading of the manuscript.

References

  1. Albouy G, Sterpenich V, Balteau E, Vandewalle G, Desseilles M, ng-Vu T, et al. Both the hippocampus and striatum are involved in consolidation of motor sequence memory. Neuron. 2008;58:261–272. doi: 10.1016/j.neuron.2008.02.008. [DOI] [PubMed] [Google Scholar]
  2. Born J, Rasch B, Gais S. Sleep to remember. Neuroscientist. 2006;12:410–424. doi: 10.1177/1073858406292647. [DOI] [PubMed] [Google Scholar]
  3. Bueti D, Walsh V, Frith C, Rees G. Different brain circuits underlie motor and perceptual representations of temporal intervals. J Cogn Neurosci. 2008;20:204–214. doi: 10.1162/jocn.2008.20017. [DOI] [PubMed] [Google Scholar]
  4. Bugman G. Towards a neural model of timing. Biosystems. 1998;48:11–19. doi: 10.1016/s0303-2647(98)00045-8. [DOI] [PubMed] [Google Scholar]
  5. Buhusi CV, Meck WH. What makes us tick? Functional and neural mechanisms of interval timing. Nat Rev Neurosci. 2005;6:755–765. doi: 10.1038/nrn1764. [DOI] [PubMed] [Google Scholar]
  6. Clarke S, Ivry R, Grinband J, Shimizu N. In: Time, Internal Clocks, and Movement. Vroon GES, Vroon PA, editors. New York: Elsevier Science Ltd; 1996. Exploring the domain of the cerebellar timing system; pp. 257–280. [Google Scholar]
  7. Cohen DA, Pascual-Leone A, Press DZ, Robertson EM. Off-line learning of motor skill memory: a double dissociation of goal and movement. Proc Natl Acad Sci USA. 2005;102:18237–18241. doi: 10.1073/pnas.0506072102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cohen NJ. In: Neuropsychology of memory. Squire LR, Butters N, editors. New York: Guilford; 1984. Preserved learning capacity in amnesia: Evidence for multiple memory systems; pp. 83–103. [Google Scholar]
  9. Dang-Vu T, Desseilles M, Peigneux P, Maquet P. A role for sleep in brain plasticity. Pediatr Rehabil. 2006;9:98–118. doi: 10.1080/13638490500138702. [DOI] [PubMed] [Google Scholar]
  10. Doyon J, Bellec P, Amsel R, Penhune V, Monchi O, Carrier J, et al. Contributions of the basal ganglia and functionally related brain structures to motor learning. Behav Brain Res. 2009;199:61–75. doi: 10.1016/j.bbr.2008.11.012. [DOI] [PubMed] [Google Scholar]
  11. Doyon J, Penhune V, Ungerleider LG. Distinct contribution of the cortico-striatal and cortico-cerebellar systems to motor skill learning. Neuropsychologia. 2003;41:252–262. doi: 10.1016/s0028-3932(02)00158-6. [DOI] [PubMed] [Google Scholar]
  12. Eichenbaum H. Remembering: functional organization of the declarative memory system. Curr Biol. 2006;16:R643–R645. doi: 10.1016/j.cub.2006.07.026. [DOI] [PubMed] [Google Scholar]
  13. Fenn KM, Nusbaum HC, Margoliash D. Consolidation during sleep of perceptual learning of spoken language. Nature. 2003;425:614–616. doi: 10.1038/nature01951. [DOI] [PubMed] [Google Scholar]
  14. Fischer S, Drosopoulos S, Tsen J, Born J. Implicit learning -- explicit knowing: a role for sleep in memory system interaction. J Cogn Neurosci. 2006;18:311–319. [PubMed] [Google Scholar]
  15. Fischer S, Hallschmid M, Elsner AL, Born J. Sleep forms memory for finger skills. Proc Natl Acad Sci USA. 2002;99:11987–11991. doi: 10.1073/pnas.182178199. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Friston KJ, Penny WD, Glaser DE. Conjunction revisited. Neuroimage. 2005;25:661–667. doi: 10.1016/j.neuroimage.2005.01.013. [DOI] [PubMed] [Google Scholar]
  17. Gibbon J. Scalar expectancy theory and Weber’s law in animal timing. Psychological Review. 1977;84:279–325. [Google Scholar]
  18. Ivry R. Cerebellar timing systems. Int Rev Neurobiol. 1997;41:555–573. [PubMed] [Google Scholar]
  19. Ivry RB, Keele SW, Diener HC. Dissociation of the lateral and medial cerebellum in movement timing and movement execution. Exp Brain Res. 1988;73:167–180. doi: 10.1007/BF00279670. [DOI] [PubMed] [Google Scholar]
  20. Karni A, Tanne D, Rubenstein BS, Askenasy JJ, Sagi D. Dependence on REM sleep of overnight improvement of a perceptual skill. Science. 1994;265:679–682. doi: 10.1126/science.8036518. [DOI] [PubMed] [Google Scholar]
  21. Laureys S, Peigneux P, Perrin F, Maquet P. Sleep and motor skill learning. Neuron. 2002;35:5–7. doi: 10.1016/s0896-6273(02)00766-3. [DOI] [PubMed] [Google Scholar]
  22. Lewis PA. Finding the timer. Trends Cogn Sci. 2002;6:195–196. doi: 10.1016/s1364-6613(02)01906-x. [DOI] [PubMed] [Google Scholar]
  23. Lewis PA, Miall RC. Distinct systems for automatic and cognitively controlled time measurement: evidence from neuroimaging. Curr Opin Neurobiol. 2003;13:250–255. doi: 10.1016/s0959-4388(03)00036-9. [DOI] [PubMed] [Google Scholar]
  24. Lewis PA, Miall RC. A right hemispheric prefrontal system for cognitive time measurement. Behav Processes. 2006a;71:226–234. doi: 10.1016/j.beproc.2005.12.009. [DOI] [PubMed] [Google Scholar]
  25. Lewis PA, Miall RC. Remembering the time: a continuous clock. Trends Cogn Sci. 2006b;10:401–406. doi: 10.1016/j.tics.2006.07.006. [DOI] [PubMed] [Google Scholar]
  26. Lewis PA, Wing AM, Pope PA, Praamstra P, Miall RC. Brain activity correlates differentially with increasing temporal complexity of rhythms during initialisation, synchronisation, and continuation phases of paced finger tapping. Neuropsychologia. 2004;42:1301–1312. doi: 10.1016/j.neuropsychologia.2004.03.001. [DOI] [PubMed] [Google Scholar]
  27. Macar F, Vidal F, Casini L. The supplementary motor area in motor and sensory timing: evidence from slow brain potential changes. Exp Brain Res. 1999;125:271–280. doi: 10.1007/s002210050683. [DOI] [PubMed] [Google Scholar]
  28. Maquet P, Laureys S, Peigneux P, Fuchs S, Petiau C, Phillips C, et al. Experience-dependent changes in cerebral activation during human REM sleep. Nat Neurosci. 2000;3:831–836. doi: 10.1038/77744. [DOI] [PubMed] [Google Scholar]
  29. Maquet P, Peigneux P, Laureys S, Boly M, ng-Vu T, Desseilles M, et al. Memory processing during human sleep as assessed by functional neuroimaging. Rev Neurol (Paris) 2003;159:6S27–6S29. [PubMed] [Google Scholar]
  30. Mauk MD, Buonomano DV. The neural basis of temporal processing. Annu Rev Neurosci. 2004;27:307–340. doi: 10.1146/annurev.neuro.27.070203.144247. [DOI] [PubMed] [Google Scholar]
  31. Meck WH. Neuropsychology of timing and time perception. Brain Cogn. 2005;58:1–8. doi: 10.1016/j.bandc.2004.09.004. [DOI] [PubMed] [Google Scholar]
  32. Meck WH, Church RM, Olton DS. Hippocampus, time, and memory. Behav Neurosci. 1984;98:3–22. doi: 10.1037//0735-7044.98.1.3. [DOI] [PubMed] [Google Scholar]
  33. Miall RC, Weir DJ, Stein JF. Visuo-motor tracking during reversible inactivation of the cerebellum. Exp Brain Res. 1987;65:455–464. doi: 10.1007/BF00236319. [DOI] [PubMed] [Google Scholar]
  34. Miall RC, Weir DJ, Wolpert DM, Stein JF. Is the cerebellum a smith predictor? J Mot Behav. 1993;25:203–216. doi: 10.1080/00222895.1993.9942050. [DOI] [PubMed] [Google Scholar]
  35. Nixon PD, Passingham RE. The cerebellum and cognition: cerebellar lesions impair sequence learning but not conditional visuomotor learning in monkeys. Neuropsychologia. 2000;38:1054–1072. doi: 10.1016/s0028-3932(99)00138-4. [DOI] [PubMed] [Google Scholar]
  36. Nixon PD, Passingham RE. Predicting sensory events. The role of the cerebellum in motor learning. Exp Brain Res. 2001;138:251–257. doi: 10.1007/s002210100702. [DOI] [PubMed] [Google Scholar]
  37. Plihal W, Born J. Effects of early and late notcurnal sleep on declarative and procedural memory. J Cogn Neurosci. 1997;9:534–547. doi: 10.1162/jocn.1997.9.4.534. [DOI] [PubMed] [Google Scholar]
  38. Poldrack RA, Clark J, Pare-Blagoev EJ, Shohamy D, Creso MJ, Myers C, et al. Interactive memory systems in the human brain. Nature. 2001;414:546–550. doi: 10.1038/35107080. [DOI] [PubMed] [Google Scholar]
  39. Poldrack RA, Foerde K. Category learning and the memory systems debate. Neurosci BiobehavRev. 2008;32:197–205. doi: 10.1016/j.neubiorev.2007.07.007. [DOI] [PubMed] [Google Scholar]
  40. Poldrack RA, Packard MG. Competition among multiple memory systems: converging evidence from animal and human brain studies. Neuropsychologia. 2003;41:245–251. doi: 10.1016/s0028-3932(02)00157-4. [DOI] [PubMed] [Google Scholar]
  41. Rammsayer TH. Neuropharmacological evidence for different timing mechanisms in humans. Q J Exp Psychol B. 1999;52:273–286. doi: 10.1080/713932708. [DOI] [PubMed] [Google Scholar]
  42. Rauchs G, Desgranges B, Foret J, Eustache F. The relationships between memory systems and sleep stages. J Sleep Res. 2005;14:123–140. doi: 10.1111/j.1365-2869.2005.00450.x. [DOI] [PubMed] [Google Scholar]
  43. Robertson EM, Pascual-Leone A, Press DZ. Awareness modifies the skill-learning benefits of sleep. Curr Biol. 2004;14:208–212. doi: 10.1016/j.cub.2004.01.027. [DOI] [PubMed] [Google Scholar]
  44. Sakai K, Ramnani N, Passingham RE. Learning of sequences of finger movements and timing: frontal lobe and action-oriented representation. J Neurophysiol. 2002;88:2035–2046. doi: 10.1152/jn.2002.88.4.2035. [DOI] [PubMed] [Google Scholar]
  45. Spencer RM, Ivry RB. Sequence learning is preserved in individuals with cerebellar degeneration when the movements are directly cued. J Cogn Neurosci. 2009;21:1302–1310. doi: 10.1162/jocn.2009.21102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Spencer RM, Sunm M, Ivry RB. Sleep-dependent consolidation of contextual learning. Curr Biol. 2006;16:1001–1005. doi: 10.1016/j.cub.2006.03.094. [DOI] [PubMed] [Google Scholar]
  47. Spencer RM, Zelaznik HN, Diedrichsen J, Ivry RB. Disrupted timing of discontinuous but not continuous movements by cerebellar lesions. Science. 2003;300:1437–1439. doi: 10.1126/science.1083661. [DOI] [PubMed] [Google Scholar]
  48. Squire LR, Zola-Morgan S. The medial temporal lobe memory system. Science. 1991;253:1380–1386. doi: 10.1126/science.1896849. [DOI] [PubMed] [Google Scholar]
  49. Staddon JE, Higa JJ. Time and memory: towards a pacemaker-free theory of interval timing. Journal of the experimental analysis of behavior. 1999;71:215–251. doi: 10.1901/jeab.1999.71-215. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Sterpenich V, Albouy G, Boly M, Vandewalle G, Darsaud A, Balteau E, et al. Sleep-related hippocampo-cortical interplay during emotional memory recollection. PLoS Biol. 2007;5:e282. doi: 10.1371/journal.pbio.0050282. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Sterpenich V, Albouy G, Darsaud A, Schmidt C, Vandewalle G, ng Vu TT, et al. Sleep promotes the neural reorganization of remote emotional memory. JNeurosci. 2009;29:5143–5152. doi: 10.1523/JNEUROSCI.0561-09.2009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Stickgold R, Whidbee D, Schirmer B, Patel V, Hobson JA. Visual discrimination task improvement: A multi-step process occurring during sleep. J Cogn Neurosci. 2000;12:246–254. doi: 10.1162/089892900562075. [DOI] [PubMed] [Google Scholar]
  53. Ullen F, Bengtsson SL. Independent processing of the temporal and ordinal structure of movement sequences. J Neurophysiol. 2003;90:3725–3735. doi: 10.1152/jn.00458.2003. [DOI] [PubMed] [Google Scholar]
  54. 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]
  55. Walker MP, Stickgold R. Sleep-dependent learning and memory consolidation. Neuron. 2004;44:121–133. doi: 10.1016/j.neuron.2004.08.031. [DOI] [PubMed] [Google Scholar]
  56. Walker MP, Stickgold R. Sleep, memory, and plasticity. Annu Rev Psychol. 2006;57:139–166. doi: 10.1146/annurev.psych.56.091103.070307. [DOI] [PubMed] [Google Scholar]
  57. Walker MP, Stickgold R, Alsop D, Gaab N, Schlaug G. Sleep-dependent motor memory plasticity in the human brain. Neuroscience. 2005;133:911–917. doi: 10.1016/j.neuroscience.2005.04.007. [DOI] [PubMed] [Google Scholar]
  58. Wiener M, Turkeltaub P, Coslett HB. The image of time: a voxel-wise meta-analysis. Neuroimage. 2010;49:1728–1740. doi: 10.1016/j.neuroimage.2009.09.064. [DOI] [PubMed] [Google Scholar]
  59. Wing AM, Kristofferson AB. Response delays and the timing of discrete motor responses. Perception and Psychophysics. 1973;14:5–12. [Google Scholar]
  60. Zelaznik HN, Spencer RM, Doffin JG. Temporal precision in tapping and circle drawing movements at preferred rates is not correlated: further evidence against timing as a general-purpose ability. J Mot Behav. 2000;32:193–199. doi: 10.1080/00222890009601370. [DOI] [PubMed] [Google Scholar]
  61. Zelaznik HN, Spencer RM, Ivry RB. Dissociation of explicit and implicit timing in repetitive tapping and drawing movements. J Exp Psychol Hum Percept Perform. 2002;28:575–588. doi: 10.1037//0096-1523.28.3.575. [DOI] [PubMed] [Google Scholar]

RESOURCES