【作者】 唐军华；

【导师】 苗丹民；

【作者基本信息】 第四军医大学 ， 应用心理学， 2013， 博士

【摘要】 自杀是导致个体死亡的主要原因之一，也是一种复杂的人类行为。根据2012年世界卫生组织的报告，每年有近一百万人死于自杀，自杀死亡的人数远大于被谋杀的人数，与45年前相比自杀率上升了60%。许多研究者认为，自杀意念是自杀前的重要指标。因此通过对自杀意念的研究，可以更好地理解、预防和干预自杀行为。然而，由于人们有意、无意地回避报告个人自杀意念的存在，导致自杀意念的客观测量非常困难，对于自杀预防和干预也难以实施。然而，由于内隐认知测量规避了心理测验中社会赞许性的弊端，可能有助于自杀风险的评估。目前有20余种内隐认知测验的方法。其中，内隐联想测验应用最广、效果最明显，并已积累了大量研究证据；在心理健康、消费决策、选举行为等方面的评估中，内隐联想测验被用于众多心理结构的测量，如自尊、态度、偏见和自我概念等。于是，我们想探究那些有自杀意念并尝试过自杀的个体，是否将“自我”与“死亡”连接得更加紧密，即具有内隐自杀意念；探究内隐自杀意念对自杀行为的预测是否比自我报告法更好；并采用ERPs技术探讨无自杀意念者内隐联想测验的认知加工机制。逃避理论对自杀行为的发生进行了理论解释，在这个理论中自杀涉及到六个阶段。第一个阶段是由于没有达到期望目标产生的失败感或失望感。这可能是由于期望太高导致的，也可能是由于现实太坏导致的，也可能是两方面的原因同时作用的结果，本质是个体的现实和期望间存在差距。第二个阶段是将失败归咎于自己产生的自责。第三至第六阶段分别为：自我意识变化、负性情绪、认知解构和最终导致自杀意念的产生。本研究试图通过探究逃避理论与内隐自杀意念产生之间的关系，验证逃避理论能否解释内隐自杀意念在自杀形成中的作用。本研究的主要研究结果有：（1）那些曾经有过自杀行为的个体在内隐联想测验中将“自我”与“死亡”进行了正向的内隐连接（M=0.34，SE=0.1），而“自我”与“死亡”的内隐连接在仅有自杀意念的被试中较弱（M=-0.06，SE=0.06），而那些没有自杀意念和自杀行为的个体在内隐联想测验中将“自我”与“死亡”进行了负向的内隐连接（M=-0.34，SE=0.07）。通过方差分析发现，三组被试间在内隐联想测验中的表现具有显著的差异，F(2,69)=17.89，p <0.01。研究数据说明内隐联想测验测量自杀意念有效性较好。（2）根据自杀的逃避理论建立的模型经过修改后，与本研究的数据具有较好的拟合度，卡方为17.54，自由度为14，p=0.23，RMSEA=0.04，GFI=0.98。说明自杀逃避理论可以用于解释内隐自杀意念的产生和发展。（3）在启动失败认知、启动成功认知和控制组三种实验条件下，被试的内隐自杀意念强度不同，F(2,135)=13.42，p <.001，η2=.17。另外，个体的控制点倾向会影响到该效应，F(2,132)=13.38，p <.001，η2=.16。具体地，在内控的个体中该启动效应依然存在，F(2,67)=15.46，p <.001，η2=.32，而在外控的个体中该效应却消失了，F(2,65)=1.57，p=.22。这些研究数据符合根据自杀逃避理论建立的假设，即失败会导致自杀意念增加而成功会导致自杀意念减少，并且该效应会受到个体控制点倾向的影响，在内控倾向的个体中启动效应会增大，而在外控倾向的个体中启动效应会减小或消失。（4）在内隐联想测验中，不相容任务组的反应时（M=621.90ms，SD=64ms），显著长于相容任务组的反应时（M=566.38ms，SD=60ms），t(20)=3.31，p <0.01。是否相容与是否任务转换对反应时间的效应存在交互作用，F(1,20)=4.43，p <0.05。在相容条件下，任务是否转换对反应时间没有显著的影响。在不相容条件下，任务转换组的反应时（M=654.93ms，SD=71.61ms）显著长于非任务转换组的反应时（M=575.21ms，SD=79.53ms）。研究还发现相容条件下刺激诱导的P300成分的波幅（M=4.86μV，SD=1.84μV）大于不相容条件下刺激诱导的P300成分的波幅（M=3.51μV，SD=1.55μV），t(20)=4.44，p <0.01。在不相容任务下，需要任务转换的刺激诱导的P300成分的波幅（M=3.96μV，SD=1.55μV）与不需要任务转换的刺激诱导的P300成分的波幅（M=3.17μV，SD=1.75μV）不一致，t(20)=2.58，p <0.05。这些研究结果反应出在内隐联想测验中，相容任务需要较少的认知资源，个体有更多的认知资源进行刺激加工，使得个体在反应中具有更高的自信；在不相容任务中，对于那些不需要任务转换的刺激，被试主观上认为其发生的概率较高会影响到其对刺激的认知加工。更多还原

【Abstract】 Suicide is a leading cause of death worldwide and is among the most perplexing ofall human behaviors. According to the World Health Organization (2012), almost onemillion people die by suicide every year. Suicide accounts for more than all deaths frommurder. In the past45years, suicide rates have increased by60%worldwide. Manyresearchers consider that suicidal thought is the primary antecedent of suicidal behaviors.Many researchers thus believe that by understanding the origin of suicidal thought, we canbetter understand, predict, and treat suicidal behavior. However, suicide is difficult topredict and prevent because people who consider killing themselves often are unwilling orunable to report their intentions. Advances in the measurement of implicit cognition mayprovide a behavioral marker for suicide risk.There are more than20measurement procedures in the measurement of implicit cognition. Among these, the Implicit Association Test (IAT) provides the bulk ofaccumulated research evidence. Furthermore, IAT has been adapted to measure a widevariety of mental contents including stereotypes, attitudes, self-esteem, and self-concepts,across a wide variety of topics such as mental health, consumer decision making, andpolitical behavior. We tested whether individuals who have made a decision to killthemselves would reveal stronger implicit cognition associating self with death andwhether the strength of such an association would predict suicide attempts beyond thatachieved with suicidal thoughts from self-report. And study of the cognitive processesinvolved in the IAT is necessary to establish the validity of this measure and to suggestfurther refinements to its use and interpretation. The current study used ERPs to studycognitive processes associated with the IAT.Escape theory is an integrative framework proposing a causal chain that leads tosuicide. In this framework, suicide involves an orderly progression through six distinctstages. The initial step is a major failure or disappointment that is due to events fallingseverely short of standards or expectations. This may occur because events are verynegative, because standards are too high, or both. It is the discrepancy between the realityand standard that is crucial. The second stage involves self-blame: Responsibility forfailure is attributed to the self. The following steps successively involve highself-awareness, negative affect and feelings, a state of cognitive deconstruction, andfinally several consequences of cognitive deconstruction that may increase the appeal ofsuicide. The other purpose of the present study was to assess the effectiveness of theEscape Theory of Suicide in accounting for the development of implicit cognitionassociating self with death.Through the above study, the mainly results were found as following:(1) Nonsuicidal individuals show a negative association between death and oneself(M=-0.34, SE=0.07); suicide ideators show a small negative association between deathand oneself (M=-0.06, SE=0.06); and suicide attempters show a large positiveassociation between s death and oneself (M=0.34, SE=0.1). A one-way ANOVArevealed a significant group effect, F (2,69)=17.89, p <0.01. These results support the validity of implicit association test as a measure of suicide thoughts.(2) The final model from the Escape Theory of Suicide fit our data very well. Thechi-square for this model was17.54, df=14, p=0.23, RMSEA=0.04, and GFI=0.98.Our findings confirm the predictive validity of the Escape Theory of Suicide.(3) Results revealed significant differences between groups (failure-related priminggroup, success-related priming group, and control group) in accessibility to implicitsuicidal mind, F (2,135)=13.42, p <.001, η2=.17. Furthermore, priming manipulationinteracted with individual differences in locus of control (LOC), F (2,132)=13.38,p <.001, η2=.16. Significant differences in accessibility to implicit suicidal mind wereobserved in individuals with internal LOC, F (2,67)=15.46, p <.001, η2=.32, whileeffects of priming manipulation were eliminated in individuals with external LOC, F (2,65)=1.57, p=.22. The results of our experiment were in line with our predictions thatfailure increases implicit suicidal mind and success decreases implicit suicidal mind. Thisstudy supports our hypothesis that the impact of failure or success priming on associationsbetween self and death is more pronounced among individuals with internal LOC, andabsent among individuals with external LOC.(4) The fourth study found response time was significantly longer in the incongruentcondition (M=621.90ms, SD=64ms) than in the congruent condition (M=566.38ms，SD=60ms), t(20)=3.31, p <0.01. A two-way interaction between condition and task setswitching was found for response time, F (1,20)=4.43, p <0.05. Task switching had noeffect in the congruent condition. Task switching did however significantly prolongresponse time in the Incongruent condition (for task switch: M=654.93ms, SD=71.61ms; for no task switch: M=575.21ms, SD=79.53ms). And the study also foundsignificant differences in P300amplitude in the congruent (M=4.86μV, SD=1.84μV)and incongruent conditions (M=3.51μV, SD=1.55μV), t (20)=4.44, p <0.01. Therewere differences in P300amplitude between trials of the incongruent condition thatrequired (M=3.96μV, SD=1.55μV) or did not require a task switch (M=3.17μV, SD=1.75μV), t (20)=2.58, p <0.05. The less demanding nature of the congruent conditionmay have resulted in greater resources being available to process the stimuli, resulting in less uncertainty about the correct response to the stimulus. In the incongruent condition,there is greater perceived local probability on no switch trials compared to switch trials.更多还原