Mental health disorders represent a growing burden across the Middle East and North Africa (MENA) region, where depression and anxiety are highly prevalent amid conflict, displacement, and socioeconomic strain, affecting up to 40 percent of adults, yet treatment gaps remain at 80-95% due to provider shortages, financial strain, and cultural barriers. In this context, artificial intelligence (AI), in the form of large language models (LLMs) and specialized psychotherapy chatbots, may offer a scalable adjunct to help address these gaps through anonymous screening, predictive risk modeling, psychoeducation, and brief interventions. This narrative review examines current evidence of AI-driven conversational tools in mental health with a specific focus on their application, acceptance, and limitations within the MENA region. To do so, A structured search of MEDLINE and Embase (2000–2026) identified studies on conversational AI in mental health, prioritizing evidence from the MENA region and supplemented by relevant global literature. Overall, findings suggest that while these tools offer high accessibility and user engagement, particularly for low-intensity support, their effectiveness is limited by linguistic and cultural mismatches, including Arabic diglossia and poor alignment with locally grounded expressions of distress. At the same time, user acceptance reflects a paradox in which stigma and privacy concerns drive reliance on anonymous AI tools while simultaneously limiting trust in their clinical reliability, reinforcing a preference for hybrid models with human oversight. Taken together, these findings indicate that current systems remain insufficiently adapted to the MENA context, underscoring the need for culturally grounded, dialect-sensitive, and clinically supervised approaches to ensure safe and effective integration.
Beyond Theory of Mind: mentalization as a relational and developmental framework for autism
Autistic individuals and those around them often navigate social and emotional situations in which behaviors, intentions, and affects are difficult to interpret. Supporting mentalizing processes within child–caregiver interactions may help address these challenges; however, a broader conceptual shift is needed, moving beyond a narrow deficit-based perspective toward understanding mentalization as a multidimensional, relational, and developmental process. By shifting the focus from individual deficits to child–caregiver meaning-making processes, this framework may help clarify assessment and intervention targets and inform future research on psychopathological vulnerability in autism. This targeted narrative mini-review therefore aimed to summarize preliminary evidence suggests that other-related mentalizing may show greater difficulties than self-related mentalizing, although this hypothesis requires further replication. Findings also highlight caregiver mentalization, particularly parental reflective functioning, as a key relational process shaping how children’s behavior is interpreted, regulated, and responded to over time. In this light, preliminary intervention studies suggest that mentalization-based and mentalization-informed approaches may improve parental reflective functioning, cognitive reappraisal, self-efficacy, and inferential style, with potential indirect benefits for children’s emotional outcomes. We therefore propose a relational-developmental framework in which mentalization is conceptualized as a shared and dynamic process of meaning-making under conditions of social and emotional ambiguity. Adopting an individual, relational and developmentally informed perspective may contribute to the development of more precise assessment models, more targeted interventions, and a deeper understanding of mental health vulnerability in autism.
Construction and validation of multiple machine learning models for influencing factors of postpartum post-traumatic stress disorder in primiparas
ObjectiveTo analyze the multidimensional factors associated with postpartum post-traumatic stress disorder (PP-PTSD) in primiparas based on the Integrated Framework for Population Health Risk Management (IFPHRM), multiple machine learning-based predictive models were constructed and externally validated to identify high-risk individuals and to provide a robust evidence base for targeted preventive interventions.MethodsThis cross-sectional study consecutively enrolled 1, 135 primiparous women from the Department of Obstetrics at Hefei Maternal and Child Health Hospital between June 2024 and May 2025. Participants were divided chronologically into a training cohort and an independent temporal validation cohort. Women recruited from June 2024 to January 2025 were included in the training cohort (n = 794), whereas those recruited from February 2025 to May 2025 were included in the temporal validation cohort (n = 341). At six weeks postpartum, PP-PTSD symptoms were assessed using the Post-traumatic Stress Disorder Checklist-Civilian Version (PCL-C), with a score ≥38 indicating probable PP-PTSD. Multidimensional variables, including physiological and psychological factors, environmental and family-related factors, and social-behavioral factors, were collected. Candidate predictors were first screened using univariate analysis and then selected using least absolute shrinkage and selection operator (LASSO) regression. Multivariable logistic regression was used to identify independent associated factors. Seven machine learning models, including Logistic Regression, Naive Bayes, Support Vector Machine, Decision Tree, Gradient Boosting, AdaBoost, and Linear Discriminant Analysis, were constructed. Model performance was evaluated in the independent temporal validation cohort using receiver operating characteristic curves, calibration curves, decision curve analysis, and the DeLong test. SHAP analysis was used to interpret the optimal model.ResultsAmong the 794 participants in the training cohort, the incidence of PP-PTSD was 25.18%. Five key predictors were selected by LASSO regression: social support, depression, neonatal caregiving style, husband’s participation, and sleep quality. Multivariable logistic regression showed that depression and poor sleep quality were associated with an increased risk of PP-PTSD, whereas higher social support, greater husband’s participation, and parental assistance in neonatal care were associated with a reduced risk. Among the seven models, the Gradient Boosting model achieved the best overall performance in the temporal validation cohort, with an AUC of 0.939, F1 score of 0.700, specificity of 0.943, sensitivity of 0.651, and Youden index of 0.595. The DeLong test showed that Gradient Boosting performed significantly better than Logistic Regression. SHAP analysis further indicated that social support, husband’s participation, sleep quality, and depression were the major contributors to model prediction.ConclusionPostpartum PTSD (PP-PTSD) exhibits a higher incidence among primiparous women and exerts substantial adverse effects on maternal mental health, the mother–infant relationship, and overall family functioning. Guided by the Integrated Framework of Perinatal Health Risk Management (IFPHRM), this study elucidated the multidimensional mechanisms underlying PP-PTSD, encompassing physiological and psychological factors (e.g., sleep quality and depression), environmental and occupational factors (e.g., social support, paternal involvement, and infant caregiving practices), and social behavioral factors. The Gradient Boosting prediction model demonstrated robust performance and high predictive accuracy upon independent external validation, highlighting its potential utility for risk stratification and future clinical translation. Nevertheless, multicentre validation and the development of clinically implementable tools are warranted. Collectively, this study offers a theoretical foundation and methodological framework for the early identification, targeted intervention, and long-term health management of PP-PTSD in primiparous women.
ADAR-based DNA adenine base editing with single-nucleotide precision
Nature Biotechnology, Published online: 15 July 2026; doi:10.1038/s41587-026-03228-8
Adenosine deaminases acting on RNA (ADARs) have not been previously exploited for DNA adenine base editing in mammals. We developed a single-nucleotide-resolution adenine base editor (snuABE) that uses ADAR and customized guide RNAs to precisely convert adenines to guanines on DNA targets with minimal bystander editing.
Multiscale in vivo imaging of tumor evolution using a germline conditional triple-reporter mouse
Nature Biotechnology, Published online: 15 July 2026; doi:10.1038/s41587-026-03184-3
Tumors are imaged across scales—from cells to whole body—with a triple-reporter mouse.
High numerical aperture confocal volumetric mesoscope reveals mesoscale subcellular dynamics in vivo
Nature Biotechnology, Published online: 15 July 2026; doi:10.1038/s41587-026-03231-z
Volumetric mesoscale imaging reveals subcellular immune dynamics in different mammalian organs.
High-fidelity fast fluorescence lifetime imaging by event-based denoising
Nature Biotechnology, Published online: 15 July 2026; doi:10.1038/s41587-026-03222-0
Event-based first-photon fluorescence lifetime imaging enables fluorescence lifetime imaging below one photon per pixel.
When caution becomes harm
Nature Medicine, Published online: 15 July 2026; doi:10.1038/s41591-026-04526-z
Our pregnant patients deserve accurate conversations about risk.
Fertilizer scarcity as a failure of global health governance
Nature Medicine, Published online: 15 July 2026; doi:10.1038/s41591-026-04507-2
Fertilizer scarcity as a failure of global health governance
Studying vanishing dietary diversity before it is lost: the World Diet Initiative
Nature Medicine, Published online: 15 July 2026; doi:10.1038/s41591-026-04520-5
Heritage diets are disappearing as food systems homogenize; their effects on immunity, metabolism and host–microbiome biology remain poorly understood. The World Diet Initiative offers a framework to document heritage diets and investigate their effects before they are lost.