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 613 records · Page 34Linked to original sources
PubMedDate not supplied

[Application of neural networks in occupational medicine].

The medical literature offers today many contributions in which neural networks have been used. The introduction of these tools as diagnostic aid appears very interesting also into various sectors of research in the occupational field. The author proceeds to a schematic description of the functioning of a neural network, describing an applied example. It is therefore estimated the feasible advantages which result from the use in the research of neural networks. Finally a possible fields of use of the neural networks in occupational medicine is highlighted, from the detection of lung nodules in digital chest radiographs to the detection of peripheral vascular disease from upper limb pulse waveforms, from the low back disorders due to workplace design to the diagnosis of occupational asthma.

Forecasting↗
PubMed1999

Classification of asthmatic breath sounds: preliminary results of the classifying capacity of human examiners versus artificial neural networks.

For continuous monitoring of the respiratory condition of patients, e.g., at the intensive care unit, computer assistance is required. Existing mechanical devices, such as the peak expiratory flow meter, provide only with incidental measurements. Moreover, such methods require cooperation of the patient, which at, e.g., the ICU is usually not possible. The evaluation of complicated phenomena such as asthmatic respiratory sounds may be accomplished by use of artificial neural networks. To investigate the merit of artificial neural networks, the capacities of neural networks and human examiners to classify breath sounds were compared in this study. Breath sounds were in vivo recorded from 50 school-age children with asthma and from 10 controls. Sound intervals with a duration of 20 seconds were randomly sampled from asthmatics during exacerbation, asthmatics in remission, and controls. The samples were digitized and related to peak expiratory flow. From each interval, two full breath cycles were selected. Of each selected breath cycle, a Fourier power spectrum was calculated. The so-obtained set of spectral vectors was classified by means of artificial neural networks. Humans evaluated graphic displays of the spectra. Human examiners could not clearly discriminate between the three groups by inspecting the spectrograms. Classification by self-classifying neural networks confirmed the existence of at least three classes; however, discrimination of 11 classes seemed more appropriate. Good results were obtained with supervised networks: as much as 95% of the training vectors could be classified correctly, and 43% of the test vectors. The three patient groups, as discriminated in advance, do not correspond with three sharply separated sets of spectrograms. More than three classes seem to be present. Humans cannot take up the spectral complexity and showed negative classification results. Artificial neural networks, however, are able to handle classification tasks and show positive results.

Acoustics↗
PubMed2006

A gradual noisy chaotic neural network for solving the broadcast scheduling problem in packet radio networks.

In this paper, we propose a gradual noisy chaotic neural network (G-NCNN) to solve the NP-complete broadcast scheduling problem (BSP) in packet radio networks. The objective of the BSP is to design an optimal time-division multiple-access (TDMA) frame structure with minimal TDMA frame length and maximal channel utilization. A two-phase optimization is adopted to achieve the two objectives with two different energy functions, so that the G-NCNN not only finds the minimum TDMA frame length but also maximizes the total node transmissions. In the first phase, we propose a G-NCNN which combines the noisy chaotic neural network (NCNN) and the gradual expansion scheme to find a minimal TDMA frame length. In the second phase, the NCNN is used to find maximal node transmissions in the TDMA frame obtained in the first phase. The performance is evaluated through several benchmark examples and 600 randomly generated instances. The results show that the G-NCNN outperforms previous approaches, such as mean field annealing, a hybrid Hopfield network-genetic algorithm, the sequential vertex coloring algorithm, and the gradual neural network.

Electricity↗
PubMed2006

Some new stability properties of dynamic neural networks with different time-scales.

Dynamic neural networks with different time-scales include the aspects of fast and slow phenomenons. Some applications require that the equilibrium points of these networks to be stable. The main contribution of the paper is that Lyapunov function and singularly perturbed technique are combined to access several new stable properties of different time-scales neural networks. Exponential stability and asymptotic stability are obtained by sector and bound conditions. Compared to other papers, these conditions are simpler. Numerical examples are given to demonstrate the effectiveness of the theoretical results.

Mathematics↗
PubMed2004

Artificial neural network modeling to predict the plasma concentration of aminoglycosides in burn patients.

The goal was to use an artificial neural network model to predict the plasma concentration of aminoglycosides in burn patients and identify patients whose plasma antibiotic concentration would be sub-therapeutic based on the patients' physiological data and taking into account burn severity. Physiological data and some indicators of burn severity were collected from 30 burn patients who received arbekacin. A three-layer artificial neural network with five neurons in the hidden layer was used to predict the plasma concentration of arbekacin. Linear modeling for prediction of plasma concentration and logistic regression modeling for the classification of patients were also used and the predictive performance was compared to results from the artificial neural network model. Dose, body mass index, serum creatinine concentration and amount of parenteral fluid were selected as covariates for the plasma concentration of arbekacin. Area of burn after skin graft was a good covariate for indicating burn severity. Predictive performance of the artificial neural network model including burn severity was much better than linear modeling and logistic regression analysis. An artificial neural network model should be helpful for the prediction of plasma concentration using patients' physiological data, and burn severity should be included for improved prediction in burn patients. Because the relationship between burn severity and plasma concentration of aminoglycosides is thought to be nonlinear, it is not surprising that the artificial neural network model showed better predictive performance compared to the linear or logistic regression models.

Adult↗
PubMed2006

Robust stability of switched Cohen-Grossberg Neural networks with mixed time-varying delays.

By combining Cohen-Grossberg neural networks with an arbitrary switching rule, the mathematical model of a class of switched Cohen-Grossberg neural networks with mixed time-varying delays is established. Moreover, robust stability for such switched Cohen-Grossberg neural networks is analyzed based on a Lyapunov approach and linear matrix inequality (LMI) technique. Simple sufficient conditions are given to guarantee the switched Cohen-Grossberg neural networks to be globally asymptotically stable for all admissible parametric uncertainties. The proposed LMI-based results are computationally efficient as they can be solved numerically using standard commercial software. An example is given to illustrate the usefulness of the results.

Journal Article↗
PubMed2000

Potential of the back propagation neural network in the assessment of gait patterns in ankle arthrodesis.

OBJECTIVE: The purpose of this study was to recognize gait pattern in ankle arthrodesis by using a neural network trained with time domain input and compare the performance of the neural network with the statistical method. DESIGN: Three-layered feed-forward back propagation neural network and a statistical method were used to classify gait patterns of patients with ankle arthrodesis and normal subjects. BACKGROUND: Although backpropagation neural networks are very efficient in many pattern recognition tasks, they have not been used for gait pattern recognition of ankle arthrodesis. METHODS: A total of eighteen parameters measured by force platforms, including nine force parameters and their chronologic incidence of occurrence, were used to classify gait patterns. RESULTS: The results showed that the neural network model was able to classify subjects with recognition rates up to 95.8%. In contrast, the statistical method was only able to classify the subjects with recognition rates of 91.5%. CONCLUSIONS: The backpropagation neural network method has better accuracy than the statistical method in discriminating subjects and the time domain features carry important prognostic information. RELEVANCE: It is important to be able to quantify the changes in gait pattern after arthrodesis to understand the clinical implications of arthrodesis.

Adolescent↗
PubMedDate not supplied

A neural-network approach to predicting admission decisions in a psychiatric emergency room.

Clinical decision making is based on recognizing complex patterns of patients' signs and symptoms. Neural networks have been shown to be very effective at this type of pattern recognition, and in this study a neural-network approach was used to predict which patients seen in a psychiatric emergency room required admission and which did not. Data from all walk-in patients (N = 658) evaluated during normal working hours in a psychiatric emergency room during a one-year period were used either to train a neural network or to test its performance. The network had 53 input nodes, one hidden layer, and an output layer with a single node. The back-propagation method was used to train the network. The neural network's admitting decisions were in substantial agreement with those of the clinicians (kappa coefficient = 0.63). When used as a diagnostic test for admission it had a specificity of 94%, a sensitivity of 70%, and an overall accuracy of 91%. The information gain was 35% of that of a perfect diagnostic test. These results show that a neural network can be trained to make clinical decisions that are in substantial agreement with those of experienced clinicians.

Adolescent↗
PubMed2006

A new adaptive backpropagation algorithm based on Lyapunov stability theory for neural networks.

A new adaptive backpropagation (BP) algorithm based on Lyapunov stability theory for neural networks is developed in this paper. It is shown that the candidate of a Lyapunov function V(k) of the tracking error between the output of a neural network and the desired reference signal is chosen first, and the weights of the neural network are then updated, from the output layer to the input layer, in the sense that deltaV(k) = V(k) - V(k - 1) < 0. The output tracking error can then asymptotically converge to zero according to Lyapunov stability theory. Unlike gradient-based BP training algorithms, the new Lyapunov adaptive BP algorithm in this paper is not used for searching the global minimum point along the cost-function surface in the weight space, but it is aimed at constructing an energy surface with a single global minimum point through the adaptive adjustment of the weights as the time goes to infinity. Although a neural network may have bounded input disturbances, the effects of the disturbances can be eliminated, and asymptotic error convergence can be obtained. The new Lyapunov adaptive BP algorithm is then applied to the design of an adaptive filter in the simulation example to show the fast error convergence and strong robustness with respect to large bounded input disturbances.

Algorithms↗
PubMed2000

How well can radiologists using neural network software diagnose pulmonary embolism?

OBJECTIVE: This study evaluated and optimized the performance of an automated artificial neural network image interpreter in the diagnosis of pulmonary embolism on ventilation-perfusion lung scans. The computer interpretations were compared with the interpretations of three experienced observers. MATERIALS AND METHODS: Digital data were obtained from 100 patients with normal findings on chest radiographs who were undergoing both radionuclide ventilation-perfusion scanning and pulmonary angiography. Interpretations of differently trained neural networks were compared with those of three experienced nuclear medicine practitioners unaware of the clinical diagnosis. RESULTS: Machines running neural networks performed similarly to experienced scan interpreters in the detection of pulmonary embolism. Both the human observers and the networks performed best in cases with large emboli. Neural network performance was best in the right lung, when the networks were trained using only cases with large emboli and when networks were trained independently in the right and left lungs. The best predictions resulted from a collaborative interpretation incorporating both the human and computer predictions. CONCLUSION: Computers running artificial neural networks using scan data obtained directly from the anterior and posterior ventilation and perfusion images, without human involvement, perform comparably with experienced observers in patients with normal findings on chest radiographs. Human observers can improve their interpretations by incorporating computer output to formulate diagnostic prediction. The method of training the networks is critical to optimizing performance.

Adult↗
PubMed1993

Deterministic logic versus software-based artificial neural networks in the diagnosis of atrial fibrillation.

An investigation into the use of software-based neural networks for the detection of atrial fibrillation was made. At a specific point in the Glasgow 12-lead electrocardiographic interpretation program, a decision has to be made as to whether atrial fibrillation or sinus rhythm with supraventricular or ventricular extrasystoles is present. The same input parameters used for the deterministic logic at that point were also utilized to train a variety of neural networks. Results from a separate test set showed that the sensitivity of detecting atrial fibrillation could be improved using the best of the neural networks. On the other hand, it was felt that the original deterministic logic could be improved by considering adjustments in order that the presence of certain combinations of findings not previously regarded as representing atrial fibrillation would now do so. When the deterministic logic was upgraded in this way, it was found, again using a separate test set, that the revised logic was improved compared to the original, and also gave a performance similar to that of the neural network. It is concluded that the use of a neural network at a specific diagnostic decision point in a rhythm analysis program can be as effective as deterministic logic, which may take several years to perfect.

Atrial Fibrillation↗
PubMed1995

Neural network predicted peak and trough gentamicin concentrations.

Predictions of steady state peak and trough serum gentamicin concentrations were compared between a traditional population kinetic method using the computer program NONMEM to an empirical approach using neural networks. Predictions were made in 111 patients with peak concentrations between 2.5 and 6.0 micrograms/ml using the patient factors age, height, weight, dose, dose interval, body surface area, serum creatinine, and creatinine clearance. Predictions were also made on 33 observations that were outside the 2.5 and 6.0 micrograms/ml range. Neural networks made peak serum concentration predictions within the 2.5-6.0 micrograms/ml range with statistically less bias and comparable precision with paired NONMEM predictions. Trough serum concentration predictions were similar using both neural networks and NONMEM. The prediction error for peak serum concentrations averaged 16.5% for the neural networks and 18.6% for NONMEM. Average prediction errors for serum trough concentrations were 48.3% for neural networks and 59.0% for NONMEM. NONMEM provided numerically more precise and less biased predictions when extrapolating outside the 2.5 and 6.0 micrograms/ml range. The observed peak serum concentration distribution was multimodal and the neural network reproduced this distribution with less difference between the actual distribution and the predicted distribution than NONMEM. It is concluded that neural networks can predict serum drug concentrations of gentamicin. Neural networks may be useful in predicting the clinical pharmacokinetics of drugs.

Adult↗
PubMed1997

Opening the black box: the relationship between neural networks and linear discriminant functions.

Over the last ten years feed-forward neural networks have become a popular tool for statistical decision making. During this time, they have been applied in many fields, including cytological classification. Neural networks are often treated as a black box, whose inner workings are concealed from the researcher. This is unfortunate, since the inner workings of a neural network can be understood in a manner similar to that of a linear discriminant function, which is the standard tool that researchers use for decision making. This paper discusses feed-forward neural networks and some methods to improve their performance for classification problems. Their relationship to discriminant functions will be examined for a simple two-dimensional classification problem.

Algorithms↗
PubMed1991

A neural network model for the intersensory coordination involved in goal-directed movements.

A neural network model for a sensorimotor system, which was developed to simulate oriented movements in man, is presented. It is composed of a formal neural network comprising two layers: a sensory layer receiving and processing sensory inputs, and a motor layer driving a simulated arm. The sensory layer is an extension of the topological network previously proposed by Kohonen (1984). Two kinds of sensory modality, proprioceptive and exteroceptive, are used to define the arm position. Each sensory cell receives proprioceptive inputs provided by each arm-joint together with the exteroceptive inputs. This sensory layer is therefore a kind of associative layer which integrates two separate sensory signals relating to movement coding. It is connected to the motor layer by means of adaptive synapses which provide a physical link between a motor activity and its sensory consequences. After a learning period, the spatial map which emerges in the sensory layer clearly depends on the sensory inputs and an associative map of both the arm and the extra-personal space is built up if proprioceptive and exteroceptive signals are processed together. The sensorimotor transformations occurring in the junctions linking the sensory and motor layers are organized in such a manner that the simulated arm becomes able to reach towards and track a target in extra-personal space. Proprioception serves to determine the final arm posture adopted and to correct the ongoing movement in cases where changes in the target location occur. With a view of developing a sensorimotor control system with more realistic salient features, a robotic model was coupled with the formal neural network. This robotic implementation of our model shows the capacity of formal neural networks to control the displacement of mechanical devices.

Algorithms↗
PubMed2004

Opposition logic and neural network models in artificial grammar learning.

Following neural network simulations of the two experiments of, argued that the opposition logic advocated by was incapable of distinguishing between single and multiple influences on performance of artificial grammar learning and more generally. We show that their simulations do not support their conclusions. We also provide different neural network simulations that do simulate the essential results of Higham et al. (2000).

Artificial Intelligence↗
PubMed1998

Computerised electrocardiology employing bi-group neural networks.

A configuration of bi-group neural networks (BGNN) is proposed combined with an evidential reasoning framework to interpret 12-lead electrocardiograms for three mutually exclusive classes. A number of pre-processing feature selection techniques were investigated prior to application of the input feature vector to each individual BGNN. The network outputs were discounted within a belief interval of 1 based on their performance on test data prior to combination. It was found that the application of the feature selection techniques enhanced the individual performance of the BGNN, and subsequently enhanced the overall performance. The proposed framework was compared with conventional classification techniques of multi-output neural networks and linear multiple regression. The framework attained a higher level of classification in comparison with the other methods; 70.4% compared with 66.7% for both multi-output neural and statistical techniques.

Diagnosis, Differential↗
PubMed1998

Embedded neural networks: exploiting constraints.

Using concepts and tools of embodied cognitive science, we investigate the implications of embedding neural networks in a physical structure, the body of a robot. Embedding a neural network in a body provides constraints that can be exploited for learning. We show that the constraints are given by the environment and object properties, the agent's morphology, the agent's motor system and specific ways of interacting with the objects. We argue that designing embedded neural networks implies (a) understanding these constraints, and (b) exploiting them, i.e., designing neural networks such that they-one way or other-incorporate the constraints. This in turn results in cheap and simple networks that are suited for the task environment, and have real-time responses. Moreover, this constraint-based approach provides new perspectives on two fundamental problems of cognitive science: focus-of-attention and object constancy. The main arguments are illustrated with a series of case studies with simulated and physical mobile robots that are controlled by hand-designed as well as evolved neural networks.

Journal Article↗
PubMedDate not supplied

Numerical solution of elliptic partial differential equation using radial basis function neural networks.

In this paper a neural network for solving partial differential equations is described. The activation functions of the hidden nodes are the radial basis functions (RBF) whose parameters are learnt by a two-stage gradient descent strategy. A new growing RBF-node insertion strategy with different RBF is used in order to improve the net performances. The learning strategy is able to save computational time and memory space because of the selective growing of nodes whose activation functions consist of different RBFs. An analysis of the learning capabilities and a comparison of the net performances with other approaches have been performed. It is shown that the resulting network improves the approximation results.

Neural Networks, Computer↗