SEARCH · PubMed Health

Results for “Neural Networks, Computer”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

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

At least 109 records · Page 6Linked to original sources

PubMed1992

Neural computing in discovering RNA interactions.

High-order RNA structures are involved in regulating many biological processes; various algorithms have been designed to predict them. Experimental methods to probe such structures and to decipher the results are tedious. Artificial intelligence and the neural network approach can support the process of discovering RNA structures. Secondary structures of RNA molecules are probed by autoradiographing gels, separating end-labeled fragments generated by base-specific RNases. This process is performed in both conditions, denaturing (for sequencing purposes) and native. The resultant autoradiograms are scanned using line-detection techniques to identify the fragments by comparing the lines with those obtained by 'alkaline ladders'. The identified paired bases are treated by either one of two methods to find the foldings which are consistent with the RNases' 'cutting' rules. One exploits the maximum independent set algorithm; the other, the planarization algorithm. They require, respectively, n and n2 processing elements, where n is the number of base pairs. The state of the system usually converges to the near-optimum solution within about 500 iteration steps, where each processing element implements the McCulloch-Pitts binary neuron. Our simulator, based on the proposed algorithm, discovered a new structure in a sequence of 38 bases, which is more stable than that formerly proposed.

Algorithms↗

PubMed1991

Designing a neural network simulator--the MENS modelling environment for network systems: I.

During recent years, the field of neural network research has increasingly attracted the interest of workers from a large number of different disciplines. Current research topics include aspects as different as detailed simulations in brain physiology, predictions of protein structure in biochemistry, database organization in computer science, or various technical applications. The common scheme behind these different approaches is the use of distributed networks of simple computational elements that communicate with each other by means of weighted links. Computer simulations of neural networks require an appropriate software environment. Due to the computational similarities of many classes of such networks, simulation software can be structured into modular components that, to a large degree, are independent of specific applications. The aim of this and the following paper is to discuss some of the design considerations concerning software for neural network simulations. The aspects presented are interesting for both the development of new simulation software and the efficient use and modification of existing programs. Therefore, the general user as well as the software designer may hopefully benefit from this material. This paper briefly introduces some of the basic principles of neural networks. After a short discussion of different approaches to software design, two simple example applications are presented in order to demonstrate a conceptual framework common to many network simulations. The transfer of these considerations to the design of simulation software is then shown by example of the MENS network simulator developed in the Max-Planck-Institute for Brain Research. The paper gives a general introduction to the layout of data structures and different software components. Using the two introductory examples some aspects of network analysis are demonstrated. The following paper then considers further details of the design of a neural network simulator with respect to performance, implementation, and testing.

Animals↗

PubMed1992

Psychophysical experiments and a neural network model of binocular rivalry.

The alternation process in binocular rivalry was examined by psychophysical experiments. The stimulus strength to the central part of the visual field in one eye was varied while the stimulus strength in the other eye remained constant. The data obtained by the experiments indicate that the variations in stimulus strength alter the mean dominance duration, variance and predominance, and the frequency histogram of the dominance duration almost fits to the gamma distribution. To explain these results, a model which consists of neural elements receiving impulse trains with stochastic fluctuations is proposed and the experimental results are simulated on a computer using the model. Although the present model is restricted to explain only the rivalry between two lines which have different orientations, we could obtain agreement between the simulation results and the experimental results by giving appropriate values for some parameters of the model.

Humans↗

PubMed2026

Network methods for diagonal integration of unpaired single-cell multiomics data: a review.

MOTIVATION: Advances in single-cell sequencing have enabled multiomics profiling at unprecedented resolution; however, mass spectrometry-based single-cell proteomics (scMS) remains inherently destructive, precluding simultaneous transcriptomic capture. Unlike antibody-based methods such as CITE-seq, which permit paired profiling but are restricted to targeted protein panels, scMS provides unbiased, genome-scale coverage of the intracellular proteome yet necessitates post hoc integration of unpaired datasets. This diagonal integration challenge, where transcriptomes and proteomes are measured in separate cells lacking shared anchors, remains underserved by existing reviews, which focus predominantly on vertical integration strategies enabled by non-destructive assays. RESULTS: We survey the complete computational pipeline for constructing mechanistic proteogenomic networks from unpaired single-cell data, covering: (i) unimodal network inference such as knowledge-based approaches, probabilistic graphical models, temporal directionality inference, and generative and foundation model strategies that establish the transcriptomic scaffold; (ii) cross-modal integration architectures such as network propagation, graph neural networks (scMRDR, scmFormer, scCotag), and consensus frameworks designed explicitly for the unpaired proteomics setting; and (iii) benchmarking paradigms spanning network reconstruction (BEELINE, GRETA, CausalBench) and multi-task integration evaluation (scMultiBench, SCMMIB), with guidance on metric selection under network sparsity and class imbalance. We identify three principal axes of future development: generative proteomic translation from transcriptomic precursors, inductive prior embedding in next-generation architectures, and perturbation-based causal benchmarking. AVAILABILITY AND IMPLEMENTATION: This is a review article; no novel software is distributed. A curated benchmark resource table, methods starter guide, and per-method bottleneck annotations are provided in the Supplementary Material.

Multiomics↗

PubMed1991

Neuronal interconnection as a function of brain size.

The effect of increasing brain size upon the degree of interconnection between neurons is analyzed. An explicit model suggests that as the brain is scaled up there must be a corresponding fall in percent connectedness (the fraction of cells with which any one cell communicates directly). The reason for this is that if the percent connectedness is to be maintained in the face of increased neuron number, than a large fraction of any brain size increase would be spent maintaining such interconnection while the increasing axon lengths would reduce neural computational speed. One implication is that larger brains, being necessarily limited in allowable interconnectedness, may tend to show more specialization.

Animals↗

PubMed2025

Chromatin structures from integrated AI and polymer physics model.

The physical organization of the genome in three-dimensional space regulates many biological processes, including gene expression and cell differentiation. Three-dimensional characterization of genome structure is critical to understanding these biological processes. Direct experimental measurements of genome structure are challenging; computational models of chromatin structure are therefore necessary. We develop an approach that combines a particle-based chromatin polymer model, molecular simulation, and machine learning to efficiently and accurately estimate chromatin structure from indirect measures of genome structure. More specifically, we introduce a new approach where the interaction parameters of the polymer model are extracted from experimental Hi-C data using a graph neural network (GNN). We train the GNN on simulated data from the underlying polymer model, avoiding the need for large quantities of experimental data. The resulting approach accurately estimates chromatin structures across all chromosomes and across several experimental cell lines despite being trained almost exclusively on simulated data. The proposed approach can be viewed as a general framework for combining physical modeling with machine learning, and it could be extended to integrate additional biological data modalities. Ultimately, we achieve accurate and high-throughput estimations of chromatin structure from Hi-C data, which will be necessary as experimental methodologies, such as single-cell Hi-C, improve.

Chromatin↗

PubMed1992

Cerebral information processing estimated by unpredictability of the EEG.

Processing of complex information by the brain desynchronizes neuronal cells and therefore increases the unpredictability of the EEG, rather than its amplitude or power. A model of this mechanism describes the EEG as being composed of both an unpredictable fraction and a fraction that can be predicted from the past EEG. The model realistically simulates EEG of several behavioural states and suggests a simple algorithm for the computation of EEG predictability (values between 0% and 100%). Applications in sleep analysis and event-related desynchronization demonstrated that the algorithm is more accurate and artifact-resistant than the usual (spectral) power methods.

Algorithms↗

PubMed2026

GiantHost: a domain-adaptive and uncertainty-aware framework for giant virus host prediction.

MOTIVATION: Nucleocytoplasmic large DNA viruses (NCLDVs) play crucial roles in global ecosystems. Although metagenomics has vastly accelerated the discovery of novel NCLDVs, predicting their hosts from fragmented contigs remains a critical bottleneck, with no dedicated end-to-end computational tools currently available. Addressing this gap requires overcoming three fundamental challenges: the extreme scarcity of labeled reference genomes, the severe domain shift between laboratory isolates and diverse environmental metagenomes, and the inability of traditional deterministic models to quantify prediction uncertainty-a crucial requirement for reliable ecological profiling where novel, divergent viruses are prevalent. RESULTS: We present GiantHost, the first NCLDV host prediction tool with domain adaptation and uncertainlty awareness. GiantHost employs a dual-tower neural network to integrate dense genome traits and sparse GVOG profiles, allowing better integration of heterogeneous features. To overcome label scarcity and domain shift, we leverage 1400 environmental viral genomes (GVMAGs) via semi-supervised multi-task learning and Domain Adversarial Neural Networks (DANN), effectively bridging the distributional gap between RefSeq and environmental data. Additionally, GiantHost incorporates Conformal Prediction (CP) to output statistically guaranteed prediction sets rather than overconfident single labels. Evaluated under rigorous genome-level cross-validation, GiantHost demonstrates robust predictive power. Applied to the Tara Ocean dataset, GiantHost successfully captured the vertical stratification of NCLDV hosts-revealing a depth-dependent decline of phytoplankton-infecting viruses and a relative enrichment of Amoebozoa-infecting viruses in the mesopelagic zone. AVAILABILITY: The source code of GiantHost is available via: https://github.com/FuchuanQu/GiantHost.

Giant Viruses↗

PubMedDate not supplied

Digital vision theory: Boolean logic model.

Guild (1932) stated the general requirements for processing signals in color vision system and a digital format of his paradigm is developed in this paper. The disk structure generates the digital receptor pulse. The input modalities form sets, linked by intersections, joins, and complement junctions. The synapses are elements in these junctions. Complex synapses form complex junctions to create Boolean logic processing. A computer program using these Boolean logic functions calculates: Light and dark adaptation responses; Color matching and spectral coordinate functions; Chromatic adaptation and color shift responses; and dynamic neural responses. These calculations compare favorably with the experimental data.

Adaptation, Physiological↗

PubMed1992

Crossmatch prediction of highly sensitized patients.

1. A subset of negative reactions of sera from highly sensitized patients to donor lymphocytes are predicted with high accuracy (96.5% negative correct). 2. The prediction is performed by a hybrid expert system (HES) which uses multiple knowledge of stochastic (SCORES), artificial neural net (ANN), and genetic algorithm (GA) techniques. 3. All knowledge for the T-cell predictions is derived from serological reactions of the investigated sera (93) to a large panel (284). 4. When analyzing 5 HLA Class I typing sera controls, HES performs better than a standard serum analysis method in 3 measurement categories: r value; percent correct; and percent negative correct. 5. SCORES and ANN produce the strongest complementary association. SCORES is the best method with low PRA sera, while ANN is better at predicting high PRA sera. GA performs very poorly with high PRA sera. 6. HES can acquire knowledge from any of the various methods used for serum screening and crossmatch testing. Therefore, there is no need for method standardization as each laboratory will produce its own program incorporating its patients' data. High standardization of HLA Class I typing is necessary. 7. Most recipients for whom donors are never selected by HES are in the PRA range of 97-100%. 8. Certainty level categorization of a crossmatch gives clinical flexibility in judgement of potential donors. 9. All programs are written in the C language and are portable to numerous platforms. HES is implemented on an inexpensive IBM-PC compatible computer and can calculate predictions quickly. 10. HES predicts negative crossmatches with enough accuracy to initiate an organ sharing protocol to increase the chance for highly sensitized patients to obtain a transplant.

Algorithms↗

PubMed1992

Determination of eukaryotic protein coding regions using neural networks and information theory.

Our previous work applied neural network techniques to the problem of discriminating open reading frame (ORF) sequences taken from introns versus exons. The method counted the codon frequencies in an ORF of a specified length, and then used this codon frequency representation of DNA fragments to train a neural net (essentially a Perceptron with a sigmoidal, or "soft step function", output) to perform this discrimination. After training, the network was then applied to a disjoint "predict" set of data to assess accuracy. The resulting accuracy in our previous work was 98.4%, exceeding accuracies reported in the literature at that time for other algorithms. Here, we report even higher accuracies stemming from calculations of mutual information (a correlation measure) of spatially separated codons in exons, and in introns. Significant mutual information exists in exons, but not in introns, between adjacent codons. This suggests that dicodon frequencies of adjacent codons are important for intron/exon discrimination. We report that accuracies obtained using a neural net trained on the frequency of dicodons is significantly higher at smaller fragment lengths than even our original results using codon frequencies, which were already higher than simple statistical methods that also used codon frequencies. We also report accuracies obtained from including codon and dicodon statistics in all six reading frames, i.e. the three frames on the original and complement strand. Inclusion of six-frame statistics increases the accuracy still further. We also compare these neural net results to a Bayesian statistical prediction method that assumes independent codon frequencies in each position. The performance of the Bayesian scheme is poorer than any of the neural based schemes, however many methods reported in the literature either explicitly, or implicitly, use this method. Specifically, Bayesian prediction schemes based on codon frequencies achieve 90.9% accuracy on 90 codon ORFs, while our best neural net scheme reaches 99.4% accuracy on 60 codon ORFs. "Accuracy" is defined as the average of the exon and intron sensitivities. Achievement of sufficiently high accuracies on short fragment lengths can be useful in providing a computational means of finding coding regions in unannotated DNA sequences such as those arising from the mega-base sequencing efforts of the Human Genome Project. We caution that the high accuracies reported here do not represent a complete solution to the problem of identifying exons in "raw" base sequences. The accuracies are considerably lower from exons of small length, although still higher than accuracies reported in the literature for other methods. Short exon lengths are not uncommon.(ABSTRACT TRUNCATED AT 400 WORDS)

Base Sequence↗

PubMedDate not supplied

[Computer program recognition of a cDNA sequence specifying signal peptides].

An application of a computational analysis of cDNA sequences is presented in this paper. The goal is the identification of functional domains on sequence data. The results show the capability of this technique to identify a zone of DNA associated with the signal peptide coding region, whose biological function at DNA or RNA level is still unknown.

Base Sequence↗

PubMed1992

Synthetic neural modeling applied to a real-world artifact.

We describe the general design, operating principles, and performance of a neurally organized, multiply adaptive device (NOMAD) under control of a nervous system simulated in a computer. The complete system, Darwin IV, is the latest in a series of models based on the theory of neuronal group selection, which postulates that adaptive behavior is the result of selection in somatic time among synaptic populations. The simulated brain of Darwin IV includes visual and motor areas that are connected with NOMAD by telemetry. Under suitable conditions, Darwin IV can be trained to track a light moving in a random path. After such training, it can approach colored blocks and collect them to a home position. Following a series of contacts with such blocks, value signals received through a "snout" that senses conductivity allow it to sort these blocks on the basis of differences in color associated with differences in their conductivity. Darwin IV represents a new approach to synthetic neural modeling (SNM), a technique in which large-scale computer simulations are employed to analyze the interactions among the nervous system, the phenotype, and the environment of a designed organism as behavior develops. Darwin IV retains the advantages of SNM while avoiding the difficulties and pitfalls of attempting to simulate a rich environment in addition to a brain.

Behavior↗

PubMed1992

Lexical access and the brain: anatomical constraints on cognitive models of word recognition.

Recent studies in the cognitive psychology of reading and many other skilled performances have been dominated by models inspired by neural connectivity (e.g., McClelland & Rumelhart, 1986). Such models have not yet begun to consider the accumulating evidence of considerable anatomical localization of component cognitive operations in the human brain (e.g., Posner, Petersen, Fox, & Raichle, 1988). In this article we apply anatomical findings to the job of building computational models of visual word recognition. Brain imaging studies already provide important constraints on how lexical access should be defined in terms of isolable encoding operations that compute the visual form, phonology, and semantics of words. Brain imaging studies also speak to issues of modularity versus interaction between these encoding operations, distribution versus localization of processing within the operations, and orchestration of operations to accomplish different word processing tasks. We conclude that a combined cognitive and anatomical analysis may be of considerable benefit in developing more adequate models of human information processing.

Brain↗

PubMed1992

EEG classification by learning vector quantization.

EEG classification using Learning Vector Quantization (LVQ) is introduced on the basis of a Brain-Computer Interface (BCI) built in Graz, where a subject controlled a cursor in one dimension on a monitor using potentials recorded from the intact scalp. The method of classification with LVQ is described in detail along with first results on a subject who participated in four on-line cursor control sessions. Using this data, extensive off-line experiments were performed to show the influence of the various parameters of the classifier and the extracted features of the EEG on the classification results.

Algorithms↗

PubMed1990

Prediction of the disulfide-bonding state of cysteine in proteins.

The bonding states of cysteine play important functional and structural roles in proteins. In particular, disulfide bond formation is one of the most important factors influencing the three-dimensional fold of proteins. Proteins of known structure were used to teach computer-simulated neural networks rules for predicting the disulfide-bonding state of a cysteine given only its flanking amino acid sequence. Resulting networks make accurate predictions on sequences different from those used in training, suggesting that local sequence greatly influences cysteines in disulfide bond formation. The average prediction rate after seven independent network experiments is 81.4% for disulfide-bonded and 80.0% for non-disulfide-bonded scenarios. Predictive accuracy is related to the strength of network output activities. Network weights reveal interesting position-dependent amino acid preferences and provide a physical basis for understanding the correlation between the flanking sequence and a cysteine's disulfide-bonding state. Network predictions may be used to increase or decrease the stability of existing disulfide bonds or to aid the search for potential sites to introduce new disulfide bonds.

Amino Acid Sequence↗

PubMedDate not supplied

The Wisconsin Card Sorting Test: theoretical analysis and modeling in a neuronal network.

Neuropsychologists commonly use the Wisconsin Card Sorting Test as a test of the integrity of frontal lobe functions. However, an account of its range of validity and of the neuronal mechanisms involved is lacking. We analyze the test at 3 different levels. First, the different versions of the test are described, and the results obtained with normal subjects and brain-lesioned patients are reviewed. Second, a computational analysis is used to reveal what algorithms may pass the test, and to predict their respective performances. At this stage, 3 cognitive components are isolated that may critically contribute to performance: the ability to change the current rule when negative reward occurs, the capacity to memorize previously tested rules in order to avoid testing them twice, and the possibility of rejecting some rules a priori by reasoning. Third, a model neuronal network embodying these 3 components is described. The coding units are clusters of neurons organized in layers, or assemblies. A sensorimotor loop enables the network to sort the input cards according to several criteria (color, form, etc.). A higher-level assembly of rule-coding clusters codes for the currently tested rule, which shifts when negative reward is received. Internal testing of the possible rules, analogous to a reasoning process, also occurs, by means of an endogenous auto-evaluation loop. When lesioned, the model reproduces the behavior of frontal lobe patients. Plausible biological or molecular implementations are presented for several of its components.

Frontal Lobe↗

PubMed1990

Predicting surface exposure of amino acids from protein sequence.

The amino acid residues on a protein surface play a key role in interaction with other molecules, determined many physical properties, and constrain the structure of the folded protein. A database of monomeric protein crystal structures was used to teach computer-simulated neural networks rules for predicting surface exposure from local sequence. These trained networks are able to correctly predict surface exposure for 72% of residues in a testing set using a binary model, (buried/exposed) and for 54% of residues using a ternary model (buried/intermediate/exposed). In the ternary model, only 11% of the exposed residues are predicted as buried and only 5% of the buried residues are predicted as exposed. Also, since the networks are able to predict exposure with a quantitative confidence estimate, it is possible to assign exposure for over half of the residues in a binary model with greater than 80% accuracy. Even more accurate predictions are obtained by making a consensus prediction of exposure for a homologous family. The effect of the local environment of an amino acid on its accessibility, though smaller than expected, is significant and accounts for the higher success rate of prediction than obtained with previously used criteria. In the absence of a three-dimensional structure, the ability to predict surface accessibility of amino acids directly from the sequence is a valuable tool in choosing sites of chemical modification or specific mutations and in studies of molecular interaction.

Amino Acid Sequence↗