正如谚语所说，“金无足赤，人无完人”。个体会将对自我的感知(如真实自我)与内在的标准(如理想自我)进行比较，从而产生这两种不同的自我知识之间的不一致或差距，即现实与理想自我差异。较大的自我差异与失望、失败等诸多不舒服的心理体验有关，是抑郁症形成和维持的重要心理过程之一。然而，目前尚不清楚人们在各种特质维度上的现实与理想自我差异是否有所不同，以及这种认知特征在健康人群与临床患者之间有何不同。

本研究构建并验证了一个现实与理想自我差异的多维特质结构，探究了现实自我、理想自我和自我差异之间的认知联系，并探讨了自我差异和替代性自我差异的多维度认知机制。此外，通过一项功能磁共振成像(fMRI)研究，我们揭示了抑郁症患者相较于健康人群现实与理想自我差异的神经机制。

研究1中，我们首先进行了在线实验1 (177名被试) 来揭示自我差异的多维特质结构并初步探究其认知机制。我们通过探索性因子分析揭示了一个三维的因子结构 (社会性、能力性、勤勉性)，并在与之独立的在线实验2 (220名被试) 中通过一个适合度评估任务和一个分类任务进行了验证。另一方面，相比去计算现实自我和理想自我二者之间的数值差距，我们发现自我差异与对现实自我的直接感知更相关，并且，抑郁情绪的严重程度对自我差异大小的影响也是通过现实自我来起作用的。

研究2采用被试间设计，考察了对好朋友 (196名被试) 和陌生人 (147名被试)的替代性自我差异与对自己的自我差异之间的关系。验证性因子分析首先表明，研究1中的三维因子结构甚至可以合适地应用到替代性自我差异中。在此基础上我们进一步发现，自我差异和替代性自我差异都在能力性这一维度上最大、在社会性这一维度上最小，而且被试对好朋友的替代性自我差异评分最低、同时对自己的理想自我的渴求评分最高。此外，我们将目标人物 (自我、好朋友、陌生人) 的调节效应整合到研究1的中介模型后，发现目标人物对自我差异的调节效应是对现实自我的调节效应以及现实自我对自我差异的中介作用的共同结果。

在核磁研究3中，我们揭示了抑郁症患者和健康对照组在单变量激活和表征相似性方面的差异 (66名被试)。思考现实与理想自我差异在两组被试中都激活了相同的与自我参照相关的脑区，但只选择性地激活了抑郁症患者的右脑岛前叶 (Anterior insula)、左右丘脑 (Thalamus) 这些与消极情绪反应相关的脑区，而且从行为上来看，抑郁症患者整体的现实与理想自我差异的确较健康对照组更大。健康对照组的左侧额中回 (MFG) 和左侧楔前叶 (Precuneus) 的神经活动分布模式可以表征所有特质词语上的自我差异行为组成的不相似性结构，但没有在抑郁症患者中找到这种神经与行为的一致性。然而，我们强调抑郁症患者对于理想自我的渴求功能是完整的，我们还发现了抑郁症患者编码在杏仁核(Amygdala) 和后扣带回皮质 (PCC) 的自我保护机制：在全脑和感兴趣区域 (ROI)分析中，抑郁症患者的这两个脑区对自我差异越小的特质激活越强、进而共同巩固这些相对积极的自我知识, 特别是在能力性这一维度上。

本研究揭示了基于三维因子结构的现实与理想自我差异和替代性自我差异的普遍认知特征。此外，我们首次研究了抑郁症患者现实与理想自我差异的神经机制，并发现了由杏仁核和后扣带皮层编码的自我保护机制，为针对抑郁情绪的心理咨询和心理治疗提供了来自社会认知神经科学的视角，为抑郁症的个体化治疗提供了来自心理学研究的启示。

As the aphorism goes, ‘gold can’t be pure and man can’t be perfect’. Individuals compare the perceptions of their own (i.e., actual self) to an internal guide (i.e., ideal self), thus producing the inconsistency or gap called actual vs ideal self-discrepancy between these two different aspects of self-knowledge. Large self-discrepancy is associated with the psychological experience of lots of uncomfortable feelings such as disappointment and failure, and is regarded as one of the essential psychological processes in the formation and maintenance of depression. However, it remains unknown that whether people reflect about actual vs ideal self-discrepancy differently on various trait dimensions and how does this cognitive characteristic differ between healthy and depressive populations.

Here, we constructed and confirmed a multi-dimensional trait structure about actual vs ideal self-discrepancy, explored the cognitive interrelationship among actual self, ideal self and self-discrepancy and investigated the multi-dimensional cognitive characteristics of self-discrepancy and vicarious self-discrepancy. We also provided neural substrates of the differentiations between depressed patients and healthy populations on actual vs ideal self-discrepancy in a functional magnetic resonance imaging (fMRI) study.

In study 1, we first conducted online experiment 1 (N=177) to uncover the underlying trait structure of self-discrepancy and preliminarily discuss the cognitive mechanism. We revealed a three-dimension factor structure (Social, Rational, Devoted) by exploratory factor analysis, and confirmed it in an independent online experiment 2 (N=220) consisting of a fitness rating task and a classification task. On the other hand, we found that self-discrepancy was more associated with the direct perception for actual self rather than the calculation for distance between actual and ideal self, and the relationship between the severity of depressive mood and self-discrepancy is mediated by the perception of actual self.

In study 2, vicarious actual vs ideal self-discrepancy to a best friend (N=196) and a stranger (N=147), compared to self-discrepancy, were investigated with a between-subjects design. The confirmatory factor analysis suggested that the three-dimension factor structure applied well even for vicarious self-discrepancy. Based on that, we further discovered that participants consistently held largest self-discrepancy and vicarious self-discrepancy in Rational dimension and smallest in Social dimension, and the vicarious self-discrepancy for friend was smallest while the desirability for ideal self was strongest for one’s own. Furthermore, we integrated the modulation effect of target person (self, friend, stranger) to the mediation model in study 1 and demonstrated that the modulation of target person to self-discrepancy was the consequence of the modulation to actual self and the mediation by actual self.

We revealed the differences between depressed patients and healthy populations on univariate activation and representational similarity in fMRI study 3 (N=66). While similar self-referential brain regions were recruited in both groups, we found larger self-discrepancy behaviorally and selectively negative emotional responses in right anterior insula, left and right thalamus of depressed patients. The distributed patterns of neural activities in left middle frontal gyrus (MFG) and left precuneus of healthy controls represented the item-by-item behavioral dissimilarity structure, but this neurobehavioral coherence did not show in depressed patients. However, we highlighted the intact desirability for ideal self and discovered the self-protection mechanism encoded in amygdala and posterior cingulate cortex (PCC) of depressed patients, which responded more strongly to smaller self-discrepancy at both whole brain level and region of interest (ROI) level thus jointly consolidated positive self-knowledge, especially in Rational dimension.

Taking together, this research uncovered the universal cognitive characteristic of reflection on actual vs ideal self-discrepancy and vicarious self-discrepancy based on a three-dimension factor structure. Moreover, we first investigated the neural mechanism of the actual vs ideal self-discrepancy in depressed populations and suggested the self-protection mechanism encoded in amygdala and PCC, which provided a perspective from social cognitive neuroscience to future studies on clinical populations as well as empirical evidences and implications for treating moderate to major depressive disorder.