Sleep spindles and K-complexes (KCs) are a hallmark of N2 sleep. While the functional significance of spindles is comparatively well investigated, there is still ongoing debate about the role of the KC: it is unclear whether it is a cortical response to an arousing stimulus (either external or internal) or whether it has sleep-promoting properties. Invasive intracranial EEG recordings from individuals with drug-resistant epilepsy offer a unique opportunity to study in-situ human brain physiology. To better understand the function of the KC, we aimed to (i) investigate the intracranial correlates of spontaneous scalp KCs, and (ii) compare the intracranial activity of scalp KCs associated or not with arousals. Whole-night recordings from adults with drug-resistant focal epilepsy who underwent combined intracranial-scalp EEG for pre-surgical evaluation at the Montreal Neurological Institute between 2010 and 2018 were selected. KCs were visually marked in the scalp and categorized according to the presence of microarousals: (i) Pre-microarousal KCs; (ii) KCs during an ongoing microarousal; and (iii) KCs without microarousal. Power in different spectral bands was computed to compare physiological intracranial EEG activity at the time of scalp KCs relative to the background, as well as to compare microarousal subcategories. A total of 1198 scalp KCs selected from 40 subjects were analyzed, resulting in 32,504 intracranial KC segments across 992 channels. Fifty percent of KCs were without microarousal, 30% were pre-microarousal, and 20% occurred during microarousals. All scalp KCs were accompanied by widespread cortical increases in delta band power (0.3-4 Hz) relative to the background: the highest percentages were observed in the parietal (60-65%) and frontal cortices (52-58%). Compared to KCs without microarousal, pre-microarousal KCs were accompanied by increases (66%) in beta band power (16-30 Hz) in the motor cortex, which was present before the peak of the KC. In addition, spatial distribution of spectral power changes following each KC without microarousal revealed that certain brain regions were associated with increases in delta power (25-62%) or decreases in alpha/beta power (11-24%), suggesting a sleep-promoting pattern, whereas others were accompanied by increases of higher frequencies (12-27%), suggesting an arousal-related pattern. This study shows that KCs can be generated across widespread cortical areas. Interestingly, the motor cortex shows awake-like EEG activity before the onset of KCs followed by microarousals. Our findings also highlight region-specific sleep- or arousal-promoting responses following KCs, suggesting a dual role for the human KC.

中文翻译：

---

人类 K 复合体：结合头皮颅内脑电图记录的见解

睡眠纺锤波和 K 复合体 (KC) 是 N2 睡眠的标志。虽然对纺锤体的功能意义进行了比较深入的研究，但关于 KC 的作用仍然存在争论：不清楚它是对唤醒刺激（外部或内部）的皮质反应还是它是否具有促进睡眠的特性. 来自耐药性癫痫患者的侵入性颅内 EEG 记录为研究原位人类大脑生理学提供了独特的机会。为了更好地了解 KC 的功能，我们旨在 (i) 研究自发性头皮 KCs 的颅内相关性，以及 (ii) 比较头皮 KCs 的颅内活动是否与觉醒相关。选择了 2010 年至 2018 年在蒙特利尔神经病学研究所接受颅内-头皮脑电图联合术前评估的耐药局灶性癫痫患者的整夜录音。KCs 在头皮中进行视觉标记，并根据微唤醒的存在进行分类： (i) 微唤醒前 KC；(ii) 持续微觉醒期间的 KCs；(iii) 没有微唤醒的 KCs。计算不同光谱带中的功率，以比较头皮 KCs 相对于背景时的颅内生理脑电图活动，以及比较微唤醒亚类。分析了从 40 名受试者中选择的总共 1198 个头皮 KC，在 992 个通道中产生了 32,504 个颅内 KC 节段。50% 的 KCs 没有微觉醒，30% 有微觉醒前，20% 发生在微觉醒期间。所有头皮 KC 都伴随着与背景相关的 delta 波段功率 (0.3-4 Hz) 的广泛皮质增加：在顶叶 (60-65%) 和额叶皮质 (52-58%) 中观察到的百分比最高。与没有微唤醒的 KCs 相比，微唤醒前的 KCs 伴随着运动皮层中 β 波段功率（16-30 Hz）的增加（66%），它出现在 KC 的峰值之前。此外，在没有微唤醒的情况下，每次 KC 后光谱功率变化的空间分布表明，某些大脑区域与 delta 功率的增加 (25-62%) 或 alpha/beta 功率的降低 (11-24%) 相关，表明睡眠-促进模式，而其他人伴随着更高频率的增加（12-27%），表明与唤醒相关的模式。这项研究表明，KCs 可以在广泛的皮质区域产生。有趣的是，运动皮层在 KCs 出现之前显示出类似清醒的 EEG 活动，然后是微觉醒。我们的研究结果还强调了 KCs 后特定区域的睡眠或唤醒促进反应，表明人类 KC 具有双重作用。