The pathogenesis of anorexia nervosa (AN) involves multiple factors, among which childhood traumatic experience has attracted attention. Emotional abuse, as a form of trauma, may exert a predictive effect on AN. In addition, childhood traumatic experience is closely linked to impulsivity, yet the trauma-impulsivity-AN mediation pathway has not been directly validated. The purpose of this study is to compare differences in childhood traumatic experience of patients with different subtypes of AN, as well as exploring the mediating role of impulsivity between emotional abuse in childhood traumatic experience and symptom severity of AN. This study included 157 female patients with AN, including 76 with the restricting type (AN-R) and 81 with the binge-eating/purging type (AN-BP), as well as 124 matched healthy controls (HC). Childhood traumatic experience was evaluated using Early Trauma Inventory-short form (ETI-SF), impulsivity assessed by Barratt Impulsiveness Scale 11th Version (BIS-11), and clinical characteristics via Eating Disorder Examination-questionnaire (EDE-Q 6.0), Beck Depression Inventory (BDI) and Beck Anxiety Inventory (BAI). Furthermore, inter-group differences in childhood traumatic experience were determined by one-way analysis of variance and analysis of covariance, the correlation between childhood traumatic experience and AN by Pearson correlation analysis, and the mediating role of impulsivity between emotional abuse and symptom severity was clarified by the Bootstrap method. Cases in the AN-BP group had significantly higher exposure to childhood traumatic experience compared with the AN-R and HC groups (both p < 0.05). Significant differences were likewise observed in emotional-abuse scores across the three groups (F = 10.574, p= 0.000, partial η² = 0.084). Emotional abuse was positively correlated with impulsivity and symptom severity of AN (both p < 0.05). In addition, the mediation effect of impulsivity between emotional abuse and symptom severity was 0.073 (95
Objectives Suicide risk assessments currently rely on subjective clinical judgement, lacking objective measures. This study aimed to evaluate the association between biomarkers and suicide risk and to explore their predictive potential using machine learning (ML).Methods We analysed data from 2785 first-admission psychiatric inpatients across three institutions, including 103 biomarkers and 13 demographic and clinical variables. Suicide risk was assessed 1 week after admission using the Nurses’ Global Assessment of Suicide Risk.Results A total of 2785 patients met the inclusion criteria, from which 978 were selected via propensity score matching to minimise confounding from demographic factors, treatment differences and symptom severity. Multivariate random effects logistic regression identified nine biomarkers associated with elevated suicide risk and six with potential protective effects. Time-trend analyses further revealed that nine biomarkers showed significant changes following risk escalation. Integrating biomarkers with demographic data, treatment information and psychological scale scores substantially improved ML model performance, achieving an area under the receiver operating characteristic curve of 0.808 in the external testing cohort. The inclusion of biomarkers significantly enhanced predictive accuracy.Conclusion This study highlights the potential of biomarkers with ML to predict future risk, offering objective assessments and supporting early interventions.
Schizophrenia is a severe, chronic mental disorder that imposes a substantial burden on family members, and this burden is strongly correlated with caregivers’ increased levels of anxiety and depression. Conversely, perceived social support functions as a critical coping resource, thereby effectively moderating stress and facilitating better psychological adjustment. This quantitative study tested the mediating hypothesis that perceived social support accounts for the relationship between caregiver burden and these negative psychological outcomes. This study adopted a cross-sectional design and convenience sampling to recruit 393 patients with schizophrenia and their primary family caregivers in China. The data collected included demographic information, caregiver burden, anxiety and depression, and perceived social support levels. Statistical analyses included ANOVA, correlation analysis, and hierarchical regression analysis. The results showed that caregiver burden was positively correlated with both anxiety and depression, and that perceived social support played a partial mediating role in this relationship. The mediating effect of perceived social support was particularly strong in mitigating the association between burden and depression. The findings indicate that perceived social support is vital in alleviating psychological distress among these caregivers, suggesting that targeted interventions focusing on strengthening social support networks can effectively improve the mental health status of family caregivers of patients with schizophrenia.
Precise regulation of protein synthesis is critical for brain functions such as long-term memory, and its dysregulation is implicated in numerous memory disorders. While mRNA methylation such as N6-methyladenosine (m6A) that has been widely studied in memory, the role of mRNA acetylation remains largely unknown. N4-acetylcytidine (ac4C), the only known form of RNA acetylation in eukaryotes, promotes mRNA stability and translation. In this study, we identified the ac4C epitranscriptome in mouse hippocampal homogenates and synaptosomes through ac4C-RNA immunoprecipitation followed by next-generation sequencing (acRIP-seq). The Morris water maze (MWM) was employed to induce and evaluate memory acquisition and forgetting processes. We show that synaptic ac4C levels are dynamically regulated, increasing after memory formation and returning to baseline after natural forgetting. The dynamic changes of ac4C-mRNAs regulated by memory were validated by ac4C dot-blot, liquid chromatography-tandem mass spectrometry (LC-MS/MS), and acRIP-qPCR analysis. We further demonstrate that NAT10, the ac4C writer, in mouse hippocampus is important for spatial memory via regulating memory-related mRNAs, proteins, and ultimately synaptic plasticity. Lastly, we generated a freely-accessible website (http://ac4catlas.com) that included the dataset of ac4C epitranscriptome in mouse hippocampus. Altogether these results demonstrate that dynamic and localized mRNA acetylation is a critical for synaptic plasticity and memory.
Large language models (LLMs) usually lack monitor and self-reflect abilities. Although methods like Chain of Thought (CoT) guide models through a sequence of intermediate reasoning steps, they miss the human-like capacity to evaluate and correct errors at any time, which leads to reasoning failures and hallucinations. Inspired by human metacognition, we introduce Metacognitive Activation Addition (MAA), a novel method that adds the metacognitive vectors into the internal representations of an LLM. Specifically, drawing from cognitive science, which divides human metacognition into planning, monitoring, and reflection, we extract corresponding steering vectors for each module by calculating the average difference in hidden layer activations between positive and negative examples from specially crafted datasets. Furthermore, our innovative hierarchical approach embeds these vectors into sequential layers, mirroring theories of hierarchical cognitive. Our experiments demonstrate that MAA effectively induce strong metacognitive responses, leading to substantial improvements on downstream tasks, including complex reasoning and hallucination reduction.
