How does the human brain sleep: analysis of scalp-intracerebral correlates to better understand the region-specific neurophysiology of sleep

睡眠通常被认为是一种全球性的大脑现象。睡眠期间振荡活动的区域差异尚不清楚。颅内脑电图（iEEG）作为手术前癫痫评估的少数三级中心，提供了独特且唯一的高时空分辨率原位研究人类皮层的可能性。这个程序利用这种方法来产生关于人类大脑睡眠神经生理学的区域变异性的知识，并研究大脑深部结构对睡眠活动的非侵入性信号的贡献。目的1：睡眠期间规范区域特异性脑电图活动图谱。通过在每个患者中选择位于健康大脑中的少数接触点，并积累大量患者以覆盖整个皮层，我们率先绘制了第一个人类清醒期间正常脑电图图谱（Frauscher et al., brain 2018）。在一个子集中，我们可以访问有睡眠的片段。首先，数据显示脑电图在所有睡眠阶段都有尾-吻侧频率梯度。此外，深层结构显示出不同于头皮EEG的频谱峰。这种方法将被用于一个更有野心的项目，在睡眠的不同阶段和周期中，在所有的大脑区域建立一个图谱。目的2：分析睡眠微观结构和动态的脑内相关因素。尽管人们对睡眠纺锤波及其脑内相关性有广泛的了解（例如Frauscher等人，NeuroImage 2015），但对其他睡眠特异性瞬态或睡眠动力学变化的了解很少或根本不存在。这是因为在iEEG的设置中很少进行联合睡眠记录。目标1的大数据池将有助于阐明区域睡眠微观结构和动态变化。目的3：脑深部结构对睡眠活动的非侵入性神经生理信号的贡献。非侵入性源定位方法的缺点是我们无法知道源模型在什么条件下是正确的和可以信任的。我们提出了一种独特的方法，它将使我们能够直接验证，使用非侵入性测绘程序，如高密度脑电图和脑磁图，用脑电图记录的深层源的可探测性作为地面事实。这将决定哪些睡眠模式可以进行无创研究。该项目将(i)产生关于不同皮质区域和网络在局部睡眠调节中的独特参与的知识，（ii）证明哪些大脑区域在睡眠振荡或睡眠动态变化时特别参与，以及（iii）建立使用非侵入性大脑测绘技术可检测深层结构中哪种类型的睡眠活动。它将增加我们对睡眠的神经生理学相关性的认识，并为更好地解释大脑活动的非侵入性测量的相关性开辟新的途径。数据集将是开放获取的。

Sleep is typically considered a global brain phenomenon. Regional variations in oscillatory activity during sleep remain unknown. Intracranial electroencephalography (iEEG), as used in few tertiary centers during presurgical epilepsy evaluation, offers the unique and only possibility to study in situ the human cortex with high spatio-temporal resolution. This program exploits this method to generate knowledge on the regional variability of sleep neurophysiology in the human brain and to investigate the contribution of deep brain structures to the non-invasive signal of sleep activity. Objective #1: atlas of normative region-specific iEEG activity during sleep. We have pioneered the first human atlas of the normal iEEG during wakefulness by selecting in each patient the few contacts located in healthy brain and accumulating a large number of patients to cover the whole cortex (Frauscher et al., Brain 2018). In a subset, we had access to segments with sleep. First data show that the iEEG has a caudo-rostral frequency gradient across all sleep stages. Moreover, deep-seated structures show spectral peaks differing from the scalp EEG. This approach will be used in a more ambitious project, building an atlas during the different stages and cycles of sleep, in all brain regions. Objective #2: analysis of the intracerebral correlates of sleep microstructure and dynamics. Whereas there is extensive knowledge on sleep spindles and their intracerebral correlates (e.g. Frauscher et al., NeuroImage 2015), knowledge on other sleep-specific transients or changes in sleep dynamics are scarce or non-existing. This is because combined sleep recording is rarely performed in the set-up of iEEG. The large datapool of objective 1 will enable to elucidate regional sleep microstructural and dynamic changes. Objective #3: contribution of deep brain structures to the non-invasive neurophysiological signal of sleep activity. Non-invasive source localization methods suffer from our inability to know in what conditions a source model is correct and can be trusted. We propose a unique approach, which will enable us to validate directly, using non-invasive mapping procedures such as high-density EEG and magnetoencephalography, the detectability of deep sources as recorded with iEEG as ground truth. This will determine which sleep patterns can be studied non-invasively. This program will (i) generate knowledge on the distinct participation of various cortical regions and networks in the local regulation of sleep, (ii) demonstrate which brain regions are specifically involved at the time of sleep oscillations or changes in sleep dynamics, and (iii) establish which type of sleep activity in deep structures is detectable using non-invasive brain mapping techniques. It will increase our knowledge on neurophysiological correlates of sleep and open new avenues for better interpreting the correlates of non-invasive measures of brain activity. The dataset will be open-access.