SEARCH · Search PubMed

Results for “Bayesian networks”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 631 records · Page 35Linked to original sources

PubMed2003

Manhattan world: orientation and outlier detection by Bayesian inference.

This letter argues that many visual scenes are based on a "Manhattan" three-dimensional grid that imposes regularities on the image statistics. We construct a Bayesian model that implements this assumption and estimates the viewer orientation relative to the Manhattan grid. For many images, these estimates are good approximations to the viewer orientation (as estimated manually by the authors). These estimates also make it easy to detect outlier structures that are unaligned to the grid. To determine the applicability of the Manhattan world model, we implement a null hypothesis model that assumes that the image statistics are independent of any three-dimensional scene structure. We then use the log-likelihood ratio test to determine whether an image satisfies the Manhattan world assumption. Our results show that if an image is estimated to be Manhattan, then the Bayesian model's estimates of viewer direction are almost always accurate (according to our manual estimates), and vice versa.

Artificial Intelligence↗

PubMed2026

The efficacy of non-invasive brain stimulation interventions in obsessive-compulsive disorder management: A network meta-analysis of randomized controlled trials.

Non-invasive brain stimulation (NIBS) has been widely used as an alternative treatment for obsessive compulsive disorder (OCD). However, the most effective NIBS parameters are unclear. To compare the efficacy of NIBS in OCD. We conducted a systematic review and network meta-analyses (NMA) to combine direct and indirect comparisons of NIBS.Systematic searches were conducted in Cochrane CENTRAL, EMBASE, PubMed, and Web of Science from inception to June 20, 2025. Forty-two randomized sham-controlled trials (n = 1456) were included. All statistical analyses were conducted with R statistical software. Bayesian NMAs mainly using the BUGSnet package and gemtc package. Five NIBS protocols produced statistically significant reductions in Yale-Brown Obsessive Compulsive Scale (Y-BOCS) scores compared with sham stimulation: high-frequency rTMS over the FzFCz (Hf-rTMS-FzFCz; MD -11.77, 95% CrI -20.62 to -3.09), low-frequency rTMS over F3F4 (Lf-rTMS-F3F4; MD -9.93, 95% CrI -18.07 to -1.65), low-frequency rTMS over FCz (Lf-rTMS-FCz; MD -3.25, 95% CrI -6.06 to -0.40), high-frequency deep TMS over FzFC (Hf-dTMS-FzFC; MD -6.48, 95% CrI -12.32 to -0.50), and 2 mA anodal tDCS over F3 with cathodal over Fp2 (MD -9.34, 95% CrI -16.01 to -3.03).For secondary outcomes, high-frequency deep rTMS over FzFCz produced the largest reduction both in depressive symptoms (SMD -1.24, 95% CrI -1.92 to -0.55) and anxiety scores (SMD -1.88, 95% CrI -2.62 to -1.11), but had no effect on Clinical Global Impression-Severity (CGI-S) scores.Specific NIBS protocols are safe and effective adjunctive treatments for OCD, with promising yet inconclusive improvements in comorbid depressive symptoms. Further high-quality, head-to-head trials are needed.

Humans↗

PubMed2006

A Bayesian approach to determining connectivity of the human brain.

Recent work regarding the analysis of brain imaging data has focused on examining functional and effective connectivity of the brain. We develop a novel descriptive and inferential method to analyze the connectivity of the human brain using functional MRI (fMRI). We assess the relationship between pairs of distinct brain regions by comparing expected joint and marginal probabilities of elevated activity of voxel pairs through a Bayesian paradigm, which allows for the incorporation of previously known anatomical and functional information. We define the relationship between two distinct brain regions by measures of functional connectivity and ascendancy. After assessing the relationship between all pairs of brain voxels, we are able to construct hierarchical functional networks from any given brain region and assess significant functional connectivity and ascendancy in these networks. We illustrate the use of our connectivity analysis using data from an fMRI study of social cooperation among women who played an iterated "Prisoner's Dilemma" game. Our analysis reveals a functional network that includes the amygdala, anterior insula cortex, and anterior cingulate cortex, and another network that includes the ventral striatum, orbitofrontal cortex, and anterior insula. Our method can be used to develop causal brain networks for use with structural equation modeling and dynamic causal models.

Adult↗

PubMed2004

Molecular analysis of the dengue virus type 1 and 2 in Brazil based on sequences of the genomic envelope-nonstructural protein 1 junction region.

The genomic sequences of the Envelope-Non-Structural protein 1 junction region (E/NS1) of 84 DEN-1 and 22 DEN-2 isolates from Brazil were determined. Most of these strains were isolated in the period from 1995 to 2001 in endemic and regions of recent dengue transmission in São Paulo State. Sequence data for DEN-1 and DEN-2 utilized in phylogenetic and split decomposition analyses also include sequences deposited in GenBank from different regions of Brazil and of the world. Phylogenetic analyses were done using both maximum likelihood and Bayesian approaches. Results for both DEN-1 and DEN-2 data are ambiguous, and support for most tree bipartitions are generally poor, suggesting that E/NS1 region does not contain enough information for recovering phylogenetic relationships among DEN-1 and DEN-2 sequences used in this study. The network graph generated in the split decomposition analysis of DEN-1 does not show evidence of grouping sequences according to country, region and clades. While the network for DEN-2 also shows ambiguities among DEN-2 sequences, it suggests that Brazilian sequences may belong to distinct subtypes of genotype III.

Amino Acid Sequence↗

PubMed2003

Multivariate autoregressive modeling of fMRI time series.

We propose the use of multivariate autoregressive (MAR) models of functional magnetic resonance imaging time series to make inferences about functional integration within the human brain. The method is demonstrated with synthetic and real data showing how such models are able to characterize interregional dependence. We extend linear MAR models to accommodate nonlinear interactions to model top-down modulatory processes with bilinear terms. MAR models are time series models and thereby model temporal order within measured brain activity. A further benefit of the MAR approach is that connectivity maps may contain loops, yet exact inference can proceed within a linear framework. Model order selection and parameter estimation are implemented by using Bayesian methods.

Algorithms↗

PubMed2003

Neural representation of probabilistic information.

It has been proposed that populations of neurons process information in terms of probability density functions (PDFs) of analog variables. Such analog variables range, for example, from target luminance and depth on the sensory interface to eye position and joint angles on the motor output side. The requirement that analog variables must be processed leads inevitably to a probabilistic description, while the limited precision and lifetime of the neuronal processing units lead naturally to a population representation of information. We show how a time-dependent probability density rho(x; t) over variable x, residing in a specified function space of dimension D, may be decoded from the neuronal activities in a population as a linear combination of certain decoding functions phi(i)(x), with coefficients given by the N firing rates a(i)(t) (generally with D << N). We show how the neuronal encoding process may be described by projecting a set of complementary encoding functions phi;(i)(x) on the probability density rho(x; t), and passing the result through a rectifying nonlinear activation function. We show how both encoders phi;(i)(x) and decoders phi(i)(x) may be determined by minimizing cost functions that quantify the inaccuracy of the representation. Expressing a given computation in terms of manipulation and transformation of probabilities, we show how this representation leads to a neural circuit that can carry out the required computation within a consistent Bayesian framework, with the synaptic weights being explicitly generated in terms of encoders, decoders, conditional probabilities, and priors.

Models, Neurological↗

PubMed2005

Bayesian parameter estimates of nelfinavir and its active metabolite, hydroxy-tert-butylamide, in infants perinatally infected with human immunodeficiency virus type 1.

The objective of the present study was to develop a population pharmacokinetic model for nelfinavir mesylate (NFV) and nelfinavir hydroxy-tert-butylamide (M8), the most abundant metabolite of NFV, in infants vertically infected with human immunodeficiency virus type 1 and participating in the Paediatric European Network for Treatment of AIDS 7 study. Plasma NFV concentrations were determined during repeated NFV administrations (two to three times a day). Eighteen infants younger that age 2 years participated in this study. The doses administered ranged from 71 to 203 mg/kg of body weight/day. Pharmacokinetic parameter estimates were obtained by a compartmental approach by using a kinetic model to simultaneously fit NFV and M8 (active metabolite) concentrations. M8 was shown to be formation rate limited and was characterized by first-order rate constants of formation and elimination. Body weight was found to be a more appropriate predictor than age of the changes in (i) the rate of metabolism, (ii) the elimination rate constant of NFV, and (iii) NFV clearance. Population parameters were computed to account for the relationship between the rate of metabolism and body weight. The estimated NFV and M8 elimination half-lives were 4.3 and 2.04 h, respectively. The estimated NFV clearance was 2.13 liters/h/kg. The M8 concentration-to-NFV concentration ratio was 0.64 +/- 0.44. In conclusion, the population pharmacokinetic model describing the dispositions of NFV and M8 should facilitate the design of future studies to elucidate the relative contributions of the parent compound and M8 to the pharmacological and toxic effects of NFV therapy.

Aging↗

PubMed2000

Antibiotics and return visits for respiratory illness: a comparison of pooled versus hierarchical statistical methods.

BACKGROUND: Antibiotic prescribing for respiratory illness has been associated with small reductions in return visits in an analysis of a large practice-based network. In this study, we apply hierarchical analytical methods that account for the clustering of patients by practices to identify whether antibiotic prescribing by primary care physicians reduces subsequent visits for 6 acute respiratory illnesses-upper respiratory infection, pharyngitis, bronchitis, otitis media, sinusitis, and cough. METHODS: The study data came from 318 family physicians and internists in 45 practices in the Practice Partner Research Network from January 1995 through December 1996, with 255,564 active patients. Patients treated with antibiotics were compared with those who were not on the frequency of revisit within the next 14 days. A simple pooling model and 3 hierarchical statistical models (fixed-effects, random-effects, and Bayesian) were used to compare the odds-ratios for return visits. RESULTS: Statistically significant results were found only for bronchitis and sinusitis by the hierarchical models, but the simple pooling model produced statistically significant results for all study conditions. CONCLUSION: We conclude that antibiotics may reduce return visits for patients with bronchitis and sinusitis, but not for patients with other respiratory illness (upper respiratory infection, pharyngitis, otitis media, or cough). Studies of large clinical databases should use methods of analysis that account for the grouping of patients by practice to avoid false positive associations (type I errors.)

Academic Medical Centers↗

PubMedDate not supplied

Simultaneous determination of thiocyanate and salicylate by a combined UV-spectrophotometric detection principal component artificial neural network.

A modified principle component artificial neural network (PC-ANN) model is developed for simultaneous determination of thiocyanate and salycilate concentration after passing through the bulk of a liquid membrane by tri-phenyl benzyl phosphonium chloride. All calibration, and test samples data were obtained using UV-Vis spectrophotometer. In this way, a modified PC-ANN consisting of three layers of nodes was trained by combination of Bayesian-Levenberg-Marquardt as training rule. Sigmoid and liner transfer functions were used in the hidden and output layers respectively to facilitate nonlinear calibration. The model could accurately estimate the concentration of components with acceptable precision and accuracy, for mixtures. The PC-ANN model exhibits a good ability for the simultaneous determination of the thiocyanate and salycilate in concentration range 0.5 x 10(-4) mol.l(-1) up to 5.0 x 10(-4) mol.l(-1) with Root Mean square error (2.22% and 2.20%, for thiocyanate and salycilate, respectively) and high correlation coefficients (R2= 0.998 or greater). Results obtained with modified trained PC-ANN were compared with stepwise linear regression (SMLR) model. Validation of the two models shows a better ability in estimation of the modified PC-ANN as compared with the SMLR model (MSRE given are 3.12%, 6.31%.).

Neural Networks, Computer↗

PubMed2026

Comparative effects of pharmacological interventions in the prophylactic treatment of tension-type headache: systematic review and network meta-analysis.

BACKGROUND: Tension-type headache (TTH) is the most common neurological disorder. The comparative effect of pharmacological interventions for TTH prophylaxis remains unclear. We aimed to assess the comparative effects of pharmacological interventions in the prophylactic treatment of TTH. METHODS: Ovid Medline, Embase, and Cochrane were searched from inception to 12 December, 2025. Randomized controlled trials (RCTs) of medications compared to placebo or another medication for preventing TTH were included. The primary outcome was headache days per month. A Bayesian random-effect model was employed as the primary analysis of chronic TTH. RESULTS: Thirty-five RCTs were included, 33 (88.6%) RCTs involved chronic TTH patients, and 24 RCTs provided available data for meta-analysis. Amitriptyline 100 mg presented more reduction of monthly headache days than placebo at 4 and 8 weeks (4 weeks: MD -6.59, 95% CrI -11.22 to -0.64; 8 weeks: MD -6.14, 95% CrI -10.27 to -0.87). BTX-A 100 U can reduce monthly headache days (MD -3.79, 95% CrI -7.16 to -0.33). Amitriptyline 100 mg was the highest-ranked treatment for monthly headache days at 4 (SUCRA 0.85), 8 (SUCRA 0.85), and 24 (SUCRA 0.87) weeks; 12 weeks was lidocaine 25 ml (SUCRA 0.75). Amitriptyline 100 mg and BTX-A 500 U showed a higher adverse event rate than placebo. CONCLUSION: Amitriptyline 100 mg and BTX-A 100 U may be options to reduce monthly headache days in patients with chronic TTH. Given the low to very low certainty of evidence, high risk of bias, and high heterogeneity, more studies are needed. TRIAL REGISTRATION: PROSPERO (CRD42025639586).

Humans↗

PubMed1999

Matrix logarithm parametrizations for neural network covariance models.

Neural networks are commonly used to model conditional probability distributions. The idea is to represent distributional parameters as functions of conditioning events, where the function is determined by the architecture and weights of the network. An issue to be resolved is the link between distributional parameters and network outputs. The latter are unconstrained real numbers whereas distributional parameters may be required to lie in proper subsets, or be mutually constrained, e.g. by the positive definiteness requirement for a covariance matrix. The paper explores the matrix-logarithm parametrization of covariance matrices for multivariate normal distributions. From a Bayesian point of view the choice of parametrization is linked to the choice of prior. This is treated by investigating the invariance of predictive distributions, for the chosen parametrization, with respect to an important class of priors.

Journal Article↗

PubMed2000

Population reconstruction of the locomotor cycle from interneuron activity in the mammalian spinal cord.

Lesion studies have shown that neuronal networks in the ventromedial regions of the neonatal rat spinal cord are critical for the production of locomotion. We examined whether the locomotor cycle could be accurately predicted based on the activity recorded in a population of spinal interneurons located in these regions during pharmacologically induced locomotion. We used a Bayesian probabilistic reconstruction procedure to predict the most likely phase of locomotion given the observed activity in the neuronal population. The population reconstruction was able to predict the correct locomotor phase with high accuracy using a relatively small number of neurons. This result demonstrates that although the spike activity of individual spinal interneurons in the ventromedial region is weak and varies from cycle to cycle, the locomotor phase can be accurately predicted when information from the population is combined. This result is consistent with the proposed involvement of interneurons within these regions of the spinal cord in the production of locomotion.

Action Potentials↗

PubMed1994

Neural-network classification of normal and Alzheimer's disease subjects using high-resolution and low-resolution PET cameras.

UNLABELLED: Neural-network classification methods were applied to studies of FDG-PET images of the brain acquired from a total of 77 "probable" Alzheimer's disease and 124 normal subjects at two different centers. METHODS: Classification performances, as determined by relative-operating-characteristic (ROC) analyses of cross-validation experiments, were measured for FDG PET images obtained with either a 15-mm FWHM PETT V or a 6-mm FWHM Scanditronix PC-1024-7B camera for various methods of data representation. Neural networks were trained to distinguish between normal and abnormal subjects on the basis of regional metabolic patterns. For both databases, classification performance could be improved by increasing the "resolution" of the representation (decreasing the region size) and by normalizing the regional metabolic values to the value of a reference region (occipital region). RESULTS: The optimal classification performance for Scanditronix data (ROC area = 0.95) was higher than that for PETT V data (ROC area = 0.87). Under Bayesian theory, the classification performance with Scanditronix data corresponded to an ability to change a pre-test probability of disease of 50% to a post-test probability of either 90% for a positive classification or 10% for a negative classification. CONCLUSION: This classification can be used to either strongly confirm or rule out the presence of abnormalities.

Aged↗

PubMedDate not supplied

Adaptive topological tree structure for document organisation and visualisation.

The self-organising map (SOM) is finding more and more applications in a wide range of fields, such as clustering, pattern recognition and visualisation. It has also been employed in knowledge management and information retrieval. We propose an alternative to existing 2-dimensional SOM based methods for document analysis. The method, termed Adaptive Topological Tree Structure (ATTS), generates a taxonomy of underlying topics from a set of unclassified, unstructured documents. The ATTS consists of a hierarchy of adaptive self-organising chains, each of which is validated independently using a proposed entropy-based Bayesian information criterion. A node meeting the expansion criterion spans a child chain, with reduced vocabulary and increased specialisation. The ATTS creates a topological tree of topics, which can be browsed like a content hierarchy and reflects the connections between related topics at each level. A review is also given on the existing neural network based methods for document clustering and organisation. Experimental results on real-world datasets using the proposed ATTS method are presented and compared with other approaches. The results demonstrate the advantages of the proposed validation criteria and the efficiency of the ATTS approach for document organisation, visualisation and search. It shows that the proposed methods not only improve the clustering results but also boost the retrieval.

Bayes Theorem↗

PubMed1997

Late potential recognition by artificial neural networks.

Ventricular late potentials (LP's) are high-frequency low-amplitude signals obtained from signal-averaged electrocardiograms (ECG's) [SAECG's]. LP's are useful in identifying patients prone to ventricular tachycardia (VT), spontaneous or inducible during electrophysiology testing. A combination of self-organizing and supervised artificial neural network (ANN) models was developed to identify patients with a positive electrophysiology (PEP) test for inducible ventricular tachycardia from patients with a negative electrophysiology (NEP) test using LP's. We have added morphology information of vector magnitude waveform to original set of three time-domain features of LP's, which are total QRS duration (TQRSD), high-frequency low-amplitude signal duration (HFLAD), and root-mean-square voltage (RMSV). Pattern recognition results from an ANN model with this combination feature set are superior to the results from Bayesian classification model based on conventional three time-domain features of SAECG. In order to increase the robustness of the recognition, a filtered QRS offset point is randomly shifted +/- 8 ms to form a fuzzy training set, which was to simulate the possible error in detecting QRS offset point of filtered SAECG. We also found that nonlinear transformation through the hidden layer of developed ANN model could increase Euclidean distance between PEP and NEP patterns.

Algorithms↗

PubMed2026

Single-cell expression quantitative trait locus Mendelian randomization reveals immune cell-specific causal regulatory networks and actionable targets in polycystic ovary syndrome.

ObjectiveTo systematically investigate whether the pathogenesis of polycystic ovary syndrome (PCOS) is causally related to dysregulated gene expression in specific immune cell subsets, and to evaluate the potential of these causal genes as actionable drug targets.MethodsThis study employed a two-sample Mendelian randomization (MR) framework using publicly available genome-wide association study (GWAS) summary statistics. The participant data included 797 PCOS cases and 140,558 controls (no direct patient recruitment was involved). Instrumental variables were derived from high-resolution immune cell-specific single-cell expression quantitative trait locus (sc-eQTL) data (OneK1K project) across 14 immune cell types. Primary analyses utilized the inverse-variance weighted (IVW) method. Shared causal variants were validated using Bayesian colocalization. Phenome-wide association analysis (PheWAS), external transcriptomic dataset validation (GSE8157), and DrugBank database screening were conducted for pleiotropy assessment and drug repositioning.ResultsMR analysis revealed genome-wide significant causal associations for GLIPR1 in non-classical monocytes (Mono NC) and XBP1 in CD4+ effector memory T cells (CD4 ET) with PCOS risk. Higher GLIPR1 expression was associated with a decreased PCOS risk (OR = 0.669, P = 4.34×10-6), whereas higher XBP1 expression was associated with an increased risk (OR = 1.406, P = 9.53×10-8). Colocalization analysis confirmed that GLIPR1 shares a causal variant with PCOS (PP.H4 = 96.73%). PheWAS and external validation confirmed the safety profile and significant upregulation (P = 0.03) of GLIPR1. Drug repositioning identified SOT-107, a Phase III protein therapy drug, as a potential interacting agent for GLIPR1.ConclusionsThis sc-eQTL MR study reveals immune cell-specific causal regulatory networks in PCOS. GLIPR1 in non-classical monocytes represents a high-confidence protective target, while XBP1 provides suggestive evidence for immune-mediated pathogenesis. The candidate drug SOT-107 highlights theoretical repositioning opportunities, though rigorous preclinical validation remains required.

Female↗

PubMed2026

IQ-NET: fast and accurate quartet phylogenetic inference using deep learning trained on empirical DNA alignments.

Phylogenetic inference is fundamental to modern biology, with many applications including evolutionary biology, epidemiology, and comparative genomics. While maximum likelihood and Bayesian methods remain the gold standard for phylogenetic analysis, they rely on simplifying assumptions and are computationally intensive. Recent machine learning approaches for phylogenetics offer speed advantages, but have several limitations: exclusive reliance on simulated data for training, inadequate handling of gaps, and sensitivity to input sequence order. Here, we introduce IQ-NET (Intelligent Quartet NETwork), a deep learning framework that solves these limitations to infer four-taxon trees. IQ-NET estimates both tree topology and branch lengths directly from gapped alignments. IQ-NET outperforms existing machine learning methods in terms of accuracy, and obtained a 24-fold speedup compared with the widely used maximum likelihood software, IQ-TREE. We finally introduce a pipeline using IQ-NET and the ASTRAL software to reconstruct a larger species tree, i.e., with more than four taxa.

Empirical data training↗

PubMed2004

Fast fixed-point neural blind-deconvolution algorithm.

The aim of this letter is to introduce a new blind-deconvolution algorithm based on fixed-point optimization of a "Bussgang"-type cost function. The cost function relies on approximate Bayesian estimation achieved by an adaptive neuron. The main feature of the presented algorithm is fast convergence that guarantees good deconvolution performances with limited computational demand as compared with algorithms of the same class.

Algorithms↗