Clustering and temporal organization of sleep spindles underlie motor memory consolidation

Sleep benefits motor memory consolidation, which is mediated by sleep spindle activity and associated memory reactivations during non-rapid eye movement (NREM) sleep. However, the particular role of NREM2 and NREM3 sleep spindles and the mechanisms triggering this memory consolidation process remain controversial. Here, simultaneous electroencephalographic and functional magnetic resonance imaging (EEG-fMRI) recordings were collected during night-time sleep following the learning of a motor sequence task. Adopting a time-based clustering approach, we provide evidence that spindles iteratively occur within clustered and temporally organized patterns during both NREM2 and NREM3 sleep. However, the clustering of spindles in trains is related to motor memory consolidation during NREM2 sleep only. Altogether, our findings suggest that spindles’ clustering and rhythmic occurrence during NREM2 sleep may serve as an intrinsic rhythmic sleep mechanism for the timed reactivation and subsequent consolidation of motor memories, through synchronized oscillatory activity within a subcortical-cortical network involved during learning.


6
The repeated practice of a motor skill is crucial for its initial acquisition 1 . Yet, ample Figure 2B depicts the grand average TF maps zoomed in on -6 s and +6 s around in Figure 2C illustrate the finely tuned and regular occurrence of spindles in trains during 1 4 8 both NREM2 and NREM3 sleep, albeit with a more diffuse pattern for the latter.  Table 2. EEG coherence differences between grouped and isolated sleep spindles in relation to MSL 2 8 0 consolidation. Statistically significant regressions between hippocampus-seeded and putamen-seeded 2 8 1 functional connectivity during grouped and isolated spindles against overnight performance gains 2 8 2 (Grouped>Isolated and Isolated>Grouped spindle contrasts). Results are displayed separately for NREM2 2 8 3 and NREM3 sleep. Note: k = cluster size (determined at a threshold of P = .05 after small-volume correction). Minimum cluster size = 5 2 8 7 voxels. FWE-SVC = family-wise error small-volume correction. x, y, and z coordinates are specified in the MNI space.

Isolated>Grouped contrast
No voxels significantly active

Grouped>Isolated contrast
No voxels significantly active

Isolated>Grouped contrast
No voxels significantly active greater hippocampus-and putamen-seeded coherence in the EEG spindle frequency band (11-16 Hz) 2 9 7 during grouped spindles, compared to isolated spindles, was associated with higher overnight performance 2 9 8 gains. Left panel. As depicted in the scatter plot, hippocampus-seeded connectivity with the right-2 9 9 lateralized thalamus was significantly related to the magnitude of overnight performance gains. Middle hippocampus-and putamen-seeded regression maps, which include the bilateral thalamus, caudate and right 3 0 6 SMA. See Table 2 for details on significant clusters of activation. The aim of the present study was twofold: (i) to determine whether the clustering and  sleep play a more critical role than isolated ones in motor memory consolidation. Indeed, 3 1 6 through deep brain EEG coherence analyses in the 11-16 Hz frequency range, we 3 1 7 revealed that the rhythmic occurrence of spindles in trains might confer optimal neural 3 1 8 conditions for efficient reprocessing and consolidation of memory traces within a 3 1 9 functional network integrating the hippocampus, the striatum (i.e., the putamen in  From a holistic standpoint, we therefore advocate for a broader and multi-dimensional  repetitively reactivating task-relevant brain regions that were engaged during the learning 4 1 4 process per se. This finding is consistent with the "synaptic consolidation" hypothesis, Yet, in contrast to the findings observed for grouped spindles, regression analyses To conclude, our findings confirm the critical role of NREM2 over NREM3 sleep The study protocol was approved by the Research Ethics Board of the  Thirty-three young healthy volunteers (mean age: 25.5 ± 3.7 years, 24 females) were we also excluded participants exhibiting signs of excessive daytime sleepiness (≤ 9 on the 4 7 5 Epworth Sleepiness Scale 67 ). Participants were asked to maintain a regular sleep-wake cycle (bedtime between 10:00 pm-12:00 am, wake-time between 07:00 am-10:00 am) to 4 7 7 ensure they were not sleep-deprived. The schedule's compliance was assessed using sleep 4 7 8 logs and wrist-worn actigraphs (Actiwatch 2, Philips Respironics, Andover, MA, USA).

7 9
Finally, volunteers were also asked to refrain from all caffeine-and alcohol-containing  conditions were almost identical to that experienced in the actual 3.0T TIM TRIO MRI actual MR scanners (visit 2), EEG electrodes were removed and participants were 5 0 0 allowed to sleep for the rest of the night in the nearby sleep laboratory until 7:00-8:00 am.

0 1
Of the twenty-six participants who met the 15 min criteria of NREM2 sleep during polysomnographic technologist as either NREM stages 1-3, REM, or wake (see Table 1 6 0 1 for details). Spindle detection was then conducted using all artifact-free NREM sleep 13.5 Hz) and fast (e.g., ~13.5-16 Hz) spindles due to the insufficient number of events per 6 1 5 spindle category for statistical relevance (see also 7 ).  repetitive and closely spaced in time occurrence of spindles in trains over specific brain   Consolidation Is Linked to Spindle-Mediated Information Processing during Sleep.    297-308 (1996).  Consolidation. Curr. Biol. 26, 2127Biol. 26, -2136Biol. 26, (2016.