PubMed Health⌕ Search

SEARCH · PubMed Health

Results for “signal classification”

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 19 recordsLinked to original sources
PubMed1990

Self-organization of associative memory and pattern classification: recurrent signal processing on topological feature maps.

We extend the neural concepts of topological feature maps towards self-organization of auto-associative memory and hierarchical pattern classification. As is well-known, topological maps for statistical data sets store information on the associated probability densities. To extract that information we introduce a recurrent dynamics of signal processing. We show that the dynamics converts a topological map into an auto-associative memory for real-valued feature vectors which is capable to perform a cluster analysis. The neural network scheme thus developed represents a generalization of non-linear matrix-type associative memories. The results naturally lead to the concept of a feature atlas and an associated scheme of self-organized, hierarchical pattern classification.

Algorithms↗
PubMed1992

Magnetoencephalography: a tool for functional brain imaging.

At present, one of the most promising windows to the functional organization of the human brain is magnetoencephalography (MEG). By mapping the magnetic field distribution outside the head the sites of neural events can be located with an accuracy of a few millimeters and the temporal evolution of the activation can be traced with a millisecond resolution. This paper reviews some forward field calculation approaches suitable for the interpretation of the brain's electromagnetic signals. Inverse modelling with multiple dipoles is described in detail. An example of the analysis of the somatosensory evoked-responses illustrates the potential of multiple signal classification (MUSIC) algorithm in finding optimal dipole positions.

Algorithms↗
PubMed1992

Multiple dipole modeling and localization from spatio-temporal MEG data.

An array of biomagnetometers may be used to measure the spatio-temporal neuromagnetic field or magnetoencephalogram (MEG) produced by neural activity in the brain. A popular model for the neural activity produced in response to a given sensory stimulus is a set of current dipoles, where each dipole represents the primary current associated with the combined activation of a large number of neurons located in a small volume of the brain. An important problem in the interpretation of MEG data from evoked response experiments is the localization of these neural current dipoles. We present here a linear algebraic framework for three common spatio-temporal dipole models: i) unconstrained dipoles, ii) dipoles with a fixed location, and iii) dipoles with a fixed orientation and location. In all cases, we assume that the location, orientation, and magnitude of the dipoles are unknown. With a common model, we show how the parameter estimation problem may be decomposed into the estimation of the time invariant parameters using nonlinear least-squares minimization, followed by linear estimation of the associated time varying parameters. A subspace formulation is presented and used to derive a suboptimal least-squares subspace scanning method. The resulting algorithm is a special case of the well-known MUltiple SIgnal Classification (MUSIC) method, in which the solution (multiple dipole locations) is found by scanning potential locations using a simple one dipole model. Principal components analysis (PCA) dipole fitting has also been used to individually fit single dipoles in a multiple dipole problem. Analysis is presented here to show why PCA dipole fitting will fail in general, whereas the subspace method presented here will generally succeed. Numerically efficient means of calculating the cost functions are presented, and problems of model order selection and missing moments are discussed. Results from a simulation and a somatosensory experiment are presented.

Algorithms↗
PubMed1992

Probabilistic inference-based classification applied to myoelectric signal decomposition.

A new probabilistic inference based technique (IBC) for the classification of motor unit action potentials (MUAP's) is presented. This new technique discovers statistically significant relationships in the data and uses these relationships to generate classification rules. The technique was applied to the classification of MUAP's extracted from simulated myoelectric signals. Its performance was compared to that of classical template matching algorithms (TBC) applied to the same data. Using 32 time samples as features to represent the MUAP's it was found that the IBC based technique performed significantly better (p less than 0.005) than the TBC algorithms (83.0 +/- 2.6% versus 78.1 +/- 2.8% peak correct classification performance). As the size of the training set was reduced or as increasing numbers of random classification errors were introduced into the training data, the performance of the IBC and TBC techniques declined similarly. IBC performance remained superior until very small training sets (less than 30 MUAP's per motor unit) or training sets with large numbers of errors (greater than 50%) were used. Because the probabilistic inference technique can utilize nominal data it has the potential to use declarative problem domain knowledge which conceivably could improve its performance.

Action Potentials↗
PubMed1992

[Deep white matter hyperintensity in the occipital lobe on T2-weighted MRI in children. II. Classification based on the signal intensity].

Twenty-seven children, who had deep white matter hyperintensity in the occipital lobe (DWMH) on T2-weighted MRI, were classified into two groups, mild and severe, based on the signal intensity. The frequency of mild DWMH, which was iso-or hyperintense relative to the gray matter but hypointense relative to cerebrospinal fluid (CSF), decreased with aging; mild DWMH might result from a delayed myelination in the central nervous system. However, the frequency of severe DWMH, which was iso-or hyperintense relative to CSF, was not related to aging and was significantly high in severely retarded children. Therefore, severe DWMH might be a new indicator of mental retardation in children.

Adolescent↗
PubMed1991

Automatic segmentation and classification of ionic-channel signals.

Identification of ionic-channel types and their selectivity depends critically on the open channel current that can be resolved. In this paper, an automatic channel detection algorithm is proposed that is based on sequential minimization of an index which is usually used in cluster analysis. The algorithm consists of two stages, namely segmentation and classification. In the first stage, the signal samples are segmented based on the assumption that the samples in each segment should be sequentially connected. In the second stage, the resultant segments are classified with no regard to their connectivities. Results on synthetic and real channel currents are very encouraging and they suggest that this algorithm will substantially increase the productivity of many laboratories involved in ionic-channel research.

Algorithms↗
PubMed1976

Synaptic events and discharge patterns of cochlear nucleus cells. II. Frequency-modulated tones.

Responses of 99 cochlear nucleus cells and 24 cochlear nerve fibers were studied with FM signals; 14 cochlear nerve fibers and 57 cochlear nucleus cells were studied at four rates of modulation and several signal intensities. Classification of FM response patterns as symmetrical, asymmetrical, or unidirectional was based on the calculation of a symmetry factor (S), which compared the number of discharges evoked by the ascending and by the descending phases of the FM sweep. Certain FM response patterns could not adequately be described by the symmetry factor along and variables of modulation rate and signal intensity had significant influence. A correspondence was found between the four response classes evoked by a steady-frequency tone burst (primarylike, buildup, onset, and pause) and the FM response pattern. Cochlear nerve fibers showed symmetrical response patterns to FM stimulation. Primarylike units were similar to eighth nerve fibers and generally showed symmetrical FM responses. Occasional eighth nerve fibers and primarylike cells developed asymmetry at the fastest rate of modulation (50 sps). Buildup units showed a variety of response patterns to FM signals. Onset units generally showed asymmetrical response patterns with the greater response occurring to the ascending than to the descending phase of the FM sweep. Pause units showed a characteristic inhibition of activity at 5 sps (rate-dependent inhibition). Of the 57 cochlear nuclear cells studied in response to FM signals, 16 were symmetrical, another 16 were symmetrical except at the fastest modulation rate, 12 were asymmetrical, 3 were unidirectional, and 10 showed complex responses to certain signal rates or intensities. It is clear the the cat cochlear with its complex cytoarchitecture is involved in the recoding of acoustic information. Some units in cochlear nucleus demonstrate differential responses to the direction and to the rate of frequency movement. Other cochlear nucleus cells respond as eighth nerve fibers and may serve as simple "relays" in transmitting information from the cochlea to higher auditory centers.

Acoustic Stimulation↗
PubMed2025

Optimal Control of Directional False Discovery Rates in Large-Scale Testing.

The high-throughput biomedical technology enables measurement of thousands of gene expression levels contemporaneously. A major task in analyzing these gene expression data is to identify both over-expressed and under-expressed genes. The popular two-group models select the non-null genes without further classifying them as overexpression or underexpression. Consequently, two-group decision rules are unable to constrain the numbers of falsely discovered over-expressed or under-expressed genes respectively. We propose a general three-group model that allows dependence between the test statistics and develop a decision rule that separately controls the two types of false discoveries. We show that the optimal decision rule in our three-group model has a special monotonic structure. By making use of this monotonic structure, we can linearize the two-directional false discovery rate constraints. We prove that our decision rule optimizes the expected number of true discoveries while controlling the proportions of falsely discovered over-expressed and under-expressed genes at desired levels simultaneously. The data-driven versions of the proposed procedures are suggested, and their consistency is established. Comparisons with state-of-the-art approaches and applications to genomic studies show that our procedures work well.

Humans↗
PubMed1990

Signals in limb development: STOP, GO, STAY and POSITION.

Cell-to-cell interactions in early limb development are considered within the framework of the extracellular signals STOP, GO, STAY and POSITION, a classification which emphasises that the signals are elective rather than instructive, and that complexity arises from cells' response. Patterning in the limb is analysed in terms of signals that specify positional values along the anteroposterior axes, and retinoic acid is thought to be a positional morphogen. There is however, evidence for patterning which does not depend on a positional signal. In the early bud the mesenchyme gives POSITION signals to the apical ridge, which in turn provides a STAY signal to the mesenchyme in the progress zone. Non-ridge ectoderm produces a STOP signal with respect to cartilage differentiation. The pattern of cartilage differentiation is specified well before cartilage condensation. Growth factors affect both cartilage and muscle differentiation in culture. Pigment patterns result from feather germs providing STOP or GO signals to the melanoblasts which enter all feather germs. The pathways for the cell-to-cell signals are not known but may involve gap junctions.

Animals↗
PubMed1990

Phantom auditory perception (tinnitus): mechanisms of generation and perception.

Phantom auditory perception--tinnitus--is a symptom of many pathologies. Although there are a number of theories postulating certain mechanisms of its generation, none have been proven yet. This paper analyses the phenomenon of tinnitus from the point of view of general neurophysiology. Existing theories and their extrapolation are presented, together with some new potential mechanisms of tinnitus generation, encompassing the involvement of calcium and calcium channels in cochlear function, with implications for malfunction and aging of the auditory and vestibular systems. It is hypothesized that most tinnitus results from the perception of abnormal activity, defined as activity which cannot be induced by any combination of external sounds. Moreover, it is hypothesized that signal recognition and classification circuits, working on holographic or neuronal network-like representation, are involved in the perception of tinnitus and are subject to plastic modification. Furthermore, it is proposed that all levels of the nervous system, to varying degrees, are involved in tinnitus manifestation. These concepts are used to unravel the inexplicable, unique features of tinnitus and its masking. Some clinical implications of these theories are suggested.

Humans↗
PubMed1990

Signal analysis for ECG interpretation.

In ECG interpretation usually two main areas are discerned: the signal analysis and the diagnostic classification. This article reviews the major developments in the first area. ECG signal analysis itself is subdivided into the stages data acquisition, data transformation, feature selection, and data reduction. These stages are consecutively reviewed, while in the data transformation stage digital filtering, detection, wave typing, beat selection, and boundary recognition are discussed.

Algorithms↗
PubMed1991

Digitization of electrocardiograms by desktop optical scanner.

The fidelity of a semiautomated technique for converting paper electrocardiogram (ECG) tracings to digital form by optical scanning was examined. Sample tracings from one nonmechanical and three mechanical ECG writers (recorders) were used. The optically scanned signals were compared with the digitized version (402 Hz, 12-bit precision) of the original analog signals using time- and frequency-domain correlation coefficients and root mean square error. A total of 261 QRS complexes and 207 RR intervals were examined in 21 leads acquired from 8 patients. When data were low-pass filtered at 25 Hz, the correlation coefficients for the 261 QRS complexes were 0.997 +/- 0.005 (mean +/- SD) for the time domain data, 0.992 +/- 0.010 for the complex frequency domain (amplitude and phase) data, and 0.998 +/- 0.002 for the power spectrum. The corresponding correlations for the 207 RR intervals were 0.993 +/- 0.008, 0.992 +/- 0.008, and 0.993 +/- 0.009. The RMS errors, normalized for signal amplitude, were 2.62 +/- 1.28 (percent +/- SD) for QRS complexes and 1.82 +/- 0.87 for RR intervals. The correlations for the mechanical ECG recorder tracings were the same or better than those of the nonmechanical recorder, and the RMS errors were generally smaller. When data were low-pass filtered at 105 Hz, the correlation coefficients ranged from 0.984 to 0.996 for the QRS complexes and 0.982 to 0.988 for RR intervals. Root mean square errors were 4.54 +/- 2.03 and 2.38 +/- 1.14, respectively. For purposes of arrhythmia analysis by QRS classification, digitization of ECG signals by optical scanning appears equivalent to acquisition via standard analog-to-digital conversion.

Analog-Digital Conversion↗
PubMed1990

Methodology of the modular ECG analysis system MEANS.

The methodology, used in the Modular ECG Analysis System (MEANS) is described. MEANS consists of modules for signal analysis and diagnostic classification. The basic structure of the modular interpretation system remained intact over a period of 20 years, while all modules underwent many changes as a function of experience and insight, and the continuously changing information technology. The article describes the advantages of a modular approach to decision-support systems, the most important ones being easier maintenance of the software package and separate optimization and testing of each module. The overall evaluation of MEANS was done in the CSE study. Evaluation results for modules and for the entire system are presented.

Diagnosis, Computer-Assisted↗
PubMed1992

Sleep classification in infants based on artificial neural networks.

The study reports on the possibility of classifying sleep stages in infants using an artificial neural network. The polygraphic data from 4 babies aged 6 weeks, 6 months and 1 year recorded over 8 hours were available for classification. From each baby 22 signals were recorded, digitized and stored on an optical disc. Subsets of these signals and additional calculated parameters were used to obtain data vectors, each of which represents an interval of 30 sec. For classification, two types of neural networks were used, a Multilayer Perceptron and a Learning Vector Quantizer. The teaching input for both networks was provided by a human expert. For the 6 sleep classes in babies aged 6 months, a 65% to 80% rate of correct classification (4 babies) was obtained for the testing data not previously seen.

Cerebral Cortex↗
PubMed1990

Sequence-specific 1H-NMR assignments for the aromatic region of several biologically active, monomeric insulins including native human insulin.

The aromatic region of the 1H-FT-NMR spectrum of the biologically fully-potent, monomeric human insulin mutant, B9 Ser----Asp, B27 Thr----Glu has been investigated in D2O. At 1 to 5 mM concentrations, this mutant insulin is monomeric above pH 7.5. Coupling and amino acid classification of all aromatic signals is established via a combination of homonuclear one- and two-dimensional methods, including COSY, multiple quantum filters, selective spin decoupling and pH titrations. By comparisons with other insulin mutants and with chemically modified native insulins, all resonances in the aromatic region are given sequence-specific assignments without any reliance on the various crystal structures reported for insulin. These comparisons also give the sequence-specific assignments of most of the aromatic resonances of the mutant insulins B16 Tyr----Glu, B27 Thr----Glu and B25 Phe----Asp and the chemically modified species des-(B23-B30) insulin and monoiodo-Tyr A14 insulin. Chemical dispersion of the assigned resonances, ring current perturbations and comparisons at high pH have made possible the assignment of the aromatic resonances of human insulin, and these studies indicate that the major structural features of the human insulin monomer (including those critical to biological function) are also present in the monomeric mutant.

Amino Acid Sequence↗
PubMed1992

On AR modelling for MEG spectral estimation, data compression and classification.

The use of the autoregressive (AR) model for magnetoencephalogram (MEG) processing is examined and compared to other methods. Spectral estimation, classification and data compression of MEG signals are studied. In application to spectral estimation the AR model is compared to the classical modified periodogram method. Also, AR modelling appears to perform very successfully when used for the classification of normal and epileptic MEG signals. Finally, the 17:1 to 23:1 data compression achieved by AR modelling, along with the above-mentioned advantages, render it suitable for storage applications. For comparison, the method of feature selection via orthogonal expansion is used as a tool to achieve data reduction. It is seen that while effective, this is less drastic than the compression of data volume achieved by AR modelling.

Brain Mapping↗