Search Results (470)

Grain production resilience forms a critical foundation for national food security, and the ongoing development of land transfer provides essential momentum for establishing a more resilient grain production system. Using panel data from 30 provincial-level regions from 2013 to 2024, this study constructs a multi-dimensional evaluation system for grain production resilience and calculates the comprehensive grain production resilience index using the entropy value method. This study applies two-way fixed effects and mediation models to empirically analyze the impact of land transfer on grain production resilience and its underlying mechanisms. The results show the following: (1) Land transfer significantly enhances grain production resilience: a 1 percentage point increase in the land transfer rate leads to a 0.0014-point increase in the resilience index, equivalent to 0.64% of the sample mean, and this finding remains robust after model replacement, extreme value trimming, and variable substitution. (2) Land transfer exerts its positive effect through three mediating pathways: agricultural insurance (scale dimension), specialized farmer cooperation, and agricultural mechanization. (3) Heterogeneity analysis reveals significant regional differences: the enhancing effect is more pronounced in non-major grain-producing regions and areas with underdeveloped agricultural service systems; while in major grain-producing regions and high-service-level regions, the relationship presents an inverted U-shape, with turning points at 66.794% and 71.921% of the land transfer rate respectively. Accordingly, this study proposes that China should further improve the institutional design of land transfer to systematically support the development of grain production resilience, optimize relevant policy pathways, and implement region-specific measures for targeted and effective intervention. Full article

A modified Pitot-tube jet (PTJ) separation pump combines centrifugal phase separation with pressure buildup and enables compact oil–water treatment, where a water-rich stream can be discharged at elevated pressure. This work advances an existing laboratory PTJ configuration toward a turbomachinery-oriented rotor concept for systematic design studies and subsequent field-oriented prototypes. Starting from a centrifuge-like reference configuration without blades that prioritizes separation stability, an impeller with trimmed blades is introduced to increase pressure head while limiting blade interaction with the oil–water interface by operating primarily in the outer, water-rich annulus. Comparative experiments with and without the impeller show a pronounced increase in pressure head, up to about a factor of three at the maximum speed investigated. The results also indicate a purity penalty caused by blade-induced mixing and secondary flows. This exposes the central design trade-off of the PTJ machine. Higher specific work input increases pressure head but can reduce discharge quality. Hydraulic optimization, therefore, needs to be coupled to ppm-level purity constraints. Density-based monitoring lacks resolution in the relevant trace range, and chemical-based analyses are too slow for systematic investigations. An imaging-based fluorescence method using Nile Red as a selective tracer is, therefore, implemented as a rapid analysis tool. High-resolution imaging with automated region of interest evaluation provides a robust calibration from 5–500 ppm for safe, non-fluorescent model oils such as sunflower oil. This enables efficient operating-window mapping and comparative screening of rotor concepts under reproducible conditions. Full article

Background/Objectives: Migraine and Hashimoto’s thyroiditis (HT) are frequently comorbid, implying shared biological pathways. Selenium and myoinositol are involved in migraine pathophysiology, and their supplementation has been shown to improve thyroid function, particularly in individuals with HT. This study aimed to evaluate the impact of combined myoinositol and selenium (MYSE) supplementation on thyroid function and migraine outcomes in patients with migraine and HT. Methods: We conducted a retrospective study on a cohort of 163 adults with migraine comorbid with HT who received a 6-month MYSE supplementation. Thyroid parameters, namely thyrotropin (TSH), free thyroxine (fT4), and free triiodothyronine (fT3), and migraine features, namely monthly migraine days (MMDs), monthly migraine attacks (MMAs), and monthly symptomatic drug use (MSDs), were assessed at baseline and at follow-up. Because Shapiro–Wilk testing showed that all thyroid and migraine outcomes deviated significantly from normality, pre–post comparisons were evaluated with the Wilcoxon signed-rank test, between-group comparisons with the Mann–Whitney U test, and a three-tier non-parametric strategy (Aligned Rank Transform with ART-C contrasts, the Brunner–Langer non-parametric mixed model, and a trimmed-means between-within ANOVA) to analyze time × migraine × gender, adjusted for age and illness duration. Spearman rank correlations with percentile-bootstrap 95% confidence intervals were computed, and both robust MM-regression and rank-based Jaeckel regression were carried out. Another analysis stratified participants by baseline thyroid status: euthyroid vs. subclinical hypothyroidism (SCH). Results: After six months of MYSE supplementation, significant reductions were observed in TSH (median 3.60 → 2.80 mIU/L, Wilcoxon p < 0.001, rank-biserial r = −0.94), MMDs (14 → 11, p < 0.001, r = −0.99), and MSDs (14 → 11, p < 0.001, r = −0.99), while fT4 increased slightly (1.30 → 1.50 ng/dL, p < 0.001) and fT3 remained stable. For MMAs, a small effect was detected by the paired Wilcoxon test (p = 0.002) but the main effect of time did not survive adjustment in any of the three covariate-adjusted mixed models (ART p = 0.079; nparLD p = 0.55; WRS2 p = 0.084). Chronic migraine patients had higher baseline and follow-up headache burden but experienced greater reductions in MMDs. The percentage reduction in TSH was positively correlated with improvement in MMDs (Spearman ρ = 0.45, bootstrap 95% CI 0.31–0.57, p < 0.001) and was the only significant predictor in both robust MM-regression (β = 0.28, p < 0.001) and rank-based regression (β = 0.25, p < 0.001). The TSH–MMD association held within each thyroid-status stratum separately (ρ = 0.42 in euthyroid, ρ = 0.51 in SCH; p < 0.001 for both), indicating an individual-level signal rather than a between-group artefact. Conclusions: MYSE supplementation was associated with improved thyroid parameters and a meaningful reduction in migraine burden among patients with migraine and HT. The association between TSH reduction and headache improvement supports the hypothesis of an endocrine–metabolic contribution to migraine severity and warrants confirmation in prospective controlled trials. It also supports the clinical value of assessing and addressing thyroid function in this population. Full article

Background: Lung adenocarcinoma (LUAD) is the most common type of lung cancer and a major cause of cancer death. Zinc finger proteins (ZNFs) have been implicated in LUAD progression, functioning either as oncogenes or tumor suppressors. Therefore, an in-depth investigation of ZNFs may contribute to the development of novel diagnostic and therapeutic strategies for LUAD. Methods: Transcriptomic and clinical data were obtained from the TCGA and GEO databases. Prognosis-related ZNF genes were identified using univariate Cox, LASSO, and multivariate Cox regression analyses. An eight-gene ZNF-based prognostic signature was constructed and validated in two independent external cohorts (GSE50081 and GSE26939). A nomogram integrating independent prognostic factors was developed. Immune infiltration, somatic mutation profiles, and drug sensitivity were systematically analyzed. We further focused on FGD3, a key gene from the signature, examining its expression in LUAD cells and tissues, including lorlatinib-resistant models. Results: The prognostic signature comprising TRIM6, TRIM29, CTCFL, FGD3, GATA4, CASZ1, TRAF2, and ZNF322 effectively stratified patients into distinct risk groups with significantly different overall survival (p < 0.05). The risk score, together with T and N stage, served as independent prognostic predictors (n = 500, p < 0.05). High-risk patients exhibited an immune-desert phenotype, increased tumor mutational burden, and distinct drug sensitivity patterns. Notably, FGD3 expression was downregulated in LUAD tissues (n = 14, p < 0.0001) and lorlatinib-resistant cells, and its restoration suppressed resistant cell proliferation and partially reversed drug resistance. Conclusions: This study establishes a promising ZNF-based prognostic model for LUAD, providing a potential tool for risk stratification and individualized therapeutic decision-making. The identification of FGD3 as a potential mediator of drug resistance highlights its promise as a candidate biomarker and therapeutic target in LUAD. Full article

This study investigates the hydrodynamic performance of the KVLCC2 tanker in deep and shallow water using computational fluid dynamics (CFD) simulations, focusing on resistance and self-propulsion with both ducted and non-ducted propellers. The Reynolds-averaged Navier–Stokes (RANS) equations, coupled with the SST k-ω turbulence model, are solved using STAR-CCM+ to analyze ship resistance, open-water propeller characteristics, and self-propulsion factors. Validation against experimental data confirms the numerical accuracy, with uncertainties below acceptable thresholds. In deep water, the body force propeller and body force ducted propeller methods are validated against the discretized propeller approach, yielding errors under 5%. The ducted propeller enhances open-water efficiency but results in higher thrust deduction and lower wake fractions, leading to reduced hull and overall propulsive efficiencies compared to the non-ducted case. In shallow water, as the depth-to-draft ratio (H/T) decreases to 1.5, added resistance, sinkage, and trim increase sharply due to blockage effects. The ducted configuration mitigates these penalties, achieving a 20.8% power reduction at H/T = 1.5. Added self-propulsion factors reveal that the duct improves hull efficiency and offsets shallow-water losses, enhancing propulsive efficiency. Flow field analysis shows accelerated stern wakes and asymmetric structures in shallow water, with the body force methods providing consistent predictions despite minor discrepancies in extreme conditions. This research highlights the efficacy of ducted propellers in shallow water and the reliability of body force methods for efficient simulations, offering insights for ship design in restricted depths. Full article

This work is devoted to the investigation of an actual scientific and applied problem in the identification of exit points for early-exit DNNs based on the analysis of neural chains, which is one of the complex tasks related to the scientific and applied problems of DNN optimization, including, in particular, those based on the existing early-exit concept. The obtained computational complexity of the developed method is not limited by the latter itself, but instead, it mainly depends on the chosen algorithm for analyzing the occurrences of particular substrings (i.e., trimmed neural chains) into a defined list of strings (i.e., full neural chains). For example, in the framework of the conducted research, the Python operator “in” has been used (for this purpose), which uses an in-built optimized algorithm based on the combination of the Boyer–Moore and Horspool algorithms with a linear scalability, and computational complexity that approaches the arithmetic product of the total number of strings (i.e., full neural chains) in the array by the average length of the string in the same array. The performed practical approbation of the developed method gave positive results in decreasing the overall time for obtaining the final result of the considered DNN, as well as significantly decreasing the following timing parameters of the considered DNN: the minimal time to obtain the final result (reduced by more than 5 times); the average time to obtain the final result (reduced by ~1.4 times); and the total time spent processing all 22,500 modeling cases in total (reduced by ~1.39 times). In terms of the main positive aspects and advantages of the developed method, we could highlight its maximal versatility (in terms of the studied DNNs, their architectural and/or structural features, application areas, and input data representation, as well as further software implementation of the proposed method), together with its maximal simplicity of representation and understanding, which ensures the possibility of working with this method even for novice and inexperienced researchers and users who have only basic knowledge of DNNs. In addition, the main results and conclusions of the conducted research are given, and the prospects for further research are considered. Full article

To evaluate the techno-economic performance of supercritical carbon dioxide (sCO2) power cycles in particle-based solar tower systems, thermodynamic and techno-economic models were established for four configurations: RC-ORC, RC, RE-ORC, and RE. A one-dimensional design method was used for key printed circuit heat exchangers, and multiple cost correlations with a trimmed-mean treatment were adopted to reduce the influence of extreme cost estimates. The results show that the primary heat exchanger (PHX) dominates system investment, accounting for more than 50% of total cost in all configurations. After screening 48 pure ORC working fluids, Cyclopropane and Trans-butene were identified as the economically preferable fluids for RC-ORC and RE-ORC, respectively. ORC working-fluid selection should therefore consider not only net power output, but also the effect of heat transfer and flow characteristics on intermediate heat exchanger cost. Scale analysis shows that the specific investment cost decreases rapidly over 50–300 MW, while the reduction becomes much smaller above 300 MW. At large scales, RC-ORC and RE-ORC gradually approach 1756.64 $/kW. These results highlight the importance of PHX cost reduction, heat-exchanger-oriented ORC fluid selection, and appropriate system scaling. Full article

Background/Objectives: Evidence linking vitamins D and B₁₂ to psychiatric outcomes remains heterogeneous across designs, populations, phenotypes, exposures, and outcome formats. Methods: We conducted a PRISMA 2020 systematic review and exploratory meta-analysis of nutrient-specific status and supplementation evidence. PubMed/MEDLINE, APA PsycInfo, Cochrane Library, Google Scholar, ClinicalTrials.gov, and ProQuest were searched for human studies published in 2016–2025, with a final update on 1 March 2026. Forty-six studies were included (24 randomized controlled trials, 22 observational studies; N = 69,902), and 44 contributed quantitative data. Effects were harmonized to odds ratios (ORs) for cross-family comparability and pooled using Hartung–Knapp random-effects models; supplementation evidence was additionally interpreted on the standardized mean difference (SMD) scale. Results: Across the main evidence families, pooled estimates showed substantial heterogeneity and limited generalizability. Vitamin D supplementation showed an initial inverse estimate on the secondary harmonized OR scale (OR = 0.439, 95% CI 0.272–0.710) and a clinically interpretable SMD of −0.454 (95% CI −0.718 to −0.189), but heterogeneity was high (I² = 84.2%) and trim-and-fill attenuated the OR estimate to the null (OR = 0.88, 95% CI 0.48–1.63). Vitamin D status showed a similar pattern (primary OR = 0.615, 95% CI 0.424–0.890; trim-and-fill OR = 0.90, 95% CI 0.54–1.49). Vitamin B₁₂ status was inversely associated with outcomes (OR = 0.310, 95% CI 0.115–0.834), but heterogeneity was extreme (I² = 94.8%). B₁₂ supplementation evidence was sparse and null. Conclusions: The evidence supports targeted deficiency assessment, not routine supplementation. Full article

Background: Preoperative CT is part of the routine diagnostic work-up for colorectal cancer (CRC), and CT-based biomarkers have been linked to oncological and surgical outcomes in CRC patients. This review aims to evaluate the association between preoperative CT-derived vascular disease markers (calcification and stenosis) and postoperative outcomes after curative CRC resection with anastomosis. Methods: Following PRISMA, we conducted a systematic review and meta-analysis of studies identified from MEDLINE/PubMed, Embase, Web of Science Core Collection and Google Scholar from inception until 1st of January 2026, with the protocol registered in PROSPERO (CRD420251248044). Eligible studies examined CT-derived abdominal vascular disease markers in CRC patients treated with curative resection and anastomosis and reported postoperative outcomes (anastomotic leakage (AL) grade C, major morbidity, and mortality). Risk of bias was assessed using the Newcastle–Ottawa Scale. We pooled odds ratios using random-effects models when ≥3 studies reported comparable exposure–outcome comparisons. Results: Fourteen studies (6712 participants) were included, and 12 contributed to quantitative synthesis. Higher calcification burden was associated with increased odds of any AL (OR 3.08, 95% CI 2.09–4.54; k = 11; n = 5005) and severe/grade C AL (OR 2.68, 95% CI 1.03–6.97; k = 3; n = 3418). Evidence for major morbidity was imprecise (OR 1.99, 95% CI 0.86–4.59; k = 3; n = 841), and data for mesenteric stenosis outcomes and mortality were limited. Sensitivity analyses indicate attenuation without loss of significance after trim-and-fill (adjusted OR 2.39) and that non-ROC cut points yield a smaller effect size (OR 2.42). Conclusions: CT-derived vascular disease markers are associated with higher odds of AL after CRC surgery. Prospective studies should standardize methods and test clinical utility. Full article

Background: The accuracy of tumor marker testing is critical for clinical decision-making. Patient-based real-time quality control (PBRTQC), as a complementary approach to traditional internal quality control (IQC), has been widely adopted in clinical laboratories. With the rapid advancement of automation and artificial intelligence (AI) in recent years, a large number of AI-based PBRTQC optimization algorithms have emerged. This study compared Patient-based real-time quality control integrating neural networks and joint probability analysis (NN-PBRTQC), Patient-Based Pre-Classified Real-Time Quality Control (PCRTQC), and traditional PBRTQC to identify the optimal method for quality control of tumor marker testing. Methods: The study utilized clinical tumor marker testing data from Peking University First Hospital. Six common tumor markers were selected, and constant error (CE) and proportional error (PE) were introduced as measures of analytical error. The False Alarm Rate (FAR) was used to reflect the specificity of the algorithms, while the Trimmed Average Number of Patient Results Affected Before Error Detection (tANPed) was used to reflect their sensitivity, in order to compare the clinical performance of the different models. Results: Under the same desired FAR (DFAR) of 0.1%, NN-PBRTQC reduced the tANPed for the six tumor markers by an average of 62% compared to the traditional PBRTQC while maintaining the same FAR, which demonstrated superior sensitivity of error detection. Meanwhile, although PCRTQC strictly controlled the FAR, its tANPed was 23% higher on average than that of the traditional PBRTQC, which indicated insufficient sensitivity of error detection. Conclusions: NN-PBRTQC demonstrated superior comprehensive quality control performance in the comparison of six common tumor markers. While ensuring that the FAR does not deviate from the DFAR, it significantly reduces tANPed, such that it could meet the specificity and sensitivity requirements of clinical testing. It is expected to enable more efficient and accurate detection of tumor marker errors. Full article

In addition to inducing pathogen-specific adaptive immune responses, vaccines can train the innate immune system, thereby providing broader host protection. This concept of trained immunity (TRIM) is well-established in benchtop laboratory science. This review aims to evaluate the current evidence of vaccine-induced TRIM and translate these findings into a clinical context. Various laboratory methods are used to assess TRIM; however, inconsistent results have been reported across non-BCG vaccine studies. Existing analyses lack exploration of the mechanistic basis of vaccine-induced TRIM, particularly epigenetic reprogramming and metabolic rewiring. Patterns emerge between vaccines: live-attenuated vaccines generally induce TRIM, as evidenced by increased inflammatory cytokine production upon restimulation, whereas non-live vaccines tend to demonstrate reduced trained immunity. Such findings are not consistently observed for mRNA vaccines, which show heterogeneous patterns. The limited variety of studies on non-BCG vaccines impacts the reliability of findings. A more comprehensive understanding of the mechanisms and outputs of TRIM induced by specific vaccines could better inform rational vaccine design. Furthermore, various modifiers can alter vaccine-induced TRIM responses, including sequence and route of administration, sex, and age. Consideration of these modifiers has important clinical implications in optimising vaccine administration for enhanced immune protection. Full article

Clinical prediction from electronic health records (EHRs) is complicated by pervasive missingness and label scarcity, which make performance sensitive to the match between data conditions and pipeline choice. Choosing the best pipeline for a new incomplete dataset still requires costly trial-and-error. We cast this as an algorithm selection problem and address two bottlenecks—instance scarcity and distance quality—that have so far prevented meta-learning from reaching clinical settings. Graph neural networks offer diverse strategies (patient similarity networks, bipartite imputation graphs, attention-driven feature interaction), yet no single architecture dominates across missingness patterns, and selecting the best pipeline for a new dataset remains a trial-and-error approach. Formal algorithm selection could automate this choice but requires many characterized meta-instances—more than clinical settings typically provide. We propose two solutions: (1) constructive instance augmentation, applying controlled quality perturbations (MCAR and MNAR missingness injection, label trimming) to 20 base EHR datasets to expand the meta-knowledge base to 83 characterized meta-instances, each described by a 10-dimensional missingness fingerprint, without additional model training; and (2) dynamic-supervised metric learning, using differential evolution to optimize fingerprint feature weights so that static distances preserve method-performance similarity captured by dynamic fingerprints, which require model sweeps and are unavailable at deployment. Under base-dataset-level leave-one-dataset-out cross-validation over 21 pipelines, the resulting metric-learned kNN recommender attains the highest win rate (20.5%) among non-oracle strategies on the augmented store, selecting the correct pipeline more often than any fixed default. At deployment, the recommender needs only the 10-dimensional static fingerprint with pre-learned weights; no sweep data is required for new datasets. Cross-domain evaluation on 25 external subsets (colorectal cancer, kidney disease, MIMIC-IV) demonstrates framework modularity: when the fingerprint module is adapted (standard meta-features in place of the missingness-specific set), the recommender achieves regret of 0.025 (55% below random selection). Full article

Background/Objectives: Pterygoid implant placement represents a valuable alternative to conventional bone grafting procedures in the rehabilitation of the atrophic posterior maxilla; however, the procedure remains technically demanding because of limited visibility, difficult access, complex pterygomaxillary anatomy, and the need for precise angulation and distal bicortical anchorage. Although digital guidance has increasingly been applied in implant dentistry, a clearly described workflow integrating automatic segmentation, selective virtual trimming of the posterior maxillary anatomy, and direct three-dimensional planning for static-guided pterygoid implant placement remains insufficiently detailed in the literature. The aim of this report was to describe and illustrate such a workflow in a proof-of-concept clinical application. Methods: This work was designed as a methodological proof-of-concept with a single clinical illustration. A CBCT dataset was imported into BlueSkyPlan, where automatic segmentation was used to generate three-dimensional models of the maxilla, teeth, and pterygoid process. The segmented volumes were then selectively trimmed to expose the relevant pterygomaxillary anatomy and to support direct three-dimensional planning of the implant axis in the rendered model. A static surgical guide with combined tooth and mucosal support was subsequently designed, positioned on a printed jaw model derived from the intraoral scan, and assessed by CBCT-based internal verification. Results: In this proof-of-concept application, the workflow enabled three-dimensional visualization of the pterygomaxillary trajectory, supported implant axis planning in the rendered model, and facilitated guide design and radiographic verification of the planned trajectory. The verification step provided an internal methodological consistency check between the planned implant axis and the drill-guided direction visible on CBCT. Conclusions: The present report describes a segmentation-assisted three-dimensional planning workflow for static-guided pterygoid implant placement in a single proof-of-concept clinical application. The workflow should be interpreted as a methodological illustration rather than a quantitative validation study. Further investigations are required to evaluate accuracy, inter-operator reproducibility, and broader clinical applicability. Full article

Objectives: This study aims to investigate the potential molecular mechanisms by which α-hederin modulates HSC activation to alleviate liver fibrosis. Methods: An in vitro model of liver fibrosis was established by inducing LX-2 cells with TGF-β1. These cells were then treated with α-hederin (10 μg/mL) before undergoing phenotypic analysis and molecular-level detection. A mouse model of liver fibrosis induced by CCl₄ was established in vivo to further evaluate the expression levels of fibrosis markers, including TRIM38. Results: In TGF-β1-induced liver fibrosis in LX-2 cells, α-hederin treatment significantly inhibited HSCs activation, as evidenced by down-regulation of α-SMA and suppressed proliferation capacity. At the same time, α-hederin significantly reduced the levels of COL1A1, COL3A1, fibronectin, and MMP-2. Transcriptome sequencing analysis revealed that α-hederin treatment significantly upregulated TRIM38 expression. Differentially expressed genes (DEGs) were significantly enriched in endoplasmic reticulum stress-related pathways. TRIM38 up-regulation inhibits HSC activation and proliferation, reducing the expression of ERS marker proteins (GRP78, p-PERK, and CHOP); Co-IP experiments further confirmed that TRIM38 and GRP78 interact directly. Further rescue experiments demonstrated that TRIM38 knockdown significantly attenuated the inhibitory effects of α-hederin on these processes. In a CCl₄-induced mouse model of liver fibrosis, α-hederin (4 mg/kg) significantly reduced the liver index and serum ALT and AST levels, improved histopathological damage to the liver, upregulated TRIM38 expression in liver tissue, and inhibited the endoplasmic reticulum stress response (ERS). Conclusions: α-hederin exerts its anti-fibrotic effect by upregulating TRIM38, thereby alleviating endoplasmic reticulum stress and ultimately inhibiting the activation and proliferation of HSCs. Full article

Background: The global burden of mental disorders continues to escalate, necessitating scalable, evidence-based interventions. Artificial intelligence (AI)-delivered health promotion programs represent a promising approach to addressing treatment gaps by targeting the neuropsychological mechanisms that underlie mental health outcomes. This meta-analysis synthesizes evidence on the effectiveness of AI-delivered interventions in improving executive function, emotion regulation, and clinical outcomes across diverse populations. Methods: A systematic search identified 186 studies (n = 22,755 participants) published between 2020 and 2025. Random-effects meta-analyses estimated pooled effect sizes (Hedges’ g, calculated as between-group standardized mean differences with small-sample correction [J = 1 − 3/(4df − 1)]) for primary outcomes. Between-study heterogeneity was quantified using I² and τ² statistics. To address dependency among effect sizes from studies reporting multiple outcomes, robust variance estimation (RVE) was employed. Subgroup analyses examined intervention modalities, delivery formats, and clinical populations. Moderator analyses explored sources of heterogeneity, including publication year, sample size, intervention duration, control condition type, risk-of-bias rating, geographic region, and AI sophistication tier, and mediational models tested putative therapeutic mechanisms. Results: AI-delivered interventions demonstrated a significant overall effect on health outcomes (g = 0.68, 95% CI [0.58, 0.78]; τ² = 0.12; I² = 73.4%). Executive function outcomes showed moderate effects (g = 0.61, τ² = 0.08), with working memory improvements being strongest (g = 0.72). Emotion regulation outcomes demonstrated moderate-to-large effects (g = 0.61, 95% CI [0.51, 0.70], τ² = 0.006); formal subgroup pooled estimates by emotion regulation strategy were not calculated due to insufficient studies per strategy (k < 3 per category); individual study effect sizes ranged from g = 0.27 to g = 1.11. Among 41 studies examining neuropsychological mechanisms, convergent patterns suggested involvement of prefrontal neural circuits (DLPFC), enhanced alpha-band activity, and improved heart rate variability; however, formal mediation was tested in only 18 studies (9.7%). Among clinical populations, interventions for cognitive impairment yielded the largest effects (g = 1.02; this finding should be interpreted cautiously given modest cumulative sample size [n = 482], potential small-study effects [Egger’s p = 0.08], and trim-and-fill adjusted estimate of g = 0.85), followed by mental health conditions (g = 0.72), while other clinical populations showed smaller but significant improvements (g = 0.19). Mobile applications (g = 0.78) and chatbot-based interventions (g = 0.74) demonstrated the strongest effects among delivery formats. Among studies testing formal mediation, analyses suggested mindfulness (β = 0.42), decentering (β = 0.38), and cognitive reappraisal (β = 0.45) as processes associated with therapeutic outcomes. Conclusions: AI-delivered health promotion programs demonstrate significant effectiveness across executive function, emotion regulation, and clinical outcomes, though substantial heterogeneity (I² = 45–82%) indicates meaningful variability warranting attention to subgroup-specific effects. Given the diversity of intervention types included (chatbots, mobile apps, VR systems, neuromodulation), pooled estimates should be interpreted as characterizing the average effect across this heterogeneous landscape; subgroup-specific estimates provide more precise guidance for clinical decision-making regarding specific modalities. Effects are associated with convergent patterns of neuropsychological mechanisms, though mechanistic conclusions remain preliminary given that only 22% of studies (41/186) examined neuropsychological mechanisms, and formal mediation analyses were conducted in only 18 studies (9.7%); most of the mechanistic evidence is correlational rather than causal. Future research should establish standardized AI taxonomies, optimize adaptive algorithms, conduct adequately powered replication studies in populations with cognitive impairment, prioritize experimental mediation designs to establish causal pathways, and evaluate long-term maintenance effects with a minimum of 6–12-month follow-up periods. Full article