睡眠對於許多動物來說是不可或缺的生理節律，其功能除了調整生理狀態外，亦會重整大腦的認知功能，例如情緒及專注力調節；先前文獻亦指出深層睡眠（非快速動眼期第三階段）更與記憶鞏固息息相關。但過往睡眠相關研究大多專注於特定睡眠階段或是特定睡眠指標進行探討，鮮少有跨睡眠階段之全頻帶研究。因此為了了解睡眠中大腦網路的動態變化，我們首次嘗試使用同步腦電圖與功能性磁振造影來觀察正常受試者大腦網路在清醒狀態與非快速動眼睡眠（尤其是深層睡眠）間之差異。本論文著重於兩大目標：(1)探討腦波各頻帶在跨睡眠階段的頻率能量以及其網路連結之變化，以及(2)觀察有無深層睡眠對大腦活動的潛在影響。結果發現，腦波各頻帶在不同睡眠階段中有其特定的空間分佈以及功能性連結變化的差異；例如低頻Delta波在睡眠深度越深時除了能量逐漸增強外，其網路連結亦會從清醒時的縱向連結逐漸轉變為橫向連結。此外，缺少深層睡眠發現對清醒後的頻譜能量與橫向連結並無顯著影響，但對清醒後的縱向連結有顯著變化，暗示慢波睡眠的功用可能著重在改變大腦網路的縱向連結。不過在腦電圖以及功能性磁振造影都的結果中都發現頻率表現和網路連結不見得會呈現相同趨勢。整體而言，本論文證明腦波各頻帶在不同睡眠階段具有頻率響應及網路連結的動態特異性，而深層睡眠不僅在大腦網路調節上扮演著重要角色，有可能也會對甦醒後的腦部活動產生衝擊。;Sleep is an important physiological rhythm for many creatures, not only associated with physical regulation but also associated with mental reorganization. It has been proven that sleep correlates with lots of cognitive functions such as the emotion regulation, memory consolidation and attention. Previous sleep studies were typically conducted using electroencephalography (EEG), targeting on the changes of specific frequency bands or certain sleep stage; however, dynamic changes was rarely reported across all frequency bands and sleep stages. Therefore, to reveal the dynamicity within sleeping brain, we first investigate the sleeping EEG power and connectivity across frequency bands and non-rapid-eye-movement (NREM) sleep stages using simultaneous EEG-fMRI. Our second goal is to observe the impact of slow wave sleep (SWS) on brain activity and connectivity. Our results indicated spectral and regional disparities in EEG power and connectivity analysis across sleep stages. For example, the delta band showed increasing power along with deep sleep stages, meanwhile the delta connectivity pattern converted from waking longitudinal connection into a transverse connection during sleep. Secondly, the low-frequency bands did not show significant SWS effect on awakening in both power and transverse connectivity, but presented strong changes in longitudinal connectivity, which implied that slow wave sleep modulates the brain longitudinal connectivity. Furthermore, it should be noted that both EEG and fMRI presented a dissimilar mismatch between the spectral power and functional connectivity across sleep stages. Conclusively, this study indicate that during NREM sleep, EEG frequency bands had their own dynamic features in both power and connectivity, and lack of slow wave sleep might affect the consecutive functional connectivity upon awakening.
