Forgiveness is essential for maintaining cooperation and resolving conflicts in social interactions. As artificial agents become more prevalent in interpersonal contexts, human-robot interactions (HRI) are becoming increasingly common. Nowadays, people are confronted with the decision to forgive not only human transgressors, but also robotic ones. To address this emerging social challenge, this research investigates how different post-offense behavioral responses from humans versus robots influence forgiveness, and explores the underlying psychological mechanisms through a dual-path framework that involves cognitive and affective factors. Across two experiments (each with a total sample size of N = 320), participants engaged in scenarios in which either a human or a robot committed an offense and subsequently exhibited one of four behavioral responses: no response, apology, self-deprecating humor, or self-punishment. Both experiments systematically manipulated agent type and behavioral response types.
In Experiment 1, the offense and behavioral response were presented textually. By employing a 2 (agent type: human vs. robot)×4 (behavioral response type: no response vs. apology vs. self-deprecating humor vs. self-punishment) between-subjects design, Experiment 1 revealed significant main effects of both the agent type and behavioral response type on forgiveness. Specifically, participants were more likely to forgive robots than humans. Confirming our mechanistic hypotheses, mediation analyses showed that perceived net benefit and net positive affect significantly mediated the effect of agent type on forgiveness, highlighting the role of cognitive and affective factors in shaping forgiveness.
Experiment 2 partially replicated and extended the findings of Experiment 1 using more ecologically valid video stimuli, which made the behavioral responses of human and robot offenders visible and thus more vivid and realistic. The main effects were validated once again. Mediation analysis revealed that people tend to forgive robots more readily, primarily due to higher levels of perceived net positive affect. Additionally, Experiment 2 revealed a significant interaction effect: the same behavioral response can have different effects on forgiveness depending on whether it is performed by a human or a robot. Specifically, compared to human, robots demonstrated a relative advantage in promoting forgiveness when adopting a self-punishment strategy. Building on this significant interaction, Experiment 2 further explored the psychological mechanisms through which agent type influenced forgiveness under different behavioral responses. Mediation analysis revealed that in the self-punishment condition, both net benefit and net positive affect significantly mediated the effect of agent type on forgiveness. These results suggest that when engaging in self-punishment, robots elicited greater perceptions of net benefit and net positive affect, thereby increasing their likelihood of being forgiven. Additionally, as a supplementary analysis, we conducted computational modeling. The results indicated that agent type and behavioral response type influenced individuals’ aversion to advantageous inequity, rather than altering their underlying decision algorithm in the context of forgiveness-related decisions.
The main novel contribution of this study lies in extending the scope of forgiveness research to include robots, an area that has been largely overlooked before. Most previous studies have focused on forgiveness toward humans, typically explaining it using the dual-path framework of cognition and affect. In response to the growing relevance of artificial intelligence in contemporary society, the present study examines forgiveness toward robots. The findings provide preliminary support for the applicability of the cognition-affect dual-process framework in explaining forgiveness toward robots, thereby extending its theoretical scope to encompass human-robot interactions.
In conclusion, this study revealed two key findings. First, in the context of social offenses, individuals exhibited greater forgiveness toward robots than toward humans - a pattern primarily mediated by perceived net positive affect, with perceived net benefit also potentially contributing. Second, robots demonstrated a relative advantage over humans in eliciting forgiveness when using a self-punishment strategy.
With the rapid advancement of artificial intelligence (AI) technology, public acceptance of its cognitive capabilities has grown, particularly in domains like medical data analysis where AI may rival or surpass human experts. Despite this growing acceptance, public resistance remains high when AI leads decision-making processes, such as in healthcare, which may stem from concerns about ethical boundaries (Longoni et al., 2019; Bigman & Gray, 2018). This highlights the need to understand how the public evaluates AI behavior in complex moral scenarios, a question that cannot be addressed solely by technical regulations. Moral dilemmas, a classic paradigm in moral psychology, provide a framework to explore how individuals judge AI versus human agents when faced with conflicting choices, such as utilitarian (outcome-focused) versus deontological (principle-focused) actions. Although moral dilemma research on human agents is well established, whether these findings extend to AI agents remains to be determined, as mind perception and responsibility attribution may differ for non-human agents (Waytz et al., 2010; Moor, 2006). This study investigates whether public moral judgments of AI diverge from those of humans, and how gender and AI-induced anxiety moderate these differences.
Three online experiments were conducted using an embodied AI agent (robot) to enhance ecological validity (Malle & Scheutz,
Study 1 showed a significant agent × moral behavior interaction. Compared with human agents, AI robots received higher moral ratings for utilitarian actions (M_AI_utilitarian =.26 ± 1.67; M_human_utilitarian = −.67 ± 2.01; t(473) = 16.74, p <.001, Cohen’s d =.77), whereas moral ratings for deontological actions did not differ between AI and human agents (p >.05). Gender moderated this effect: males exhibited a stronger deontological tendency when evaluating human agents (p<.05), whereas females showed comparatively more flexible judgments. Study 2 replicated the agent × moral behavior interaction and confirmed successful anxiety induction (anxiety: M_anxiety = 5.70 ±.81; non-anxiety: M_anxiety = 3.34 ± 1.63; p <.001). Induced anxiety was associated with more conservative moral evaluations and slightly reduced the positive evaluation of AI’s utilitarian actions, but the relevant interaction effect was not significant. Study 3 further indicated a gender-specific effect of induced anxiety: relative to the control condition, females in both the anxiety and non-anxiety groups showed reduced moral acceptance of the AI robot’s utilitarian actions, whereas males maintained higher moral acceptance of AI utilitarian actions. In the control group, baseline anxiety was low, and for both males and females, moral ratings of AI utilitarian actions were significantly higher than ratings of AI deontological actions.
Taken together, the public appears to regard AI as a relatively rational and efficient executor and to apply an evaluation orientation that prioritizes utility maximization and overall efficiency for AI. Relative to humans, outcome-maximizing decisions made by AI are more likely to be judged as morally acceptable, resulting in more favorable evaluations of AI’s utilitarian choices. The results also reveal gender differences: females are more susceptible to the influence of anxiety, which may relate to higher empathy and risk sensitivity, whereas males’ judgments are comparatively more stable and place greater emphasis on utilitarian considerations. Overall, moral judgments of AI are context-dependent and dynamically moderated by emotional states and individual differences. These findings provide evidence for the psychological mechanisms underlying moral evaluations in human-AI interaction and may inform the design of responsible AI systems that align with public moral expectations.
The Beck’s cognitive model of depression posits that the formation and maintenance of negative depressive cognitive schemas are key factors underlying the development and manifestation of depressive disorders. Among these schemas, negative attentional bias and negative memory bias are typical manifestations. The former usually manifests as a depressive individual's tendency to focus on negative stimuli, difficulty in disengaging attention, facilitated attention, or avoidance of attention. Meanwhile, the latter manifests as individuals with depressive tendencies displaying poorer suppression and better retrieval of negative memories compared to healthy controls. Moreover, since attention selection regulates the gateway for sensory information to enter perceptual and memory processing channels, and existing memory experiences in turn guide an individual's attention search and selection, the combined cognitive bias hypothesis posits that negative attention bias and negative memory bias can mutually influence each other and are closely related.
Existing research has established the presence of negative attentional bias and negative memory bias in individuals with depressive disorders as well as their subclinical counterparts (i.e., those with depressive tendencies). However, there is a divergence of views regarding the relationship between negative attentional bias and negative memory bias in individuals with depression. While some studies have demonstrated interrelations among these cognitive biases, others have found no significant associations between them.
To address the current situation, this study adopts the research paradigm proposed by Jia et al. (2023), which integrates the directed forgetting (DF) task with the emotion-induced blindness (EIB) task. In this experiment, we recruited both individuals with depressive tendencies and generally healthy controls to test whether the memory bias observed in the former would bias emotional attentional selection, thereby giving rise to an attentional bias. The experiment is divided into three stages. In the first stage, both groups engaged in intentional memory and intentional forgetting of negative pictures. In the second stage, the negative pictures from the previous stage are used as emotional distractors in the EIB task. Finally, in the third stage, participants complete a recognition test for the negative pictures. In this experiment, the Chinese version of the Beck Depression Inventory-II (BDI-II) was used to screen participants. Based on the BDI-II scores, college students with scores ≤13 were assigned to the healthy controls, while those with scores ≥19 were assigned to the Depression tendencies. A total of 75 valid participants were recruited, including 41 individuals with depressive tendencies and 34 healthy controls.
Our results indicated that both groups exhibited higher recognition accuracy for the negative to-be-remembered (TBR) pictures compared to those negative to-be-forgotten (TBF) pictures, thereby demonstrating a comparable DF effect. However, the impact of intentional memory encoding on the EIB effect revealed significant intergroup differences. Specifically, a significant between-group difference in target-identification accuracy on the EIB task emerged only when negative TBR pictures served as distractors; no comparable group differences were observed for negative TBF, novel negative, or neutral distractors. Moreover, for individuals with depressive tendencies, negative TBF and novel negative distractors elicited a clear EIB effect relative to the neutral-distractor condition, whereas negative “to-be-remembered” distractors did not. But this pattern was contrary to the EIB effects observed in the healthy controls.
These findings indicate that the memory-encoding task successfully induced a memory bias in individuals with depressive tendencies. Although this memory bias was not evident in the picture-recognition task, it reshaped emotional attentional selection during the EIB task by triggering attentional avoidance of negative to-be-remembered stimuli, thereby giving rise to a memory-driven bias. These findings not only deepen our understanding of memory biases in individuals with depressive tendencies by demonstrating that the picture-recognition task within the directed-forgetting paradigm is unsuitable for stably capturing such biases, but also provide robust support for the integrated cognitive bias hypothesis of depression, showing that memory biases in depression can influence and elicit attentional biases.
Emotionally salient stimuli preferentially capture attention, a phenomenon termed emotional attentional bias. This effect occurs not only with emotional faces and images but also in language processing. Unlike emotional faces and images, which provide explicit visual cues, emotional attentional bias in language relies on semantic processing of emotional words. Specifically, it involves extracting emotional information from words via semantic access. Research indicates that this process is closely linked to the representation of emotional semantics in words, that is, how words convey emotional meanings. Based on their semantic representation, researchers classify emotional words into two types: emotion-label words (e.g., "joy," "anger"), which directly denote emotions, and emotion-laden words (e.g., "success," "failure"), which convey emotions indirectly. These two types of words differ significantly in semantic processing, especially regarding the activation and extraction of emotional information. However, these findings are primarily based on studies of isolated word processing. In natural language processing, especially in Chinese reading, it remains unclear whether these two categories of emotional words show semantic differences and whether these representational differences affect reading processes. Moreover, it is unclear whether emotional semantic activation during reading depends on the integration of lexical contextual information. In this study, we employed natural reading tasks combined with eye-tracking technology, embedding emotional words within sentences to investigate the mechanisms underlying differences in emotional semantic processing and the influence of contextual diversity.
To investigate differences in emotional semantic processing across valences during natural reading, participants were first recruited to rate the emotional prototypicality of two types of emotional words, defined as the degree to which a word represents a specific emotion. The study comprised two experiments. Experiment 1 targeted positive valence, whereas Experiment 2 targeted negative valence. Both experiments used a 2 (emotion word type: emotion-label vs. emotion-laden) × 2 (contextual diversity: high vs. low) within-subjects design for natural sentence reading. Forty-eight participants took part, and their reading was recorded using an EyeLink 1000 Plus eye-tracker. Participants read one set of four experimental materials comprising 128 sentences: 80 experimental, 40 fillers, and 8 practice sentences. Approximately 30% of the sentences were followed by comprehension questions, to which participants responded with "yes" or "no."
The rating results revealed significant differences in emotional prototypicality between the two types of emotional words. Emotion-label words exhibited higher emotional prototypicality and more accurately represented emotional meanings. In positive valence (Experiment 1), emotion-label words showed longer fixation times, indicating greater attentional capture, whereas words with high contextual diversity had shorter fixation times, suggesting faster processing. No interaction effect was observed. In negative valence (Experiment 2), the main effect of emotion word type persisted, with emotion-label words showing longer fixation times during both early and late stages. Notably, the processing difference between emotion word types in negative valence was moderated by contextual diversity, affecting the entire processing span from early to late stages. Specifically, emotion-label words exhibited longer fixation times only under low contextual diversity, whereas no fixation time differences were observed between emotion-label and emotion-laden words under high contextual diversity.
The results reveal significant differences in the semantic representation of emotional words. Compared with emotion-laden words, emotion-label words more typically represent emotional meanings and exhibit higher emotional prototypicality. Furthermore, these findings suggest stable differences in emotional semantic processing during Chinese reading, with emotion-label words demonstrating a semantic advantage over emotion-laden words. Emotion-label words more effectively convey emotional information, capturing and sustaining readers' attention, emphasizing their ecological validity and supporting the embodied theory of semantic representation. These findings further demonstrate cross-linguistic consistency in emotional semantic differences and activation patterns. Additionally, contextual diversity moderated the semantic advantage of emotion-label words differently across valences, indicating that emotional semantics encompass both depth and breadth, and revealing a dissociation mechanism for activating positive versus negative emotional semantics during reading.
A scarcity mindset refers to a psychological state triggered by the perception of insufficient resources, such as money or time. It can capture attention, consume cognitive resources, and impair the quality of decision-making. Yet, its influence on risk-taking decisions remains controversial. Some findings suggest that scarcity promotes risk aversion through heightened sensitivity to loss and limited cognitive capacity. Contrastingly, others show reversed effects. These findings demonstrate the necessity to examine potential moderators in the relationship between scarcity and risk decision-making and to further investigate the underlying mechanisms.
The present study combined behavioral paradigms and event-related potential (ERP) measures to investigate how scarcity mindset shapes risky choices and their temporal neurocognitive processes. A modified Stage Game paradigm was used to induce scarcity or abundance mindsets by manipulating participants’ initial endowments. Risk preference was measured using the Balloon Analogue Risk Task (BART), which requires participants to inflate a virtual balloon that can explode unpredictably and result in the loss of accumulated gains. The task allows continuous learning and adjustment of decisions based on feedback, capturing the dynamic and iterative nature of real-world risk-taking.
Experiment 1 employed a 2 (mindset: scarcity vs. abundance) × 2 (reward magnitude: low vs. high) mixed design. Scarcity was induced by a small endowment, creating a constant sense of resource insufficiency, while abundance was induced by a larger endowment. Risk preference was assessed using the average number of inflations for balloons that did not explode. Results showed a significant main effect of mindset. Participants inflated balloons fewer times under scarcity (M = 36.30, SD = 14.43) than under abundance (M = 39.95, SD = 15.15), F(1, 67) = 8.38, p <.01, ηp2 =.11, indicating stronger risk aversion. Reward magnitude had no significant main effect, but the interaction between mindset and reward magnitude was significant, F (1, 67) = 5.34, p <.05, ηp2 =.07. Simple effects analysis revealed that scarcity reduced risk-taking under high reward magnitude, F (1, 67) = 13.75, p<.001, ηp2 =.17, but not under low reward magnitude, F (1, 67) =.17, p>.05. These findings suggest that scarcity increases loss aversion when potential rewards, and thus potential losses, are substantial.
Experiment 2 examined the neural dynamics of scarcity’s influence on risk evaluation using a 2 (mindset: scarcity vs. abundance) × 2 (feedback type: positive vs. negative) within-subject design. Behavioral results replicated Experiment 1, confirming that scarcity reduced risk-taking. ERP analyses focused on feedback-related negativity (FRN) and P300 components. FRN, measured 200-280ms after feedback at fronto-central electrodes, reflects early evaluation of outcomes and reward prediction errors. P300, measured 300-400ms after feedback at centro-parietal electrodes, reflects attention allocation and motivational significance of feedback. Scarcity produced significantly more positive FRN amplitudes than abundance, F (1, 25) = 6.20, p <.05, ηp2 =.20, and significantly more positive P300 amplitudes, F (1, 25) = 6.28, p <.05, ηp2 =.20. Both effects were moderated by feedback type. Under negative feedback, scarcity elicited more positive FRN and P300 amplitudes than abundance (FRN: F (1, 25) = 10.02, p <.01, ηp2 =.29; P300: F (1, 25) = 7.98, p <.01,ηp2 =.24). Under positive feedback, FRN and P300 did not differ significantly between mindsets. These results indicate that scarcity reduces the neural encoding of reward prediction errors while increasing attentional allocation to negative outcomes.
Together, these findings provide robust evidence that a scarcity mindset reduces risk-taking and alters the processing of negative feedback. The moderating effect of reward magnitude suggests that expected outcomes carry heightened significance under scarcity. More positive FRN amplitudes and P300 amplitudes point to a neurocognitive mechanism through which scarcity promotes cautious and loss-averse decisions. This research makes several contributions. It extends scarcity research to dynamic, ecologically valid tasks, addressing limitations of static, one-shot paradigms. It identifies reward magnitude as a key moderator that helps reconcile prior mixed findings. It integrates behavioral and neurophysiological evidence to reveal the temporal stages through which scarcity shapes risk evaluation. The findings enrich the scarcity theory by showing that scarcity reduces the capacity for adaptive reinforcement learning and fosters reliance on rigid, loss-avoidant decision strategies. These results provide theoretical insights and practical implications for interventions that support better decision-making among resource-constrained populations and help mitigate the long-term disadvantages associated with scarcity.
Major depressive disorder (MDD) and anxiety disorders (AD) exhibit high rates of comorbidity, presenting a significant challenge for clinical diagnosis and treatment. Traditional diagnostic frameworks often conceptualize these conditions as distinct entities, potentially overlooking the complex, interactive nature of their symptomatology and the shared cognitive mechanisms that may underlie both. Cognitive control, the ability to regulate thoughts and actions in accordance with internal goals, is a core deficit transdiagnostically associated with both MDD and AD. The contemporary dual-mechanisms of control theory posits two distinct modes: proactive control and reactive control. Understanding the profile of these control modes in comorbid MDD and AD (MDD-AD) could provide crucial insights into the disorder's mechanisms and inform more precise assessment tools. This study aimed to (1) investigate the specific characteristics of proactive and reactive cognitive control in individuals with MDD-AD compared to those with single diagnoses and healthy controls, and (2) construct and validate a machine learning-based model capable of predicting overall symptom severity using behavioral indices from a cognitive control paradigm.
The study design involved four groups of participants: 50 individuals in the MDD-AD group, 50 in the MDD-only group (MDD-o), 51 in the AD-only group (AD-o), and 50 in the healthy control group (HC). Participants were screened using the Zung Self-Rating Depression Scale (SDS) and Zung Self-Rating Anxiety Scale (SAS). Through the AX-CPT task, the reaction times, accuracy, and the relative strength of proactive and reactive control modes (Proactive Behavioral Index, PBI) were assessed. Additionally, a Random Forest (RF) regression model was used to predict symptom severity based on reaction times and accuracy, evaluating the applicability of the model across different groups.
The results showed that the MDD-AD group had significantly longer reaction times compared to other groups in the AX-CPT task, and significantly lower accuracy on BX trials compared to the HC group. Specifically, the reaction times in AX trials were 646 ms for the MDD-AD group, compared to 401 ms for the HC group (p <.001). In BX trials, MDD-AD participants exhibited significantly lower accuracy (.85) compared to the HC group (.99, p <.001). The Proactive Behavioral Index (PBI) indicated that MDD-AD participants showed significantly lower proactive control (PBIRT =.10) compared to HC participants (PBIRT =.20, p <.001). Moreover, MDD-AD participants exhibited compensatory enhancement of reactive control in the accuracy-based PBI, with PBIACC values of -.18, which was significantly lower than that of the AD-o group (-.40, p <.05), and higher than the MDD-o group (.22, p <.001). These findings suggest that in MDD-AD, proactive control is impaired while reactive control is enhanced.
Machine learning analysis demonstrated that the RF regression model, particularly the comprehensive diagnostic model, outperformed single-diagnosis models in predicting symptom severity. The comprehensive diagnostic model achieved a mean absolute error (MAE) of 5.23, with an R2 of.78, indicating strong predictive accuracy. SHAP (SHapley Additive exPlanations) analysis revealed that reaction times in AY and BX trials were key features in predicting symptom severity. The model's feature importance ranking was stable, with the most significant features being the behavioral performance in these trials ( ρ =.89).
In conclusion, individuals with MDD-AD exhibit a distinct cognitive control mode characterized by reduced proactive control and compensatory enhancement of reactive control. The machine learning-based prediction model, utilizing the AX-CPT task, effectively predicts symptom severity and provides theoretical support for the precise assessment and intervention of comorbid disorders. Future research may further integrate multimodal data and expand sample sizes to validate the generalizability of the dual cognitive control model in other psychiatric conditions and explore its underlying neural mechanisms.
Interference inhibition is a core component of cognitive control, referring to the ability to complete goal-directed tasks by suppressing distracting or competing cognitive processes. Previous research on the effects of reward on interference inhibition has yielded inconsistent findings (facilitative vs. null effects). The expected value of control (EVC) theory provides an integrative framework. It proposes that the allocation of cognitive control resources depends on a cost-benefit trade-off, with efficacy playing a key moderating role. Efficacy refers to the degree to which task performance predicts reward. However, existing research generally treats efficacy as a static property and overlooks its distinct temporal forms. Efficacy is fundamentally an expectation formed through accumulated feedback and may comprise both a short-term, trial-level component and a more stable, cross-trial component. To date, no study has systematically examined whether the temporal dimension of efficacy determines the depth at which reward influences different levels of cognitive control. Accordingly, the present study aimed to investigate how immediate efficacy and cumulative efficacy, as two temporal forms of efficacy, modulated the impact of reward on overall cognitive performance and on interference inhibition as a higher-level control process.
The present study employed the classic color-word Stroop task and rigorously controlled relative task difficulty using individualized reaction-time thresholds to eliminate confounds arising from differences in expected cost. Across two experiments, we separately manipulated immediate efficacy and cumulative efficacy in order to systematically examine how the temporal dimension of efficacy shaped the level at which reward influenced cognitive control. Reward levels were set as high (￥1) and low (￥0.1). In the high-efficacy condition, rewards were contingent on task performance, whereas in the low-efficacy condition, rewards were randomly assigned. Experiment 1 (n= 39) used a 2 (immediate reward: high, low) × 2 (immediate efficacy: high, low) × 3 (congruency: incongruent, congruent, neutral) within-subject design. Reward and efficacy cues were presented before each trial to investigate the effects of immediate reward and immediate efficacy on interference inhibition. Results indicated that immediate efficacy moderated the effect of immediate reward on overall cognitive performance, with reward producing greater improvements in reaction speed under high efficacy conditions (F(1, 38) = 22.25, p <.001, ηp2 =.37). However, neither factor significantly influenced interference inhibition (Stroop effect).
Previous research suggests that efficacy exhibits cumulative learning characteristics, requiring multiple interactions with the environment to integrate task feedback and prediction errors. Psychological momentum theory further supports this view, proposing that repeated positive feedback enhances an individual’s psychological energy, thereby improving behavioral performance. Based on these considerations, experiment 2 (n= 33) employed a 2 (immediate reward: high, low) × 2 (cumulative efficacy: high, low) × 3 (congruency: incongruent, congruent, neutral) within-subject design. Efficacy cues were presented at the beginning of each block, and reward cues were presented before each trial, to examine the effects of immediate reward and cumulative efficacy on interference inhibition. Results demonstrated that cumulative efficacy moderated the effects of immediate reward on both overall cognitive performance (F(1, 32) = 4.57, p <.05, ηp2 =.13) and interference inhibition (F(1, 32) = 10.46, p <.01, ηp2 =.25). Under high efficacy conditions, high reward enhanced interference inhibition, whereas under low efficacy conditions, high reward reduced interference inhibition.
The present study introduced the temporal dimension of efficacy for the first time and demonstrated that this dimension modulated the level at which reward influenced cognitive control. Immediate efficacy modulated only the effect of reward on overall cognitive performance, whereas cumulative efficacy allowed the influence of reward to extend to the level of interference inhibition. This work provides more fine-grained and explanatory behavioral evidence for the EVC framework and deepens our understanding of how motivation shapes cognitive control.
In the process of individual moral judgment, group identity significantly influences decision-making, leading to two distinct phenomena: in-group favoritism and the "Black Sheep Effect." While children often show favoritism toward in-group members, they sometimes treat them more harshly. The precise conditions under which children shift from in-group favoritism to the stricter "Black Sheep Effect" remain unclear, particularly regarding how specific features of moral violations—such as the repetition of violations and the representativeness of violators—moderate this shift. Guided by Social Identity Theory and Subjective Group Dynamics Theory, this study systematically investigated the moral judgments and punishment behaviors of 5- to 8-year-old children under different violation contexts—specifically, repetitive violations and violations by representative group members—and examined the underlying motivations.
Three experiments were conducted. In all experiments, children watched animated video scenarios depicting moral violations. Experiment 1 employed a 2 (Group Identity: In-group vs. Out-group) × 2 (Violation Type: Single vs. Repeated) × 2 (Age: 5-6 vs. 7-8 years) mixed design. A total of 128 children watched videos depicting in-group or out-group members committing either a single or repeated moral violation (taking pencils from a donation box). Punishment intensity was measured by the number of reward stickers deducted. Experiment 2 adopted a 2 (Group Identity: In-group vs. Out-group) × 2 (Representativeness: Representative vs. Non-representative) × 2 (Age: 5-6 vs. 7-8 years) mixed design with 124 participants. The representativeness of a puppet character (e.g., a class monitor and role model vs. an ordinary member) was manipulated via verbal descriptions prior to the violation. This experiment compared how children reacted to violations committed by highly representative group members (e.g., class monitors) versus ordinary group members. Experiment 3 used a within-subjects design with 61 children aged 7-8 years to investigate whether punishment motivations (consequentialist vs. retributive) differ between in-group and out-group violators. After making their punishment decisions, children rated their agreement with statements reflecting consequentialist (e.g., hoping for improvement) and retributive (e.g., believing they deserved it) motives. Data were analyzed using paired-samples t-tests, chi-square tests, and ANOVA.
The results revealed that: First, the interaction between violation context and group identity significantly influenced punishment. In single-violation contexts, 5- to 6-year-olds displayed in-group favoritism, punishing out-group members more severely. In repeated-violation contexts, children aged 5-8 consistently demonstrated the "Black Sheep Effect," punishing in-group violators more harshly, with the difference increasing with age. Notably, 7- to 8-year-olds showed no in-group bias in single-violation contexts, punishing both groups equally. Second, the representativeness of the violator amplified this effect. Children imposed the strongest punishment on representative in-group violators, with over 80% such violations as more severe. Third, punishment of in-group members was primarily driven by consequentialist motives (hoping for improvement), with significantly higher scores for in-group than out-group targets, while retributive motivations did not differ significantly between groups.
In conclusion, moral judgment and punishment in 5- to 8-year-olds are context dependent and develop with age. Repeated violations and the representativeness of the violator are key triggers of the "Black Sheep Effect." The stricter punishment of in-group violators is essentially driven by a consequentialist desire to "improve the group," reflecting a mindset of "strict with one's own, lenient with others." Younger children (5-6 years) are more influenced by in-group favoritism in single-violation contexts, whereas older children (7-8 years) pay greater attention to the nature of the act itself, overcoming group bias. This developmental pattern suggests a transition from a simpler "in-group preference" strategy to a more nuanced, norm-upholding approach to group membership. These findings demonstrate a developmental shift in children's moral reasoning, from a simple "in-group preference" to a more sophisticated enforcement of group norms, offering new theoretical insights and empirical evidence for moral education and the fostering of group responsibility.
Self-Regulated Learning (SRL), the capacity of learners to actively plan, monitor, and adapt their learning behaviors to achieve academic goals, is essential for academic success and the development of lifelong learning competencies. Yet many learners struggle to develop SRL. Traditional interventions are limited by scarce resources and delayed feedback, preventing them from providing sustained and individualized support. Artificial Intelligence (AI) offers potential solutions, but its specific positive and negative effects on students’ SRL remain underexplored. This paper reviews SRL theories and components, proposes three human-AI collaborative models centered on learner agency, analyzes AI’s mechanisms and implementation pathways across SRL stages, and discusses challenges and future directions.
Building on Bandura’s foundational work, researchers have developed influential models such as Zimmerman’s cyclical model, Winne and Hadwin’s model, and Pintrich’s four-phase model. Despite different emphases, these models view SRL as a dynamic, cyclical process involving learning goals and strategies, with cognition, metacognition, and motivation as its core components. Drawing on perspectives such as hybrid human-machine regulation and the “collective brain,” this paper examines how these components are externalized and reconfigured in AI-enabled environments. In such contexts, some metacognitive monitoring and cognitive strategy selection are transferred to AI as an external regulatory system. Learners must still selectively interpret and integrate AI feedback according to their goals and values, transforming SRL from an internal into a multi-source regulatory process. The central challenge is balancing human and machine regulation so that AI enhances rather than undermines SRL.
The paper proposes an agency-centered human-AI collaborative SRL framework with three modes: (1) AI-dominated scaffolded regulation, in which AI largely handles task analysis, path planning, and process monitoring while learners follow recommended pathways; (2) Human-AI co-regulated decision-making, in which learners and AI negotiate goals, strategies, and evaluations through dialogue while learners retain primary decision authority; and (3) Self-directed tool-supported regulation, in which learners with clear goals use AI on demand as an information and problem-solving resource while maintaining autonomy over planning, monitoring, and reflection. These modes intersect with the phases of forethought, performance, and reflection to form a “Collaborative Mode × SRL Phase” matrix showing the human-AI division of labor across phases and its implications for SRL development.
The paper then reviews empirical studies on AI-enhanced SRL in higher education from the past five years. AI can support SRL through personalized learning pathways, real-time feedback, process-data visualization, and the externalization of metacognitive activities. AI-scaffolded regulation can reduce cognitive load and improve short-term performance, especially for learners with weaker SRL skills. However, prolonged outsourcing of planning, monitoring, and evaluation may encourage dependency, superficial processing, and metacognitive laziness, ultimately weakening learner agency. Human-AI co-regulation can strengthen metacognitive monitoring, motivation, and autonomy when roles and decision boundaries are clear and dialogue is well structured. Poor design, however, may blur responsibility, encourage direct-answer seeking, and overwhelm learners with frequent prompts, increasing emotional burden and weakening sustained motivation. Tool-supported regulation is most conducive to transferable SRL competencies when learners possess strong self-regulatory abilities and AI literacy, but it may also widen competency gaps and digital inequality.
Effective human-AI collaborative SRL depends on feedback that is descriptive rather than directive, scaffolding that fades rather than persists, interaction that is bidirectional rather than unidirectional, and support that is available on demand rather than imposed. The core principle is that AI should strengthen, not replace, learners’ monitoring and regulatory capacities. This principle supports the integration of adaptive mode switching, scaffolding fading, transparency, and controllability into AI-supported SRL interventions and tools. Nevertheless, the three modes are theoretical ideal types and may not fully capture dynamic changes caused by tasks and cognitive load in authentic contexts. Because the framework is designed specifically to enhance university students’ SRL, it should not be automatically generalized to efficiency-oriented or other human-computer collaboration contexts.
Finally, the paper identifies three research directions. First, future studies should validate the three collaborative models and their transition mechanisms, test the framework’s generalizability, develop context-sensitive theoretical frameworks and agency measures, and examine interactions between cognition and affect. Second, multimodal learning analytics should be combined with neuroimaging techniques, including eye-tracking, fNIRS, and fMRI, to investigate the cognitive and neural mechanisms of human-AI co-regulation and detect early signs of overreliance. Third, longitudinal studies should examine the development, transfer, and possible regression of AI-supported SRL competencies, while expanding evaluation to include process indicators such as reflective depth.
In recent years, “lying flat” (tang ping) has attracted growing scholarly attention as a passive coping orientation among some young adults, characterized by reduced striving and disengagement from demanding goals. Within this broader phenomenon, an avoidant lying-flat tendency—marked by withdrawal from challenges, low initiative, and preference for short-term relief over effortful coping—may be particularly relevant to university students’ adjustment and technology-related behaviors. In parallel, internet addiction (or problematic, uncontrolled internet use) remains prevalent in college settings and is linked to impaired functioning. Prior work has reported cross-sectional associations between avoidant coping styles and problematic internet use; however, most studies have not adopted longitudinal approaches capable of separating within-person fluctuations (i.e., changes around one’s own typical level over time) from stable between-person differences (i.e., trait-like tendencies). Moreover, potential psychological processes explaining these temporal links—such as self-efficacy as a mechanism and depressive symptoms as a contextual vulnerability—have received limited longitudinal testing within a framework that is sensitive to within-person dynamics.
Using a three-wave design and a random-intercept cross-lagged panel model (RI-CLPM), the study examined: (a) Whether avoidant lying-flat tendency and internet addiction show reciprocal, cross-lagged associations at the within-person level; (b) Whether self-efficacy functions as a time-varying mediator linking the two constructs across adjacent waves; and (c) Whether depressive symptoms moderate the hypothesized within-person pathways. RI-CLPM was selected because it explicitly decomposes observed variance into stable person-specific intercepts and wave-specific within-person deviations, thereby reducing the risk of conflating trait differences with temporal processes.
Participants were undergraduate students from a university in Sichuan Province, China. Data were collected at three time points: March 2024 (T1, n = 739), September 2024 (T2, n = 685), and March 2025 (T3, n = 656). Standardized self-report measures assessed avoidant lying-flat tendency, internet addiction, self-efficacy, and depressive symptoms. RI-CLPM analyses were estimated with robust maximum likelihood (MLR) and full information maximum likelihood (FIML) for missing data. Gender, age, parental education, family income, and concurrent time spent online were controlled to reduce confounding by key sociodemographic factors and internet exposure. Longitudinal indirect effects were tested using bias-corrected bootstrapping. To evaluate moderation, multi-group RI-CLPM analyses compared participants classified into low- versus high-depressive-symptom subgroups.
At the within-person level, cross-lagged paths indicated a significant reciprocal pattern across waves. Specifically, when a student reported higher-than-usual avoidant lying-flat tendency at one time point, the student tended to report increased internet addiction at the subsequent time point. Conversely, above-usual internet addiction also predicted subsequent increases in avoidant lying-flat tendency, suggesting a mutually reinforcing loop. These bidirectional cross-lagged associations remained significant after accounting for stable between-person differences via random intercepts and after controlling for covariates. At the between-person (intercept) level, the random intercepts of avoidant lying-flat tendency, internet addiction, and self-efficacy were correlated in theoretically coherent directions, consistent with the view that individuals with more persistent avoidance tendencies also tend to show more stable problematic internet use and lower enduring self-efficacy.
Mediation analyses further supported self-efficacy as a within-person mechanism. Higher-than-usual avoidant lying-flat tendency predicted subsequent reductions in self-efficacy, which in turn predicted later increases in internet addiction. A parallel indirect pathway was observed in the opposite direction: higher-than-usual internet addiction predicted subsequent decreases in self-efficacy, which then predicted later increases in avoidant lying-flat tendency. These findings suggest that self-efficacy may act as a time-varying psychological “hinge” through which avoidance and problematic internet use perpetuate one another over time. In contrast, multi-group analyses showed no robust evidence that depressive symptoms moderated the key within-person cross-lagged paths, indicating that the reciprocal within-person loop may operate relatively similarly across depression levels in this non-clinical student sample.
Overall, the results provide longitudinal, within-person evidence that avoidant lying-flat tendency and internet addiction are dynamically intertwined and may mutually amplify one another over time, with self-efficacy serving as a central mediating process. The findings align with social cognitive perspectives suggesting that diminished efficacy beliefs can increase reliance on maladaptive coping strategies and short-term relief behaviors. Practically, campus prevention and counseling programs may benefit from targeting self-efficacy enhancement (e.g., mastery experiences, adaptive goal setting, and cognitive-behavioral skill training) to disrupt the avoidance-internet addiction cycle. Future research should replicate these findings across more diverse university contexts, incorporate multi-method assessments (e.g., behavioral logs), and test intervention-based designs to further clarify temporal mechanisms.
The increase in employee mobility across organizational boundaries presents challenges and opportunities for knowledge acquisition beyond formal employment frameworks. In this context, this study advances the concept of alumni voice—defined as former employees’ promotive behavior involving the expression of constructive challenges intended to improve their former employers—to investigate two pivotal questions: whether and how former organizational position influences alumni voice, and how personal sense of power moderates this relationship. Building on psychological ownership theory, we explore how employees' organizational positions in the former employer foster alumni voice. Senior employees accumulate valuable skills and knowledge at the former employer, forming a foundation for alumni voice. Psychological ownership mediates this relationship by reflecting continued feelings of control, knowing and investment toward the former organization, while sense of power amplifies this link, further enhancing alumni voice. Thus, this study develops a moderated mediation model that examines how former organizational positions foster alumni voice via enhanced psychological ownership, with personal sense of power amplifying this relationship.
A two-wave online survey involved 633 participants, all of whom were former employees now working as entrepreneurs or in top executive roles. The first wave assessed former organizational position, psychological ownership, and sense of power. The second wave, conducted one month later, measured alumni voice. This study controlled demographic data, including gender, age, education, and tenure. Moreover, unmeasured latent method construct (ULMC) was used to test the potential common method bias (CMB), and the moderated mediation hypothesis was tested using hierarchical regression and PROCESS macro. Additionally, this study employed a 2 (former organizational position: high/low) × 2 (sense of power: high/low) between-subjects experimental design with 267 participants. Using role-playing scenarios to simulate employees transitioning into founding team members post-departure, we successfully validated and confirmed the robustness of our initial survey findings.
The data analysis revealed that a higher former organizational position is associated with a greater willingness to engage in alumni voice, and this link is explained by psychological ownership: individuals in more senior roles reported stronger feelings of “ownership” toward their former organization, which in turn fueled their willingness to engage in alumni voice. Importantly, personal sense of power moderated this relationship: those with a stronger personal sense of power showed a steeper increase in psychological ownership as former organizational position rose, and consequently exhibited a stronger indirect effect of position on alumni voice through ownership. The results of data analysis demonstrated significant direct and indirect pathways, and the indirect effect of former organizational position on alumni voice via psychological ownership was significantly greater at higher levels of sense of power. Together, these results illustrate a moderated-mediation process by which former organizational position drives alumni voice through psychological ownership, with personal power amplifying this indirect effect through psychological ownership.
The theoretical contributions of this study are threefold. First, this study defines and introduces the concept of alumni voice. It positions such voice as a cross-organizational knowledge backflow from former employees, thereby transcending the traditional organizational boundaries that have limited the voice literature to current employees. Second, this study challenges the employment constraint in traditional voice literature that roots voice in employment relationships as an insider obligation. In doing so, it reveals that alumni voice is primarily motivated by affective bonds, representing voluntary knowledge feedback rather than role-based duty. Finally, this study deepens the understanding of psychological ownership by validating its dynamic continuity after employment ends, challenging static perspectives. Practically, organizations can leverage this mechanism through three strategies: (1) Establish formal channels for alumni to contribute insights via dedicated platforms (e.g., alumni forums) to maximize their value; actively solicit feedback through structured initiatives. (2) Nurture alumni networks to access external resources and collaboration opportunities; foster an alumni-friendly culture and implement connection-focused recruitment; and host open days and visits to strengthen ties and enable relationship spillover. (3) Cultivate psychological ownership among employees to deepen their organizational attachment; prompt alumni to serve as external advocates who promote organizational interests and generate tangible benefits.
The complexity of the external environment exposes leaders and subordinates to situational disparities and information asymmetries, increasing the likelihood of emotional conflicts and unpredictable leader expressions. However, existing research has primarily focused on the effects of leaders’ predictable emotional expressions on subordinates, assuming that emotional cues align with situational information. In reality, emotional expressions do not always correspond with situational logic, making them appear unpredictable to observers. This study explores how leader emotional unpredictability shapes subordinates’ reactions. Grounded in the emotion-as-social-information (EASI) theory, we propose that leader emotional unpredictability heightens subordinates’ cognitive confusion and anxiety, subsequently influencing their interpersonal behavior and task performance. Furthermore, we expect these effects to be more pronounced in newer leader-subordinate relationships.
To test our hypotheses, we conducted three studies. A priori power analysis using G*Power indicated that a minimum sample size of 85 was required for the proposed model. Study 1 employed the critical incident method, collecting 427 valid responses. Participants were asked to recall and describe incidents in which they directly perceived leaders’ unpredictable emotional expressions, including the cause, process, and outcome of the event, before completing relevant questionnaires. Study 2 adopted a 2 (leader emotional predictability: predictable vs. unpredictable) × 2 (leader-subordinate relationship type: new vs. old) between-subjects experimental design. We obtained 248 valid samples via a professional online platform. Study 3 involved a two-wave field survey conducted one month apart among 386 leader-subordinate dyads from companies in the pharmaceutical, chemical, and food service industries. At Time 1, subordinates reported leaders’ emotional unpredictability, team procedural fairness, relationship length, and demographics. At Time 2, subordinates reported their anxiety, cognitive confusion, and interpersonal withdrawal, while leaders evaluated subordinates’ task performance.
Linear regression analyses and independent-samples t-tests were conducted to test Hypotheses 1 and 3, while the bootstrap method was employed to examine the mediating effects proposed in Hypotheses 2 and 4. In addition, simple slope analyses and ANOVA were used to assess the moderating role of relationship type in testing Hypotheses 5 and 6. The findings indicated that: (a) Leader emotional unpredictability was positively associated with subordinates’ cognitive confusion and anxiety; (b) Both cognitive confusion and anxiety mediated the relationship between leader emotional unpredictability and subordinates’ interpersonal withdrawal, whereas the indirect effect on task performance through anxiety was only supported in Studies 1 and 3; and (c) The type of leader-subordinate relationship moderated the indirect effects of leader emotional unpredictability via cognitive confusion on both withdrawal behavior and task performance. Specifically, in newer leader-subordinate relationships, the negative impact of leader emotional unpredictability on subordinates was most pronounced.
Our research makes several theoretical contributions. First, it extends the boundaries of the EASI framework by demonstrating how leader emotional unpredictability affects subordinates, providing a foundation for future studies that jointly consider emotional cues and situational information. Second, by uncovering the underlying dimensions of emotional unpredictability, the study enriches the theoretical understanding of complex leader emotional expressions and offers new insights into how such expressions relate to subordinate behavior in ambiguous contexts. Third, by introducing the mechanism of emotional complementarity, this research highlights how leaders’ unpredictable emotions can elicit complementary anxiety in subordinates, thereby expanding the conceptual scope of emotional complementarity within EASI. Finally, by examining the moderating role of leader-subordinate relationship type, our research takes a broader perspective on the interpersonal effects of leaders’ episodic emotional expressions, underscoring the critical role of relationship stage in emotion interpretation.
Perceived overqualification, characterized by the perception that one’s knowledge, skills, and abilities exceed job requirements, is becoming increasingly prevalent. While perceived overqualification is commonly conceptualized as person-job misfit associated with adverse outcomes, it may also constitute surplus resources that can be converted into constructive behaviors through job crafting. Mixed findings regarding the overqualification-job crafting association indicate substantial heterogeneity in employees’ responses. We propose that motivational orientations account for this heterogeneity by directing overqualified employees toward either promotion-oriented or prevention-oriented job crafting. To address these issues, this study integrates the conservation of resources theory and regulatory focus theory to examine how perceived overqualification influences promotion-oriented, and prevention-oriented job crafting, while also exploring the mediating role of job insecurity and the moderating role of vocational delay of gratification.
The present study employed two complementary studies to provide evidence for our research model. Study 1 adopted a scenario-based experiment with a 2 (overqualification: high vs. low) × 2 (vocational delay of gratification: high vs. low) between-subjects design (N = 152). Participants were instructed to immerse themselves in carefully constructed workplace scenarios and then complete measures of job insecurity, promotion-oriented and prevention-oriented job crafting. The results revealed a significant interaction between overqualification and vocational delay of gratification on job insecurity. Under low vocational delay of gratification, high overqualification heightened job insecurity, whereas under high vocational delay of gratification, high overqualification reduced job insecurity. Moderated mediation analyses further showed that among individuals high in vocational delay of gratification, perceived overqualification indirectly promoted promotion-oriented job crafting via reduced job insecurity. Conversely, among individuals low in vocational delay of gratification, perceived overqualification increased prevention-oriented job crafting through elevated job insecurity.
Study 2 extended the experimental findings to a real organizational context using a three-wave field survey (N = 231). The results replicated the moderated mediation patterns observed in Study 1, thereby confirming that vocational delay of gratification shapes whether perceived overqualification functions as a resource that attenuates job insecurity or as a threat that amplifies it, thereby leading to divergent forms of job crafting.
Our study offers significant theoretical contributions. First, it contributes to the literature on job crafting by proposing its potential negative effect. By linking overqualification to both promotion-oriented and prevention-oriented job crafting, we provide a novel explanatory lens that reconciles previously inconsistent findings. Second, drawing on the COR theory, we demonstrate that employees are not passive victims of person-job misfit; rather, they actively regulate their resources by investing or conserving them. This finding enriches person-job fit theory by underscoring the importance of self-regulatory processes in shaping outcomes of misfit experiences. Finally, by integrating the COR theory with Regulatory Focus Theory, we highlight job insecurity as a key mediating mechanism and vocational delay of gratification as a pivotal boundary condition. These insights reveal divergent pathways—resource investment versus resource conservation—through which perceived overqualification manifests in the workplace.
The research provides important practical implications for organizations in managing overqualified employees. First, organizations should approach the recruitment of highly qualified candidates with an awareness of both risks and opportunities. Instead of dismissing overqualified applicants outright, managers should consider whether roles can be structured to harness employees’ surplus potential. Second, organizations should prioritize the creation of a supportive and stable work climate that enhances psychological safety, thereby alleviating overqualified employees’ concerns about future uncertainties. Moreover, during the selection stages, organizations may benefit from assessing candidates’ vocational delay of gratification as an indicator of their ability to adapt constructively to short-term person-job misfit. Finally, day-to-day managerial practices should be sensitive to the motivational orientations of overqualified employees. Supervisors can tailor developmental opportunities to foster promotion crafting among those with high delay of gratification, while offering targeted support to help employees with a short-term orientation to avoid maladaptive prevention strategies.
Recent developments in the self-control theory suggest that the successful regulation of desires and impulses does not always depend on effortful willpower. Instead, individuals can achieve self-control through the use of less effortful, strategic approaches that proactively minimize the emergence or impact of temptations. This theoretical shift has prompted growing interest in how to effectively promote the use of self-control strategies. However, most research to date has been conducted in Western contexts and has primarily focused on individual-level factors, such as traits, motivation, or regulatory capacity. Less attention has been paid to whether culturally rooted ways of thinking are associated with the use and effectiveness of self-control strategies in daily life. To address this gap, the present study explored self-control strategy use within a Chinese cultural context and further examined whether zhongyong thinking, a culturally rooted thinking style emphasizing holistic consideration, dialectical coordination, and harmony orientation, predicts both the use and effectiveness of self-control strategies. Two key questions guided this study: (1) Are all types of self-control strategies equally effective in mitigating temptations and impulses within a Chinese cultural context? (2) Is zhongyong thinking associated with self-control success through the use of specific self-control strategies?
Ecological momentary assessment (EMA) was employed to track participants’ self-control experiences and strategy use in daily life. After completing a baseline survey measuring zhongyong thinking, trait self-control, and demographic variables, participants took part in a seven-day EMA procedure, with seven assessments per day at 8:00, 10:00, 12:00, 14:00, 16:00, 18:00, and 20:00. In each assessment, participants reported whether they had experienced a self-control situation during the current or previous two hours. If they had attempted self-control, they further reported the type of situation, desire strength, strategies used, and self-control success. Self-control strategies were coded into four categories: situation, attention, cognition, and inhibition strategies. A total of 127 participants contributed 5,886 valid assessments. Given the nested structure of the data, multilevel models were estimated using the lme4 package in R. Multilevel linear models were used to examine whether different strategies predicted self-control success, whereas multilevel logistic models were used to examine whether zhongyong thinking predicted the use of different strategies. Bayesian multilevel mediation analyses were further conducted using the brms package to test the indirect effects of zhongyong thinking on self-control success through strategy use.
The results (see
By employing EMA, this study examined the use and effectiveness of self-control strategies in everyday life within a Chinese cultural context. Furthermore, it investigated how zhongyong thinking, as a culturally rooted thinking style, is associated with individuals’ use of strategies. The findings suggest that, in contrast to inhibition strategies, situation, attention, and cognition strategies are more consistently associated with self-control success in Chinese contexts. This research contributes to a more culturally inclusive understanding of self-control by highlighting how macro-level cultural values and thinking styles can shape micro-level behavioral strategies. These findings provide preliminary evidence for understanding how culturally related psychological constructs may be associated with self-control strategy use and effectiveness in a Chinese cultural context.
Cognitive impairment has become an increasingly pressing public health concern in aging populations, particularly in low- and middle-income countries. Emerging evidence indicates that adverse childhood experiences (ACEs) and childhood socioeconomic status (SES) may leave enduring biological and psychosocial imprints, shaping cognitive trajectories throughout the life course. However, the independent and combined contributions of ACEs and childhood SES to cognitive impairment in later life remain inadequately characterized, and the underlying mechanisms linking early-life adversity to subsequent cognitive decline are poorly understood. Addressing these gaps is critical for identifying modifiable pathways and informing interventions aimed at preserving cognitive health across the lifespan.
This longitudinal study used data from the China Health and Retirement Longitudinal Study (CHARLS). A total of 4964 adults aged ≥45 years without baseline cognitive impairment were included and followed for a median of 8.9 years. ACEs were retrospectively assessed using ten indicators from the 2014 life history survey, and childhood SES was measured based on parental education and childhood family economic conditions. Cognitive impairment was identified using a validated composite measure covering episodic memory, orientation, numerical ability, and visuospatial function. Lifestyle factors, including smoking, alcohol consumption, sleep duration, and social participation, were combined into a composite lifestyle score. Cox proportional hazards models were applied to estimate the independent and joint associations of ACEs and childhood SES with incident cognitive impairment. Both additive and multiplicative interactions were tested. Additionally, subgroup analyses were conducted according to sex, age, residence, marital status, educational level, and activities of daily living to explore potential effect modification. Causal mediation analysis under a counterfactual framework was further performed to assess the mediating role of lifestyle behaviors.
During follow-up, 889 participants (17.9%) developed cognitive impairment. Compared with individuals without ACEs, exposure to 1-2 ACEs was associated with an 18% higher risk of cognitive impairment (HR = 1.18, 95% CI: 1.01-1.37, p <.05), whereas exposure to three or more ACEs was associated with a 36% higher risk (HR = 1.36, 95% CI: 1.10-1.68, p <.01). Relative to high childhood SES, middle childhood SES was associated with a 22% increased risk of cognitive impairment (HR = 1.22, 95% CI: 1.04-1.42, p <.05), and low childhood SES conferred an even greater increase in risk of 38% (HR = 1.38, 95% CI: 1.15-1.66, p =.001). Joint exposure to ≥3 ACEs and low childhood SES was associated with the highest risk of cognitive impairment (HR = 2.00, 95% CI: 1.38-2.90, p = <.001), although no significant additive or multiplicative interactions were observed. Mediation analyses indicated that lifestyle partially mediated the associations of both ACEs and childhood SES with cognitive impairment, with sleep duration emerging as a significant individual mediator in the ACEs pathway. Subgroup analyses revealed that the association between ACEs and cognitive impairment did not differ significantly across any subgroup (all p for interaction >.05). In contrast, a significant interaction was observed for childhood SES in the sex subgroup (p for interaction =.036), indicating that low childhood SES was more strongly associated with cognitive impairment in females. Both ACEs and low childhood SES independently increase the risk of cognitive impairment among middle-aged and older Chinese adults. Lifestyle behaviors were found to partially mediate these associations, indicating that adopting healthier behaviors in adulthood—such as maintaining adequate sleep, engaging in social activities, and avoiding harmful habits—may attenuate, though not fully eliminate, the long-term cognitive consequences of early-life adversity.
These results underscore the dual importance of addressing early-life risk factors and promoting life-course-oriented lifestyle interventions. Integrating early preventive strategies with sustained behavioral modifications throughout adulthood may provide a comprehensive approach to mitigating cognitive decline and supporting healthy aging in populations at risk.
Prosocial risky behavior refers to voluntary actions undertaken to benefit others or collective welfare, while knowingly bearing potential risks of personal loss, including physical harm, economic cost, foregone developmental opportunities, or social interpersonal penalties. As a unique prosocial construct that integrates altruistic motivation and voluntary risk-taking, it is a cornerstone of social cohesion, emergency mutual aid, and collective well-being in both daily risky scenarios and large-scale public crises. However, existing research has predominantly focused on the predictors of low-risk, cost-free conventional prosocial behaviors (e.g., charitable donation, informal helping) or separately explored the determinants of individual self-interested risk-taking. Few studies have examined the socio-cognitive drivers of prosocial risky behavior, which requires the simultaneous activation of altruistic motivation and a willingness to bear personal costs, leaving a critical gap in both theoretical explanations and empirical evidence. In particular, no systematic research has investigated how social mobility beliefs shape individuals’ subjective positive expectations of upward socioeconomic mobility through personal effort and hard work, influencing their willingness to engage in altruistic acts that involve risk. To address this gap, an integrated theoretical framework synthesizing the social comparison theory, conservation of resources theory, and social identity theory is constructed. Two hypotheses were tested: (1) social mobility belief significantly and positively predicts prosocial risky behavior; and (2) personal relative deprivation, the negative cognitive-affective experience arising from unfavorable upward social comparisons, mediates the positive association between social mobility belief and prosocial risky behavior.
A sequential multi-study design with multi-method triangulation was adopted to establish a rigorous evidence chain from macro correlational association to causal effect identification, and finally to underlying psychological mechanism verification. For Study 1, data were extracted from the 2013 Chinese General Social Survey (CGSS), a nationally representative large-scale dataset, with a final valid sample of 3800 community-dwelling adults across China. Social mobility belief was measured via items capturing its two core dimensions (perceived opportunity fairness and social class boundary permeability); prosocial risky behavior was assessed through a validated high-stakes situational indicator; and the predictive effect was tested via binary logistic regression, with key demographic and socioeconomic covariates controlled. Study 2 adopted a between-subjects experimental design (n = 260) to verify the causal effect of social mobility belief, where belief was manipulated via a well-validated self-persuasion writing task, and prosocial risky behavior was measured by a computerized 20-scenario daily-life prosocial risk dilemma task. Study 3 was a cross-sectional questionnaire survey (n = 600), where standardized and well-validated scales were used to measure all core variables, and bootstrap mediation analysis was performed to preliminarily test the hypothesized mediating role of relative deprivation. Study 4 used an experimental causal chain design (n = 260) with an alternative text-comic reading manipulation paradigm to validate the full causal mediation pathway by sequential measurement of the mediator and outcome variable after experimental manipulation.
The results from the four studies were highly consistent and robust. Study 1 showed that social mobility belief significantly and positively predicted prosocial risky behavior after controlling for all covariates. Study 2 confirmed the causal effect: participants in the high social mobility belief condition exhibited a significantly higher rate of helping choices than those in the low belief condition. Study 3 revealed a significant partial mediating effect of relative deprivation, accounting for 9% of the total effect. Most critically, Study 4 validated the full causal mediation chain: experimentally induced high social mobility belief significantly reduced participants’ relative deprivation, and lower relative deprivation in turn significantly increased prosocial risky behavior. The indirect effect via relative deprivation was robustly significant, explaining 69% of the total effect, and the direct effect of social mobility belief became non-significant after including the mediator.
In conclusion, social mobility beliefs can effectively promote prosocial risky behavior, with the positive effect primarily channeled through reducing individual relative deprivation. The findings yield key theoretical contributions: the explanatory scope of social mobility belief is extended from conventional low-cost prosocial behavior to the domain of risky altruism; the affective psychological mechanism linking macro-societal structural belief to micro individual prosocial action is clarified; and an integrated theoretical explanation is provided for the dual requirements of altruism and risk-taking in prosocial risky behavior. The findings also offer evidence-based practical insights for social governance, especially in the context of China’s common prosperity initiative, indicating that fostering positive social mobility beliefs via equal opportunity policies can help mitigate relative deprivation, enhance social cohesion, and encourage courageous mutual aid in risk situations.
This paper aims to explore the roles of experiential avoidance and self-differentiation in the relationship between adverse childhood experiences and self-injurious behavior in adolescents. It also seeks to examine the effectiveness of dialectical behavior therapy (DBT) incorporating self-differentiation in intervening in self-injurious behavior. In recent years, self-injurious behavior has become a serious problem among adolescents. According to the gene-environment interaction model, adverse childhood experiences are significant predictors of self-injurious behavior. Furthermore, the experiential avoidance model suggests that experiential avoidance plays a key role in the relationship between adverse childhood experiences and self-injurious behavior in adolescents. In addition, several studies have found that self-differentiation can reduce non-suicidal self-injurious behaviors in adolescents. As a traditional intervention therapy for self-injurious behavior, DBT still has limitations. Therefore, the present study investigates whether DBT that integrates self-differentiation is able to reduce self-injurious behaviors in adolescents.
A total of 783 adolescents were selected for Study 1, including 405 boys and 378 girls. The average age was 13.16 years. The relationships between adverse childhood experiences, experiential avoidance, self-differentiation, and self-injurious behavior were tested using a questionnaire method. Study 2 was an intervention study in which a total of 98 adolescents were selected and randomly divided into three groups: the Self-Differentiation-Focused Integrated Training Group (SDFIT group), the Dialectical Behavior Therapy Control Group (DBT group), and the Waitlist Control Group for intervention. The levels of self-differentiation, self-injurious behavior, distress tolerance, and well-being were measured at three waves (baseline, post-intervention, and follow-up). Data were analyzed using SPSS 22.0 for both Study 1 and Study 2.
The results showed that: (1) Adverse childhood experiences had a significant direct effect on self-injurious behavior (β =.22, p <.001); adverse childhood experiences had a significant direct effect on experiential avoidance (β =.43, p <.001); experiential avoidance had a significant direct effect on self-injurious behavior (β =.17, p <.001). Adverse childhood experiences had a significant indirect effect on self-injurious behavior through experiential avoidance, with a 95% confidence interval of [.07,.19]. The interaction between self-differentiation and experiential avoidance had a significant effect on self-injurious behavior (β = -.25, p <.001). At the low self-differentiation level, there was a positive correlation between adolescents' levels of experiential avoidance and non-suicidal self-injurious behavior. However, at the high level of self-differentiation, this relationship was not significant. (2) After the intervention, the level of self-differentiation was significantly higher in the SDFIT group than in the DBT group (p<.05) and the control group (p <.01). Furthermore, the score of self-injurious behavior in the SDFIT group was significantly lower than that in the control group (p <.001) after the intervention. Additionally, the level of well-being in the SDFIT group was significantly higher than that in the control group (p <.05) at post-intervention, and was significantly higher than that in the other two groups during the follow-up period (p <.001; p <.001). At post-intervention, distress tolerance was significantly lower in the SDFIT group than in the DBT group (p <.001) and the waitlist group (p<.05). During follow-up, it remained significantly lower in the SDFIT group than in the waitlist group (p <.05). These findings suggest that the DBT intervention integrating self-differentiation is effective in increasing adolescents' self-differentiation level and reducing self-injurious behavior.
The findings suggest that experiential avoidance plays a mediating role in the relationship between adverse childhood experiences and self-injurious behavior, and self-differentiation moderates the relationship between experiential avoidance and self-injurious behavior. Furthermore, dialectical behavior therapy integrating self-differentiation not only serves as an effective intervention for self-injurious behavior but also provides theoretical and practical support for the prevention of self-injurious behavior.
School bullying remains a prevalent global concern, posing significant risks to the positive development of adolescents. The vulnerability of adolescents’ internal meaning systems is closely linked to the risk of bullying victimization. When individuals lack a sense of life value and a clear self-concept, they are less likely to develop self-protective awareness or proactive coping, increasing their likelihood of being bullied. Strengthening adolescents’ internal meaning resources plays a pivotal role in the prevention of school bullying.
Life awe and self-esteem represent two such resources. According to meaning-making theory, when individuals encounter negative events, their global meaning system, which encompasses core beliefs about life and self-worth, interacts dynamically with the situational meaning of the event. A stable meaning system provides a coherent framework for integrating stressors and adapting positively. Building on this foundation, life awe fosters a profound appreciation of the value of life, which strengthens individuals’ valuation of their own lives and enhances self-protective awareness. Self-esteem reflects a global evaluation of one’ s worth. High self-esteem reinforces positive self-beliefs, leading to greater confidence and resistance in dealing with bullying. Together, they help adolescents maintain meaning-system stability under threat, foster confidence and proactivity in self-protection, and thereby reduce the risk of victimization. However, bullying victimization, as a peer-interpersonal stressor characterized by aggression and social exclusion, may challenge the global meaning system, thereby potentially undermining life awe and self-esteem.
Life awe and self-esteem may have a mutually reinforcing relationship. Life awe can enhance self-esteem by promoting positive cognitive restructuring. In turn, self-esteem helps individuals better understand appreciate the finiteness of life, fosters more positive life experiences, and thereby enhances life awe. Accordingly, this study employed the cross-lagged panel analysis to examine the dynamic relationships among life awe, self-esteem, and bullying victimization.
A longitudinal design was adopted, with 3,911 adolescents (Mage = 14.7 years, 55.7% boys) surveyed twice over a six-month interval in April 2024 (T1) and October 2024 (T2). This study used the Life Awe Scale for Chinese Adolescents, the Self-Esteem Scale, and the Chinese version of the Delaware Bullying Victimization Scale-Student. Descriptive statistics, correlation analyses, and common method bias tests were conducted using SPSS 26.0, while longitudinal measurement invariance testing and latent cross-lagged modeling were performed using Mplus 8.3. The results revealed that: (1) Cross-lagged paths indicated that both T1 self-esteem and T1 life awe significantly and negatively predicted T2 bullying victimization (β = -.11, -.12), and T1 bullying victimization negatively predicted T2 self-esteem (β = -.04); (2) A bidirectional predictive relationship existed between life awe and self-esteem, with the effect of self-esteem on life awe being stronger (β =.22 vs..10); (3) Multi-group analyses showed that the associations among life awe, self-esteem, and bullying victimization were consistent across gender, whereas the negative predictive effect of self-esteem on bullying victimization differed across grade levels. Specifically, the negative predictive effect was significant in junior secondary students only; (4) Longitudinal mediation analysis showed that, after controlling for T1 life awe and T1 bullying victimization. T2 life awe partially mediated the relationship between T1 self-esteem and T2 bullying victimization, with an indirect effect of -.04 (95% CI= [-.051, -.021]), accounting for 30.8% of the total effect.
This study suggests that life awe and self-esteem function as key internal meaning resources that protect adolescents against bullying victimization. Beyond their direct protective effects, self-esteem may also exert an indirect effect by facilitating the development of life awe. More importantly, self-esteem and life awe may jointly contribute to an integrated psychological defense mechanism through a meaning-development process that extends from self-worth to the value of life, thereby enhancing adolescents’ capacity to cope with the risk of bullying. The interaction between these two resources may contribute to the formation of a meaning-oriented self-protection system, helping adolescents shift from passive defense to more proactive prevention.
These findings highlight important practical implications for bullying prevention. Effective intervention requires not only external support, but also the cultivation of stable and enduring internal motivation. Educators and practitioners should emphasize the development of adolescents’ internal meaning systems, especially through life education and the promotion of healthy self-concept. Strengthening adolescents’ recognition of life value and fostering a stable sense of self-worth can activate intrinsic motivation for self-protection and provide more sustainable psychological and social resources to support their adaptation.
Although the comorbidity of post-traumatic stress symptoms (PTSS) and depressive symptoms is well-documented across various traumatic events, the comorbidity network among school bullying victims remains understudied, particularly concerning potential gender differences. The overlapping symptoms hypothesis posits that shared symptoms between PTSS and depressive symptoms drive their comorbidity, which has been rarely tested using network analysis. Therefore, this study aimed to examine the comorbidity network characteristics of PTSS and depressive symptoms among adolescent victims of school bullying, focusing on how these symptoms manifest differently across gender groups through the application of network analysis method. A total of 724 adolescents who were victims of school bullying participated in the survey, and this study was conducted using the Chinese version of the Delaware Bullying Victimization Scale, the PTSD Checklist for DSM-5, and the Chinese version of the Center for Epidemiologic Studies Depression Scale for Children.
The results showed that in the overall sample of this study, the nodes with higher expected influence (EI) in the network were grief, flashback, sadness, no positive emotion, detachment, and concentration problems. The primary bridge symptoms connecting the PTSS and depressive symptom networks were no positive emotion, detachment, concentration problems, cannot concentrate, feeling fearful, and negative beliefs. However, attentional problems were the overlapping symptoms shared by PTSS and depressive symptoms.
Furthermore, the results revealed that there were no significant differences in the structure or overall strength of the PTSS-depressive symptoms networks between male and female adolescents. However, there were similarities and differences in the central symptoms and bridge symptoms of the comorbidity networks. Notably, overlapping symptoms of PTSS and depressive symptoms, specifically attention problems and negative emotions, were identified only in the male group, with these symptoms demonstrating stronger bridge centrality.
Specifically, among boys, the symptoms of grief, sadness, flashback, detachment, and irritability and anger exhibited stronger EI. The main bridge symptoms in the comorbidity network of PTSS and depressive symptoms among boys were concentration problems, cannot concentrate detachment, no positive emotion, sadness, and people dislike me. Therefore, attention problems and negative emotions were overlapping symptoms of PTSS and depressive symptoms in boys who have experienced school bullying. In the PTSS-depressive symptoms comorbidity network among bullying victimized girls, flashback, grief, no positive emotion, concentration problems, and feeling tired were identified as central symptoms. The main bridge symptoms in the girls’ PTSS and depressive symptoms network were no positive emotion, feeling fearful, detachment, hard to do something and sleep problems. It could be seen that grief and flashback were central symptoms common to both genders, while detachment and no positive emotion served as shared bridge symptoms.
Theoretically, these findings provide gender-specific support for the overlapping symptoms hypothesis. While the boys' comorbidity network was driven by both overlapping (e.g., attention problems, negative emotions) and non-overlapping bridge symptoms (e.g., people dislike me), the girls’ network was driven exclusively by non-overlapping symptoms (e.g., feeling fearful, hard to do something). Practically, this study offers a prioritized intervention framework for bullying-victimized adolescents. First, targeting shared bridge symptoms (e.g., detachment, no positive emotion) can disrupt cross-diagnostic connectors, yielding broad effectiveness across both genders. Second, treatments should be tailored to gender-specific dynamics—prioritizing overlapping symptoms in boys and distinct non-overlapping symptoms in girls. Taken together, these findings highlight the value of integrating transdiagnostic and gender-specific strategies in treating PTSS and depressive symptoms.
Over the past three decades, routine outcome monitoring (ROM), or formal feedback (also called measurement-based care), has emerged as a pivotal advance in psychotherapy. ROM entails the systematic collection of client-reported outcomes to track therapeutic progress and guide clinical decision-making. Since the introduction of “patient focused research” in 1996, ROM has garnered sustained scholarly attention due to its demonstrated efficacy in improving treatment outcomes, reducing client dropout, and bridging the research-practice gap.
In China, preliminary studies have reviewed ROM’s clinical effectiveness and applicability. However, research has largely focused on early-stage developments, with limited attention to its overall effectiveness, practical implementation challenges, and strategies for its adoption. Advances in information technology, accumulating meta-analytic evidence, and the global dissemination of ROM have created both opportunities and challenges, motivating this systematic review of three decades of research. This review addresses three core questions: “When and for whom is ROM effective?”, “What implementation challenges exist?”, and “How can ROM research and practice be effectively promoted in China?”
Evidence from over 60 clinical trials and 11 meta-analyses confirms the positive effect of ROM on treatment outcomes, with effect sizes ranging from small (g =.10) to moderate (g =.49). Variability is influenced by several factors including client feedback system, implementation process, client symptom severity, and clinical context. Only two feedback systems have extensive randomized clinical trial (RCT) support: the Outcome Questionnaire−45 System and the Partners for Change Outcome Management System (PCOMS). These two systems account for 74% of the studies included in the recent meta-analysis. The evidence shows that the PCOMS and the use of additional clinical support tools (CST) yield stronger effects, a larger effect for not-on-track clients and those with moderate baseline symptoms, and studies conducted in the U.S. Conversely, feedback may be less effective, or potentially counterproductive, for clients with severe baseline symptoms.
Challenges in ROM research and practice are multifaceted. Cross-cultural and population-specific evidence is limited, with 97% of trials conducted in Western contexts and sparse evidence for children, adolescents, or group therapy. Quantitative studies have primarily focused on therapists, leaving client contributions underexplored despite ROM’s current client-centered foundation. Research design issues, including researcher allegiance and dual use of measures for intervention and evaluation, may introduce bias, although large-sample trials controlling for these factors still support ROM’s effectiveness.
At the clinical level, therapist fidelity remains a recurrent challenge. Effective ROM requires three interrelated clinical processes: collecting client-reported outcome data, sharing feedback with clients, and acting on the data to guide shared decision-making. Evidence suggests that merely collecting or reviewing feedback without collaborative discussion limits its impact. High-quality application, characterized by integrating feedback data into therapeutic dialogue, is critical for enhancing treatment outcomes.
Barriers to ROM adoption operate across individual, organizational, and cultural levels. Individual-level challenges include client concerns about privacy, perceived burdens and therapists’ attitudes, knowledge, and confidence in using feedback. Organizational constraints encompass limited resources, lack of electronic feedback systems, insufficient training, and weak leadership support. Cultural factors, such as “face” concerns in Chinese contexts, may inhibit therapist engagement, particularly when negative outcomes are perceived to threaten professional authority.
Strategies to enhance ROM’s effectiveness and uptake include rigorous, culturally sensitive research, flexible implementation in group or long-term therapy, and technology-assisted data collection. Adoption is further shaped by culture: in non-Western societies like China, therapists’ concerns about perceived risks or “face concern” may be a more salient barrier to its use, indicating training should prioritize mitigating perceived risks. Finally, systemic policy and clinical guideline support are also essential for routine, institutionally supported ROM implementation, as evidenced by initiatives in the United States.
In summary, ROM represents an evidence-based intervention with substantial potential to enhance psychotherapy outcomes, facilitate early identification of stagnating cases, support individualized treatment, and strengthen therapeutic alliances. Despite preliminary domestic research, widespread implementation faces challenges related to cultural adaptation, therapist training, and organizational support. Future work should consolidate empirical evidence, examine client-centered mechanisms, and develop culturally appropriate strategies and policy frameworks to promote routine feedback-based psychotherapy in China. This review provides a comprehensive overview of ROM’s development, current challenges, and future directions, aiming to advance research and standardize evidence-based psychotherapy practice.
The efficient-coding constrained Bayesian inference model, developed by Stocker and colleagues (Wei & Stocker, 2012, 2015, 2017; Sun, Xu & Stocker, 2025), has emerged as a pivotal theoretical framework in visual cognition research. This model delineates a comprehensive perceptual process comprising three key stages: efficient sensory encoding, Bayesian decoding, and perception-action mapping. It formally characterizes the interplay between neural resource allocation (Fisher Information, $ J(\theta)$), prior ($P(\theta)$), width of neural tuning curves/likelihood distribution $\left(\sigma^{2}\right)$, perception biases ($b(\theta)$), as expressed by the following relationships:$ p(\theta) \propto \sqrt{J(\theta)} ; \sigma^{2} \geq 1 / \sqrt{J(\theta)} \propto 1 / p(\theta) ; b(\theta) \propto\left(\frac{1}{p(\theta)^{2}}\right)^{\prime}$
Collectively, these equations unveil the computational principles governing visual perception under environmental constraints.
Existing research has explored the model’s implications from three primary perspectives:
1. Unified Account of Perceptual Biases: The model provides a parsimonious computational framework explaining diverse phenomena, including noise-dependent biases (Wei & Stocker,
2. Multi-Timescale Prior Integration: The framework reveals how long-term priors (reflecting learned statistical regularities) stabilize perception, while short-term priors (driven by recent stimuli) induce serial dependence (Fischer & Whitney,
3. Computational Basis of ASD Perceptual Anomalies: Compared to neurotypical individuals, populations with autism spectrum disorder (ASD) exhibit reduced encoding efficiency and impaired statistical adaptation (Noel et al.,
Building on these advances, four critical directions can be considered in future studies:
1. Multisensory and Cross-Modal Applications: Current research indicates that the efficient coding principle is universally observed across unimodal sensory systems, such as vision and audition, while Bayesian decoding mechanisms are also applicable to cross-modal integration (e.g., Ernst & Banks,
2. Extensions to Social Cognition: Social cognitive processes are similarly influenced by prior experience and exhibit Bayesian inference properties (e.g., the Prevalence-Induced Concept Change [PICC] phenomenon, Levari et al.,
3. Computational Specificity in Special Populations: Individuals with ADHD, dyslexia, and other neurodevelopmental conditions exhibit atypical perceptual processing, including impaired encoding efficiency or aberrant prior formation (e.g., Nigg & Casey,
4. Elucidating Neural Mechanisms: Although existing studies demonstrate that neural activity aligns with both efficient coding (Ganguli & Simoncelli,
Addressing these questions will not only advance theoretical paradigms in perceptual computation but also deepen our understanding of human perception. Moreover, these insights may inspire biologically grounded designs for brain-like artificial intelligence systems, with transformative implications for clinical practice and machine perception.
