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 73 records · Page 4Linked to original sources
PubMed1996

Performance of a computer simulated neural network trained to categorise normal, premalignant and malignant oral smears.

The accurate detection of malignant neoplasms whilst they are still small is recognised as one of the main factors increasing chances of survival. Neural networks have many biomedical applications and they have been applied to neoplasia but their use in oral pathology has only recently been documented. The objectives of this study were to train networks to discriminate between normal and dysplastic mucosa. Each network was trained by back propagation, internal cross validation and tested on additional data. The data were derived by analysing 348 intra-oral smears and included mean nuclear and mean cytoplasmic areas of the smears measured by image analysis. A neural network differentiated between normal/non-dysplastic mucosa and dysplastic/malignant mucosa (specificity 0.82, sensitivity 0.76). These early results suggest that integrating neural networks and image analysis, as well as investigating additional criteria, could enhance automation and accuracy of smear techniques in diagnosing oral malignancy.

Automation↗
PubMed1997

Removing irrelevant features in neural network classification using evolutionary computations.

Evolutionary artificial neural networks (EANN) are a new paradigm that refers to a special class of artificial neural networks (ANN) in which evolution is another fundamental form of adaptation in addition to learning. Evolution can be introduced at various levels of ANN. It can be used to evolve weights, architectures and learning parameters. Evolutionary computations are population-based search methods that have shown promise in many similarly complex tasks. This paper presents an application of evolutionary programming for simultaneously inducing the input structure and weights evolving for multilayer feed-forward perceptrons (MLP) with standard sigmoidal activation function.

Algorithms↗
PubMed2007

Breast MRI lesion classification: improved performance of human readers with a backpropagation neural network computer-aided diagnosis (CAD) system.

PURPOSE: To develop and test a computer-aided diagnosis (CAD) system to improve the performance of radiologists in classifying lesions on breast MRI (BMRI). MATERIALS AND METHODS: A CAD system was developed that uses a semiautomated segmentation method. After segmentation, 42 features based on lesion shape, texture, and enhancement kinetics were computed, and the 13 best features were selected and used as inputs to a backpropagation neural network (BNN). The BNN was trained and tested using the leave-one-out method on 80 BMRI lesions (37 benign, 43 malignant). Lesion histopathology was used as the reference standard. Five human readers classified the 80 lesions first without and then with CAD assistance. The performance of the computer classifier and the human readers was assessed using receiver operating characteristic curves; the performance of the human readers was also evaluated using multireader multicase (MRMC) analysis. RESULTS: The performance of the human readers significantly improved when aided by the CAD system (P < 0.05). MRMC analysis showed that human reader performance with and without CAD system assistance can be generalized to the population of cases (P < 0.001). CONCLUSION: A CAD system based on lesion morphology and enhancement kinetics can improve the performance of human readers in classifying lesions on breast MRI.

Breast Neoplasms↗
PubMedDate not supplied

A comparison of neural networks for computing predicted probability of survival for trauma victims.

TRISS is a statistical method for predicting the probability of survival of trauma victims. Analysis of data from the Trauma Registry at Charleston Area Medical Center showed that only 48% of the trauma fatalities in the 5-year period 1992-1996 were correctly predicted by TRISS. Trauma practitioners from other Trauma Centers report similar problems with TRISS. Researchers have suggested improvements that range from simply changing the input variables and/or regression coefficients in TRISS to using an entirely different model. In this study we describe a method of calculating survival probabilities using Artificial Neural Networks (ANN). This method was chosen because of the similarity of the ANN output function to the function that produces the TRISS probability of survival. Additional variables were added based on the results of other research efforts as well as analysis of the CAMC Trauma Registry. A comparison was made between the abilities of TRISS to predict fatalities and to approximate probability of survival. The ANN outperformed TRISS in predicting fatalities in a training set (68.1% correct vs. 47.9% correct) and in a testing set (61.3% correct vs. 51.3% correct). More importantly, the ANN produced better estimates of predicted deaths. Using a data set that included 119 deaths, the ANN model predicted 125 deaths for a 5% relative error. The predicted number using TRISS was 86 for a relative error of 27.7%. Since effective quality improvement for trauma care depends on accurately identifying cases that fall outside the expected results, a more accurate predictive tool allows a more focused review of those significant cases, thus conserving resources without compromising quality. Neural Networks appear to be a predictive tool that can provide probability of survival estimates that are more accurate than TRISS.

Humans↗
PubMed1993

Computer-simulated neural networks: an appropriate model for motor development?

The idea of an artificial neural network is introduced in a historical context, and the essential aspect of it, viz., the modifiable synapse, is compared to the aspect of plasticity in the natural nervous system. Based on such an artificial neural network, a model is presented for the way in which (the path along which) the connectivity in the spinal cord is modified during the period that a newborn 'learns' to control the movement of his forearm. In this way an automatic calibration of the receptors and the antagonists' recruitment of motor units is represented. The learning process is described in non-mathematical terminology. The model is then shown to be able after learning to reach target angles outside the training set of angles, and to be able to relearn when an important receptor has been made inoperative. In this way it is shown that the model is able to generalize, and that it is robust against at least some damage.

Computer Simulation↗
PubMed2006

Novel correlation for predicting impact sensitivity of nitroheterocyclic energetic molecules.

A novel correlation is introduced for predicting impact sensitivity of a variety nitroheterocyclic molecular types, such as nitropyridines, nitroimidazoles, nitropyrazoles, nitrofurazanes, nitrotriazoles and nitropyrimidines. This approach is based on elemental composition and two structural parameters of C(a)H(b)N(c)O(d) nitroheterocyclic energetic compounds. The results for mentioned compounds are compared with complex neural networks computations which use compositional and topological descriptors. Root mean square (rms) of deviation of different nitroheterocyclic molecules including nitropyridines, nitroimidazoles, nitropyrazoles, nitrofurazanes, nitrotriazoles and nitropyrimidines are 58 and 71cm for new correlation and neural networks computations methods, respectively.

Explosive Agents↗
PubMed1999

New methods for studying hallucinated 'voices' in schizophrenia.

The mechanism of hallucinated speech or 'voices', a symptom commonly reported by schizophrenic patients, is poorly understood. We have undertaken two types of study to explore the hypothesis that this symptom arises from pathologically altered speech perception networks. The first consists of neural network computer simulations of narrative speech perception. We have shown that if these networks are partially disconnected or undergo a 'monamine' neuromodulatory disturbance, 'hallucinated speech' (speech percepts occurring in the absence of phonetic input) are simulated as well as specific speech perception impairments. The latter finding prompted us to conduct parallel studies of actual schizophrenic patients using a 'masked speech tracking' (MST) task. MST requires subjects to track narrative speech the phonetic clarity of which is reduced with superimposed multispeaker 'babble.' Hallucinators demonstrated speech perception impairments and experimentally induced perceptual 'illusions' which clearly differentiated them from non-hallucinators and normal controls. Our results support the hypothesis that 'voices' are spurious products of altered speech perception networks.

Adolescent↗
PubMed1998

Reduced corticocortical connectivity can induce speech perception pathology and hallucinated 'voices'.

Reciprocal interactions between multiple cortical areas are required for higher-order cognitive processes in animals and humans. A growing body of evidence suggests that reduced cortical connectivity is associated with schizophrenia. We have used neural network computer simulations to characterize symptoms and cognitive pathology that could arise from connectivity reductions. When this pathology was reproduced in simulations of speech perception networks, spontaneous percepts analogous to hallucinated speech were induced due to the emergence of 'locked-in' or 'parasitic' states. We propose that these simulations provide a model for hallucinated 'voices' reported by actual schizophrenic patients. The 'hallucinogenic' model of speech perception also demonstrated subtle speech perception impairments. We have demonstrated similar impairments when assessing schizophrenic patients reporting 'voices', but not when assessing similarly diagnosed patients not reporting this symptom. The model also demonstrates how a 'hypodopaminergic' state can reverse cognitive pathology arising from connectivity disturbances.

Cerebral Cortex↗
PubMedDate not supplied

Comparison between neural networks (NN) and principal component analysis (PCA): structure activity relationships of 1,4-dihydropyridine calcium channel antagonists (nifedipine analogues).

The applicability of the neural network computer package PSDD (Perceptron Simulator for Drug Design/Perceptron-type Neural Network Simulator) in structure-activity relationship (SAR) studies was investigated. A group of 1,4-dihydropyridine derivatives was used in order to compare the PSDD results with those obtained previously with PCA. Calculated atomic and molecular descriptors using the semiempirical AM1 method were mainly used. It was shown that the predictive capability demonstrated by PSDD in SAR analysis were almost equivalent to that of PCA.

Algorithms↗
PubMedDate not supplied

Automatic computerized diagnosis of fetal sinusoidal heart rate.

OBJECTIVES: Computerized automatic detection of pathologic fetal sinusoidal heart rate (FSHR) and its differentiation from physiologic FSHR is the purpose of this study. The results will be applied in the objective evaluation of fetal heart rate (FHR) with artificial neural network computer. METHODS: FHR tracings of pathologic FSHR of 9 cases of fetal-neonatal anemia, death, or severe asphyxia, those of 7 physiologic FSHR followed by normal outcome, and those of 5 normal FHR are processed with fast Fourier transform (FFT) analysis after digitization, and their power spectrums are obtained. The peak power spectrum frequency (PPSF), peak power spectrum density (PPSD), the area under the power spectrum of 0.03125-0.1 Hz (La), the area under the whole power spectrum (Ta), and the ratio of La/Ta (%) of pathologic FSHR are compared to those of physiologic FSHR and normal FHR. RESULTS: The La/Ta ratio and PPSD are significantly larger in the pathologic FSHR than those of physiologic FSHR and normal FHR. The true positive rate is 100%, false negative and false positive rates are 0%, respectively, when the pathologic FSHR is diagnosed by such combined criteria as 39% or more of La/Ta ratio and 300 or more of PPSD. CONCLUSION: Pathologic FSHR is clearly separated from physiologic FSHR and normal FHR by the La/Ta ratio and PPSD obtained by FFT frequency analysis of FHR. Consequently, it is capable to automatically diagnose pathologic FSHR, and to apply it to neural network computer evaluation of FHR.

Anemia↗
PubMed1998

Artificial neural networks for computer-based molecular design.

The theory of artificial neural networks is briefly reviewed focusing on supervised and unsupervised techniques which have great impact on current chemical applications. An introduction to molecular descriptors and representation schemes is given. In addition, worked examples of recent advances in this field are highlighted and pioneering publications are discussed. Applications of several types of artificial neural networks to compound classification, modelling of structure-activity relationships, biological target identification, and feature extraction from biopolymers are presented and compared to other techniques. Advantages and limitations of neural networks for computer-aided molecular design and sequence analysis are discussed.

Computer-Aided Design↗
PubMed2000

Testing neural network models of memory with behavioral experiments.

In recent years, a number of computational neural networks have been proposed aimed at describing memory functions associated with different subregions of the hippocampus, namely dentate gyrus, CA3 and CA1. Recent evidence suggests that indeed specific subregions of the hippocampus may subserve different computational functions, such as spatial and temporal pattern separation, short-term or working memory, pattern association, and temporal pattern completion.

Animals↗
PubMed1990

A neural network simulation of simultaneous single-unit activity recorded from the dragonfly ganglia.

Techniques are described that allow the use of multiple neuron spike data in a computational neural network architecture. The network architecture was devised to match the number of actual neurons from which data were obtained. The network was successfully trained to accurately predict the multiple neuron spike trains. Simultaneous spike histories of 44 neurons were modeled by a network architecture consisting of 44 input units, 88 hidden units with recurrent connections and 44 output units. The activation function of each unit was determined by data unique to a single neuron. These data were coupled with an analog gradient that preserved both the exact spiking times and the relative spiking tendency of each neuron. The input activation values were compared to network output target values calculated to occur 5 msec forward in the composite spiking records of all neurons. Following 2000 training cycles with the gradient data, the average error of each unit in the network was 0.0016. Discrete output values for each network unit were correlated with those of all other units. These correlations were comparable to those done using the actual neuron data. Both correlations reveal a functional connectivity pattern among the units and neurons. These connectivity patterns indicate that the networks may synthesize patterns of activity needed for biological function; in this case, flight patterns carried out in the mesothoracic ganglion of the dragonfly. This model represents, to the best of our knowledge, the first computer based network simulation using actual experimental neural data obtained from a large number of spontaneously active cells in a small intact ganglion.

Action Potentials↗
PubMed1999

SPIKEII: an integrate-and-fire AVLSI chip.

We present the SPIKEII chip, an integrate-and-fire neural network chip with programmable synapses implemented in analogue VLSI. It is the successor to the SPIKEI chip. We describe the circuitry, and show some results using networks of integrate-and-fire neurons.

Action Potentials↗
PubMed1997

Neural network classification and quantification of organic vapors based on fluorescence data from a fiber-optic sensor array.

Computational neural networks have been developed to classify and quantify nine organic vapors. The neural network analyses used data that consisted of the change in fluorescence from a sensor array that consisted of 19 fiber optics with immobilized dye in polymer matrices. Plots of change in fluorescence intensity versus time were measured as pulses of analyte were presented to the sensor array. Descriptors were calculated from the intensity vs time plots, and they were used to build neural network models that accurately classified and quantified each of the nine analytes. Most of the data were used to train the neural networks (training set members), some were used to assist termination of training (cross-validation set members), and some were used to validate the models (prediction set members). Classification rates approaching 100% were achieved for the training set data, and 90% of the members in the prediction set were correctly classified. In addition, 97% of the prediction set observations were assigned a correct relative concentration.

Acetates↗
PubMed2003

A neural network model of spiral-planar motion tuning in MSTd.

Neurophysiological studies in MSTd report the existence of motion pattern selective cells whose visual motion properties span a continuum of values, suggesting a role in estimates of self-motion from optic flow. Biologically motivated models of heading estimation support this view, having identified similar visual motion properties within their "neural" structures. While such models have addressed the computational sufficiency of their respective feed-forward designs they have not explicitly examined the underlying computational structures, particularly as they relate to the interaction between planar and spiral motion responses within MSTd. Here we use an expanded stimulus training set that includes planar motions to extend the range of neurophysiological properties identified within an existing network structure [Network: Comput. Neural Syst. 9 (1998) 467]. In doing so, we quantify the emergent planar motion properties within the network hidden layer and examine how they interact, functionally and computationally, with cardinal/spiral motion pattern responses. Throughout the hidden layer we demonstrate that the input activation associated with a unit's preferred planar motion is consistent with an overlapping gradient hypothesis [J. Neurophysiol. 65(6) (1991) 1346]. Together with the change to a peripheral excitation profile in the presence of a unit's preferred spiral motion these results suggest a more complex computational architecture in which the cell's 'classical' receptive field properties are dependent on the type of stimulus used to map them. Based on the computational model we propose an experimental paradigm to investigate the existence of equivalent computational structures in MSTd.

Animals↗