Neural coupling is a fundamental mechanism in neuroscience that facilitates the emergence of cognitive functions through dynamic interactions and synchronization among distributed brain regions. Inspired by this principle, we pose the question: Might the biological mechanism of neural oscillatory synchronization inspire the feature representation learning for neuroscience? By addressing this question through the Kuramoto model, renowned for simulating oscillatory dynamics, we present a novel physics-informed deep model, `SyncBrain`, it models brain regions as interacting oscillatory units and simulates their temporal dynamics and synchronization patterns to distinguish cognitive states. Furthermore, inspired by the brain's inherent ability to dynamically attend to critical temporal information, we incorporate an adaptive control module that introduces an attention-like mechanism to guide information flow. We evaluate our model on multiple functional neuroimaging datasets, it demonstrates promising performance and enhanced interpretability in both cognitive state decoding and early disease diagnosis, outperforming existing computational methods. These results demonstrate the effectiveness of neural oscillatory mechanisms in shaping robust and interpretable machine learning models for neuroscience applications.

A reliable foundation model of functional neuroimages is critical to promote clinical applications where the performance of current AI models is significantly impeded by a limited sample size. To that end, tremendous efforts have been made to pretraining large models on extensive unlabeled fMRI data using scalable self-supervised learning. Since self-supervision is not necessarily aligned with the brain-to-outcome relationship, most foundation models are suboptimal to the downstream task, such as predicting disease outcomes. By capitalizing on rich environmental variables and demographic data along with an unprecedented amount of functional neuroimages, we form the brain modeling as a multitask learning and present a scalable model architecture for (i) multitask pretraining by tokenizing multiple brain-environment interactions (BEI) and (ii) semi-supervised finetuning by assigning pseudo-labels of default BEI. We have evaluated our foundation model on a variety of applications, including sex prediction, human behavior recognition, and disease early diagnosis of Autism, Parkinson's disease, Alzheimer's disease, and Schizophrenia, where promising results indicate the great potential to facilitate current neuroimaging applications in clinical routines.

Modeling the progression of neurodegenerative diseases such as Alzheimer’s disease (AD) is crucial for early detection and prevention given their irreversible nature. However, the scarcity of longitudinal data and complex disease dynamics make the analysis highly challenging. Moreover, longitudinal samples often contain irregular and large intervals between subject visits, which underscore the necessity for advanced data generation techniques that can accurately simulate disease progression over time. In this regime, we propose a novel conditional generative model for synthesizing longitudinal sequences and present its application to neurodegenerative disease data generation conditioned on multiple time-dependent ordinal factors, such as age and disease severity. Our method sequentially generates continuous data by bridging gaps between sparse data points with a diffusion model, ensuring a realistic representation of disease progression. The synthetic data are curated to integrate both cohort-level and individual-specific characteristics, where the cohort-level representations are modeled with an ordinal regression to capture longitudinally monotonic behavior. Extensive experiments on four AD biomarkers validate the superiority of our method over nine baseline approaches, highlighting its potential to be applied to a variety of longitudinal data generation.

State‑space models (SSMs) have become a cornerstone for unraveling brain dynamics, capturing how latent neural states evolve over time and give rise to observed signals. By combining deep learning’s flexibility with SSMs’ principled dynamical structure, recent studies have achieved powerful fits to functional neuroimaging data. However, most approaches still view the brain as a set of loosely connected regions or impose oversimplified network priors, falling short of a truly holistic, self‐organized dynamical system perspective. Brain functional connectivity (FC) at each time point naturally forms a symmetric positive definite (SPD) matrix, which lives on a curved Riemannian manifold rather than in Euclidean space. Capturing the trajectories of these SPD matrices is key to understanding how coordinated networks support cognition and behavior. To this end, we introduce *GeoDynamics*, a geometric state space neural network that tracks latent brain state trajectories directly on the high‑dimensional SPD manifold. *GeoDynamics* embeds each connectivity matrix into a manifold‑aware recurrent framework, learning smooth, geometry‑respecting transitions that reveal task‐driven state changes and early markers of Alzheimer’s, Parkinson’s, and autism. Beyond neuroscience, we validate *GeoDynamics* on human action recognition benchmarks (UTKinect, Florence, HDM05), demonstrating its scalability and robustness in modeling complex spatiotemporal dynamics across diverse domains.

Alzheimer’s disease (AD) is marked by cognitive decline along with the widespread of tau aggregates across the brain cortex. Due to the challenges of imaging pathology spreading flows \textit{in vivo}, however, quantitative analysis on the cortical pathways of tau propagation and its interaction with the cascade of amyloid-beta (A$\beta$) plaques lags behind the experimental insights of underlying pathophysiological mechanisms. To address this challenge, we present a physics-informed neural network, empowered by mean-field theory, to uncover the biologically meaningful spreading pathways of tau aggregates between two longitudinal snapshots. Following the notion of `prion-like' mechanism in AD, we first formulate the dynamics of tau propagation as a mean-field game (MFG), where the spread of tau aggregate at each location (aka. agent) depends on the collective behavior of the surrounding agents as well as the potential field formed by amyloid burden. Given the governing equation of propagation dynamics, MFG reaches an equilibrium that allows us to model the evolution of tau aggregates as an optimal transport with the lowest cost in \textit{Wasserstein} space. By leveraging the variational primal-dual structure in MFG, we propose a \textit{Wasserstein}-1 Lagrangian generative adversarial network (GAN), in which a Lipschitz critic seeks the appropriate transport cost at the population level and a generator parameterizes the flow fields of optimal transport across individuals. Additionally, we incorporate a symbolic regression module to derive an explicit formulation capturing the A$\beta$-tau crosstalk. Experimental results on public neuroimaging datasets demonstrate that our explainable deep model not only yields precise and reliable predictions of future tau progression for unseen new subjects but also provides a new window to uncover new understanding of pathology propagation in AD through learning-based approaches.

In typical graph neural networks (GNNs), feature representation learning naturally evolves through iteratively updating node features and exchanging information based on graph topology. In this context, we conceptualize that the learning process in GNNs is a mean-field game (MFG), where each graph node is an agent, interacting with its topologically connected neighbors. However, current GNNs often employ the identical MFG strategy across different graph datasets, regardless of whether the graph exhibits homophilic or heterophilic characteristics. To address this challenge, we propose to formulate the learning mechanism into a variational framework of the MFG inverse problem, introducing an in-context selective message passing paradigm for each agent, which promotes the best overall outcome for the graph. Specifically, we seek for the application-adaptive transportation function (controlling information exchange throughout the graph) and reaction function (controlling feature representation learning on each agent), \textit{on the fly}, which allows us to uncover the most suitable selective mechanism of message passing by solving an MFG variational problem through the lens of Hamiltonian flows. Taken together, our variational framework unifies existing GNN models into various mean-field games with distinct equilibrium states, each characterized by the learned in-context message passing operators. Furthermore, we present an agnostic end-to-end deep model, coined \textit{Game-of-GNN}, to jointly identify the message passing mechanism and fine-tune the GNN hyper-parameters on top of the elucidated message passing operators. \textit{Game-of-GNN} has achieved SOTA performance on diverse graph data, including popular benchmark datasets and human connectomes. More importantly, the mathematical insight of MFG framework provides a new window to understand the foundational principles of graph learning as an interactive dynamical system, which allows us to reshape the idea of designing next-generation GNN models.

In both neuroscience and artificial intelligence (AI), it is well-established that neural “coupling” gives rise to dynamically distributed systems. These systems exhibit self-organized spatiotemporal patterns of synchronized neural oscillations, enabling the representation of abstract concepts. By capitalizing on the unprecedented amount of human neuroimaging data, we propose that advancing the theoretical understanding of rhythmic coordination in neural circuits can offer powerful design principles for the next generation of machine learning models with improved efficiency and robustness. To this end, we introduce a physics-informed deep learning framework for \underline{B}rain \underline{R}hythm \underline{I}dentification by \underline{K}uramoto and \underline{C}ontrol (coined \textit{BRICK}) to characterize the synchronization of neural oscillations that shapes the dynamics of evolving cognitive states. Recognizing that brain networks are structurally connected yet behaviorally dynamic, we further conceptualize rhythmic neural activity as an artificial dynamical system of coupled oscillators, offering a shared mechanistic bridge to brain-inspired machine intelligence. By treating each node as an oscillator interacting with its neighbors, this approach moves beyond the conventional paradigm of graph heat diffusion and establishes a new regime of representation compression through oscillatory synchronization. Empirical evaluations demonstrate that this synchronization-driven mechanism not only mitigates over-smoothing in deep GNNs but also enhances the model’s capacity for reasoning and solving complex graph-based problems.

Given the large scale of public functional Magnetic Resonance Imaging (fMRI), e.g., UK Biobank (UKB) and Human Connectome Projects (HCP), brain foundation models are emerging. Although the amount of samples under rich environmental variables is unprecedented, existing brain foundation models learn from fMRI derived from a narrow range of cognitive states stimulated by similar environments, causing the limited robustness demonstrated in various applications and datasets acquired with different pipelines and limited sample size. By capitalizing on the variety of cognitive status as subjects performing explicit tasks, we present the mixture of brain experts, namely BrainMoE, pre-training on tasking fMRI with rich behavioral tasks in addition to resting fMRI for a robust brain foundation model. Brain experts are designed to produce embeddings for different behavioral tasks related to cognition. Afterward, these cognition embeddings are mixed by a cognition adapter via cross-attention so that BrainMoE can handle orthogonal embeddings and be robust on those boutique downstream datasets. We have pre-trained two existing self-regressive architectures and one new supervised architecture as brain experts on 68,251 fMRI scans among UKB and HCP, containing 12 different cognitive states. Then, BrainMoE is evaluated on a variety of applications, including sex, age prediction, human behavior recognition, disease early diagnosis of Autism, Parkinson's disease, Alzheimer's disease, and Schizophrenia, and fMRI-EEG multimodal applications, where promising results in eight datasets from three different pipelines indicate great potential to facilitate current neuroimaging applications in clinical routines.

A fundamental challenge in cognitive neuroscience is understanding how cognition emerges from the interplay between structural connectivity (SC) and dynamic functional connectivity (FC) in the brain. Network neuroscience has emerged as a powerful framework to understand brain function through a holistic perspective on structure-function relationships. In this context, current machine learning approaches typically seek to establish direct mappings between structural connectivity (SC) and functional connectivity (FC) associated with specific cognitive states. However, these state-independent methods often yield inconsistent results due to overlapping brain networks across cognitive states. To address this limitation, we conceptualize to uncover the dendritic coupling mechanism between one static SC and multiple FCs by solving a flow problem that bridges the distribution of SC to a mixed distribution of FCs, conditioned on various cognitive states, along a Riemannian manifold of symmetric positive-definite (SPD) manifold. We further prove the equivalence between flow matching on the SPD manifold and on the computationally efficient Cholesky manifold. Since a spare of functional connections is shared across cognitive states, we introduce the notion of consensus control to promote the shared kinetic structures between multiple FC-to-SC pathways via synchronized coordination, yielding a biologically meaningful underpinning on SC-FC coupling mechanism. Together, we present BrainFlow, a reversible generative model that achieves state-of-the-art performance on not only synthetic data but also large-scale neuroimaging datasets from UK Biobank and Human Connectome Project.

Positron Emission Tomography (PET) is essential for understanding the pathophysiological mechanisms underlying neurodegenerative diseases like Alzheimer's disease (AD). However, existing approaches primarily focus on stereotypical patterns of pathology burden, lacking the ability to elucidate the underlying propagation mechanisms by which pathologies spread throughout the brain over time. Given that many neurodegenerative diseases exhibit prion-like pathology spread, it is essential to uncover the spot-to-spot flow field between consecutive PET snapshots. To address this, we reformulate the problem of identifying latent cortical propagation pathways of neuropathological burden within the well-established framework of optimal mass transport (OMT). In this formulation, the dynamic spreading of pathology across longitudinal PET scans is inherently constrained by the geometry of the brain cortex. To solve this problem, we introduce a variational framework that characterizes the dynamical system of pathology propagation in the brain, ultimately reducing to a Wasserstein geodesic between two density distributions of pathology accumulation. Furthermore, we hypothesize that a well-characterized mechanism of pathology propagation will enable the prediction of future pathology accumulation at the individual level, paving the way for personalized disease progression modeling. Building on the principles of physics-informed deep models, we derive the governing equation of the underlying OMT model and introduce an explainable, generative adversarial network-inspired framework. Our approach (1) parameterizes population-level OMT dynamics through a flow adjuster and (2) predicts the spreading flow in unseen subjects using a trained flow driver. We validate the accuracy of our model on publicly available datasets, demonstrating its effectiveness in forecasting future pathology accumulation. Since our deep model adheres to the second law of thermodynamics, we further explore the propagation dynamics of tau aggregates throughout the progression of AD. In contrast to traditional methods, our physics-informed approach enhances both accuracy and interpretability, demonstrating its potential to reveal novel neurobiological mechanisms driving disease progression.

One of the fundamental scientific problems in neuroscience is to have a good understanding of how cognition and behavior emerge from brain function. Since the neuroscience concept of cognitive control parallels the notion of system control in engineering, many computational models formulate the dynamics neural process into a dynamical system, where the hidden states of the complex neural system are modulated by energetic simulations. However, the human brain is a quintessential complex biological system. Current computation models either use neural networks to approximate the underlying dynamics, which makes it difficult to fully understand the system mechanics, or compromise to simplified linear models with very limited power to characterize non-linear and self-organized dynamics along with complex neural activities. To address this challenge, we devise an end-to-end deep model to identify the underlying brain dynamics based on Koopman operator theory, which allows us to model a complex non-linear system in an infinite-dimensional linear space. In the context of reverse engineering, we further propose a biology-inspired control module that adjusts the input (neural activity data) based on feedback to align brain dynamics with the underlying cognitive task. We have applied our deep model to predict cognitive states from a large scale of existing neuroimaging data by identifying the latent dynamic system of functional fluctuations. Promising results demonstrate the potential of establishing a system-level understanding of the intricate relationship between brain function and cognition through the landscape of explainable deep models.

The human brain is a complex inter-wired system that emerges spontaneous functional fluctuations. In spite of tremendous success in the experimental neuroscience field, a system-level understanding of how brain anatomy supports various neural activities remains elusive. Capitalizing on the unprecedented amount of neuroimaging data, we present a physics-informed deep model to uncover the coupling mechanism between brain structure and function through the lens of data geometry that is rooted in the widespread wiring topology of connections between distant brain regions. Since deciphering the puzzle of self-organized patterns in functional fluctuations is the gateway to understanding the emergence of cognition and behavior, we devise a geometric deep model to uncover manifold mapping functions that characterize the intrinsic feature representations of evolving functional fluctuations on the Riemannian manifold. In lieu of learning unconstrained mapping functions, we introduce a set of graph-harmonic scattering transforms to impose the brain-wide geometry on top of manifold mapping functions, which allows us to cast the manifold-based deep learning into a reminiscent of *MLP-Mixer* architecture (in computer vision) for Riemannian manifold. As a proof-of-concept approach, we explore a neural-manifold perspective to understand the relationship between (static) brain structure and (dynamic) function, challenging the prevailing notion in cognitive neuroscience by proposing that neural activities are essentially excited by brain-wide oscillation waves living on the geometry of human connectomes, instead of being confined to focal areas.

Human brain is a complex system composed of many components that interact with each other. A well-designed computational model, usually in the format of partial differential equations (PDEs), is vital to understand the working mechanisms that can explain dynamic and self-organized behaviors. However, the model formulation and parameters are often tuned empirically based on the predefined domain-specific knowledge, which lags behind the emerging paradigm of discovering novel mechanisms from the unprecedented amount of spatiotemporal data. To address this limitation, we sought to link the power of deep neural networks and physics principles of complex systems, which allows us to design explainable deep models for uncovering the mechanistic role of how human brain (the most sophisticated complex system) maintains controllable functions while interacting with external stimulations. In the spirit of optimal control, we present a unified framework to design an explainable deep model that describes the dynamic behaviors of underlying neurobiological processes, allowing us to understand the latent control mechanism at a system level. We have uncovered the pathophysiological mechanism of Alzheimer's disease to the extent of controllability of disease progression, where the dissected system-level understanding enables higher prediction accuracy for disease progression and better explainability for disease etiology than conventional (black box) deep models.

Graphs are ubiquitous in various domains, such as social networks and biological systems. Despite the great successes of graph neural networks (GNNs) in modeling and analyzing complex graph data, the inductive bias of locality assumption, which involves exchanging information only within neighboring connected nodes, restricts GNNs in capturing long-range dependencies and global patterns in graphs. Inspired by the classic Brachistochrone problem, we seek how to devise a new inductive bias for cutting-edge graph application and present a general framework through the lens of variational analysis. The backbone of our framework is a two-way mapping between the discrete GNN model and continuous diffusion functional, which allows us to design application-specific objective function in the continuous domain and engineer discrete deep model with mathematical guarantees. First, we address over-smoothing in current GNNs. Specifically, our inference reveals that the existing layer-by-layer models of graph embedding learning are equivalent to a ${\ell _2}$-norm integral functional of graph gradients, which is the underlying cause of the over-smoothing problem. Similar to edge-preserving filters in image denoising, we introduce the total variation (TV) to promote alignment of the graph diffusion pattern with the global information present in community topologies. On top of this, we devise a new selective mechanism for inductive bias that can be easily integrated into existing GNNs and effectively address the trade-off between model depth and over-smoothing. Second, we devise a novel generative adversarial network (GAN) to predict the spreading flows in the graph through a neural transport equation. To avoid the potential issue of vanishing flows, we tailor the objective function to minimize the transportation within each community while maximizing the inter-community flows. Our new GNN models achieve state-of-the-art (SOTA) performance on graph learning benchmarks such as Cora, Citeseer, and Pubmed.

Unsupervised feature selection is a vital yet challenging topic for effective data learning. Recently, 2-D feature selection methods show good performance on image analysis by utilizing the structure information of image. Current 2-D methods usually adopt a sparse regularization to spotlight the key features. However, such scheme introduces additional hyperparameter needed for pruning, limiting the applicability of unsupervised algorithms. To overcome these challenges, we design a feature filter to estimate the weight of image features for unsupervised feature selection. Theoretical analysis shows that a sparse regularization can be derived from the feature filter by transformation, indicating that the filter plays the same role as the popular sparse regularization does. We deploy two distinct strategies in terms of feature selection, called multiple feature filters and single common feature filter. The former divides the optimization problem into multiple independent subproblems and selects features that meet the respective interests of each subproblem. The latter selects features that are in the interest of the overall optimization problem. Extensive experiments on seven benchmark datasets show that our unsupervised 2-D weight-based feature selection methods achieve superior performance over the state-of-the-art methods.

Multiview clustering seeks to partition objects via leveraging cross-view relations to provide a comprehensive description of the same objects. Most existing methods assume that different views are linear transformable or merely sampling from a common latent space. Such rigid assumptions betray reality, thus leading to unsatisfactory performance. To tackle the issue, we propose to learn both common and specific sampling spaces for each view to fully exploit their collaborative representations. The common space corresponds to the universal self-representation basis for all views, while the specific spaces are the view-specific basis accordingly. An iterative self-supervision scheme is conducted to strengthen the learned affinity matrix. The clustering is modeled by a convex optimization. We first solve its linear formulation by the popular scheme. Then, we employ the deep autoencoder structure to exploit its deep nonlinear formulation. The extensive experimental results on six real-world datasets demonstrate that the proposed model achieves uniform superiority over the benchmark methods.

Functional connectivities (FC) of brain network manifest remarkable geometric patterns, which is the gateway to understanding brain dynamics. In this work, we present a novel geometric-attention neural network to characterize the time-evolving brain state change from the functional neuroimages by tracking the trajectory of functional dynamics on high-dimension Riemannian manifold of symmetric positive definite (SPD) matrices. Specifically, we put the spotlight on learning the common state-specific manifold signatures that represent the underlying cognition. In this context, the driving force of our neural network is tied up with the learning of the evolution functionals on the Riemannian manifold of SPD matrix that underlies the known evolving brain states. To do so, we train a convolution neural network (CNN) on the Riemannian manifold of SPD matrices to seek for the putative low-dimension feature representations, followed by an end-to-end recurrent neural network (RNN) to yield the time-varying mapping function of SPD matrices which fits the evolutionary trajectories of the underlying states. Furthermore, we devise a geometric attention mechanism in CNN, allowing us to discover the latent geometric patterns in SPD matrices that are associated with the underlying states. Notably, our work has the potential to understand how brain function emerges behavior by investigating the geometrical patterns from functional brain networks, which is essentially a correlation matrix of neuronal activity signals. Our proposed manifold-based neural network achieves promising results in predicting brain state changes on both simulated data and task functional neuroimaging data from Human Connectome Project, which implies great applicability in neuroscience studies.

Nasopharyngeal carcinoma (NPC) is a malignant tumor whose survivability is greatly improved if early diagnosis and timely treatment are provided. Accurate segmentation of both the primary NPC tumors and metastatic lymph nodes (MLNs) is crucial for patient staging and radiotherapy scheduling. However, existing studies mainly focus on the segmentation of primary tumors, eliding the recognition of MLNs, and thus fail to comprehensively provide a landscape for tumor identification. There are three main challenges in segmenting primary NPC tumors and MLNs: variable location, variable size, and irregular boundary. To address these challenges, we propose an automatic segmentation network, named by NPCNet, to achieve segmentation of primary NPC tumors and MLNs simultaneously. Specifically, we design three modules, including position enhancement module (PEM), scale enhancement module (SEM), and boundary enhancement module (BEM), to address the above challenges. First, the PEM enhances the feature representations of the most suspicious regions. Subsequently, the SEM captures multiscale context information and target context information. Finally, the BEM rectifies the unreliable predictions in the segmentation mask. To that end, extensive experiments are conducted on our dataset of 9124 samples collected from 754 patients. Empirical results demonstrate that each module realizes its designed functionalities and is complementary to the others. By incorporating the three proposed modules together, our model achieves state-of-the-art performance compared with nine popular models.

The topology of human functional networks is assumed to oscillate during brain states changes. The functional neuroimage is employed to offer a non-invasive window to understand cognition and behaviors by characterizing the functional connections between spatially distinct brain regions. Consequently, identifying the transitions of functional connectivities is the critical step to understanding the mechanism of cognition that might be underlined with neurological disorders. However, little attention has been paid to studying the geometry of the entire functional brain network. To tackle this issue, this paper models the cognition changes on functional brain networks as a set of landmarks residing on a Riemannian manifold. Accordingly, we propose a Riemannian manifold mean shift method to detect cognition changes by identifying the representative function networks of the distribution of functional networks. The manifold mean shift (MMS) method is applied on both simulated data and real functional neuroimaging data, downloaded from Human Connectome Project (HCP). Experimental results demonstrated the MMS achieved highly accurate and consistent cognition change, by comparing three state-of-the-art methods.

The human retinal surface resembles to a sphere while it is captured by two-dimensional (2D) planar imaging to have a stereo sequence in clinical practice. Reconstructing its three-dimensional (3D) structure from the 2D planar retinal images is crucial for analyzing the relationship between the topological morphology and clinical implication. In this regard, we propose to reconstruct the 3D retina structure from 2D stereo fundus images via dynamic registration. The fundus images from different viewpoints are first co-registrated by using multi-scale deep convolutional feature and geometric structure feature by building their transformation function. The aligned images are then mosaicked together and a 3D reconstruction is obtained by a learned weighted smoothing project the registered images onto 3D coordinates. We compare the proposed registration method with five state-of-the-art methods. Extensive experimental results demonstrate that the proposed framework achieves superior performances, even with challenging scenarios in which the tested images are severely degraded by illness, large eyeball rotation and low resolutions.

COVID-19 causes burdens to the ICU. Evidence-based planning and optimal allocation of the scarce ICU resources is urgently needed but remains unaddressed. This study aims to identify variables and test the accuracy to predict the need for ICU admission, death despite ICU care, and among survivors, length of ICU stay, before patients were admitted to ICU. Retrospective data from 733 in-patients confirmed with COVD-19 in Wuhan, China, as of March 18, 2020. Demographic, clinical and laboratory were collected and analyzed using machine learning to build the predictive models. The built machine learning model can accurately assess ICU admission, length of ICU stay, and mortality in COVID-19 patients toward optimal allocation of ICU resources. The prediction can be done by using the clinical data collected within 1-15 days before the actual ICU admission. Lymphocyte absolute value involved in all prediction tasks with a higher AUC. The online predictive system is freely available to the public (http://212.64.70.65:8000/).

Accurate nasopharyngeal carcinoma (NPC) segmentation in magnetic resonance image (MRI) is crucial for diagnosis and treatment. However, most existing deep learning methods performed unsatisfactorily, since NPC is infiltrative and typically has a small or even tiny volume with indistinguishable boundary, making it indiscernible from tightly connected surrounding tissue in immense and complex background. To address the background dominant problem, this paper proposes a coarse-to-fine deep model. The proposed model starts with predicting a coarse mask with a well-designed segmentation module, followed by a boundary rendering module, which exploits semantic information from different layers of feature maps to refine the boundary of the coarse mask. The designed rendering module is shown to achieve superior performance with dramatically fewer parameters by operating only on the segmented mask, rather than on the whole feature maps as the popular methods do. Extensive experiments are conducted on a collected dataset consisting of 2000 MRI slices from 596 patients. Experimental results demonstrate that the proposed model not only outperforms six popular segmentation models but also has a considerable generalization capability on existing models.