Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Neural Networks, Computer”

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.

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

At least 289 records · Page 16Linked to original sources
PubMedDate not supplied

A pulsed neural network model of bursting in the basal ganglia.

We present new techniques for extending the functionality of spiking neurons which allow the incorporation of several aspects of neuron function previously confined to the domain of low level ion-channel based models. These aspects include spontaneous (or endogenous) firing, the complex interaction of multiple ion-species and the spatial distribution of synaptic contacts over the cell membrane. These ideas are applied to a neural circuit consisting of the cortex and a subset of the nuclei in the basal ganglia-the subthalamic nucleus (STN) and the external segment of the globus pallidus (GPe). This circuit has been studied extensively in vitro by Plenz and Kitai [Plenz, D., & Kitai, S. T. (1999). A basal ganglia pacemaker formed by the subthalamic nucleus and external globus pallidus. Nature, 400 677-682] whose data we use to constrain our model. With respect to this circuit, we have obtained three main results. First, that its characteristic burst firing is due to a Ca2+ current mediated mechanism. Second, that noise can assist in the generation of bursting and, paradoxically, stabilise the network behaviour under synaptic weight variations. Third, that a variety of dendritic processing is necessary in order to obtain the full range of bursting behaviour.

Action Potentials↗
PubMed2001

Information processing in dendrites II. Information theoretic complexity.

In the companion paper, we established a rationale for exploring the general principles of dendritic processing using a class of Boolean functions-the Multi-Cube Units (MCUs). Here, we use this approach to further characterise dendritic processing using ideas from information theory and studies in complexity. The starting point is a novel decomposition of a Boolean function's total mutual information (between input variables and the output). Each component of the decomposition is a mutual information measure with respect to a single input, conditioned on a subset of the remaining inputs. We call this decomposition the information spectrum and conceive of it as a re-representation of the function in the information domain. Furthermore, the information spectrum of a Boolean function may be assigned a complexity value using the approximate entropy introduced by Pincus (Pincus, S. M. (1991). Approximate entropy as a measure of system complexity. Proc. Natl. Acad. Sci. USA, 88, 2297-2301). Using Monte Carlo methods, we provide evidence that the information spectral complexity of MCUs is larger than that of any other class of Boolean function. We explain this phenomenon in terms of information flow through the 2-stage MCU architecture. Under our modelling assumptions, the implication for biological neural processing is that dendrites implement functions that have maximal information spectral complexity with respect to the class of multivariate functions from which they are drawn.

Animals↗
PubMedDate not supplied

Computer aided diagnosis of breast cancer in digitized mammograms.

A computer aided neural network classification of regions of suspicion (ROS) on digitized mammograms is presented in this paper which employs features extracted by a new technique based on independent component analysis. Our approach is concentrated in finding a set of independent source regions that generate the observed mammograms. The coefficients of the linear transformation of the source regions are used as features that describe effectively any normal and abnormal region in digital mammograms as well as benign and malignant ROS in the latter case. Extensive experiments in the MIAS Database have shown a recognition accuracy of 88.23% in the detection of all kinds of abnormalities and 79.31% in the task of distinguishing between benign and malignant regions, outperforming in both cases standard textural features, widely used for cancer detection in mammograms.

Breast Diseases↗
PubMed2002

Attentional control of the processing of neural and emotional stimuli.

A typical scene contains many different objects that compete for neural representation due to the limited processing capacity of the visual system. At the neural level, competition among multiple stimuli is evidenced by the mutual suppression of their visually evoked responses and occurs most strongly at the level of the receptive field. The competition among multiple objects can be biased by both bottom-up sensory-driven mechanisms and top-down influences, such as selective attention. Functional brain imaging studies reveal that biasing signals due to selective attention can modulate neural activity in visual cortex not only in the presence but also in the absence of visual stimulation. Although the competition among stimuli for representation is ultimately resolved within visual cortex, the source of top-down biasing signals likely derives from a distributed network of areas in frontal and parietal cortex. Competition suggests that once attentional resources are depleted, no further processing is possible. Yet, existing data suggest that emotional stimuli activate brain regions "automatically," largely immune from attentional control. We tested the alternative possibility, namely, that the neural processing of stimuli with emotional content is not automatic and instead requires some degree of attention. Our results revealed that, contrary to the prevailing view, all brain regions responding differentially to emotional faces, including the amygdala, did so only when sufficient attentional resources were available to process the faces. Thus, similar to the processing of other stimulus categories, the processing of facial expression is under top-down control.

Attention↗
PubMed2003

A report of the functional connectivity workshop, Dusseldorf 2002.

This report provides a commentary on the issues presented and discussed at the recent "Functional Brain Connectivity" workshop, held in Dusseldorf, Germany. The workshop brought together researchers using different approaches to study connectivity in the brain, providing them with an opportunity to share conceptual, mathematical, and experimental ideas and to develop strategies and collaborations for future work on functional integration. The main themes that emerged included: (1) the importance of anatomical knowledge in understanding functional interactions the brain; (2) the need to establish common definitions for terms used across disciplines; (3) the need to develop a satisfactory framework for inferring causality from functional imaging and EEG/MEG data; (4) the importance of analytic tools that capture the dynamics of neural interactions; and (5) the role of experimental paradigms that exploit the functional imaging of perturbations to cortical interactions.

Animals↗
PubMed1997

Display technique using time-dependent scatter diagrams for the study of changing relative timings of three spike trains: evaluation of functional connectivity in a local area of the nervous system.

The relative timings of impulses caused by two neurons can be studied using cross-correlation techniques or the related techniques of the joint peristimulus-time histogram and the time-dependent correlation function. Analysis of the relative timings of several related neurons is important to the study of the dynamics, since we consider functional connectivity to be partially related to the dynamics. Functional connectivity changes of short duration reflect physiological functions rather than changes of synaptic connections. The relative timings of impulses are considered to reflect functional connectivity. Using the joint impulse configuration scatter diagram (JICSD) proposed by Perkel, a time-dependent scatter diagram (TDSD) is calculated to analyze the change of functional connectivity. The JICSD displays the long-time average of relative timings of impulses or the functional connectivity of three interconnected neurons. We generated JICSDs for short durations and we subsequently generated TDSDs by arranging JICSDs in chronological order. The change of a "snowflake pattern" of each JICSD was interpreted in terms of change of the functional connectivity. We propose a method for analyzing the changing functional connectivity among three neurons.

Action Potentials↗
PubMed2004

Central mechanisms of pain perception.

Mechanisms of human nociception can be studied by the use of CO2 laser stimulation, which selectively activates nociceptive receptors, and by the use of various noninvasive techniques. In addition to the contralateral thalamus, at least several cortical areas including the contralateral SI, bilateral SII, anterior cingulated cortex, and insular cortices are involved in the pain sensation/perception. Pain perception (Fig. 8) is unique because these cortical structures seem to be activated in parallel at nearly the same latency after the stimulus presentation. SI seems to play a role in basic pain processing while SII and insula are involved in higher functions of pain perception. Emotional aspects of pain perception are mediated by anterior cingulate cortex and posterior insula/parietal operculum.

Brain Mapping↗
PubMed2002

Prediction of peptide ion collision cross sections from topological molecular structure and amino acid parameters.

Quantitative structure-property relationships (QSPRs) have been developed to predict the ion mobility spectrometry (IMS) collision cross sections of singly protonated lysine-terminated peptides using information derived from topological molecular structure and various amino acid parameters. The primary amino acid sequence alone is sufficient to accurately predict the collision cross section. The models were built using multiple linear regression (MLR) and computational neural networks (CNNs). The best MLR model found contains six descriptors and predicts 94 of 113 peptides (83%) to within 2% of their experimentally determined values. The best CNN model using the same six descriptors predicts 105 of the 113 peptides (93%) to within 2% of their experimentally determined values. The best overall CNN model, using a different set of six descriptors, predicts 109 of the 113 peptides (96%) to within 2% of their experimentally determined values. In addition, this model can discriminate among peptides having identical amino acid composition, but differing in primary amino acid sequence. This represents a capability not found in previously described models. The descriptors used in the models presented may provide some insight into the nature of peptide ion folding in the gas phase.

Amino Acid Sequence↗
PubMedDate not supplied

Prediction of surface tension, viscosity, and thermal conductivity for common organic solvents using quantitative structure-property relationships.

Predictive models for the surface tension, viscosity, and thermal conductivity of 213 common organic solvents are reported. The models are derived from numerical descriptors which encode information about the topology, geometry, and electronics of each compound in the data set. Multiple linear regression and computational neural networks are used to train and evaluate models based on statistical indices and overall root-mean-square error. Eight-descriptor models were developed for both surface tension and viscosity, while a nine-descriptor model was developed for thermal conductivity. In addition, a single nine-descriptor model was developed for prediction of all three properties. The results of this study compare favorably to previously reported prediction methods for these three properties.

Journal Article↗
PubMedDate not supplied

Prediction of C60 solubilities from solvent molecular structures.

Models predicting fullerene solubility in 96 solvents at 298 K were developed using multiple linear regression and feed-forward computational neural networks (CNN). The data set consisted of a diverse set of solvents with solubilities ranging from -3.00 to 2.12 log (solubility) where solubility = (1 x 10(4))(mole fraction of C60 in saturated solution). Each solvent was represented by calculated molecular structure descriptors. A pool of the best linear models, as determined by rms error, was developed, and a CNN model was developed for each of the linear models. The best CNN model was chosen based on the lowest value of a specified cost function and had an architecture of 9-3-1. The 76-compound training set for this model had a root-mean-square error of 0.255 log solubility units, while the 10-compound cross-validation set had an rms error of 0.253. The 10-compound external prediction set had an rms error of 0.346 log solubility units.

Journal Article↗
PubMedDate not supplied

A QSPR study of the 31P NMR chemical shifts of phosphines.

Quantitative structure-property relationships (QSPR) on a large set of descriptors are developed for the 31P NMR chemical shifts of a large set of phosphines. The data set was composed of 291 primary, secondary, and tertiary phosphines, PH3-nRn, including substituents with different steric and electronic characteristics. Multiple linear regression and computational neural networks (CNN) were employed to create the models best suited for the prediction of 31P NMR chemical shifts. A correlation equation including seven descriptors (R2 = 0.8619) is reported. A 7-5-1 CNN was developed that produced a root-mean-error of 9.6 ppm (R2 = 0.9513) for the training set, of 11.7 ppm (R2 = 0.8986) for the cross-validation set, and of 11.3 ppm (R2 = 0.9527) for an external prediction set. The CNN methods give significantly better predictions of 31P NMR chemical shifts for phosphines than the multiple linear regression approach.

Journal Article↗
PubMedDate not supplied

Prediction of aqueous solubility of heteroatom-containing organic compounds from molecular structure.

The use of quantitative structure-property relationships (QSPRs) to predict aqueous solubilities (log S) of heteroatom-containing organic compounds from their molecular structure is presented. Three data sets are examined. Data set 1 contains 176 compounds having one or more nitrogen atoms with some oxygen (log S[mol/L] range is -7.41 to 0.96). Data set 2 contains 223 compounds having one or more oxygen atoms, with no nitrogen (log S[mol/L] range is -8.77 to 1.57). Data set 3 contains all 399 compounds from sets 1 and 2 (log S/mol/L] range is -8.77 to 1.57). After descriptor generation and feature selection, multiple linear regression (MLR) and computational neural network (CNN) models are developed for aqueous solubility prediction. The best results were obtained with nonlinear CNN models. Root-mean-square (rms) errors for training with the three data sets ranged from 0.3 to 0.6 log units. All models were validated with external prediction sets, with the rms errors ranging from 0.6 log units to 1.5 log units.

Journal Article↗
PubMedDate not supplied

QSARs for 6-azasteroids as inhibitors of human type 1 5alpha-reductase: prediction of binding affinity and selectivity relative to 3-BHSD.

Quantitative structure-activity relationships (QSARs) are developed to describe the ability of 6-azasteroids to inhibit human type 1 5alpha-reductase. Models are generated using a set of 93 compounds with known binding affinities (K(i)) to 5alpha-reductase and 3beta-hydroxy-Delta(5)-steroid dehydrogenase/3-keto-Delta(5)-steroid isomerase (3-BHSD). QSARs are generated to predict K(i) values for inhibitors of 5alpha-reductase and to predict selectivity (S(i)) of compound binding to 3-BHSD relative to 5alpha-reductase. Log(K(i)) values range from -0.70 log units to 4.69 log units, and log(S(i)) values range from -3.00 log units to 3.84 log units. Topological, geometric, electronic, and polar surface descriptors are used to encode molecular structure. Information-rich subsets of descriptors are identified using evolutionary optimization procedures. Predictive models are generated using linear regression, computational neural networks (CNNs), principal components regression, and partial least squares. Compounds in an external prediction set are used for model validation. A 10-3-1 CNN is developed for prediction of binding affinity to 5alpha-reductase that produces root-mean-square error (RMSE) of 0.293 log units (R(2) = 0.97) for compounds in the external prediction set. Additionally, an 8-3-1 CNN is generated for prediction of inhibitor selectivity that produces RMSE = 0.513 log units (R(2) = 0.89) for the external prediction set. Models are further validated through Monte Carlo experiments in which models are generated after dependent variable values have been scrambled.

Azasteroids↗
PubMedDate not supplied

Determination of lithium cation basicity from molecular structure.

A quantitative structure-property relationship (QSPR) is developed to calculate the Lithium Cationic Basicity (LCB) of a large set of 229 compounds, of very different chemical nature. The proposed models derived from multiple linear regression analysis (MLRA) and computational neural networks (CNN) contain seven descriptors calculated solely from the molecular structure of compounds. The models were validated by an external prediction set. Good results were obtained from both methodologies, being the best those from CNN, that give a rms error of 6.54 (R2 = 0.954) and an average error of 3.57% for the training set; for the prediction set the rms error is 8.61 (R2 = 0.914) and the average error 4.39%. The models derived from the two approaches contain descriptors that belong to the same classes, constitutional and electrostatic. The comparison with the results obtained from high level theoretical methods shows that the values obtained from the QSPR approach are very similar and even better, especially when the sets compared are large and contain compounds of different chemical structure. These good results shows that, despite the complexity of Li+-base interactions, the proposed models contain descriptors which encode properly the characteristics of the molecules directly related to their gas-phase basicity against the Li cations.

Journal Article↗
PubMedDate not supplied

Development of linear, ensemble, and nonlinear models for the prediction and interpretation of the biological activity of a set of PDGFR inhibitors.

A QSAR modeling study has been done with a set of 79 piperazyinylquinazoline analogues which exhibit PDGFR inhibition. Linear regression and nonlinear computational neural network models were developed. The regression model was developed with a focus on interpretative ability using a PLS technique. However, it also exhibits a good predictive ability after outlier removal. The nonlinear CNN model had superior predictive ability compared to the linear model with a training set error of 0.22 log(IC50) units (R2 = 0.93) and a prediction set error of 0.32 log(IC50) units (R2 = 0.61). A random forest model was also developed to provide an alternate measure of descriptor importance. This approach ranks descriptors, and its results confirm the importance of specific descriptors as characterized by the PLS technique. In addition the neural network model contains the two most important descriptors indicated by the random forest model.

Linear Models↗
PubMedDate not supplied

Determination of Abraham solute parameters from molecular structure.

The Abraham solute parameters are well-known factors for the quantitative description of solute/solvent interactions. A quantitative structure-property relationship (QSPR) is reported for the E, S, A, and B parameters of a large set of 457 solutes, of very different chemical nature. The proposed models, derived from multilinear regression analysis (MLRA) and computational neural networks (CNN), contain five descriptors calculated solely from the molecular structure of compounds. Good correlations were obtained for the four parameters studied, and the corresponding values of R(2) and standard deviations are better or similar than those derived from other theoretical bases. All models were validated by external prediction sets. The proposed QSPR models, both by MLRA and CNN, contain analogous descriptors encoding similar information, that agree with the accepted physicochemical meaning of the Abraham parameters; however, some descriptors which encode information that is not associated with this physicochemical meaning are also included in the QSPR models.

Journal Article↗
PubMedDate not supplied

Prediction of inhibition of the sodium ion-proton antiporter by benzoylguanidine derivatives from molecular structure.

The use of quantitative structure-activity relationships to predict IC50 values of 113 potential Na+/H+ antiporter inhibitors is reported. Multiple linear regression and computational neural networks (CNNs) are used to develop models using a set of information-rich descriptors. The descriptors encode information about topology, geometry, electronics, and combination hybrids. A five-descriptor CNN model with root-mean-square (rms) errors of 0.278 log units for the training set and 0.377 log units for the prediction set was developed. Examination of data set subclasses showed that systematic structural variations were also well-encoded resulting in 100% accuracy of prediction trends. An experiment involving a committee of five CNNs was also performed to examine the effect of network output averaging. This showed improved results decreasing the training and cross-validation set rms error to 0.228 log units and the prediction set rms error to 0.296 log units.

Animals↗
PubMedDate not supplied

Prediction of IC50 values for ACAT inhibitors from molecular structure.

A quantitative structure-activity study is performed on several series of compounds derived from N-chlorosulfonyl isocyanate to develop models that relate their structures to IC50 activity for inhibition of acyl-CoA:cholesterol O-acyltransferase (ACAT). Numerical descriptors are used to encode topological, electronic, and geometric information from the molecular structures of the inhibitors. A data set of 157 compounds showing triglyceride- and cholesterol-lowering activity is used to develop successful linear regression models and nonlinear computational neural network models. The models are validated using an external prediction set.

Enzyme Inhibitors↗