The dynamic programming high-order Dynamic Bayesian Networks learning for identifying effective connectivity in human brain from fMRI

Neuroimaging-based drug discovery for amyloid clearance therapy in Alzheimer's disease using validated causation analysis

Aging-induced hepatic dysfunction can impair cholesterol metabolism, reducing the availability of cholic acid (CA, bile-acid) in brain. CA is reported to have neuroprotective characteristics in preclinical investigations of Alzheimer's disease (AD). Our aim is to probe the causal-connectivity between the players: amyloid, cholic acid and cerebral-blood-flow, and thereby explore therapeutic applicability in AD. From AD neuroimaging initiative biospecimen platform, we evaluated serum cholic-acid (182 healthy/136 AD individuals). We also assessed 50 healthy/50 Alzheimer's subjects containing MRI-ASL scanning (cerebral blood-flow, CBF) and PET-AV45 scanning (amyloid-load). We performed computational causal connectivity to determine the cause-effect relationship among the parameters. Serum cholic acid in AD subjects substantially decreased to half of controls. Causal-connectivity revealed two novel causative pathways: (i) Decreasing serum CA markedly increased amyloid-load; (ii) Increasing amyloid-load distinctly decreased CBF. We substantiated these two causation pathways respectively with collateral available preclinical observations: (a) increased cholic acid reduces amyloid formation by diminishing gamma-secretase; (b) this decreased amyloid induces capillary-flow enhancement by relaxing vascular pericytes. Indeed, cholic acid can increase amyloid-clearance factor. Neuroimaging-based causal connectivity analysis showed that repositioned pharmacological modulation by cholate derivatives may have appreciable potential as novel window for therapeutic approach to AD. Indicative clinical validation is furnished from available therapeutic trial leads.

Improved brain effective connectivity modelling by dynamic Bayesian networks

BACKGROUND

If brain effective connectivity network modelling (ECN) could be accurately achieved, early diagnosis of neurodegenerative diseases would be possible. It has been observed in the literature that Dynamic Bayesian Network (DBN) based methods are more successful than others. However, DBNs have not been applied easily and tested much due to computational complexity problems in structure learning.

NEW METHOD

This study introduces an advanced method for modelling brain ECNs using improved discrete DBN (Improved- dDBN) which addresses the computational challenges previously limiting DBN application, offering solutions that enable accurate and fast structure modelling.

RESULTS

The practical data and prior sizes needed for the convergence to the globally correct network structure are proved to be much smaller than the theoretical ones using simulated dDBN data. Besides, Hill Climbing is shown to converge to the true structure at a reasonable iteration step size when the appropriate data and prior sizes are used. Finally, importance of data quantization methods are analysed.

COMPARISON WITH EXISTING METHODS

The Improved-dDBN method performs better and robust, when compared to the existing methods for realistic scenarios such as varying graph complexity, various input conditions, noise cases and non-stationary connections. The data used in these tests is the simulated fMRI BOLD time series proposed in the literature.

CONCLUSIONS

Improved-dDBN is a good candidate to be used on real datasets to accelerate developments in brain ECN modelling and neuroscience. Appropriate data and prior sizes can be identified based on the approach proposed in this study for global and fast convergence.

Dynamic Bayesian network structure learning based on an improved bacterial foraging optimization algorithm

With the rapid development of artificial intelligence and data science, Dynamic Bayesian Network (DBN), as an effective probabilistic graphical model, has been widely used in many engineering fields. And swarm intelligence algorithm is an optimization algorithm based on natural selection with the characteristics of distributed, self-organization and robustness. By applying the high-performance swarm intelligence algorithm to DBN structure learning, we can fully utilize the algorithm's global search capability to effectively process time-based data, improve the efficiency of network generation and the accuracy of network structure. This study proposes an improved bacterial foraging optimization algorithm (IBFO-A) to solve the problems of random step size, limited group communication, and the inability to maintain a balance between global and local searching. The IBFO-A algorithm framework comprises four layers. First, population initialization is achieved using a logistics-sine chaotic mapping strategy as the basis for global optimization. Second, the activity strategy of a colony foraging trend is constructed by combining the exploration phase of the Osprey optimization algorithm. Subsequently, the strategy of bacterial colony propagation is improved using a "genetic" approach and the Multi-point crossover operator. Finally, the elimination-dispersal activity strategy is employed to escape the local optimal solution. To solve the problem of complex DBN learning structures due to the introduction of time information, a DBN structure learning method called IBFO-D, which is based on the IBFO-A algorithm framework, is proposed. IBFO-D determines the edge direction of the structure by combining the dynamic K2 scoring function, the designed V-structure orientation rule, and the trend activity strategy. Then, according to the improved reproductive activity strategy, the concept of "survival of the fittest" is applied to the network candidate solution while maintaining species diversity. Finally, the global optimal network structure with the highest score is obtained based on the elimination-dispersal activity strategy. Multiple tests and comparison experiments were conducted on 10 sets of benchmark test functions, two non-temporal and temporal data types, and six data samples of two benchmark 2T-BN networks to evaluate and analyze the optimization performance and structure learning ability of the proposed algorithm under various data types. The experimental results demonstrated that IBFO-A exhibits good convergence, stability, and accuracy, whereas IBFO-D is an effective approach for learning DBN structures from data and has practical value for engineering applications.

A Survey on Brain Effective Connectivity Network Learning

Human brain effective connectivity characterizes the causal effects of neural activities among different brain regions. Studies of brain effective connectivity networks (ECNs) for different populations contribute significantly to the understanding of the pathological mechanism associated with neuropsychiatric diseases and facilitate finding new brain network imaging markers for the early diagnosis and evaluation for the treatment of cerebral diseases. A deeper understanding of brain ECNs also greatly promotes brain-inspired artificial intelligence (AI) research in the context of brain-like neural networks and machine learning. Thus, how to picture and grasp deeper features of brain ECNs from functional magnetic resonance imaging (fMRI) data is currently an important and active research area of the human brain connectome. In this survey, we first show some typical applications and analyze existing challenging problems in learning brain ECNs from fMRI data. Second, we give a taxonomy of ECN learning methods from the perspective of computational science and describe some representative methods in each category. Third, we summarize commonly used evaluation metrics and conduct a performance comparison of several typical algorithms both on simulated and real datasets. Finally, we present the prospects and references for researchers engaged in learning ECNs.

An explainable deep learning framework for characterizing and interpreting human brain states

Deep learning approaches have been widely adopted in the medical image analysis field. However, a most of existing deep learning approaches focus on achieving promising performances such as classification, detection, and segmentation, and much less effort is devoted to the explanation of the designed models. Similarly, in the brain imaging field, many deep learning approaches have been designed and applied to characterize and predict human brain states. However, these models lack interpretation. In response, we propose a novel domain knowledge informed self-attention graph pooling-based (SAGPool) graph convolutional neural network to study human brain states. Specifically, the dense individualized and common connectivity-based cortical landmarks system (DICCCOL, structural brain connectivity profiles) and holistic atlases of functional networks and interactions system (HAFNI, functional brain connectivity profiles) are integrated with the SAGPool model to better characterize and interpret the brain states. Extensive experiments are designed and carried out on the large-scale human connectome project (HCP) Q1 and S1200 dataset. Promising brain state classification performances are observed (e.g., an average of 93.7% for seven-task classification and 100% for binary classification). In addition, the importance of the brain regions, which contributes most to the accurate classification, is successfully quantified and visualized. A thorough neuroscientific interpretation suggests that these extracted brain regions and their importance calculated from self-attention graph pooling layer offer substantial explainability.

Human Health Activity Recognition Algorithm in Wireless Sensor Networks Based on Metric Learning

Wireless sensor network is an ad hoc network with sensing capability. Usually, a large number of sensor nodes are randomly deployed in an unreachable environment or complex area for data collection and transmission, which can realize the perception and monitoring of the target area or specific objects and transmit the obtained data to the remote end of the system. Human health activity recognition algorithm is a hot topic in the field of computer. Based on the small sample problem and the linear indivisibility of real samples encountered in metric learning, this paper proposes a human activity recognition algorithm for wireless sensor networks. Human activity recognition algorithm for wireless sensor networks uses human activity recognition algorithm to solve the singularity of intraclass divergence matrix, so as to reduce the impact of small sample problem. The algorithm maps two different feature spaces to the high-dimensional linearly separable kernel space through the corresponding kernel function, calculates the distance between samples in the two projected feature subspaces to obtain two distance measurement functions, and finally linearly combines them with weights to obtain the final distance measurement function.

Dynamic Bayesian network in infectious diseases surveillance: a simulation study

The surveillance of infectious diseases relies on the identification of dynamic relations between the infectious diseases and corresponding influencing factors. However, the identification task confronts with two practical challenges: small sample size and delayed effect. To overcome both challenges to imporve the identification results, this study evaluated the performance of dynamic Bayesian network(DBN) in infectious diseases surveillance. Specifically, the evaluation was conducted by two simulations. The first simulation was to evaluate the performance of DBN by comparing it with the Granger causality test and the least absolute shrinkage and selection operator (LASSO) method; and the second simulation was to assess how the DBN could improve the forecasting ability of infectious diseases. In order to make both simulations close to the real-world situation as much as possible, their simulation scenarios were adapted from real-world studies, and practical issues such as nonlinearity and nuisance variables were also considered. The main simulation results were: ① When the sample size was large (n = 340), the true positive rates (TPRs) of DBN (≥98%) were slightly higher than those of the Granger causality method and approximately the same as those of the LASSO method; the false positive rates (FPRs) of DBN were averagely 46% less than those of the Granger causality test, and 22% less than those of the LASSO method. ② When the sample size was small, the main problem was low TPR, which would be further aggravated by the issues of nonlinearity and nuisance variables. In the worst situation (i.e., small sample size, nonlinearity and existence of nuisance variables), the TPR of DBN declined to 43.30%. However, it was worth noting that such decline could also be found in the corresponding results of Granger causality test and LASSO method. ③ Sample size was important for identifying the dynamic relations among multiple variables, in this case, at least three years of weekly historical data were needed to guarantee the quality of infectious diseases surveillance. ④ DBN could improve the foresting results through reducing forecasting errors by 7%. According to the above results, DBN is recommended to improve the quality of infectious diseases surveillance.

Genome-wide trait-trait dynamics correlation study dissects the gene regulation pattern in maize kernels

BACKGROUND

Dissecting the genetic basis and regulatory mechanisms for the biosynthesis and accumulation of nutrients in maize could lead to the improved nutritional quality of this crop. Gene expression is regulated at the genomic, transcriptional, and post-transcriptional levels, all of which can produce diversity among traits. However, the expression of most genes connected with a particular trait usually does not have a direct association with the variation of that trait. In addition, expression profiles of genes involved in a single pathway may vary as the intrinsic cellular state changes. To work around these issues, we utilized a statistical method, liquid association (LA) to investigate the complex pattern of gene regulation in maize kernels.

RESULTS

We applied LA to the expression profiles of 28,769 genes to dissect dynamic trait-trait correlation patterns in maize kernels. Among the 1000 LA pairs (LAPs) with the largest LA scores, 686 LAPs were identified conditional correlation. We also identified 830 and 215 LA-scouting leaders based on the positive and negative LA scores, which were significantly enriched for some biological processes and molecular functions. Our analysis of the dynamic co-expression patterns in the carotene biosynthetic pathway clearly indicated the important role of lcyE, CYP97A, ZEP1, and VDE in this pathway, which may change the direction of carotene biosynthesis by controlling the influx and efflux of the substrate. The dynamic trait-trait correlation patterns between gene expression and oil concentration in the fatty acid metabolic pathway and its complex regulatory network were also assessed. 23 of 26 oil-associated genes were correlated with oil concentration conditioning on 580 LA-scoutinggenes, and 5% of these LA-scouting genes were annotated as enzymes in the oil metabolic pathway.

CONCLUSIONS

By focusing on the carotenoid and oil biosynthetic pathways in maize, we showed that a genome-wide LA analysis provides a novel and effective way to detect transcriptional regulatory relationships. This method will help us understand the biological role of maize kernel genes and will benefit maize breeding programs.