Background: Brain age derived from structural MRI is a potent biomarker for neuro-integrity, yet its clinical adoption is stalled by a critical failure to generalize. Models optimized for high-resolution research scans degrade precipitously when applied to routine hospital protocols, which often utilize thicker slices or different sequences (e.g., FLAIR). Furthermore, while effective for detecting massive atrophy in dementia, standard models often lack the sensitivity required to distinguish subtle psychiatric conditions in real-world workflows. We introduce ATUN, a deep learning framework designed to bridge this translational gap using a novel simulation-based training strategy. Methods: We trained a Transformer U-Net on 10,748 healthy individuals. To ensure clinical robustness without requiring massive labeled clinical datasets, we employed a “simulation-based” domain adaptation strategy, training the model on mathematically degraded research data to mimic the imperfections of routine acquisition. We validated the model’s clinical utility on a real-world psychiatric cohort (n = 769) and its biological interpretability by spatially aligning the model’s attention maps with Amyloid-β and FDG-PET molecular pathology in the ADNI cohort. Findings: On high-resolution research benchmarks, ATUN achieved state-of-the-art accuracy (MAE = 2.50 years), significantly outperforming existing baselines. Crucially, the simulation-trained variant (ATUN-S) demonstrated exceptional robustness: it adapted seamlessly to T1w-FLAIR images (detecting Alzheimer’s with AUC = 0.904, equivalent to the performance using research-grade images) and maintained high longitudinal reliability (ICC > 0.90). Biological validation revealed that the model’s attention maps spontaneously converged onto regions of high Amyloid-β burden and glucose hypometabolism, confirming that predictions are driven by underlying neurodegenerative biology. In a challenging clinical workflow, the model successfully distinguished Major Depressive Disorder (MDD) from transitional Depressive States (AUC = 0.741) using routine hospital scans, establishing its sensitivity to subtle psychiatric pathology. Interpretations: By leveraging simulation to overcome resolution and modality barriers, ATUN unlocks the latent potential of standard MRI, enabling precise and biologically interpretable brain health tracking from historical and routine hospital scans. This framework offers a scalable, zero-cost solution for transdiagnostic screening and psychiatric stratification, effectively moving brain age estimation from the research laboratory to clinical reality.

Background: Brain age is a potent biomarker of cumulative neuroanatomical decline and transdiagnostic vulnerability. However, a critical "lab-to-clinic gap" prevents its deployment: models optimized for high-resolution research scans fail precipitously on the heterogeneous, low-resolution (thick-slice), and diverse-contrast (e.g., T2-FLAIR) sequences ubiquitous in real-world clinical workflows. Methods: To overcome these barriers, we developed a simulation-trained deep learning framework, the Attention-gate and Transformer U-Net (ATUN). Initially trained on high-resolution T1-weighted MRIs from 10,748 healthy participants, a variant (ATUN-S) was subsequently developed to address spatial and contrast disparities. ATUN-S was pre-trained using mathematically downsampled data mimicking clinical slice-thickness, followed by lightweight cross-modality fine-tuning. It was rigorously evaluated on independent clinical T2-FLAIR datasets for Alzheimer’s disease (AD) and major depressive disorder (MDD). Results: ATUN-S achieved state-of-the-art accuracy (mean absolute error= 2.63 years) on simulated data, maintaining high concordance with the high-resolution baseline. Longitudinal tracking confirmed exceptional test-retest reliability across 197 subjects (intra-class correlation coefficients >0.90). Opening the AI "black box," voxel-wise spatial analyses revealed that model attention maps tightly converged with PET-derived patterns of amyloid-beta deposition and glucose hypometabolism, validating its biological fidelity. Clinically, ATUN-S demonstrated robust transdiagnostic utility exclusively using routine FLAIR scans, accurately classifying AD (AUC=0.904) and successfully stratifying recurrent MDD from single depressive episodes (AUC=0.727). Conclusions: By computationally resolving intrinsic MRI resolution and modality barriers without collecting massive new clinical datasets, this simulation-based framework enables scalable, biologically interpretable, and automated transdiagnostic brain health screening over millions of archived routine clinical scans.

Chest X-Ray Vision-Language pretraining (VLP) leverages large-scale radiograph-report pairs to develop joint image-text representations, demonstrating significant potential for medical image diagnosis. However, existing VLP approaches often overlook the multi-view nature of chest X-Rays, and some multi-view methods apply uniform feature fusion, neglecting view-key semantic contributions. Moreover, random cross-modal Masked Language Modeling (MLM) fails to facilitate effective interactions, impeding representation alignment. Additionally, global alignment in VLP may lead to the false-negative problem. To address these limitations, we propose a novel medical VLP framework comprising three core components. First, a Key Semantics-enhanced Multi-view MLM module aggregates pathology-relevant patches across views, providing semantically rich supervision for MLM. A local semantics enhancing approach, which identifies and aggregates pathology-relevant key patches across views to guide MLM. Second, a Frontal-Lateral Alignment module extracts view-specific pathological features, ensuring semantic consistency and preserving critical information during aggregation. This module independently extracts pathological features from both views to preserve view-specific information while ensuring semantic consistency, which mitigates the loss of crucial information during aggregation. Third, a High-order Semantic Alignment approach mitigates false-negative issues by aligning features with semantically consistent clusters, enhancing global alignment through prototype-level semantics. Extensive experiments across seven public datasets demonstrate that our framework outperforms state-of-the-art methods in four downstream tasks, validating its efficacy. The code is available at https://github.com/sajiutea/F-L.

Brain age has emerged as a critical biomarker for assessing neurodevelopmental health and aging trajectories, demonstrating significant potential in early detection and monitoring of neurological and psychiatric disorders. While existing brain age prediction models predominantly rely on structural MRI (sMRI) due to their rich anatomical detail and predictive accuracy, resting-state functional MRI (rs-fMRI)—which captures dynamic brain connectivity—remains underutilized despite offering complementary insights. This study proposes a highly accurate and generalizable brain age prediction framework that effectively fuses sMRI and rs-fMRI modalities. Specifically, we integrate a Squeeze-and-Excitation Transformer for structural feature extraction with a Graph Frequency Recurrent Network for modeling functional dynamics. Our hybrid model achieves a MAE of 1.31 years and Pearson's R of 0.976 on the ABIDE I dataset while generalizing effectively across independent datasets, thus demonstrating the utility of multimodal fusion for robust brain age estimation.

Diffusion models show great potential in solving inverse problems, including MRI reconstruction. With its unique characteristics, medical imaging demands both efficiency and accuracy in the reconstruction process. However, existing MRI reconstruction methods based on diffusion models often fall short of fully leveraging the available measurements during sampling. Consequently, these methods suffer from compromised reconstruction quality and elevated bias, especially when dealing with large acceleration factors. In response to these challenges, we propose Dual Manifold Constraints (DMC), a fast MRI reconstruction method based on diffusion models. We treat the sampling process as a combination of denoising and adding noise processes, and we constrain these two processes using both pristine measurements and their noisy counterparts to adapt to the geometry of diffusion. It's worth noting that we propose a method to estimate the noisy measurement that satisfies the sub-sampling process to maintain the current data manifold when performing data consistency constraints. Experimental results show that our method outperforms the latest diffusion-based methods regarding both reconstruction speed and accuracy, and exhibits strong out-of-distribution generalization performance.

Modeling the biological changes of the human brain is crucial for identifying brain-related diseases and health monitoring. The brain age predicted from MRI data is one useful biomarker for quantifying the maturation and ageing process of human brain. However, the acquisition and preprocessing of MRI data can introduce significant variations between datasets, making it essential to develop models with higher accuracy and robustness for cross-dataset evaluation. To achieve this goal, our paper combines the strengths of CNNs and transformers, proposing the Pyramid Squeeze and Excitation Transformer (PSET) as a novel approach for brain age estimation. In the PSET framework, 3D inception blocks function as an advanced CNN module to capture localized features while the self-attention mechanism is integrated with a squeeze-and-excitation module to extract global features across disparate patches. In particular, a dataset-adaptive and bias-constrained (DABC) model training strategy is proposed to improve the robustness for cross-dataset situations and reduce the bias by introducing self-supervised pre-training, meta-learning and novel loss functions. Experiment results on the dataset of 15,437 healthy brain T1-MRIs (MAE=2.342), demonstrated that the proposed method outperforms both classic visual models and existing brain age estimation models, in the aspect of accuracy, generality and unbiasedness. Additionally, through visualization analysis, we identified the key brain regions that play significant roles in brain age estimation, including the occipital lobe. We compared the brain age gap between patients with diseases and healthy control groups, demonstrating the phenomenon of abnormal aging in conditions such as Alzheimer’s disease and mild cognitive impairment.

Bot accounts on microblogging platforms significantly impact information reliability and cyberspace security. Accurately identifying these bots is essential for effective community governance and opinion management. This article introduces a category of online social behavior features (OSBF), derived from microblog behaviors such as emotional expression, language organization, and self-description. Through a series of experiments, OSBF has demonstrated the stable and robust performance in characterizing and detecting microblog bots on Twitter and Chinese Weibo. By identifying significant differences in OSBF between bot and human accounts, we established an OSBF-based detection model. This model showed excellent performance across multitask and multiscale challenges in two English Twitter datasets. Additionally, we explored cross-language and cross-dataset applications using two Chinese Weibo datasets, further affirming the model's effectiveness and robustness. The experimental results confirm that our OSBF-based model surpasses existing methods in detecting microblog bots.

Hand gesture recognition is pivotal in facilitating human–machine interaction within the Internet of Things. Nevertheless, it encounters challenges, including labeling expenses and robustness. To tackle these issues, we propose a semi-supervised learning framework guided by pseudo-label consistency. This framework utilizes a dual-branch structure with a mean-teacher network. Within this setup, a global and locally guided self-supervised learning encoder acts as a feature extractor in a teacher–student network to efficiently extract features, maximizing data utilization to enhance feature representation. Additionally, we introduce a pseudo-label Consistency-Guided Mean-Teacher model, where simulated noise is incorporated to generate newly unlabeled samples for the teacher model before advancing to the subsequent stage. By enforcing consistency constraints between the outputs of the teacher and student models, we alleviate accuracy degradation resulting from individual differences and interference from other body parts, thereby bolstering the network’s robustness. Ultimately, the teacher model undergoes refinement through exponential moving averages to achieve stable weights. We evaluate our semi-supervised method on two publicly available hand gesture datasets and compare it with several state-of-the-art fully-supervised algorithms. The results demonstrate the robustness of our method, achieving an accuracy rate exceeding 99% across both datasets.

With the explosive popularity of social media, more and more people, including those with depressive symptoms, are starting to express their emotions online through vlogs recently, which makes it important for video-based depression recognition. As video data contains rich acoustical and visual information, the main challenges faced by existing methods include (1) how to accurately mine features associated with depression in massive data and (2) how to effectively fuse various features from different modalities. In this paper, a multi-domain acoustical-visual information fusion network (MDAVIF) is designed to extract depressive spatio-temporal features from image sequences and audios, and an adaptive feature interaction module is proposed to mix these features. Combined with two autoencoders to retain information and prevent overfitting, the proposed method obtains the state-of-the-art result with the precision of 74.25% and the F1-Score of 75.25% when evaluated on the D-vlog dataset.

Unsupervised Domain Adaptation (UDA) is a promising strategy for representing unlabeled data through domain alignment. Nonetheless, a considerable number of whole-domain alignment techniques often neglect the essential interconnections between pixels and patches across distinct domains that exhibit analogous semantic characteristics. This oversight can hinder their ability to manage semantic variations across domains and to create a discriminative embedding for different classes, ultimately leading to reduced discrimination and poor generalization. This paper presents a novel UDA method for medical image analysis, termed CMRVAE. The proposed method is composed of a margin-restrained variational auto-encoder (MR-VAE) and a class-separation patch-level manifold clustering (CPMC) module. The MR-VAE embeds an adaptive margin-based enhancement technique through an innovative variational inference for optimal encoder mapping in UDA. The CPMC module integrates multi-granularity class information into the manifold for improved preparatory work before UDA. Experimental results on three cardiac datasets show that the proposed method achieves substantially enhanced accuracy compared to the state-of-the-art unsupervised approaches.

With the rise of large-scale model applications, leveraging these models as the base network for efficient transfer learning has garnered increasing attention. Currently, parameter-efficient transfer learning methods have made significant improvements in reducing the number of trainable parameters but introduce latency during inference. In this study, we propose an enhanced adaptation of the adapter using a reparameterization technique, revamping the activating layers into linear layers. This modification retains the high-dimensional fine-tuning capability of the adapter for visual tasks while avoiding additional inference latency. We name this plug-and-play module the Re 3 adapter, which optimizes the model with only 0.26% of the parameters and introduces no inference latency. Experimental results demonstrate its clear advantages in traditional classification and medical tasks.

BackgroundIn recent years, the number of people with anxiety disorders has increased worldwide. Methods for identifying anxiety through objective clues are not yet mature, and the reliability and validity of existing modeling methods have not been tested. The objective of this paper is to propose an automatic anxiety assessment model with good reliability and validity.MethodsThis study collected 2D gait videos and Generalized Anxiety Disorder (GAD-7) scale data from 150 participants. We extracted static and dynamic time-domain features and frequency-domain features from the gait videos and used various machine learning approaches to build anxiety assessment models. We evaluated the reliability and validity of the models by comparing the influence of factors such as the frequency-domain feature construction method, training data size, time-frequency features, gender, and odd and even frame data on the model.ResultsThe results show that the number of wavelet decomposition layers has a significant impact on the frequency-domain feature modeling, while the size of the gait training data has little impact on the modeling effect. In this study, the time-frequency features contributed to the modeling, with the dynamic features contributing more than the static features. Our model predicts anxiety significantly better in women than in men (rMale = 0.666, rFemale = 0.763, p < 0.001). The best correlation coefficient between the model prediction scores and scale scores for all participants is 0.725 (p < 0.001). The correlation coefficient between the model prediction scores for odd and even frame data is 0.801~0.883 (p < 0.001).ConclusionThis study shows that anxiety assessment based on 2D gait video modeling is reliable and effective. Moreover, we provide a basis for the development of a real-time, convenient and non-invasive automatic anxiety assessment method.

The brain age is one of important biomarkers for identifying neurodegenerative diseases. Most existing works construct prediction models based on brain MRIs by deep learning methods while multi-scale features of MRI have been paid less attention. So, this paper proposes a new brain age predicting model, ERU-Net, which introduces an adaptive feature fusion (AFF) block into U-Net, for exploiting multi-scale features for high estimating accuracy. Moreover, the attention gate is also applied to enhance this model’s ability to extract important features among various brain regions. On a dataset of 6318 healthy people, ERU-Net achieves MAE = 2.85 and Pearson correlation coefficient R = 0.98, which outperforms other existing models for brain age estimation. In addition, by the results on the neurodegenerative disease datasets, it is clear that for the index of brain age gap (BAG), there are significantly statistical differences between healthy subjects and neurodegenerative patients such as Aizheimer’s disease, mild cognitive impairment, schizophrenia and Parkinson disease.

With the increasing number of depressed patients and the development of computer vision technology, the study of individual automatic depression estimation (ADE) methods based on facial images has attracted much attention in recent years. Most existing works focus on obtaining informative features from the whole images with advancing deep learning models, while the local tiny changes and the fusion of different features have been paid less attention. In this paper, a two-branch predicting model with a elaborate transfomer block NFFT is designed to combine global and local features extracted from whole images and image patches for predicting the depression score precisely. Besides, a classification head is added to guide the regression results for improving accuracy. Experiment results on the AVEC2014 dataset (MAE=5.81, RMSE=7.49) demonstrate that the proposed model outperforms other methods, and the extended experiments on one new dataset (CCPL) are conducted to validate the generalizability and robustness of our model.

Social anxiety disorder (SAD) is a common anxiety disorder in childhood and adolescence. Studies on SAD in adults have reported both structural and functional aberrancies of the brain at the group level. However, evidence has shown differences in anxiety-related brain abnormalities between adolescents and adults. Since children and adolescents can afford limited scan time, optimizing the scan tasks is essential for SAD research in children and adolescents. Thus, we need to address whether brain structure, resting-state fMRI, and naturalistic imaging enable individualized identification of SAD in children and adolescents, which measurement is more effective, and whether pooling multi-modal features can improve the identification of SAD. We comprehensively addressed these questions by building machine learning models based on parcel-wise brain features. We found that naturalistic fMRI yielded higher classification accuracy (69.17%) than the other modalities and the classification performance showed dependence on the contents of the movie. The classification models also identified contributing brain regions, some of which exhibited correlations with the symptoms scores of SAD. However, pooling brain features from the three modalities did not help enhance the classification accuracy. These results support the application of carefully designed naturalistic imaging in recognizing children and adolescents at risk of SAD.

Personality affects an individual's academic achievements, occupational tendencies, marriage quality and physical health, so more convenient and objective personality assessment methods are needed. Gait is a natural, stable, and easy-to-observe body movement that is closely related to personality. The purpose of this paper is to propose a personality assessment model based on gait video and evaluate the reliability and validity of the multidimensional model. This study recruited 152 participants and used cameras to record their gait videos. Each participant completed a 44-item Big Five Inventory (BFI-44) assessment. We constructed diverse static and dynamic time-frequency features based on gait skeleton coordinates, interframe differences, distances between joints, angles between joints, and wavelet decomposition coefficient arrays. We established multidimensional personality trait assessment models through machine learning algorithms and evaluated the criterion validity, split-half reliability, convergent validity, and discriminant validity of these models. The results showed that the reliability and validity of the Gaussian process regression (GPR) and linear regression (LR) models were best. The mean values of their criterion validity were 0.478 and 0.508, respectively, and the mean values of their split-half reliability were all greater than 0.8. In the formed multitrait-multimethod matrix, these methods also had higher convergent and discriminative validity. The proposed approach shows that gait video can be effectively used to evaluate personality traits, providing a new idea for the formation of convenient and non-invasive personality assessment methods.

The aging process of human brain is complex, which can result in brain structural changes. One promising way to gain a deep understanding of aging process is using machine learning, typically convolutional neural network (CNN), to predict brain age based on magnetic resonance imaging data. Though CNN has a strong ability to capture features from a small local region of the input image, it lacks the ability to capture global features of the surrounding neighbors. Thus, in this paper, we propose the squeeze and excitation transformer (SQET) for pursuing high-accuracy brain age estimation, in which a squeeze and excitation module is designed and fused in conventional self-attention in the transformer structure to capture global features among different localities even if they are spatially far apart. In particular, for 9 public datasets with 6,318 healthy brain Tl-MRIs with an age range of 6-88, our proposed SQET can achieve the result of 2.55 MAE and the correlation coefficient r=0.983, which has significantly outperformed all other reported models up to now.

Chaotic maps are very important for establishing chaos-based image encryption systems. This paper introduces a coupling chaotic system based on a certain unit transform, which can combine any two 1D chaotic maps to generate a new one with excellent performance. The chaotic behavior analysis has verified this coupling system's effectiveness and progress. In particular, we give a specific strategy about selecting an appropriate unit transform function to enhance chaos of generated maps. Besides, a new chaos based pseudo-random number generator, shorted as CBPRNG, is designed to improve the distribution of chaotic sequences. We give a mathematical illustration on the uniformity of CBPRNG, and test the randomness of it. Moreover, based on CBPRNG, an image encryption algorithm is introduced. Simulation results and security analysis indicate that the proposed image encryption scheme is competitive with some advanced existing methods.

Chaotic maps are found to be a promising external entropy source for image encryption. However, many existing chaotic systems have relatively narrow parameter space and cannot generate uniformly distributed sequences, which may reduce the reliability of cryptosystem. In this paper, a novel chaotification model named as UCS is proposed to improve the statistical property and expand the parameter range of existing maps. The new chaotic maps generated from UCS exhibit more complex chaotic behaviors than seed maps, and the two independent parameters can take almost any value in $$\mathbf {R}^2$$ . Moreover, the iterated chaotic sequences are approximately uniformly distributed and thus have better randomness. Based on the improved chaotic maps, a new image encryption algorithm is designed, which encodes digital images by derangement-based confusion, row–column bidirectional diffusion and 2D zigzag diffusion. Experimental results indicate the high performance of the proposed image encryption scheme, since it has obtained excellent results in various tests when compared with several existing methods.

The brain age, estimated from the brain MRI data, is found to be a promising biomarker for human brain development and neuroanatomical aging processes. A well-performed brain age predicting model is in great demand for many applications like healthcare and disease diagnosis. In this paper, we proposed a dimensional-attention-based 3D convolutional neural network (DACNN) to estimate the biological age for developing normal brain from T1-weighted MRI, in which a dimensional attention module was designed and applied to restrain noises and increase the weights of effective voxels for feature maps. Experimental results indicated that our model significantly outperformed the best reported methods up to now. In particular, with the dilated convolution, the proposed DACNN achieved the state-of-the-art result of 1.01 MAE on a combined dataset consisted of 880 healthy children and adolescents.