Trustworthy process monitoring based on canonical correlation analysis considering epistemic uncertainty

Shared and sex-differentiated brain-behavior associations at rest: A canonical correlation analysis of fMRI and resting-state cognition.

Accurate preoperative identification of true positive white matter pathways involved in critical eloquent functions such as motor, language, and vision plays a vital role in minimizing the risk of postoperative functional deficits and improving postoperative functional outcomes in pediatric epilepsy surgery. This study proposes a novel deep learning model: "ESM-AnatTractNet" that can accurately classify true positive eloquent white matter pathways across preoperative diffusion weighted imaging tractography data of 85 drug-resistant epilepsy patients (age: 10.70 ± 4.41 years). To enhance geometric and anatomical consistency of true positive tract classification, the ESM-AnatTractNet integrated two features in a point-cloud-based framework, 1) electro-physiologically confirmed spatial coordinates using electrical stimulation mapping (ESM) and 2) anatomically-contexted labels of the end-to-end neural connection using a standard brain atlas. Its overall performance was validated by accurately classifying 14 eloquent functional areas in whole brain, objectively optimizing resection margins to preserve eloquent functions using Kalman filter, and precisely predicting postoperative language outcomes using canonical correlation. Our ESM-AnatTractNet outperformed other baseline models, achieving an accuracy of 97% in correctly classifying eloquent areas within 10mm spatial resolution of clinical subdural grid electroencephalography. The Kalman filter analysis achieved 94% accuracy in predicting no deficits when the ESM-AnatTractNet-defined preservation zones were not resected. Postoperative decrease in language-related white matter connection efficacy defined by the ESM-AnatTractNet analysis was significantly associated with worse postoperative language outcome (R=0.73, p < 0.001). Our findings demonstrate that the ESM-AnatTractNet improves non-invasive localization of true positive eloquent white matter pathways, supporting its potential to enhance current preoperative evaluation of pediatric epilepsy surgery.

Streamlined foreign finances accelerate sustainable transformative economies and reducing ecological footprints. Sub-Saharan African (SSA) countries are reeling from heavy debt coupled with absence of eco-investments and foreign aid overdependence. Limited transition towards low-carbon future risked dignified life, quality health, water and zero poverty, among sustainable development goals.This paper investigated the role of foreign aid and remittance inflows on ecological footprints and pollution levels from a census of SSA from 1990 to 2023. Regressions involved canonical correlations and GMM estimations. Persistent lagged values of EFPRD and AQI over current pollutions resulted from past regulatory challenges, clean energy costs and limited green investments. Increasing ihs_REM_IGDP and ODA_NPCP caused high ecological footprints and air pollutions while ODA_NGDP caused a reduction with negligible effect due to their statistical insignificance. Moreover, Instrument variables from lagged values were valid given Arellano-Bond tests at AR (1) were significant at 0.05, AR (2) were insignificant, while instrument counts and Hansen tests were valid. We recommend policymakers to develop blended sustainable financial models to spur economic resilience through private sector sustainable entrepreneurships while green finance social bonds earmarked for diaspora remiitances should be facilitated for attainment of SDG’s through collaborative funding of healthcare systems and resilient community empowerment projects.

The performance and operational lifespan of aircraft engines are critical factors influencing flight safety and economic viability. Traditional methods for predicting aircraft engine service life often overlook microstructural changes, resulting in inaccurate predictions. To address this, a deep learning approach combining the Bidirectional Gated Recurrent Unit (BiGRU) model and self-Attention Mechanism (AM) for more accurate and reliable predictions is proposed. The method constructs a Health Index (HI) curve using a stacked denoising autoencoder and Kernel Canonical Correlation Analysis (KCCA), integrating self-AM for improved pattern recognition. The results were compared with the Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) methods, showing superior efficiency and accuracy. The integrated BiGRU method demonstrates rapid fitness reduction, superior computational speed, and consistent prediction accuracy across various prediction horizons. The overall findings suggest that this method holds significant promise for enhancing the prediction accuracy of aircraft engine combustion chambers' Remaining Useful Life (RUL), with potential applications in aviation engine maintenance and control protocols. Further optimization and testing under extreme conditions are warranted for future research.

Understanding how aging reshapes balance is essential for targeted assessment and rehabilitation. Because wobble boards impose continuous instability, their kinematics may provide a noninvasive readout of long-latency (transcortical) reflex function in ecologically valid conditions. We analyzed wobble-board dynamics in healthy younger adults ( y, ) and community-dwelling older adults ( y, ) performing with and without a concurrent cognitive load (Trail Making Task). Spatial behavior was indexed by peak angular excursion magnitudes and their RMS; temporal behavior by cycle durations. For each series, we estimated Hurst exponents ( ) using detrended moving-average analysis. Across ages and task conditions, cycle-duration series clustered near , indicating emergent timing, and peak-angular-excursion series exhibited persistence ( ). Age did not affect for either measure, nor the number of cycles per trial. By contrast, older adults showed substantially larger peak angular excursions and higher root mean square (RMS) peak angular excursions than younger adults. Multivariate analyses (principal components and canonical correlations) revealed that wobble-board metrics capture variance not explained by center-of-pressure and center-of-mass measures, yielding distinct signatures differentiating individuals, conditions, and-via amplitude metrics-age groups. These results indicate that wobble-board dynamics in aging primarily differ in amplitude rather than in temporal organization or cadence of corrections. This pattern suggests that the timing of corrective responses remains intact with aging, while their effectiveness is reduced. These findings support wobble-board dynamics as a promising paradigm for assessing age-related changes in balance control. Direct validation using electromyography (EMG) and controlled perturbations is warranted to confirm that wobble-board metrics reflect long-latency reflex function.

The retina provides a unique window into systemic health, yet molecular mechanisms linking retinal features (oculomics) to clinical traits in aging remain unclear. In this study, we leveraged the homogeneous Canton 70 s Alumni Cohort (N = 258 females aged ~70 years) to minimize socio-demographic confounders and extracted oculomic features from fundus images using AutoMorph. Linear mixed-effects models identified 129 significant associations between oculomic and clinical features (p < 0.05). Sparse canonical correlation analysis indicated a key phenotypic retina-body axis (r = 0.538, p = 0.047) primarily driven by central retinal venular equivalent (Hubbard, zone b) and mean corpuscular hemoglobin concentration. Permutational Multivariate Analysis of Variance revealed that oculomic categories were significantly associated with systemic conditions like chest pain, dyslipidemia, and stroke. Age differentially impacted retinal features across clinical condition status (adjusted p for interaction < 0.2), with pronounced trends in individuals with health problems but relative stability in healthy controls. Plasma proteomics was integrated to explore potential molecular mechanisms. Weighted gene co-expression network analysis identified shared proteomic modules associated with both oculomic and clinical features. These modules were enriched in pathways including complement and coagulation cascades, cholesterol metabolism, and cytokine-cytokine receptor interaction. This study establishes both phenotypic and molecular connections underlying the retina-body axis in a female aging cohort.

Integrating high-dimensional multi-omics data is essential for uncovering the coordinated molecular mechanisms underlying cancer progression and improving survival prediction. DNA methylation and mRNA expression represent two tightly coupled regulatory layers; however, many existing approaches either model them independently or rely on linear assumptions that fail to capture the nonlinear cross-omics structure. Here, we propose MT-ADSCCA, a multitask adaptive deep sparse canonical correlation analysis framework that jointly learns correlated latent representations, selects interpretable multi-omics biomarkers, and supports downstream survival modeling. MT-ADSCCA embeds sparse CCA into a nonlinear encoder architecture and uses uncertainty-guided adaptive weighting to stabilize multi-objective training. The selected features were subsequently modeled using a BiLSTM-Cox survival network with genes ordered by chromosomal coordinates to capture local genomic dependencies. We evaluated MT-ADSCCA using event-stratified nested 10-fold cross-validation across three TCGA cohorts: breast invasive carcinoma (BRCA), glioma (GBMLGG), and pan-kidney carcinoma (KIPAN), including 485, 563, and 652 matched multi-omics samples, respectively. MT-ADSCCA achieved the highest concordance indices across all cohorts, outperforming six feature-selection baselines (DA, WGCNA, lmQCM, CCA, OSCCA, DeepCorrSurv) and four survival-model baselines (LASSO-Cox, RSF, MTLSA, DeepSurv). Kaplan-Meier analyses further confirmed a clear separation between the predicted high- and low-risk groups. The selected canonical features were enriched in biologically coherent functional categories, supporting the interpretability of the learned patterns. Together, these results demonstrate that MT-ADSCCA provides a robust and interpretable framework for multi-omics integration and cancer prognosis prediction.

Brain functions involve processing in local networks as well as modulation from brainwide signals, such as arousal. Dissecting the contributions of populations of neurons to these functions requires knowledge of interactions between brain areas. We investigated these interactions using dual hemisphere recordings of prefrontal cortex in monkeys performing a spatial memory task. To tease apart global processing from local interactions, we applied a novel statistical approach called pCCA-FA (a combination of probabilistic canonical correlation analysis and factor analysis) to analyze trial-to-trial variability in neuronal activity. We found substantial shared variability among neurons within each population, much of which was actually shared across populations and linked to an arousal process. Our work presents a path by which we can leverage multi-area recordings to reveal aspects of brain functions that are hidden in single-area recordings.

Triple-negative breast cancer (TNBC) is an aggressive malignancy lacking effective targeted therapies, underscoring the need for robust and interpretable biomarkers for personalized treatment. Here, we propose a graph attention network (GAT)-based multimodal framework that integrates scRNA-seq, scATAC-seq, and radiomics to capture cross-modal regulatory interactions underlying TNBC heterogeneity. Transcriptional, chromatin accessibility, and imaging features are aligned via canonical correlation analysis, with intercellular communication-derived gene relationships and transcription factor-binding-guided edges incorporated into a multimodal graph. Multi-head attention enables adaptive weighting of omics-specific interactions, while an ensemble multilayer perceptron with variational dropout stratifies patient prognosis. The model demonstrates strong predictive performance in an external TCGA-TNBC cohort (log-rank p<0.01), outperforming single-omics and alternative graph-based approaches (AUC-ROC = 0.839, 95% CI: 0.81-0.87). Pathway analysis validates canonical TNBC drivers, including PI3K, ERBB2, PTK6, and EGFR signaling, while revealing previously underappreciated regulatory programs involving complement-coagulation cascades, ECM-integrin-focal adhesion signaling, leukocyte transendothelial migration, sphingolipid-mediated metabolic-immune coupling, and nanoparticle-receptor interactions. Collectively, this framework provides an interpretable strategy for multimodal biomarker discovery in TNBC, uncovering both established and novel therapeutic vulnerabilities and offering a scalable approach toward precision oncology.

Objective: To investigate the integrated relationship between Cerebral Small Vessel Disease (CSVD) markers and quantitative facial soft-tissue measurements in Alzheimer's disease (AD) continuum, utilizing peripheral muscle health as a potential biomarker for systemic frailty and neurodegeneration. Methods: Retrospective analysis of 3T brain MRI data from 67 patients (AD, N = 45; Mild Cognitive Impairment [MCI], N = 22). CSVD markers were assessed using STRIVE and standardized scales (Fazekas, Potter). Facial soft-tissue metrics, including masseter and tongue volume, temporal muscle thickness (TMT), and fat infiltration (Mercuri Scale), were quantified via semi-automatic segmentation on T1-weighted sequences. Group comparisons (AD vs. MCI) used regression models adjusted for age and sex. The overall central-peripheral relationship was explored via Canonical Correlation Analysis (CCA). Results: The AD group showed a highly significant cognitive decline (MMSE: 23.2 ± 4.1 vs. 28.2 ± 1.4, p < 0.0001). Centrally, the presence of PVSs in the mesencephalic region was the most robust predictor for AD (p = 0.003). Peripherally, average masseter muscle volume was significantly lower in the AD group (p = 0.0273), and masseter fat infiltration was significantly higher (p = 0.025), supporting localized sarcopenia. The CCA demonstrated a statistically significant positive multivariate relationship (r = 0.51, Roy's Largest Root p = 0.015) between a higher combined CSVD burden and a worse soft tissue profile across the cohort. Conclusions: Quantitative indices of facial soft tissues, particularly masseter muscle volume and quality, reflect systemic frailty and cognitive deterioration along the AD continuum. The strong central-peripheral correlation suggests that sarcopenia and CSVD are interconnected manifestations of a shared pathobiological process. These easily measurable facial markers could serve as valuable, non-invasive peripheral biomarkers, complementing traditional neuroimaging risk stratification in AD.

A brain-computer interface (BCI) enables direct communication between the brain and external devices by translating neural activity into executable control commands. Among electroencephalography (EEG)-based paradigms, steady-state visual evoked potential (SSVEP) is widely adopted due to its high signal-to-noise ratio, robustness, and minimal calibration requirements. While SSVEP-based spellers have been extensively investigated, many existing systems rely on high-channel-density EEG recordings and computationally complex processing pipelines, and are primarily designed for alphabetic input structures. In this study, we present an SSVEP-based Korean speller that integrates the Cheonjiin keyboard layout to support intuitive composition of Hangul syllables. The proposed system adopts a simple configuration, employing only five visual stimulation frequencies (6.67-12 Hz) and two occipital EEG channels (O1 and O2), with real-time frequency recognition performed using canonical correlation analysis (CCA) within a 1.5 s sliding window. EEG signals were acquired at 200 Hz using an OpenBCI Ganglion board, band-pass filtered (5-45 Hz), and processed with harmonic sinusoidal reference templates for multi-frequency classification. The proposed interface generates five control commands (up, down, left, right, and select), enabling directional cursor navigation and character confirmation on a 4 × 4 virtual Cheonjiin keyboard. Experimental validation with three healthy participants demonstrated an average classification accuracy of approximately 82% and an information transfer rate (ITR) of 31.2 bits/min. Frequency-domain analysis revealed clear spectral peaks at the stimulation frequencies and their harmonics, indicating reliable SSVEP responses. The proposed system employs a simple two-channel configuration integrated with a Korean language-specific input structure, demonstrating that reliable SSVEP-based communication can be realized without computationally intensive algorithms or high-cost EEG acquisition systems. These findings demonstrate that reliable SSVEP-based communication can be achieved using a low-channel configuration without reliance on high-cost EEG equipment.

Amyloid-β (Aβ) accumulation is a continuous process central to pathological aging that begins decades before cognitive impairment emerges. While subthreshold Aβ levels have been linked to future decline in cognitive control, the neural mechanisms connecting this early accumulation to its neurocognitive impact are poorly understood. Brain circuit dynamics, which are essential for cognitive function, may offer a sensitive lens into these initial pathological changes. Here, we tested whether brain state dynamics could serve as sensitive markers for cognitive impairment at an early stage of Aβ burden. Using the Bayesian Switching Dynamic System (BSDS) model, we identified 4 distinct latent brain states from high-temporal-resolution (800 ms) fMRI data acquired from 116 older adults, including 72 cognitively normal (CN) individuals and 44 with mild cognitive impairment (MCI), during an N-back working-memory task. Adopting a dimensional approach, we examined how latent brain state dynamics relate to early amyloid burden, cognitive performance, and clinical symptoms. While Aβ levels failed to differentiate clinical groups or predict clinical symptoms and task performance, the dynamics of latent brain states proved highly sensitive to both early Aβ accumulation and cognition. Canonical correlation analysis revealed a significant relationship between brain state dynamics and early Aβ burden. Furthermore, the temporal properties of brain states were significantly predictive of working memory performance in CN individuals, a relationship that was selectively disrupted in the MCI group. The features of brain dynamics can also successfully predict cognitive impairment. Our findings establish brain state dynamics as sensitive neural markers of initial Aβ accumulation and early cognitive impairment, offering a new framework for developing predictive models to identify individuals at risk for future cognitive decline.

Trace element with cytokine interactions reveal a magnesium interaction with inflammation-insulin-like growth factor 1 axis in neural tube defects.

Opt Deep CSSAN: Optimized Deep Convolutional Spectral-Spatial Attention Network for hyperspectral image classification.

Machine Learning and Artificial Intelligence in Nutrition Research: Analytical Methods, Applications, and Key Considerations.

Shell gland RNA-seq reveals key genes regulating eggshell quality and potential links between eggshell quality and hatchability across different laying stages.

Accurate numerical cognition relies on representing and comparing quantities, a fundamental skill for adaptive behavior. In nonhuman primates, the parieto-frontal network, including the ventral intraparietal area (VIP) and prefrontal cortex (PFC), supports this process, but the functional dynamics and directionality of communication between these regions remain unclear. We examine population-level interactions of simultaneously recorded neuronal activity between VIP and PFC in two male macaques performing a numerosity task. Using time-lagged canonical correlation analyses, we show that correct trials exhibit sustained correlations driven by numerosity-selective neurons, with early feedforward dominance from VIP to PFC following sample onset. By contrast, error trials show weaker correlations, reduced VIP-to-PFC feedforward signaling, and transient breakdowns during the late working memory period. These findings show that coordinated parieto-frontal population activity enables accurate numerical judgments, whereas disrupted interactions impair performance, highlighting the critical role of dynamic, task-dependent interareal communication in categorical decisions.

Abstract. Wind turbines are complex electromechanical systems that require continuous monitoring to ensure operational efficiency, reduce maintenance costs, and prevent critical failures. Machine learning has shown great promise in structural health monitoring (SHM) by enabling automated fault detection through data-driven approaches. However, challenges remain in adapting SHM methods to complex environmental conditions while maintaining reliable fault detection and classification. This work proposes a hybrid model that combines supervised and unsupervised learning techniques for classifying operational failures in wind turbines. The proposed framework integrates multimodal data, combining structural and environmental information to monitor four distinct operational states. The approach begins with analysing sensor signals and extracting descriptive features that capture the turbine's dynamic behaviour, accounting for the effects of temperature and wind speed. The unsupervised k-means is applied to label and cluster the dataset, while feature and sensor selection are performed using canonical correlation analysis to rank the most informative variables. A novel relative change damage index is introduced to normalize and scale features based on their relative variability, enhancing the accuracy of clustering and fault classification. Classification is performed using six machine learning algorithms, and the best model is identified. Experimental results, also considering environmental conditions and sensor failures, demonstrate strong performance across both binary and multi-class tasks, including the detection of pitch drive faults and the accurate identification of rotor icing and aerodynamic imbalance. The model achieved classification accuracies ranging from 85 % to 98 %, highlighting its effectiveness in diagnosing wind turbine conditions, and improving the overall reliability and operational analysis of these systems.

This study presents large-scale normative models of white matter (WM) organization across the lifespan, using diffusion magnetic resonance imaging data from over 25,000 healthy individuals ages 0 to 100 years from multiple cohorts including the Human Connectome Project (HCP) Lifespan and UK Biobank. These models capture lifespan trajectories and interindividual variation in fractional anisotropy (FA), a marker of WM integrity. By addressing non-Gaussian data distributions, self-reported race, and site effects, the models offer reference baselines across diverse ages and scanning conditions. We applied these FA models to the HCP Early Psychosis cohort and performed a multivariate analysis to map symptoms onto deviations from multimodal normative models using multiview sparse canonical correlation analysis. Our results reveal extensive WM heterogeneity in psychosis, which is not captured by group-level analyses, with key regions identified, including the right uncinate fasciculus and thalami. These normative models offer valuable tools for individualized WM deviation identification, improving precision in psychiatric assessments. All models are publicly available for community use.