Frontiers in Computational Neuroscience最新文献

Robustness, defined as a system's ability to maintain functional reliability in the face of perturbations, is achieved through its capacity to filter external disturbances using internal priors encoded in its structure and states. While biophysical neural networks are widely recognized for their robustness, the precise mechanisms underlying this resilience remain poorly understood. In this study, we explore how orientation-selective neurons arranged in a one-dimensional ring network respond to perturbations, with the aim of uncovering insights into the robustness of visual subsystems in the brain. By analyzing the steady-state dynamics of a rate-based network, we characterize how the activation state of neurons influences the network's response to disturbances. Our results demonstrate that the activation state of neurons, rather than their firing rates alone, governs the network's sensitivity to perturbations. We further show that lateral connectivity modulates this effect by shaping the response profile across spatial frequency components. These findings suggest a state-dependent filtering mechanism that contributes to the robustness of visual circuits, offering theoretical insight into how different components of perturbations are selectively modulated within the network.

Introduction: Accurate and early identification of brain tumors is essential for improving therapeutic planning and clinical outcomes. Manual segmentation of Magnetic Resonance Imaging (MRI) remains time-consuming and subject to inter-observer variability. Computational models that combine Artificial Intelligence and biomedical imaging offer a pathway toward objective and efficient tumor delineation. The present study introduces a deep learning framework designed to enhance brain tumor segmentation performance.

Methods: A comprehensive ensemble architecture was developed by integrating Generative Autoencoders with Attention Mechanisms (GAME), Convolutional Neural Networks, and attention-augmented U-Net segmentation modules. The dataset comprised 5,880 MRI images sourced from the BraTS 2023 benchmark distribution accessed via Kaggle, partitioned into training, validation, and testing subsets. Preprocessing included intensity normalization, augmentation, and unsupervised feature extraction. Tumor segmentation employed an attention-based U-Net, while tumor classification utilized a CNN coupled with Transformer-style self-attention. The Generative Autoencoder performed unsupervised representation learning to refine feature separability and enhance robustness to MRI variability.

Results: The proposed framework achieved notable performance improvements across multiple evaluation metrics. The segmentation module produced a Dice Coefficient of 0.85 and a Jaccard Index of 0.78. The classification component yielded an accuracy of 87.18 percent, sensitivity of 88.3 percent, specificity of 86.5 percent, and an AUC-ROC of 0.91. The combined use of generative modeling, attention mechanisms, and ensemble learning improved tumor localization, boundary delineation, and false positive suppression compared with conventional architectures.

Discussion: The findings indicate that enriched representation learning and attention-driven feature refinement substantially elevate segmentation accuracy on heterogeneous MRI data. The integration of unsupervised learning within the pipeline supported improved generalization across variable imaging conditions. The demonstrated performance suggests strong potential for clinical utility, although broader validation across external datasets is recommended to further substantiate generalizability.

In learning goal-directed behavior, state representation is important for adapting to the environment and achieving goals. A predictive state representation called successive representation (SR) has recently attracted attention as a candidate for state representation in animal brains, especially in the hippocampus. The relationship between the SR and the animal brain has been studied, and several neural network models for computing the SR have been proposed based on the findings. However, studies on implementation of the SR involving action selection have not yet advanced significantly. Therefore, we explore possible mechanisms by which the SR is utilized biologically for action selection and learning optimal action policies. The actor-critic architecture is a promising model of animal behavioral learning in terms of its correspondence to the anatomy and function of the basal ganglia, so it is suitable for our purpose. In this study, we construct neural network models for behavioral learning using the SR. By using them to perform reinforcement learning, we investigate their properties. Specifically, we investigated the effect of using different state representations for the actor and critic in the actor-critic method, and also compared the actor-critic method with Q-learning and SARSA. We found the difference between the effect of using the SR for the actor and the effect of using the SR for the critic in the actor-critic method, and observed that using the SR in conjunction with one-hot encoding makes it possible to learn with the benefits of both representations. These results suggest the possibility that the striatum can learn using multiple state representations complementarily.

In algebraic topology, a k-dimensional simplex is defined as a convex polytope consisting of k + 1 vertices. If spatial dimensionality is not considered, it corresponds to the complete graph with k + 1 vertices in graph theory. The alternating sum of the number of simplices across dimensions yields a topological invariant known as the Euler characteristic, which has gained significant attention due to its widespread application in fields such as topology, homology theory, complex systems, and biology. The most common method for calculating the Euler characteristic is through simplicial decomposition and the Euler-Poincaré formula. In this study, we introduce a new "subgraph" polynomial, termed the simplex polynomial, and explore some of its properties. Using those properties, we provide a new method for computing the Euler characteristic and prove the existence of the Euler characteristic as an arbitrary integer by constructing the corresponding simplicial complex structure. When the Euler characteristic is 1, we determined a class of corresponding simplicial complex structures. Moreover, for three common network structures, we present the recurrence relations for their simplex polynomials and their corresponding Euler characteristics. Finally, at the end of this study, three basic questions are raised for the interested readers to study deeply.

The big strides seen in generative AI are not based on somewhat obscure algorithms, but due to clearly defined generative principles. The resulting concrete implementations have proven themselves in large numbers of applications. We suggest that it is imperative to thoroughly investigate which of these generative principles may be operative also in the brain, and hence relevant for cognitive neuroscience. In addition, ML research led to a range of interesting characterizations of neural information processing systems. We discuss five examples, the shortcomings of world modeling, the generation of thought processes, attention, neural scaling laws, and quantization, that illustrate how much neuroscience could potentially learn from ML research.

Recent advances in self-supervised learning have attracted significant attention from both machine learning and neuroscience. This is primarily because self-supervised methods do not require annotated supervisory information, making them applicable to training artificial networks without relying on large amounts of curated data, and potentially offering insights into how the brain adapts to its environment in an unsupervised manner. Although several previous studies have elucidated the correspondence between neural representations in deep convolutional neural networks (DCNNs) and biological systems, the extent to which unsupervised or self-supervised learning can explain the human-like acquisition of categorically structured information remains less explored. In this study, we investigate the correspondence between the internal representations of DCNNs trained using a self-supervised contrastive learning algorithm and human semantics and recognition. To this end, we employ a few-shot learning evaluation procedure, which measures the ability of DCNNs to recognize novel concepts from limited exposure, to examine the inter-categorical structure of the learned representations. Two comparative approaches are used to relate the few-shot learning outcomes to human semantics and recognition, with results suggesting that the representations acquired through contrastive learning are well aligned with human cognition. These findings underscore the potential of self-supervised contrastive learning frameworks to model learning mechanisms similar to those of the human brain, particularly in scenarios where explicit supervision is unavailable, such as in human infants prior to language acquisition.

Brain activities often follow an exponential family of distributions. The exponential distribution is the maximum entropy distribution of continuous random variables in the presence of a mean. The memoryless and peakless properties of an exponential distribution impose difficulties for data analysis methods. To estimate the rate parameter of multivariate exponential distribution from a time series of sensory inputs (i.e., observations), we constructed a hierarchical Bayesian inference model based on a variant of general hierarchical Brownian filter (GHBF). To account for the complex interactions among multivariate exponential random variables, the model estimates the second-order interaction of the rate intensity parameter in logarithmic space. Using variational Bayesian scheme, a family of closed-form and analytical update equations are introduced. These update equations also constitute a complete predictive coding framework. The simulation study shows that our model has the ability to evaluate the time-varying rate parameters and the underlying correlation structure of volatile multivariate exponentially distributed signals. The proposed hierarchical Bayesian inference model is of practical utility in analyzing high-dimensional neural activities.

Six variants associated with autism spectrum disorder (ASD) abnormally activate the WASP-family Verprolin-homologous protein (WAVE) regulatory complex (WRC), a critical regulator of actin dynamics. This abnormal activation may contribute to the pathogenesis of this disorder. Using molecular dynamics (MD) simulations, we recently investigated the structural dynamics of wild-type (WT) WRC and R87C, A455P, and Q725R WRC disease-linked variants. Here, by extending MD simulations to I664M, E665K, and D724H WRC, we suggest that all of the mutations weaken the interactions and affect intra-complex allosteric communication between the WAVE1 active C-terminal region (ACR) and the rest of the complex. This might contribute to an abnormal complex activation, a hallmark of WRC-linked ASD. In addition, all mutants but I664M destabilize the ACR V-helix and increase the participation of ACR in large-scale movements. All these features may also abnormally influence the inactive WRC toward a dysfunctional state. We hypothesize that small-molecule ligands counteracting these effects may help restore normal WRC regulation in ASD-related variants.