Long-distance Connections Research Articles

Medical Image Classification with a Hybrid SSM Model Based on CNN and Transformer

Medical image classification, a pivotal task for diagnostic accuracy, poses unique challenges due to the intricate and variable nature of medical images compared to their natural counterparts. While Convolutional Neural Networks (CNNs) and Transformers are prevalent in this domain, each architecture has its drawbacks. CNNs, despite their strength in local feature extraction, fall short in capturing global context, whereas Transformers excel at global information but can overlook fine-grained details. The integration of CNNs and Transformers in a hybrid model aims to bridge this gap by enabling simultaneous local and global feature extraction. However, this approach remains constrained in its capacity to model long-range dependencies, thereby hindering the efficient extraction of distant features. To address these issues, we introduce the MambaConvT model, which employs a state-space approach. It begins by locally processing input features through multi-core convolution, enhancing the extraction of deep, discriminative local details. Next, depth-separable convolution with a 2D selective scanning module (SS2D) is employed to maintain a global receptive field and establish long-distance connections, capturing the fine-grained features. The model then combines hybrid features for comprehensive feature extraction, followed by global feature modeling to emphasize on global detail information and optimize feature representation. This paper conducts thorough performance experiments on different algorithms across four publicly available datasets and two private datasets. The results demonstrate that MambaConvT outperforms the latest classification algorithms in terms of accuracy, precision, recall, F1 score, and AUC value ratings, achieving superior performance in the precise classification of medical images.

Exploring Complicated Search Spaces With Interleaving-Free Sampling.

Conventional neural architecture search (NAS) algorithms typically work on search spaces with short-distance node connections. We argue that such designs, though safe and stable, are obstacles to exploring more effective network architectures. In this brief, we explore the search algorithm upon a complicated search space with long-distance connections and show that existing weight-sharing search algorithms fail due to the existence of interleaved connections (ICs). Based on the observation, we present a simple-yet-effective algorithm, termed interleaving-free neural architecture search (IF-NAS). We further design a periodic sampling strategy to construct subnetworks during the search procedure, avoiding the ICs to emerge in any of them. In the proposed search space, IF-NAS outperforms both random sampling and previous weight-sharing search algorithms by significant margins. It can also be well-generalized to the microcell-based spaces. This study emphasizes the importance of macrostructure and we look forward to further efforts in this direction. The code is available at github.com/sunsmarterjie/IFNAS.

Pottery décor as networks on the Middle Niger, Mali

ABSTRACT This paper reports on an attempt to analyse decorative techniques on archaeological pottery from the Middle Niger region of West Africa as knowledge exchange networks. The twelfth-century AD state of this network is modelled and analysed and then compared to its ninth-century counterpart. The analysis shows how knowledge about pottery decoration was propagated within networks of different sizes and intensities, from local, well-connected networks to more loosely tied long-distance connections. The article shows how the perspective of network analysis differs from, and can fruitfully complement, previous treatments of ceramic décor distribution in the region. It adds a novel perspective to what the distribution of archaeological ceramics in this region reflects and contributes to the generation of hypotheses that can be further tested by fieldwork.

Click train elicited local gamma synchrony: differing performance and pharmacological responsivity of primary auditory and prefrontal cortices

Auditory neural networks in the brain naturally entrain to rhythmic stimuli. Such synchronization is an accessible index of local network performance as captured by EEG. Across species, click trains delivered ∼ 40 Hz show strong entrainment with primary auditory cortex (Actx) being a principal source. Imaging studies have revealed additional cortical sources, but it is unclear if they are functionally distinct. Since auditory processing evolves hierarchically, we hypothesized that local synchrony would differ between between primary and association cortices. In female SD rats (N = 12), we recorded 40 Hz click train-elicited gamma oscillations using epidural electrodes situated at two distinct sites; one above the prefrontal cortex (PFC) and another above the Actx, after dosing with saline (1 ml/kg, sc) or the NMDA antagonist, MK801 (0.025, 0.05 or 0.1 mpk), in a blocked crossover design. Post-saline, both regions showed a strong 40 Hz auditory steady state response (ASSR). The latencies for the N1 response were ∼ 16 ms (Actx) and ∼ 34 ms (PFC). Narrow band (38–42 Hz) gamma oscillations appeared rapidly (<40 ms from stim onset at Actx but in a more delayed fashion (∼200 ms) at PFC. MK801 augmented gamma synchrony at Actx while dose-dependently disrupting at the PFC. Event-related gamma (but not beta) coherence, an index of long-distance connectivity, was disrupted by MK801. In conclusion, local network gamma synchrony in a higher order association cortex performs differently from that of the primary auditory cortex. We discuss these findings in the context of evolving sound processing across the cortical hierarchy.

Human movement and environmental barriers shape the emergence of dengue

Understanding how emerging infectious diseases spread within and between countries is essential to contain future pandemics. Spread to new areas requires connectivity between one or more sources and a suitable local environment, but how these two factors interact at different stages of disease emergence remains largely unknown. Further, no analytical framework exists to examine their roles. Here we develop a dynamic modelling approach for infectious diseases that explicitly models both connectivity via human movement and environmental suitability interactions. We apply it to better understand recently observed (1995-2019) patterns as well as predict past unobserved (1983-2000) and future (2020-2039) spread of dengue in Mexico and Brazil. We find that these models can accurately reconstruct long-term spread pathways, determine historical origins, and identify specific routes of invasion. We find early dengue invasion is more heavily influenced by environmental factors, resulting in patchy non-contiguous spread, while short and long-distance connectivity becomes more important in later stages. Our results have immediate practical applications for forecasting and containing the spread of dengue and emergence of new serotypes. Given current and future trends in human mobility, climate, and zoonotic spillover, understanding the interplay between connectivity and environmental suitability will be increasingly necessary to contain emerging and re-emerging pathogens.

A Fast Feedforward Small-World Neural Network for Nonlinear System Modeling.

It is well-documented that cross-layer connections in feedforward small-world neural networks (FSWNNs) enhance the efficient transmission for gradients, thus improving its generalization ability with a fast learning. However, the merits of long-distance cross-layer connections are not fully utilized due to the random rewiring. In this study, aiming to further improve the learning efficiency, a fast FSWNN (FFSWNN) is proposed by taking into account the positive effects of long-distance cross-layer connections, and applied to nonlinear system modeling. First, a novel rewiring rule by giving priority to long-distance cross-layer connections is proposed to increase the gradient transmission efficiency when constructing FFSWNN. Second, an improved ridge regression method is put forward to determine the initial weights with high activation for the sigmoidal neurons in FFSWNN. Finally, to further improve the learning efficiency, an asynchronous learning algorithm is designed to train FFSWNN, with the weights connected to the output layer updated by the ridge regression method and other weights by the gradient descent method. Several experiments are conducted on four benchmark datasets from the University of California Irvine (UCI) machine learning repository and two datasets from real-life problems to evaluate the performance of FFSWNN on nonlinear system modeling. The results show that FFSWNN has significantly faster convergence speed and higher modeling accuracy than the comparative models, and the positive effects of the novel rewiring rule, the improved weight initialization, and the asynchronous learning algorithm on learning efficiency are demonstrated.

Lithic projectile technology in the western Late Epigravettian: The case study of north-eastern Italy

Projectile implements are one of the most frequent transformed lithic blanks within the Upper Paleolithic assemblages. Since the first taxonomic studies, their morphological variability has been used as primary fossil director to define specific regional trends and diachronic phases. However, an exhaustive analysis of manufacturing methods and techniques applied to shape this gear category is extremely rare. In fact, during the Late Epigravettian, except for a certain typological variability, the lack of detailed studies resulted in an apparent homogeneity in armatures production. In this paper, a technological and traceological approach applied to reconstruct the whole chaîne opératoire of armatures manufacture in three Northern Italian sites dated between 17,000 and 12,000 cal BP demonstrates important variations in how projectile implements are fashioned, concerning shape, production methods and retouch techniques. Moreover, a large-scale comparison suggests how renewals in projectile implements production in North-Eastern Italy may result from long-distance connections between human groups over the western Late Epigravettian territory.

Distributed representations of prediction error signals across the cortical hierarchy are synergistic

A relevant question concerning inter-areal communication in the cortex is whether these interactions are synergistic. Synergy refers to the complementary effect of multiple brain signals conveying more information than the sum of each isolated signal. Redundancy, on the other hand, refers to the common information shared between brain signals. Here, we dissociated cortical interactions encoding complementary information (synergy) from those sharing common information (redundancy) during prediction error (PE) processing. We analyzed auditory and frontal electrocorticography (ECoG) signals in five common awake marmosets performing two distinct auditory oddball tasks and investigated to what extent event-related potentials (ERP) and broadband (BB) dynamics encoded synergistic and redundant information about PE processing. The information conveyed by ERPs and BB signals was synergistic even at lower stages of the hierarchy in the auditory cortex and between auditory and frontal regions. Using a brain-constrained neural network, we simulated the synergy and redundancy observed in the experimental results and demonstrated that the emergence of synergy between auditory and frontal regions requires the presence of strong, long-distance, feedback, and feedforward connections. These results indicate that distributed representations of PE signals across the cortical hierarchy can be highly synergistic.

CSDFormer: A cloud and shadow detection method for landsat images based on transformer

Cloud and shadow (CS) detection is crucial prerequisite for application of remote sensing images. Current deep learning-based detection algorithms mainly employ Convolutional Neural Networks (CNNs). However, the local receptive field in CNNs cannot effectively capture global contextual information, which hinders accurate characterization of the dependency between clouds and shadows. In vision Transformers, self-attention mechanisms can effectively capture the long-distance dependencies between different regions in an image. Inspired by this, this paper proposed a new CS Detection algorithm based on a Transformer, called CSDFormer. Specifically, we exclusively employed a hierarchical Transformer structure in the encoder stage to extract features of CS. Each Transformer layer contains several multi-head self-attention mechanisms for calculating pixel-wise long-distance connectivity. The designed structure enables the Transformer to better extract global context information, which helps to strengthen the comprehension of the semantic relationships between clouds and shadows. Benefiting from the global feature extraction capability of the encoder stage, we employed several simple multilayer perceptron layers for multi-scale feature map fusion and pixel classification in the decoder stage. The proposed CSDFormer was validated using 898 Landsat 8 Biome images with 512 × 512 pixels, producing an overall accuracy of 95.28 % and a mean intersection over union of 84.08 %, outperforming three state-of-the-art CNN-based algorithms. CSDFormer is consistently more accurate in detection of both clouds and shadows. Owing to the parallel computing capability of the self-attention mechanism, CSDFormer is computationally more efficient than the three CNN-based benchmark methods. For the input spectral bands, the performance of CSDFormer produced can be further enhanced with additional thermal infrared bands.

Social Organization, Intersections, and Interactions in Bronze Age Sardinia. Reading Settlement Patterns in the Area of Sarrala with the Contribution of Applied Sciences

Abstract Bronze Age sites in the coastal area of Sarrala, in Eastern Sardinia, have been subjected to survey and excavation over the last half-century. The study area, whose social and economic evolution and changing scales of interactions are traced through settlement patterns and building analysis, is interpreted in light of more general trends suggested by stable isotopes, archaeogenetics, and paleoclimatology. The local picture of progressive demographic growth and infilling of the landscape, with a subsequent concentration of population and labor, follows a sequence widely detected in Sardinia. More specific identifiable aspects include a comparatively higher fragmentation/competition (ratio of complex vs simple nuraghes; ratio of tombs vs nuraghes) and a consistent pattern in the distribution of non-local building materials in the latest phase at the sites showing archaic features, taken as a clue of a long-lasting authority at select sites. These elements are compatible with organized pastoral exploitation of the available territory, structured according to patrilocality and closeness to ancestral lineages and residences. The interplay of internal dynamics relative to constraints and opportunities is discussed, such as climate change and long-distance trade connections, with possible implications for interpreting Nuragic society.

A phase-shifting anterior-posterior network organizes global phase relations.

Prior research has identified a variety of task-dependent networks that form through inter-regional phase-locking of oscillatory activity that are neural correlates of specific behaviors. Despite ample knowledge of task-specific functional networks, general rules governing global phase relations have not been investigated. To discover such general rules, we focused on phase modularity, measured as the degree to which global phase relations in EEG comprised distinct synchronized clusters interacting with one another at large phase lags. Synchronized clusters were detected with a standard community-detection algorithm, and the degree of phase modularity was quantified by the index q. Notably, we found that the mechanism controlling phase modularity is remarkably simple. A network comprising anterior-posterior long-distance connectivity coherently shifted phase relations from low-angles (|Δθ| < π/4) in low-modularity states (bottom 5% in q) to high-angles (|Δθ| > 3π/4) in high-modularity states (top 5% in q), accounting for fluctuations in phase modularity. This anterior-posterior network may play a fundamental functional role as (1) it controls phase modularity across a broad range of frequencies (3-50 Hz examined) in different behavioral conditions (resting with the eyes closed or watching a silent nature video) and (2) neural interactions (measured as power correlations) in beta-to-gamma bands were consistently elevated in high-modularity states. These results may motivate future investigations into the functional roles of phase modularity as well as the anterior-posterior network that controls it.

Molecular and thermodynamic determinants of self-assembly and hetero-oligomerization in the enterobacterial thermo-osmo-regulatory protein H-NS.

Environmentally regulated gene expression is critical for bacterial survival under stress conditions, including extremes in temperature, osmolarity and nutrient availability. Here, we dissect the thermo- and osmo-responsory behavior of the transcriptional repressor H-NS, an archetypal nucleoid-condensing sensory protein, ubiquitous in enterobacteria that infect the mammalian gut. Through experiments and thermodynamic modeling, we show that H-NS exhibits osmolarity, temperature and concentration dependent self-association, with a highly polydisperse native ensemble dominated by monomers, dimers, tetramers and octamers. The relative population of these oligomeric states is determined by an interplay between dimerization and higher-order oligomerization, which in turn drives a competition between weak homo- versus hetero-oligomerization of protein-protein and protein-DNA complexes. A phosphomimetic mutation, Y61E, fully eliminates higher-order self-assembly and preserves only dimerization while weakening DNA binding, highlighting that oligomerization is a prerequisite for strong DNA binding. We further demonstrate the presence of long-distance thermodynamic connectivity between dimerization and oligomerization sites on H-NS which influences the binding of the co-repressor Cnu, and switches the DNA binding mode of the hetero-oligomeric H-NS:Cnu complex. Our work thus uncovers important organizational principles in H-NS including a multi-layered thermodynamic control, and provides a molecular framework broadly applicable to other thermo-osmo sensory proteins that employ similar mechanisms to regulate gene expression.

Multigraph approximate-representation learning for hyperspectral band selection

ABSTRACT Unsupervised hyperspectral image (HSI) band selection methods have been attracting ever-increasing attention. However, the local structural features captured by most of the existing methods suffer from a certain degree of risk of feature ambiguity. Moreover, these methods typically have difficulty preserving the latent manifold structure of the HSI in selected subsets of the bands. To address these problems, firstly, the graph approximate representation learning (GARL) model is proposed in this article by incorporating graph-regularized sparse coding into self-representation learning to preserve the local manifold structure information of HSI in the selected subset of bands. Meanwhile, to alleviate the ambiguity of the local graph structures, we design a local-global feature relationship expression model to construct the latent long-distance contextual connectivity (LDCC) graphs between the obtained local graph structures. Then, to preserve HSI’s local and non-local manifold structure information as much as possible in the selected subset of bands, a multigraph approximate representation learning (MGARL) model is proposed in this article by incorporating the obtained LDCC graph into the proposed GARL model. Next, we design a solution method to solve the proposed formulation. Finally, extensive experimental results on four datasets reveal the promising performance of MGARL over some state-of-the-art methods.

An application of small-world network on predicting the behavior of infectious disease on campus

Networks haven been widely used to understand the spread of infectious disease. This study examines the properties of small-world networks in modeling infectious disease on campus. Two different small-world models are developed and the behaviors of infectious disease in the models are observed through numerical simulations. The results show that the behavior pattern of infectious disease in a small-world network is different from those in a regular network or a random network. The spread of the infectious disease increases as the proportion of long-distance connections p increasing, which indicates that reducing the contact among people is an effective measure to control the spread of infectious disease. The probability of node position exchange in a network (p2) had no significant effect on the spreading speed, which suggests that reducing human mobility in closed environments does not help control infectious disease. However, the spreading speed is proportional to the number of shared nodes (s), which means reducing connections between different groups and dividing students into separate sections will help to control infectious disease. In the end, the simulating speed of the small-world network is tested and the quadratic relationship between simulation time and the number of nodes may limit the application of the SW network in areas with large populations.

Long-range connections damage in white matter hyperintensities affects information processing speed.

White matter hyperintensities, one of the major markers of cerebral small vessel disease, disrupt the integrity of neuronal networks and ultimately contribute to cognitive dysfunction. However, a deeper understanding of how white matter hyperintensities related to the connectivity patterns of brain hubs at the neural network level could provide valuable insights into the relationship between white matter hyperintensities and cognitive dysfunction. A total of 36 patients with moderate to severe white matter hyperintensities (Fazekas score ≥ 3) and 34 healthy controls underwent comprehensive neuropsychological assessments and resting-state functional MRI scans. The voxel-based graph-theory approach-functional connectivity strength was employed to systematically investigate the topological organization of the whole-brain networks. The white matter hyperintensities patients performed significantly worse than the healthy controls in episodic memory, executive function and information processing speed. Additionally, we found that white matter hyperintensities selectively affected highly connected hub regions, predominantly involving the medial and lateral prefrontal, precuneus, inferior parietal lobule, insula and thalamus. Intriguingly, this impairment was connectivity distance-dependent, with the most prominent disruptions observed in long-range connections (e.g. 100-150 mm). Finally, these disruptions of hub connectivity (e.g. the long-range functional connectivity strength in the left dorsolateral prefrontal cortex) positively correlated with the cognitive performance in white matter hyperintensities patients. Our findings emphasize that the disrupted hub connectivity patterns in white matter hyperintensities are dependent on connection distance, especially longer-distance connections, which in turn predispose white matter hyperintensities patients to worse cognitive function.

Disrupted topological organization of functional brain networks in traumatic axonal injury

Traumatic axonal injury (TAI) may result in the disruption of brain functional networks and is strongly associated with cognitive impairment. However, the neural mechanisms affecting the neurocognitive function after TAI remain to be elucidated. We collected the resting-state functional magnetic resonance imaging data from 28 patients with TAI and 28 matched healthy controls. An automated anatomical labeling atlas was used to construct a functional brain connectome. We utilized a graph theoretical approach to investigate the alterations in global and regional network topologies, and network-based statistics analysis was utilized to localize the connected networks more precisely. The current study revealed that patients with TAI and healthy controls both showed a typical small-world topology of the functional brain networks. However, patients with TAI exhibited a significantly lower local efficiency compared to healthy controls, whereas no significant difference emerged in other small-world properties (Cp, Lp, γ, λ, and σ) and global efficiency. Moreover, patients with TAI exhibited aberrant nodal centralities in some regions, including the frontal lobes, parietal lobes, caudate nucleus, and cerebellum bilaterally, and right olfactory cortex. The network-based statistics results showed alterations in the long-distance functional connections in the subnetwork in patients with TAI, involving these brain regions with significantly altered nodal centralities. These alterations suggest that brain networks of individuals with TAI present aberrant topological attributes that are associated with cognitive impairment, which could be potential biomarkers for predicting cognitive dysfunction and help understanding the neuropathological mechanisms in patients with TAI.

Functional connectivity in autism spectrum disorder evaluated using rs-fMRI and DKI.

We evaluated functional connectivity (FC) in patients with adult autism spectrum disorder (ASD) using resting-state functional MRI (rs-fMRI) and diffusion kurtosis imaging (DKI). We acquired rs-fMRI data from 33 individuals with ASD and 33 healthy controls (HC) and DKI data from 18 individuals with ASD and 17 HC. ASD showed attenuated FC between the right frontal pole (FP) and the bilateral temporal fusiform cortex (TFusC) and enhanced FC between the right thalamus and the bilateral inferior division of lateral occipital cortex, and between the cerebellar vermis and the right occipital fusiform gyrus (OFusG) and the right lingual gyrus, compared with HC. ASD demonstrated increased axial kurtosis (AK) and mean kurtosis (MK) in white matter (WM) tracts, including the right anterior corona radiata (ACR), forceps minor (FM), and right superior longitudinal fasciculus (SLF). In ASD, there was also a significant negative correlation between MK and FC between the cerebellar vermis and the right OFusG in the corpus callosum, FM, right SLF and right ACR. Increased DKI metrics might represent neuroinflammation, increased complexity, or disrupted WM tissue integrity that alters long-distance connectivity. Nonetheless, protective or compensating adaptations of inflammation might lead to more abundant glial cells and cytokine activation effectively alleviating the degeneration of neurons, resulting in increased complexity. FC abnormality in ASD observed in rs-fMRI may be attributed to microstructural alterations of the commissural and long-range association tracts in WM as indicated by DKI.

Evolutionary algorithm with a regression model for multiobjective minimization of systemic risk in financial systems

AbstractThis paper addresses a problem of systemic risk minimization in which the optimization algorithm has to simultaneously minimize the number of companies affected by a wave of bankruptcies simulated on a graph as well as the level of reserves the companies keep to avoid going bankrupt. A MOEA/D-NN algorithm (where NN stands for a neural network) is proposed, which optimizes parameters of a machine learning model (a neural network) used in turn to determine the level of reserves the companies keep, based on several attributes describing each node in the graph. In the experiments, the proposed MOEA/D-NN algorithm was found to outperform comparison methods: evolutionary algorithms optimizing the level of reserves for all companies and a method based on the training of neural networks on a dataset previously collected by an evolutionary algorithm solving “training” instances of the optimization problem. The neural networks optimized by MOEA/D-NN were also tested on problem instances based on REDS graphs generated using varying values of R, E, and S parameters and were found to be applicable to these instances for certain ranges of parameters. The R parameter controlling the possibility of generating long-distance connections was found to have a bigger impact on the performance of the optimized neural networks than the other two parameters.

In the footsteps of Ohthere: biomolecular analysis of early Viking Age hair combs from Hedeby (Haithabu)

Hedeby was the largest town in the Viking North. Investigations have identified imports at the site from central and northern Scandinavia revealing long-distance connections. The chronology of this trade, however, is unclear. Here, the authors use a typological-biomolecular approach to examine connections during the early Viking Age. The application of ZooMS to an assemblage of antler combs, stylistically dated to the ninth century AD, reveals nearly all were made of reindeer antler. As most craft production waste from Hedeby comprises red deer antler, it is argued that these combs were manufactured elsewhere, perhaps hundreds of kilometres further north. The results have implications for understanding of production and regional connectivity in early medieval Scandinavia.

Recent Developments in the Archaeology of Long-Distance Connections Across the Ancient Indian Ocean

Scholarship on the ancient Indian Ocean, which stretches deep into the previous century, is available from an array of academic disciplines including but not limited to history, archaeology, epigraphy, numismatics, art history, and materials science. It spans from prehistory to the present era and includes evidence ranging from the Mediterranean to East Asia. What binds together the world of Indian Ocean research is an enduring interest in the complex maritime-based links crosscutting this space and—for archaeologists—the movements of cultural elements (objects, ideas, people, etc.) that have left behind some material trace. Recent field projects and materials science studies have greatly expanded this material database, refining (and sometimes challenging) traditional interpretations about Indian Ocean maritime relations. This review presents a streamlined perspective, focusing on recent archaeological contributions about long-distance interregional connections across the Indian Ocean from 500 bce to 1000 ce.