Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Neural network”

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 1,783 records · Page 99Linked to original sources
PubMed1998

Analysis of DNA curvature using artificial neural networks.

MOTIVATION: Our aim is to utilize an artificial neural network (ANN) for the prediction of DNA curvature in terms of retardation anomaly. RESULTS: An ANN capturing the role of phasing, increased helix flexibility, run of poly(A) tracts and flanking base pair effects in determining the extent of DNA curvature has been developed. The network predictions validate the known experimental results and also explain how the base pairs other than ApA affect the curvature. The results suggest that ANN can be used as a model-free tool for studying DNA curvature. AVAILABILITY: The optimal weights and the procedure to compute the retardation anomaly value are available on request from the authors. CONTACT: bdk@ems. ncl.res.in

Base Sequence↗
PubMed1996

Fast Second Order Learning Algorithm for Feedforward Multilayer Neural Networks and its Applications.

The paper presents the efficient training program of multilayer feedforward neural networks. It is based on the best second order optimization algorithms including variable metric and conjugate gradient as well as application of directional minimization in each step. Its efficiency is proved on the standard tests, including parity, dichotomy, logistic and two-spiral problems. The application of the algorithm to the solution of higher dimensionality problems like deconvolution, separation of sources and identification of nonlinear dynamic plant are also given and discussed. It is shown that the appropriately trained neural network can be used for the nonconventional solution of these standard signal processing tasks with satisfactory accuracy. The results of numerical experiments are included and discussed in the paper. Copyright 1996 Elsevier Science Ltd.

Journal Article↗
PubMed1997

Application of the fuzzy ART/MAP and MINMAX/MAP neural network models to radiographic image classification.

This paper concerns the classification analysis of exercise-induced lower leg pain by applying competitive neural network clustering and mapping techniques to type 1 and type 2 fuzzy descriptions of bone scan images of the tibia. The clusters are described and compared with each other and with the experts known classes that would be expected from medical findings. The discovered clusters provide training sets for supervised learning by an ARTMAP and similar neural network. These were used to classify the previously unclassified images and hence improve the classification process. The overall conclusion is that the use of the neural clustering methods has improved the classification process of the shin images despite the paucity of data and its inherent uncertainty.

Athletic Injuries↗
PubMed2003

Predicting cranial computed tomography results of head injury patients using an artificial neural network.

We conducted this study to determine if artificial neural network modeling would predict cranial computed tomography results in head injury patients using different combinations of clinical variables. 150 consecutive patients admitted to a regional trauma center with head injury were enrolled in the study. The CT was labeled with presence of surgically significant intracranial hematoma (SSIH), if midline shift, obliteration of ambient cistern or basal cistern were found. The best performance of our models to differentiate normal from abnormal cranial CT and detection of SSIH was ideal.

Craniocerebral Trauma↗
PubMed2004

An artificial neural network for predicting the presence of spermatozoa in the testes of men with nonobstructive azoospermia.

PURPOSE: We developed an artificial neural network (ANN) for predicting spermatozoa prior to testicular biopsy in men with nonobstructive azoospermia. The performance of this ANN was compared to that of the standard logistic regression (LR) model. MATERIALS AND METHODS: Data were retrospectively collected from the physical examination and laboratory records of 303 patients who presented with infertility due to nonobstructive azoospermia. Input factors were patient age, duration of infertility, serum follicle-stimulating hormone, luteinizing hormone, total testosterone and prolactin, and left and right testicular volume. The ANN and LR models were constructed based on data on cases in which spermatozoa were and were not detected on testicular sperm extraction. The ANN was trained and validated with the data in the training (200 cases) and validation (30 cases) sets, and the model was then used to predict findings in the test set (73 cases). The LR model was constructed using the same data on the 230 training and validation cases. The same 73 patients served as the test set. RESULTS: The sensitivity of the ANN model was significantly higher than that of the LR model (68% vs 28%, p < 0.0001). The neural network correctly predicted the outcome in 59 of the 73 test set patients (80.8%), whereas LR correctly predicted the outcome in 48 (65.7%, p = 0.07). CONCLUSIONS: This ANN model, which is based on age, duration of infertility, serum hormone levels and testicular volumes, has clinically acceptable sensitivity. It may be of value for predicting spermatozoa in men with nonobstructive azoospermia.

Adult↗
PubMed2001

An independent evaluation of a new method for automated interpretation of lung scintigrams using artificial neural networks.

The purpose of this study was to evaluate a new automated method for the interpretation of lung perfusion scintigrams using patients from a hospital other than that where the method was developed, and then to compare the performance of the technique against that of experienced physicians. A total of 1,087 scintigrams from patients with suspected pulmonary embolism comprised the training group. The test group consisted of scintigrams from 140 patients collected in a hospital different to that from which the training group had been drawn. An artificial neural network was trained using 18 automatically obtained features from each set of perfusion scintigrams. The image processing techniques included alignment to templates, construction of quotient images based on the perfusion/template images, and finally calculation of features describing segmental perfusion defects in the quotient images. The templates represented lungs of normal size and shape without any pathological changes. The performance of the neural network was compared with that of three experienced physicians who read the same test scintigrams according to the modified PIOPED criteria using, in addition to perfusion images, ventilation images when available and chest radiographs for all patients. Performances were measured as area under the receiver operating characteristic curve. The performance of the neural network evaluated in the test group was 0.88 (95% confidence limits 0.81-0.94). The performance of the three experienced experts was in the range 0.87-0.93 when using the perfusion images, chest radiographs and ventilation images when available. Perfusion scintigrams can be interpreted regarding the diagnosis of pulmonary embolism by the use of an automated method also in a hospital other than that where it was developed. The performance of this method is similar to that of experienced physicians even though the physicians, in addition to perfusion images, also had access to ventilation images for most patients and chest radiographs for all patients. These results show the high potential for the method as a clinical decision support system.

Adolescent↗
PubMed1999

Prediction of chiral separations using a combination of experimental design and artificial neural networks.

In this work the advantages of using artificial neural networks (ANNs) combined with experimental design (ED) to optimize the separation of amino acids enantiomers, with alpha-cyclodextrin as chiral selector, were demonstrated. The results obtained with the ED-ANN approach were compared with those of either the partial least-squares (PLS) method or the response surface methodology where experimental design and the regression equation were used. The ANN approach is quite general, no explicit model is needed, and the amount of experimental work can be decreased considerably.

Amino Acids↗
PubMed1997

Discrete State Neural Networks and Energies.

In this paper we give under an appropriate theoretical framework a characterization about neural networks (evolving in a binary set of states) which admit an energy. We prove that a neural network, iterated sequentially, admits an energy if and only if the weight matrix verifies two conditions: the diagonal elements are non-negative and the associated incidence graph does not admit non-quasi-symmetric circuits. In this situation the dynamics are robust with respect to a class of small changes of the weight matrix. Further, for the parallel update we prove that a necessary and sufficient condition to admit an energy is that the incidence graph does not contain non-quasi-symmetric circuits. Copyright 1997 Elsevier Science Ltd. All Rights Reserved.

Journal Article↗
PubMed2001

Recurrent neural networks for predicting outcomes after liver transplantation: representing temporal sequence of clinical observations.

OBJECTIVES: This paper investigates a version of recurrent neural network with the backpropagation through time (BPTT) algorithm for predicting liver transplant graft failure based on a time series sequence of clinical observations. The objective is to improve upon the current approaches to liver transplant outcome prediction by developing a more complete model that takes into account not only the preoperative risk assessment, but also the early postoperative history. METHODS: A 6-fold cross-validation procedure was used to measure the performance of the networks. The data set was divided into a learning set and a test set by maintaining the same proportion of positive and negative cases in the original set. The effects of network complexity on overfitting were investigated by constructing two types of networks with different numbers of hidden units. For each type of network, 10 individual networks were trained on the learning set and used to form a committee. The performance of the networks was measured exhaustively with respect to both the entire training and test sets. RESULTS: The networks were capable of learning the time series problem and achieved good performances of 90% correct classification on the learning set and 78% on the test set. The prediction accuracy increases as more information becomes progressively available after the operation with the daily improvement of 10% on the learning set and 5% on the test set. CONCLUSIONS: Recurrent neural networks trained with BPTT algorithm are capable of learning to represent temporal behavior of the time series prediction task. This model is an improvement upon the current model that does not take into account postoperative temporal information.

Adult↗
PubMed1997

Application of artificial neural network for micro-crack and damage evaluation of bone.

This paper presents the reasoning and adaptive learning method of artificial neural network (ANN) for micro-crack assessment and damage accumulation due to stiffness loss of dog bone. The importance of using the alternative approach of ANN is that it avoids the complexity of modeling problems, overrides the consideration of simplified assumptions and can be developed directly from experimental data using adaptive learning mechanisms. The proposed artificial neural network model provides a relationship between microdamage accumulation, stiffness loss and number of fatigue cycles (Nf) to failure from an experimental study where stiffness loss and crack area (Cr.Ar., mm2/mm2) are evaluated. This preliminary study using ANN for microdamage evaluation shows that ANN accurately predicts the amount of damage accumulation from stiffness loss.

Animals↗
PubMed2002

Digital dermoscopy analysis and artificial neural network for the differentiation of clinically atypical pigmented skin lesions: a retrospective study.

Noninvasive diagnostic methods such as dermoscopy or epiluminescence light microscopy have been developed in an attempt to improve diagnostic accuracy of pigmented skin lesions. The evaluation of the many morphologic characteristics of pigmented skin lesions observable by epiluminescence light microscopy, however, is often extremely complex and subjective. With the aim of obviating these problems of qualitative interpretation, methods based on mathematical analysis of pigmented skin lesions have recently been designed. These methods are based on computerized analysis of digital images obtained by epiluminescence light microscopy. In this study we used a digital dermoscopy analyzer with 147 clinically atypical pigmented skin lesions (90 nevi and 57 melanomas) to determine its discriminating power with respect to histologic diagnosis. The system evaluated 48 objective parameters used to train an artificial neural network. Using the artificial neural network with 10 variables selected by a stepwise procedure, we obtained a maximum accuracy in distinguishing melanoma from benign lesions of about 93%. Comparing this result with those of the many studies using classical epiluminescence light microscopy, it emerges that the method proposed is equal or even superior in diagnostic accuracy and has the advantage of not depending on the expertise of the clinician who examines the lesion.

Humans↗
PubMed1999

The effects of learning and intention on the neural network involved in the perception of meaningless actions.

PET was used to explore the neural network involved in the perception of meaningless action. In two conditions, subjects observed learned and unknown meaningless actions without any purpose. In two other conditions, subjects observed the same type of stimuli for later imitation. The control condition, which consisted of the presention of stationary hands, served as a baseline. Unsurprisingly, a common network that forms part of the dorsal pathway was engaged in all conditions when compared with stationary hands, and this was interpreted as being devoted to the analysis of hand movements. One of the most striking results of the present study was that some brain areas were strongly modulated by the learning level, independent of the subject's intention. Two different effects were observed: a reduced activity in posterior regions within the common network, which correlated with specific increases in the frontopolar area 10 and in the angular gyrus during the perception of learned meaningless actions compared with the perception of unknown actions. Finally, the major effect of the subject's intention to imitate was a strong increase in the dorsal pathway extending to the lateral premotor cortex and to the dorsolateral prefrontal cortex, which reflects the information processing needed for prospective action. Overall, our results provide evidence for both an effect of the visuomotor learning level and of the subject's intention on the neural network involved during the perception of human meaningless actions.

Adult↗
PubMed2004

A method to estimate emission rates from industrial stacks based on neural networks.

This paper presents a technique based on artificial neural networks (ANN) to estimate pollutant rates of emission from industrial stacks, on the basis of pollutant concentrations measured on the ground. The ANN is trained on data generated by the ISCST3 model, widely accepted for evaluation of dispersion of primary pollutants as a part of an environmental impact study. Simulations using theoretical values and comparison with field data are done, obtaining good results in both cases at predicting emission rates. The application of this technique would allow the local environment authority to control emissions from industrial plants without need of performing direct measurements inside the plant.

Air Pollutants↗
PubMed1995

Speech sound acquisition, coarticulation, and rate effects in a neural network model of speech production.

This article describes a neural network model of speech motor skill acquisition and speech production that explains a wide range of data on variability, motor equivalence, coarticulation, and rate effects. Model parameters are learned during a babbling phase. To explain how infants learn language-specific variability limits, speech sound targets take the form of convex regions, rather than points, in orosensory coordinates. Reducing target size for better accuracy during slower speech leads to differential effects for vowels and consonants, as seen in experiments previously used as evidence for separate control processes for the 2 sound types. Anticipatory coarticulation arises when targets are reduced in size on the basis of context; this generalizes the well-known look-ahead model of coarticulation. Computer simulations verify the model's properties.

Humans↗
PubMed1998

Self-organizing neural networks bridge the biomolecular resolution gap.

Topology-representing neural networks are employed to generate pseudo-atomic structures of large-scale protein assemblies by combining high-resolution data with volumetric data at lower resolution. As an application example, actin monomers and structural subdomains are located in a three-dimensional (3D) image reconstruction from electron micrographs. To test the reliability of the method, the resolution of the atomic model of an actin polymer is lowered to a level typically encountered in electron microscopic reconstructions. The atomic model is restored with a precision nine times the nominal resolution of the corresponding low-resolution density. The presented self-organizing computing method may be used as an information-processing tool for the synthesis of structural data from a variety of biophysical sources.

Actins↗
PubMedDate not supplied

Use of neural networks in medical expert systems.

A prototype Expert System, able to assist gynaecologists in diagnosing and classifying cancer of the neck of the womb is designed and implemented. The system is implemented using NeuroShell, an Expert System Building Tool, based on Neural Networks. A neural model, which predicts the possible outcomes of colposcopy and biopsy, based on Pap test results, as well as on information from the patient record, is developed. Such a system is a very useful tool for physicians both as a training tool as well as a forecasting system, while it also has the capability of incorporating new knowledge, provided by each new case it is presented with. Furthermore, a number of different implementations are provided in order to evaluate a comparative study of their performance.

Artificial Intelligence↗
PubMed1997

Three-dimensional quantitative structure-activity relationships from molecular similarity matrices and genetic neural networks. 1. Method and validations.

The utility of genetic neural network (GNN) to obtain quantitative structure-activity relationships (QSAR) from molecular similarity matrices is described. In this application, the corticosteroid-binding globulin (CBG) binding affinity of the well-known steroid data set is examined. Excellent predictivity can be obtained through the use of either electrostatic or shape properties alone. Statistical validation using a standard randomization test indicates that the results are not due to chance correlations. Application of GNN on the combined electrostatic and shape matrix produces a six-descriptor model with a cross-validated r2 value of 0.94. The model is superior to those obtained from partial least-squares and genetic regressions, and it also compares favorably with the results for the same data set from other established 3D QSAR methods. The theoretical basis for the use of molecular similarity in QSAR is discussed.

Algorithms↗
PubMed1995

Application of neural networks for the prediction of human pharmacokinetic parameters.

Artificial neural network (ANN) is a method used in the prediction of response variables from a set of input and target parameters. The most commonly used network in the area of pattern recognition is the feed forward/back propagation (BPN) network. A method to predict human pharmacokinetic parameters has been proposed using BPN with a combination of physicochemical properties and animal pharmacokinetic parameters. The results were compared with in vitro estimation of the same pharmacokinetic parameters. Fourteen network models, using a variety of input variables, were developed. Protein binding, partition coefficients, dissociation constants, and the total clearance (Cltot) and volume of distribution (Vz) in rat and dog species of 41 drugs were evaluated for prediction of human total clearance and volume of distribution using the EDBD algorithm. The observations showed highest prediction for Cltot and Vz when rat and dog pharmacokinetics, combined with protein binding and partition coefficients of the drugs, were used as input parameters. Drugs with a partition coefficient (log P) < 1.17 showed predictability of 63.41% for Cltot and 48.78% for Vz. Drugs with low protein binding (approximately 20%) showed predictability of 19.51% for Cltot and 41.46% for Vz. Comparison with in vitro estimation showed no bias in the prediction of either clearance (p < 0.2) or volume of distribution (p < 0.5) by the two methods.

Animals↗