生物谷報(bào)道：人在睡眠時(shí)大部分時(shí)間內(nèi)出現(xiàn)的是慢波睡眠,，它的生理意義是什么呢？一直不是很清楚,，以前有人認(rèn)為它可能與記憶的存儲(chǔ)有關(guān),，但沒有實(shí)驗(yàn)依據(jù),。最新的研究顯示，慢波睡眠能顯著提高睡眠后（白天）工作時(shí)的工作效率,。這一成果揭示了睡眠與工作關(guān)系的機(jī)理,。如果提高慢波睡眠，可以使睡眠后的工作效率提高,，反之則效率降低,。請(qǐng)看詳細(xì)報(bào)道。

To investigate local aspects of sleep regulation, we asked subjects to perform a motor learning task just before going to sleep. In this task, subjects reach for visual targets using a handheld cursor while unconsciously adapting to systematic rotations imposed on the perceived cursor trajectory (Fig. 1a; see also ref. 4). This rotation adaptation task was chosen because it is an implicit learning paradigm; it is suitable for extended sessions; it permits accurate parameterization of both performance improvement and noise reduction; it can be contrasted with a no-rotation task that has the same kinematic requirements and is subjectively indistinguishable; and in contrast with the no-rotation task, it activates circumscribed brain regions (that is, right parietal areas 40 and 7 (ref. 4)). Over a number of movements, all our subjects adapted to the imposed rotation by progressively reducing the directional error of their trajectory as well as its variance (Fig. 1b).

 Figure 1 Rotation adaptation.   Full legend
 
High resolution image and legend (32k)
 

Immediately after the rotation adaptation task, we recorded the sleep electroencephalogram (EEG) using a 256-channel system (Electrical Geodesics) inside a soundproofed room. As whole-night recordings with the high-density EEGs are not comfortable, the cap was removed after 2 h and subjects were then allowed to sleep undisturbed for the rest of the night (all reported satisfactory, restful sleep). As a control condition, 1 week earlier or later, subjects performed the no-rotation task, where the requirements were kinematically identical but the cursor was not rotated4, 5. High-density recordings after both tasks showed the usual progression of sleep stages. Values (mean  s.e.m.) for rotation and no-rotation tasks were, respectively: sleep latency, 6.1  0.8 and 6.5  1.4 min; wakefulness, 14.4  4.8% and 14.2  3.8%; rapid eye movement (REM) sleep, 2.3  0.9% and 2.9  1.2%; non-REM sleep (stages 2–4), 65.2  6.4% and 65.7  7.0% (% of recording time).

Average power spectra of consecutive 4-s epochs during non-REM sleep showed that SWA was prevalent in anterior regions, in accord with previous studies2, 6, and that the topographic pattern was highly reproducible in both the rotation and no-rotation conditions (Fig. 2a). However, when we compared the two conditions, we found that the rotation condition produced a local increase in SWA at a cluster of six right parietal electrodes, which averaged 13.1  6.1% and was statistically significant7 (P < 0.01, Fig. 2b). The peak SWA increase, whose precise location varied slightly between subjects, amounted to 25.3  4.6%. Thus, the rotation adaptation task elicited a response in the sleep EEG, and this response was local rather than global.

 Figure 2 Local SWA homeostasis during sleep after rotation adaptation.   Full legend
 
High resolution image and legend (86k)
 

We then examined whether the local EEG response after rotation adaptation shared key features with the global homeostatic response observed in the sleep EEG after prolonged wakefulness2, 6. The EEG signature of sleep pressure is an increase in power predominantly in the frequency range of SWA (1–4 Hz)2, 6 and a slight reduction in power in the sigma band (12–15 Hz). Figure 3a shows the percentage changes observed in sleep EEG power after the rotation adaptation task compared with the no-rotation condition. Consistent with a homeostatic response, there was an increase in power predominantly in the SWA frequency range, as well as a slight decrease in power in the sigma band. The increase was especially evident within the low delta band (< 2 Hz) and at frequencies corresponding to the slow oscillation (< 1 Hz)1. Period amplitude analysis revealed that the increase of SWA at the site of the peak response was primarily due to an increased amplitude of slow waves (22.1  9.7% at 1–2 Hz), whereas their period was unchanged. Another feature of the global homeostatic response of SWA is its decline within and across consecutive sleep cycles2. The local SWA response also revealed a decreasing trend across the first 90 min after lights off (Fig. 3b). Thus, both the spectral signature and the time course of local sleep changes resembled the global homeostatic response of sleep.

 Figure 3 Frequency specificity, time course and anatomical localization of local SWA homeostasis.   Full legend
 
High resolution image and legend (90k)
 

To localize the region of SWA increase, we used a positioning system (Nexstim) to digitize the 256 electrodes and co-register them with each subject's magnetic resonance images. The increase of SWA was localized to regions of the right parietal lobe encompassing Brodmann areas 40 and 7 (Fig. 3c). These regions corresponded to those highlighted by positron emission tomography (PET) subtractions between the rotation and no-rotation conditions in subjects scanned after training4. Both areas 40 and 7 receive converging visual and proprioceptive inputs, and are involved in processing sensory information relevant for spatial attention8. Area 7 is a station in the dorsal visual pathways and processes spatial aspects of vision related to skilled actions8. The right hemisphere specificity in both EEG and PET experiments is in line with the right hemisphere specialization for spatial tasks9 and for spatial coordinate transformation8.

Recent work has shown that performance in certain perceptual, motor and categorization tasks improves after a night of sleep10-18. These studies have demonstrated that the improvement is specifically due to sleep rather than to the mere passage of time or to circadian factors11-13. We therefore asked whether performance in the rotation adaptation task would also be enhanced by sleep. As shown in Fig. 4a, when subjects were tested after a night of sleep, the directional error had decreased further, corresponding to a performance enhancement of 11.1  3.0% above and beyond the level achieved at the end of awake training. By contrast, an independent group of subjects who trained in the morning and were re-tested after 8 h of wakefulness showed no such improvement (Fig. 4a). Because other movement parameters that also improved with training, such as movement time, did not differ between the two groups, the improvement in directional error was probably due to sleep and not to circadian differences. In summary, the EEG changes observed by comparing sleep after rotation adaptation with sleep after the kinematically identical no-rotation task were: (1) local, suggesting a cellular substrate; (2) characterized by an increase in SWA that behaved as expected of homeostatic changes in the sleep EEG; (3) localized to cortical regions known to be involved in learning rotation adaptation; and (4) followed the next day by an enhancement of performance.

 Figure 4 Enhancement of performance after sleep and its relationship to SWA.   Full legend
 
High resolution image and legend (44k)
 

Finally, we asked whether the post-sleep enhancement of performance and the increase of SWA in right parietal areas 40 and 7 were correlated. We found that the decrease of directional error was positively correlated with the peak increase in SWA (Fig. 4b; r = 0.86, P < 0.005). The positive correlation of post-sleep performance enhancement with the local increase in EEG power was specific to the SWA frequency range (1–4 Hz, data not shown). By contrast, there was no correlation between the improvement in rotation adaptation and SWA changes at other electrodes, or between the local increase of SWA and changes in other movement parameters such as movement time. Therefore, local SWA homeostasis in right parietal areas 40 and 7 is probably related to local neural processes specific to rotation adaptation and to their post-sleep enhancement. We did not find any correlation between the increase in SWA and the mean decrease of directional error at the end of the training session. Instead, we found a high correlation between the increase in SWA and the standard deviation of directional error at the beginning minus the end of training (Fig. 4c; r = 0.83, P < 0.005). Therefore, subjects who began rotation adaptation with much more variability or 'noise' in directional error compared to the end of training not only displayed a marked local increase in SWA during sleep but also benefited most from sleep.

These findings provide compelling evidence that the electrophysiological marker of sleep homeostasis, SWA, can be selectively induced in circumscribed regions of the cerebral cortex. Thus, they make a strong case for the local regulation of sleep2, 6, 19-22 and support a role for sleep at the cellular level3, 23, 24. Moreover, they show that local SWA induction is triggered by a learning task, suggesting that local plastic changes associated with learning may be involved, directly or indirectly. Finally, they show that local SWA homeostasis is strongly correlated with improved performance in the task after sleep. This suggests that SWA homeostasis may be related to cellular processes underlying learning rather than to metabolic fatigue or depletion. Thus, together with evidence from intracellular studies25, our results support the notion that slow oscillations might help synaptic consolidation25-30 or produce synaptic downscaling and increase signal-to-noise ratios in relevant neural circuits3. Most importantly, these results connect two fields that had thus far remained separate—the study of sleep homeostasis and that of sleep and plasticity.