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 1,279 records · Page 71Linked to original sources
PubMed1998

Neuromodulatory control of hippocampal function: towards a model of Alzheimer's disease.

Alzheimer's disease (AD) is a progressive neurodegenerative disorder of cognitive function whose cellular pathology and molecular etiology have been increasingly and dramatically unraveled over the last several years. Despite this substantial knowledge base, the disease remains poorly understood due to a basic lack of understanding of how memories are stored and recalled in the brain. We describe a preliminary attempt at constructing a detailed model of these basic neural mechanisms; in particular, the natural dynamics of neuronal activity in hippocampal region CA3 and the modulation and control of these dynamics by subcortical cholinergic and GABAergic input to the hippocampus. We view the construction of such a model, with sufficient detail at the cellular and subcellular level, to be a necessary first step in understanding the effect of AD pathology on the functional behavior of the underlying neural circuitry. The network is based on the 66-compartment hippocampal pyramidal cell model of Traub and colleagues and their 51-compartment interneuron interconnected with realistic AMPA-, NMDA-, and GABA(A)-mediated synapses. Traub and others have shown that a network composed of these modeled cells is capable of synchronization in the gamma frequency range. We demonstrate here that this synchronization mechanism can implement an attractor-based autoassociative memory. A new input pattern arrives at the beginning of each theta cycle (comprised of 5-10 gamma cycles), and the pattern of activity across the network converges, over several gamma cycles, to a stable attractor that represents the stored memory. In this model, cholinergic deprivation, one of the hallmarks of AD, leads to a slowing of the gamma frequency which reduces the number of "cycles" available to reach an attractor state. We suggest that this may be one mechanism underlying the memory loss and cognitive slowing seen in AD. Our results also support the idea that acetylcholine acts on individual neurons to induce and maintain a transition from intrinsic bursting to spiking in pyramidal cells. These results are consistent with the hypothesis that spiking and bursting in CA3 pyramidal cells mediate separate behavioral functions, and that cholinergic input is required for the transition to and support of behavioral states associated with the online processing and recall of information.

Acetylcholine↗
PubMed1993

Digital neural networks for high-speed divisions and root extractions.

The instruction set of digital integer-valued VLSIs implementing neural networks is typically restricted to weighted sums, threshold activation functions and weight assignments. However, neural networks also need divisions (for data normalizations) and root extractions (for distance or vector norm computations). This paper presents neural networks which perform the latter operations by using only the above basic instructions. These networks thus provide "macros" to the overall network which calls them. This yields a homogeneous environment. Moreover, a "virtual base" is introduced to exploit the available parallelism. It is freely chosen so as to achieve the desired trade-off between the speed and complexity of the proposed networks (with a minimum of only two neurons). This high-speed capability also makes these structures attractive independently from neural applications.

Algorithms↗
PubMed2004

Neural network classification of autoregressive features from electroencephalogram signals for brain-computer interface design.

In this paper, we have designed a two-state brain-computer interface (BCI) using neural network (NN) classification of autoregressive (AR) features from electroencephalogram (EEG) signals extracted during mental tasks. The main purpose of the study is to use Keirn and Aunon's data to investigate the performance of different mental task combinations and different AR features for BCI design for individual subjects. In the experimental study, EEG signals from five mental tasks were recorded from four subjects. Different combinations of two mental tasks were studied for each subject. Six different feature extraction methods were used to extract the features from the EEG signals: AR coefficients computed with Burg's algorithm, AR coefficients computed with a least-squares (LS) algorithm and adaptive autoregressive (AAR) coefficients computed with a least-mean-square (LMS) algorithm. All the methods used order six applied to 125 data points and these three methods were repeated with the same data but with segmentation into five segments in increments of 25 data points. The multilayer perceptron NN trained by the back-propagation algorithm (MLP-BP) and linear discriminant analysis (LDA) were used to classify the computed features into different categories that represent the mental tasks. We compared the classification performances among the six different feature extraction methods. The results showed that sixth-order AR coefficients with the LS algorithm without segmentation gave the best performance (93.10%) using MLP-BP and (97.00%) using LDA. The results also showed that the segmentation and AAR methods are not suitable for this set of EEG signals. We conclude that, for different subjects, the best mental task combinations are different and proper selection of mental tasks and feature extraction methods are essential for the BCI design.

Algorithms↗
PubMed2004

Network transitivity and matrix models.

This paper is a step towards a systematic theory of the transitivity (clustering) phenomenon in random networks. A static framework is used, with adjacency matrix playing the role of the dynamical variable. Hence, our model is a matrix model, where matrices are random, but their elements take values 0 and 1 only. Confusion present in some papers where earlier attempts to incorporate transitivity in a similar framework have been made is hopefully dissipated. Inspired by more conventional matrix models, analytic techniques to develop a static model with nontrivial clustering are introduced. Computer simulations complete the analytic discussion.

Cluster Analysis↗
PubMed1996

Detection of seizure activity in EEG by an artificial neural network: a preliminary study.

Neural networks, inspired by the organizational principles of the human brain, have recently been used in various fields of application such as pattern recognition, identification, classification, speech, vision, signal processing, and control systems. In this study, a two-layered neural network has been trained for the recognition of temporal patterns of the electroencephalogram (EEG). This network is called a Learning Vector Quantization (LVQ) neural network since it learns the characteristics of the signal presented to it as a vector. The first layer is a competitive layer which learns to classify the input vectors. The second, linear, layer transforms the output of the competitive layer to target classes defined by the user. We have tested and evaluated the LVQ network. The network successfully detects epileptiform discharges (EDs) when trained using EEG records scored by a neurologist. Epochs of EEG containing EDs from one subject have been used for training the network, and EEGs of other subjects have been used for testing the network. The results demonstrate that the LVQ detector can generalize the learning to previously "unseen" records of subjects. This study shows that the LVQ network offers a practical solution for ED detection which is easily adjusted to an individual neurologist's style and is as sensitive and specific as an expert visual analysis.

Adolescent↗
PubMed2002

Neural network modeling to predict the hypnotic effect of propofol bolus induction.

Dose requirements of propofol to achieve loss of consciousness depend on the interindividual variability. Until now when propofol was administered by a single bolus, how to define the optimal individual dose and to assess its hypnotic effect have not been clearly studied. The goal of this study is to develop an artificial neural network model to predict the hypnotic effect of propofol on the basis of common clinical parameters. Ten parameters were chosen as the input factors based on the related literatures and clinical experiences. The bispectral index of EEG was used to record the consciousness level of patients and served as the output factor. The predictive results of neural net models were superior to that of clinician. This model could potentially help determine the optimal dose of propofol and thus reduce the anesthetic cost.

Anesthesia, Intravenous↗
PubMedDate not supplied

Adaptive neural network that subserves optimal homeostatic control of breathing.

An adaptive neural network model that exhibits the optimality and homeostasis characteristics of the respiratory control system is described. Based upon the Hopfield network structure and a postulated Hebb-like respiratory synapse with correlational short-term potentiation, the model is capable of mimicking the normal ventilatory responses to exercise and CO2 inputs without the need for an explicit exercise stimulus. Results suggest the possibility of an adaptive neuronal mechanism that effects optimal homeostatic regulation of respiration in mammals.

Adaptation, Physiological↗
PubMed2004

The combined technique for detection of artifacts in clinical electroencephalograms of sleeping newborns.

In this paper, we describe a new method combining the polynomial neural network and decision tree techniques in order to derive comprehensible classification rules from clinical electroencephalograms (EEGs) recorded from sleeping newborns. These EEGs are heavily corrupted by cardiac, eye movement, muscle, and noise artifacts and, as a consequence, some EEG features are irrelevant to classification problems. Combining the polynomial network and decision tree techniques, we discover comprehensible classification rules while also attempting to keep their classification error down. This technique is shown to out-perform a number of commonly used machine learning technique applied to automatically recognize artifacts in the sleep EEGs.

Algorithms↗
PubMed2005

Constructive feedforward neural networks using hermite polynomial activation functions.

In this paper, a constructive one-hidden-layer network is introduced where each hidden unit employs a polynomial function for its activation function that is different from other units. Specifically, both a structure level as well as a function level adaptation methodologies are utilized in constructing the network. The functional level adaptation scheme ensures that the "growing" or constructive network has different activation functions for each neuron such that the network may be able to capture the underlying input-output map more effectively. The activation functions considered consist of orthonormal Hermite polynomials. It is shown through extensive simulations that the proposed network yields improved performance when compared to networks having identical sigmoidal activation functions.

Algorithms↗
PubMed2000

[Computer-assisted cardiotocogram analysis--from descriptive to perinatal expert system].

The interpretation of cardiotocograms is essentially for perinatal management, it remains labour intensive and depends on training and experience. With the development of modern analysing systems with conventional classification techniques, it seems possible to introduce expert systems and neuronal networks for a better interpretation and classification of cardiotocograms. Techniques of artificial intelligence are powerful tools and should be integrated in future systems. Because CTG-analysis--as a single procedure--is only of limited value for sufficient fetal prognosis, we conclude that computer aided CTG-analysis in combination with additional fetal parameters, for example partial oxygen pressure, will allow the development of future sufficient perinatal expert systems. Nothing should hinder future integration of computerized cardiotocogram analysing systems into everyday clinical routine and to connect it to obstetrical databases.

Cardiotocography↗
PubMed1997

Two complementary methods for predicting peptides binding major histocompatibility complex molecules.

Peptides that bind to major histocompatibility complex products (MHC) are known to exhibit certain sequence motifs which, though common, are neither necessary nor sufficient for binding: MHCs bind certain peptides that do not have the characteristic motifs and only about 30% of the peptides having the required motif, bind. In order to develop and test more accurate methods we measured the binding affinity of 463 nonamer peptides to HLA-A2.1. We describe two methods for predicting whether a given peptide will bind to an MHC and apply them to these peptides. One method is based on simulating a neural network and another, called the polynomial method, is based on statistical parameter estimation assuming independent binding of the side-chains of residues. We compare these methods with each other and with standard motif-based methods. The two methods are complementary, and both are superior to sequence motifs. The neural net is superior to simple motif searches in eliminating false positives. Its behavior can be coarsely tuned to the strength of binding desired and it is extendable in a straightforward fashion to other alleles. The polynomial method, on the other hand, has high sensitivity and is a superior method for eliminating false negatives. We discuss the validity of the independent binding assumption in such predictions.

Artificial Intelligence↗
PubMed2000

Mean-field solution of the small-world network model.

The small-world network model is a simple model of the structure of social networks, which possesses characteristics of both regular lattices and random graphs. The model consists of a one-dimensional lattice with a low density of shortcuts added between randomly selected pairs of points. These shortcuts greatly reduce the typical path length between any two points on the lattice. We present a mean-field solution for the average path length and for the distribution of path lengths in the model. This solution is exact in the limit of large system size and either a large or small number of shortcuts.

Computer Simulation↗
PubMed1998

Using neural networks as an aid in the determination of disease status: comparison of clinical diagnosis to neural-network predictions in a pedigree with autosomal dominant limb-girdle muscular dystrophy.

Studies of the genetics of certain inherited diseases require expertise in the determination of disease status even for single-locus traits. For example, in the diagnosis of autosomal dominant limb-girdle muscular dystrophy (LGMD1A), it is not always possible to make a clear-cut determination of disease, because of variability in the diagnostic criteria, age at onset, and differential presentation of disease. Mapping such diseases is greatly simplified if the data present a homogeneous genetic trait and if disease status can be reliably determined. Here, we present an approach to determination of disease status, using methods of artificial neural-network analysis. The method entails "training" an artificial neural network, with input facts (based on diagnostic criteria) and related results (based on disease diagnosis). The network contains weight factors connecting input "neurons" to output "neurons," and these connections are adjusted until the network can reliably produce the appropriate outputs for the given input facts. The trained network can be "tested" with a second set of facts, in which the outcomes are known but not provided to the network, to see how well the training has worked. The method was applied to members of a pedigree with LGMD1A, now mapped to chromosome 5q. We used diagnostic criteria and disease status to train a neural network to classify individuals as "affected" or "not affected." The trained network reproduced the disease diagnosis of all individuals of known phenotype, with 98% reliability. This approach defined an appropriate choice of clinical factors for determination of disease status. Additionally, it provided insight into disease classification of those considered to have an "unknown" phenotype on the basis of standard clinical diagnostic methods.

Adolescent↗
PubMedDate not supplied

Do type and token effects reflect different mechanisms? Connectionist modeling of Dutch past-tense formation and final devoicing.

In this paper, we show that both token and type-based effects in lexical processing can result from a single, token-based, system, and therefore, do not necessarily reflect different levels of processing. We report three Simple Recurrent Networks modeling Dutch past-tense formation. These networks show token-based frequency effects and type-based analogical effects closely matching the behavior of human participants when producing past-tense forms for both existing verbs and pseudo-verbs. The third network covers the full vocabulary of Dutch, without imposing predefined linguistic structure on the input or output words.

Computer Simulation↗
PubMed2004

Computing and stability in cortical networks.

Cortical neurons are predominantly excitatory and highly interconnected. In spite of this, the cortex is remarkably stable: normal brains do not exhibit the kind of runaway excitation one might expect of such a system. How does the cortex maintain stability in the face of this massive excitatory feedback? More importantly, how does it do so during computations, which necessarily involve elevated firing rates? Here we address these questions in the context of attractor networks-networks that exhibit multiple stable states, or memories. We find that such networks can be stabilized at the relatively low firing rates observed in vivo if two conditions are met: (1) the background state, where all neurons are firing at low rates, is inhibition dominated, and (2) the fraction of neurons involved in a memory is above some threshold, so that there is sufficient coupling between the memory neurons and the background. This allows "dynamical stabilization" of the attractors, meaning feedback from the pool of background neurons stabilizes what would otherwise be an unstable state. We suggest that dynamical stabilization may be a strategy used for a broad range of computations, not just those involving attractors.

Action Potentials↗
PubMed2006

A neural network based method for optical patient set-up registration in breast radiotherapy.

Patient set-up optimization is required in breast-cancer radiotherapy to fill the accuracy gap between personalized treatment planning and uncertainties in the irradiation set-up. Opto-electronic systems allow implementing automatic procedures to minimize the positional mismatches of light-reflecting markers located on the patient surface with respect to a corresponding reference configuration. The same systems are used to detect the position of the irradiated body surface by means of laser spots; patient set-up is then corrected by matching the control points onto a CT based reference model through surface registration algorithms. In this paper, a non-deterministic approach based on Artificial Neural Networks is proposed for the automatic, real-time verification of geometrical set-up of breast irradiation. Unlike iterative surface registration methods, no passive fiducials are used and true real-time performance is obtained. Moreover, the non-deterministic modeling performed by the neural algorithm minimizes sensitivity to intra-fractional and inter-fractional non-rigid motion of the breast. The technique was validated through simulated activities by using reference CT data acquired on four subjects. Results show that the procedure is able to detect and reduce simulated set-up errors and revealed high reliability in patient position correction, even when the surface deformation is included in testing conditions.

Breast Neoplasms↗