PubMed Health⌕ Search

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 1,657 records · Page 92Linked to original sources
PubMed2006

Boolean network analysis of a neurotransmitter signaling pathway.

BACKGROUND: A Boolean network is a simple computational model that may provide insight into the overall behavior of genetic networks and is represented by variables with two possible states (on/off), of the individual nodes/genes of the network. In this study, a Boolean network model has been used to simulate a molecular pathway between two neurotransmitter receptor, dopamine and glutamate receptor, systems in order to understand the consequence of using logic gate rules between nodes, which have two possible states (active and inactive). RESULTS: The dynamical properties of this Boolean network model of the biochemical pathway shows that, the pathway is stable and that, deletion/knockout of certain biologically important nodes cause significant perturbation to this network. The analysis clearly shows that in addition to the expected components dopamine and dopamine receptor 2 (DRD2), Ca(2+) ions play a critical role in maintaining stability of the pathway. CONCLUSION: So this method may be useful for the identification of potential genetic targets, whose loss of function in biochemical pathways may be responsible for disease onset. The molecular pathway considered in this study has been implicated with a complex disorder like schizophrenia, which has a complex multifactorial etiology.

Animals↗
PubMed2000

Spike-driven synaptic plasticity: theory, simulation, VLSI implementation.

We present a model for spike-driven dynamics of a plastic synapse, suited for aVLSI implementation. The synaptic device behaves as a capacitor on short timescales and preserves the memory of two stable states (efficacies) on long timescales. The transitions (LTP/LTD) are stochastic because both the number and the distribution of neural spikes in any finite (stimulation) interval fluctuate, even at fixed pre- and postsynaptic spike rates. The dynamics of the single synapse is studied analytically by extending the solution to a classic problem in queuing theory (Takacs process). The model of the synapse is implemented in aVLSI and consists of only 18 transistors. It is also directly simulated. The simulations indicate that LTP/LTD probabilities versus rates are robust to fluctuations of the electronic parameters in a wide range of rates. The solutions for these probabilities are in very good agreement with both the simulations and measurements. Moreover, the probabilities are readily manipulable by variations of the chip's parameters, even in ranges where they are very small. The tests of the electronic device cover the range from spontaneous activity (3-4 Hz) to stimulus-driven rates (50 Hz). Low transition probabilities can be maintained in all ranges, even though the intrinsic time constants of the device are short (approximately 100 ms). Synaptic transitions are triggered by elevated presynaptic rates: for low presynaptic rates, there are essentially no transitions. The synaptic device can preserve its memory for years in the absence of stimulation. Stochasticity of learning is a result of the variability of interspike intervals; noise is a feature of the distributed dynamics of the network. The fact that the synapse is binary on long timescales solves the stability problem of synaptic efficacies in the absence of stimulation. Yet stochastic learning theory ensures that it does not affect the collective behavior of the network, if the transition probabilities are low and LTP is balanced against LTD.

Action Potentials↗
PubMed2002

Electrical coupling can prevent expression of adult-like properties in an embryonic neural circuit.

Electrical coupling is widespread in developing nervous systems and plays a major role in circuit formation and patterning of activity. In most reported cases, such coupling between rhythmogenic neurons tends to synchronize and enhance their oscillatory behavior, thereby producing monophasic rhythmic output. However, in many adult networks, such as those responsible for rhythmic motor behavior, oscillatory neurons are linked by synaptic inhibition to produce rhythmic output with multiple phases. The question then arises whether such networks are still able to generate multiphasic output in the early stage of development when electrical coupling is abundant. A suitable model for addressing this issue is the lobster stomatogastric nervous system (STNS). In the adult animal, the STNS consists of three discrete neural networks that are comprised of oscillatory neurons interconnected by reciprocal inhibition. These networks generate three distinct rhythmic motor patterns with large amplitude neuronal oscillations. By contrast, in the embryo the same neuronal population expresses a single multiphasic rhythm with small-amplitude oscillations. Recent findings have revealed that adult-like network properties are already present early in the embryonic system but are masked by an as yet unknown mechanism. Here we use computer simulation to test whether extensive electrical coupling may be involved in masking adult-like properties in the embryonic STNS. Our basic model consists of three different adult-like STNS networks that are built of relaxation oscillators interconnected by reciprocal synaptic inhibition. Individual model cells generate slow membrane potential oscillations without action potentials. The introduction of widespread electrical coupling between members of these networks dampens oscillation amplitudes and, at moderate coupling strengths, may coordinate neuronal activity into a single rhythm with different phases, which is strongly reminiscent of embryonic STNS output. With a further increase in coupling strength, the system reaches one of two final states depending on the relative contribution of inhibition and inherent oscillatory properties within the networks: either fully synchronized and dampened oscillations, or a complete collapse of activity. Our simulations indicate that, beginning from either of these two states, the emergence of distinct adult networks during maturation may arise from a developmental decrease in electrical coupling that unmasks preexisting adult-like network properties.

Action Potentials↗
PubMed2006

Spatial vs temporal continuity in view invariant visual object recognition learning.

We show in a 4-layer competitive neuronal network that continuous transformation learning, which uses spatial correlations and a purely associative (Hebbian) synaptic modification rule, can build view invariant representations of complex 3D objects. This occurs even when views of the different objects are interleaved, a condition where temporal trace learning fails. Human psychophysical experiments showed that view invariant object learning can occur when spatial but not temporal continuity applies because of interleaving of stimuli, although sequential presentation, which produces temporal continuity, can facilitate learning. Thus continuous transformation learning is an important principle that may contribute to view invariant object recognition.

Computer Simulation↗
PubMed1994

A comparison of Radial Basis Function and backpropagation neural networks for identification of marine phytoplankton from multivariate flow cytometry data.

Two artificial neural network classifiers, the well-known Multi-layer Perception (MLP) (also known as the 'backpropagation network'), and the more recently developed Radial Basis Function (RBF) network, were evaluated and compared for their ability to identify multivariate flow cytometric data from five North Sea plankton groups (Dinoflagellidae, Bacillariophyceae, Prymnesiomonadida, Cryptomonadida, and other flagellates). RBF networks generally performed similarly to MLPs, and slightly better in cases where the data were markedly multimodal; RBF networks also have much shorter training times. The performance of MLPs was improved greatly by the use of a symmetrical bipolar 'transfer function' as opposed to the commonly-used asymmetric form. The issues of network optimisation and computational efficiency in use are discussed.

Evaluation Studies as Topic↗
PubMed2004

Chaotic itinerancy in the oscillator neural network without Lyapunov functions.

Chaotic itinerancy (CI), which is defined as an incessant spontaneous switching phenomenon among attractor ruins in deterministic dynamical systems without Lyapunov functions, is numerically studied in the case of an oscillator neural network model. The model is the pseudoinverse-matrix version of the previous model [S. Uchiyama and H. Fujisaka, Phys. Rev. E 65, 061912 (2002)] that was studied theoretically with the aid of statistical neurodynamics. It is found that CI in neural nets can be understood as the intermittent dynamics of weakly destabilized chaotic retrieval solutions.

Animals↗
PubMed2004

Unexpected epidemic thresholds in heterogeneous networks: the role of disease transmission.

We reformulate several recent analyses of infection processes on highly heterogeneous networks (e.g., scale-free networks) which conclude that diseases will spread and persist even for vanishingly small transmission probabilities. The results of these latter studies contrast with conventional epidemiological models where there are clear threshold effects, namely, should the transmission probability fall below a critical threshold level the disease is expected to die out. Here we show that epidemic propagation depends equally on the infection scheme as well as the network structure. Connectivity-dependent infection schemes can yield threshold effects even in scale-free networks where they would otherwise be unexpected.

Animals↗
PubMed1993

Dynamic clamp: computer-generated conductances in real neurons.

1. We describe a new method, the dynamic clamp, that uses a computer as an interactive tool to introduce simulated voltage and ligand mediated conductances into real neurons. 2. We simulate a gamma-aminobutyric acid (GABA) response of a cultured stomatogastric ganglion neuron to illustrate that the dynamic clamp effectively introduces a conductance into the target neuron. 3. To demonstrate an artificial voltage-dependent conductance, we simulate the action of a voltage-dependent proctolin response on a neuron in the intact stomatogastric ganglion. We show that shifts in the activation curve and the maximal conductance of the response produce different effects on the target neuron. 4. The dynamic clamp is used to construct reciprocal inhibitory synapses between two stomatogastric ganglion neurons that are not coupled naturally, illustrating that this method can be used to form new networks at will.

Animals↗
PubMed2005

Skin capillary network recognition and analysis by means of neural algorithms.

BACKGROUND: The intra-dermal capillary network can be easily assessed by a computerized videomicroscope system. Nevertheless, finding capillary loops automatically in an image is a difficult yet important first step in order to achieve microcirculation analysis. METHODS: A detection system was tested by combining videocapillaroscopy and principal component analysis (PCA). Our goal was to build a generic detector of capillary associated with a retinally connected neural network filter. The filter examines small windows of an image, and decides with this detector whether each window contains a capillary or not. RESULTS: Comparisons with manual detections showed that the system has a detection rate of 82% on test set A containing 100 good-quality images of the scalp. A detection rate of 65% was obtained on test set B containing 50 images with noisy background and large artifacts. The performance was increased by a color detector with a detection rate of 71% on the last test. These results correspond to a false detection rate lower than or equal to 10%. CONCLUSION: This neural filter system is capable of real-time processing; it recognizes capillaries anywhere in an image, and operates successfully under wide range of lighting and noisy conditions.

Algorithms↗
PubMed2004

A nonlinear circuit architecture for magnetoencephalographic signal analysis.

OBJECTIVES: The objective of this paper was to face the complex spatio-temporal dynamics shown by Magnetoencephalography (MEG) data by applying a nonlinear distributed approach for the Blind Sources Separation. The effort was to characterize and differ-entiate the phases of a yogic respiratory exercise used in the treatment of obsessive compulsive disorders. METHODS: The patient performed a precise respiratory protocol, at one breath per minute for 31 minutes, with 10 minutes resting phase before and after. The two steps of classical Independent Component Approach have been performed by using a Cellular Neural Network with two sets of templates. The choice of the couple of suitable templates has been carried out using genetic algorithm optimization techniques. RESULTS: Performing BSS with a nonlinear distributed approach, the outputs of the CNN have been compared to the ICA ones. In all the protocol phases, the main components founded with CNN have similar trends compared with that ones obtained with ICA. Moreover, using this distributed approach, a spatial location has been associated to each component. CONCLUSIONS: To underline the spatio-temporal and the nonlinearly of the neural process a distributed nonlinear architecture has been proposed. This strategy has been designed in order to overcome the hypothesis of linear combination among the sources signals, that is characteristic of the ICA approach, taking advantage of the spatial information.

Algorithms↗
PubMed2000

A parallel algorithm for simulation of large neural networks.

The simulation of biologically realistic neural networks requires the numerical solution of very large systems of differential equations. Variables within the system can be changing at rates that vary by orders of magnitude, not only at different times of the solution, but at the same time in different parts of the network. Therefore, an efficient implementation must be able to vary the solution step size, and do so independently in different subsystems. A single processor algorithm is presented in which each neuron can be solved with its own step size by using a priority queue to integrate them in the correct order. But this leaves the problem of how communication and synchronisation between neurons should be managed when executing in parallel. The proposed solution uses an algorithm based on waveform relaxation, which allows groups of neurons on different processors to be solved independently and hence in parallel, for substantial parts of the computation. Realistic test problems were run on a distributed memory parallel computer and results show that speedups of 10 using 16 processors are achievable, and indicate that further speedups may be possible.

Algorithms↗
PubMed2001

A computational approach to identify genes for functional RNAs in genomic sequences.

Currently there is no successful computational approach for identification of genes encoding novel functional RNAs (fRNAs) in genomic sequences. We have developed a machine learning approach using neural networks and support vector machines to extract common features among known RNAs for prediction of new RNA genes in the unannotated regions of prokaryotic and archaeal genomes. The Escherichia coli genome was used for development, but we have applied this method to several other bacterial and archaeal genomes. Networks based on nucleotide composition were 80-90% accurate in jackknife testing experiments for bacteria and 90-99% for hyperthermophilic archaea. We also achieved a significant improvement in accuracy by combining these predictions with those obtained using a second set of parameters consisting of known RNA sequence motifs and the calculated free energy of folding. Several known fRNAs not included in the training datasets were identified as well as several hundred predicted novel RNAs. These studies indicate that there are many unidentified RNAs in simple genomes that can be predicted computationally as a precursor to experimental study. Public access to our RNA gene predictions and an interface for user predictions is available via the web.

Computational Biology↗
PubMed1996

Simulation of oxygen saturation of hemoglobin solution, RBC suspension and hemosome by a neural network system.

Hemoglobin-based artificial blood substitutes as oxygen carrier is advantageous over current plasma expander. In this study, oxygen saturation of hemoglobin solution, red blood cell suspension and artificial blood substitute under various conditions were measured by yeast-consuming-oxygen experiments instead of spectrophotometer. The empirical results were assigned into training feedforward back-propagation neural network system in order to simulate the oxygen saturation model modulated by those factors such as pH, [Cl-], [2,3-DPG], pO2 and pCO2. Consequently, this neural network is able to simulate accurately the oxygen saturation of Hb solution. The prediction of hemosome is not agreed well possible because of the resistance of transport of oxygen. However, the results showed neural net can offer a simple and convenient way in comparison with the conventional methods, especially in dealing with complex and ambiguous problem.

Animals↗
PubMed2003

Weight-elimination neural networks applied to coronary surgery mortality prediction.

The objective was to assess the effectiveness of the weight-elimination cost function in improving classification performance of artificial neural networks (ANNs) and to observe how changing the a priori distribution of the training set affects network performance. Backpropagation feedforward ANNs with and without weight-elimination estimated mortality for coronary artery surgery patients. The ANNs were trained and tested on cases with 32 input variables describing the patient's medical history; the output variable was in-hospital mortality (mortality rates: training 3.7%, test 3.8%). Artificial training sets with mortality rates of 20%, 50%, and 80% were created to observe the impact of training with a higher-than-normal prevalence. When the results were averaged, weight-elimination networks achieved higher sensitivity rates than those without weight-elimination. Networks trained on higher-than-normal prevalence achieved higher sensitivity rates at the cost of lower specificity and correct classification. The weight-elimination cost function can improve the classification performance when the network is trained with a higher-than-normal prevalence. A network trained with a moderately high artificial mortality rate (artificial mortality rate of 20%) can improve the sensitivity of the model without significantly affecting other aspects of the model's performance. The ANN mortality model achieved comparable performance as additive and statistical models for coronary surgery mortality estimation in the literature.

Adult↗
PubMed2003

Prediction of beta-turns in proteins from multiple alignment using neural network.

A neural network-based method has been developed for the prediction of beta-turns in proteins by using multiple sequence alignment. Two feed-forward back-propagation networks with a single hidden layer are used where the first-sequence structure network is trained with the multiple sequence alignment in the form of PSI-BLAST-generated position-specific scoring matrices. The initial predictions from the first network and PSIPRED-predicted secondary structure are used as input to the second structure-structure network to refine the predictions obtained from the first net. A significant improvement in prediction accuracy has been achieved by using evolutionary information contained in the multiple sequence alignment. The final network yields an overall prediction accuracy of 75.5% when tested by sevenfold cross-validation on a set of 426 nonhomologous protein chains. The corresponding Q(pred), Q(obs), and Matthews correlation coefficient values are 49.8%, 72.3%, and 0.43, respectively, and are the best among all the previously published beta-turn prediction methods. The Web server BetaTPred2 (http://www.imtech.res.in/raghava/betatpred2/) has been developed based on this approach.

Algorithms↗
PubMed2004

Overview of commonly used bioinformatics methods and their applications.

Bioinformatics, in its broad sense, involves application of computer processes to solve biological problems. A wide range of computational tools are needed to effectively and efficiently process large amounts of data being generated as a result of recent technological innovations in biology and medicine. A number of computational tools have been developed or adapted to deal with the experimental riches of complex and multivariate data and transition from data collection to information or knowledge. These include a wide variety of clustering and classification algorithms, including self-organized maps (SOM), artificial neural networks (ANN), support vector machines (SVM), fuzzy logic, and even hyphenated techniques as neuro-fuzzy networks. These bioinformatics tools are being evaluated and applied in various medical areas including early detection, risk assessment, classification, and prognosis of cancer. The goal of these efforts is to develop and identify bioinformatics methods with optimal sensitivity, specificity, and predictive capabilities.

Computational Biology↗
PubMed1999

Implications of a neural network model of early sensori-motor development for the field of developmental neurology.

This paper reports on a neural network model for early sensori-motor development and on the possible implications of this research for our understanding and, eventually, treatment of motor disorders like cerebral palsy. We recapitulate the results we published in detail in a series of papers [1-4]. The neural circuits in the model self-organize on the basis of rhythmic activity spontaneously generated in the model. This indicates the importance of endogenously generated activity in the developing brain. We also show that afferent feed-back from the mechanical part of the model is easily incorporated in the neural part of the model. In this way the model acquires reflex-related properties which have long been demonstrated in man. In the discussion we relate these experimental findings to the variability concept from developmental neurology and show how variable motor performance is important for motor learning. We also discuss possible implications of our modelling effort for movement disorders, specifically spastic cerebral palsy.

Central Nervous System↗