研究目的：過去的心理生理學研究顯示，個體在壓抑生氣與表達生氣情緒時，分別對於情緒調節和生理反應上有著不同的作用機制。本研究試圖以Engebretson等人（1989）提出的「適配假說」（Matching Hypothesis），進一步延伸探討其對於後續睡眠的影響，並預期在引起受試者的生氣情緒後，若其當下所採用的生氣表達方式與自身所慣用的方式是一致的，則能有效地降低個體睡前的生理激發狀態，促使其恢復到先前的水準，因此對於個體後續睡眠的負面影響也會較小。 方法：本研究共收案35人(男12人，女23人)，對象為年齡介於20到35歲之間，身心健康的正常睡眠者。研究者先透過短式華人敵意量表，將受試者分類為高壓抑敵意特質組(HS)18人、高表達敵意特質組(HE)17人，之後再將其隨機分派到壓抑生氣(AI)或表達生氣(AO)其中一種實驗情境。實驗中，係透過高難度的認知作業來引發受試者的壓力和挫敗感，再於過程中不停地打斷其作答狀態來造成其煩擾，以達到生氣情緒的誘發；隨後再經由有/無給予機會讓受試者表達心情並對主試者(激怒者)進行評價，來達到表達/壓抑生氣的實驗操弄，之後即請其入睡。期間持續以Biopac MP150生理儀器來測量受試者的心跳速率(HR)、收縮壓(SBP)、舒張壓(DBP)和膚電反應(SCR)，以了解其於入睡前的生理激發與恢復狀態；後續以多頻道睡眠記錄儀（polysomnography, PSG）來進行夜間睡眠之記錄，並以受試者的自評睡眠狀況作為主觀睡眠品質的分析指標。 研究結果：在誘發生氣情緒後，不論是在表達生氣或壓抑生氣的情境下，HE都比HS有顯著更高的DBP上升反應。在恢復期期間，HE的DBP則有低於HS的傾向；AO比AI有更多的DBP恢復程度，以及較短的SCR恢復時間。而在適配情境下，受試者後續的HR恢復時間較短，且DBP的下降程度較多，此也支持了適配假說的觀點；然而，不論適配與否，對於後續的睡眠相關參數並無顯著的影響，此未支持本研究假設。再以2(特質)x 2(情境)之二因子獨立樣本變異數分析比較各組在睡眠結構上的差異，結果顯示階段二睡眠以及慢波睡眠的比例，皆有來自特質的主要效果；其中HS的階段二睡眠時間比例高於HE，HE的慢波睡眠時間比例則高於HS。另外，AO的慢波睡眠時間比例高於AI──顯示比起壓抑生氣，表達生氣的因應方式對於深睡期的增加可能較具正向的影響。最後，研究發現睡前SBP、DBP的上升皆和主觀入睡耗時的增加呈顯著正相關；PSG的檢查結果亦顯示，入睡期HR、SCR、SBP的上升和階段二睡眠潛時的增加呈顯著正相關；此外，當SCR的恢復時間越長，階段二睡眠潛時越長，顯示個體睡前較高的生理激發狀態和較慢的恢復速率，對於主客觀睡眠品質皆具有負面的影響。 結論：在經驗生氣情緒時，表達生氣相對而言是對於生理恢復狀態較為有利的因應方式；然而，「特質」不僅可能影響著生氣情緒當下的生理激起狀態，且對於後續的生理恢復亦具有一定之影響力。本研究支持了適配假說的觀點：當採用的生氣表達方式與自身所慣用的方式一致時，最能有效地降低心血管反應；然而，再去檢驗適配假說和後續睡眠品質之間的關係，並未發現符合預期的結果，故若欲以適配假說來預測後續較佳的睡眠品質，仍須更多相關的研究加以支持和驗證，且可能尚須納入其他重要變項加以探討，例如個人的認知風格、內在情緒調節策略等。最後，本研究結果再度彰顯了睡前的生理激發和恢復狀態對於後續睡眠歷程的影響力；而生理激發狀態在情緒、因應行為和睡眠間之關係所扮演的角色，也為急性失眠的病因機制提供了一種可能的註解。 / Introduction：From past studies it is known that anger suppression and anger expression may play different roles in emotion regulation and in psychophysiological response. Engebretson and his colleagues (1989) had proposed ‘Matching Hypothesis’-- that is, if one used his/her preferred mode of anger expression style in response to the instigation, then his/her cardiovascular reactivity (CVR) would be less elevated and would have a faster recovery rate of CVR than those who were engage in a manner of anger expression that was inconsistent with their preferred mode. To do a further study, we intend to know what would bring on in the following sleep if we used the two different anger expression style during the pre-sleep period. Besides, in according to Matching Hypothesis, we predict that if one expressed his/her anger in a preferred way, then this ‘matching’ state would make one’s physiological arousal effectively declined and would help it back to the baseline level more quickly; therefore, it would have less negative effect on the following sleep, and vice versa. Method：35 healthy individuals who has psychometric characteristic of high hostility were examined. All subjects’ preferred style of anger expression-- highly suppressive (HS) or highly expressive (HE), had been assessed by a self-report questionnaire. After the anger instigation, subjects were randomly assigned to two experimental conditions: anger in (AI) or anger out (AO), and then went to sleep. Biopac instrument and software were used to measure heart rate (HR), blood pressure (BP) and skin conductance response (SCR). Besides, polysomnography (PSG) was used to record the nocturnal sleep. In addition, the subjects were asked to fill out the self-report questionnaires next morning, as a measurement of subjective sleep quality. Finally, statistic analyzed the influence of two variables (trait and condition) on the physical arousal level, recovery rate and sleep quality. Results：HE showed higher DBP elevation than HS, both in the two conditions. During the recovery phase, there was an inclination that HE showed lower DBP than HS; AO showed a higher DBP recovery level, and a shorter SCR recovery time, comparing to AI. Subjects in the matching state showed a shorter HR recovery time, and much DBP decline; however, no matter in the matching state or not, there was no difference in the nocturnal sleep quality. To view the sleep architecture, HS showed higher percentage of stage 2 sleep than HE, and HE showed higher slow-wave sleep percentage than HS. On the other hand, AO showed higher slow-wave sleep percentage than AI, suggesting that expressing anger seems to be much helpful for the promotion of slow-wave sleep. Finally, the findings demonstrated that before sleep, the elevation of SBP and DBP showed positive correlations with longer subjective sleep onset latency. Besides, the elevation of HR, SCR and SBP was positively correlated with the longer onset latency of stage 2 sleep. Furthermore, longer SCR recovery time accompanied longer onset latency of stage 2, indicating that both high arousal state and low recovery rate brought negative effects on the following sleep. Conclusions：The CVR results generally were consistent with Matching Hypothesis. However, the findings indicated that the matching/non-matching state couldn’t be merely used to predict the sleep quality. Therefore, there is still a need to do more research to figure it out. In addition, variables such as personal cognitive style and emotional regulation strategy, should be bring into the future studies. In conclusion, the findings pointed out that the physical arousal state during the pre-sleep period will impact on the following sleep process significantly, and it plays a important role in the relationships between emotion, coping behavior and sleep.