SEARCH · PubMed Health

Results for “feature selection”

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 127 records · Page 7Linked to original sources

PubMed2002

Patterns of chemokine receptor expression on peripheral blood gamma delta T lymphocytes: strong expression of CCR5 is a selective feature of V delta 2/V gamma 9 gamma delta T cells.

Gammadelta T lymphocytes play an important role in the immune defense against infection, based on the unique reactivity of human Vdelta2Vgamma9 gammadelta T cells toward bacterial phosphoantigens. Chemokines and their corresponding receptors orchestrate numerous cellular reactions, including leukocyte migration, activation, and degranulation. In this study we investigated the expression of various receptors for inflammatory and homeostatic chemokines on peripheral blood gammadelta T cells and compared their expression patterns with those on alphabeta T cells. Although several of the analyzed receptors (including CCR6, CCR7, CXCR4, and CXCR5) were not differentially expressed on gammadelta vs alphabeta T cells, gammadelta T cells expressed strongly increased levels of the RANTES/macrophage inflammatory protein-1alpha/-1beta receptor CCR5 and also enhanced levels of CCR1-3 and CXCR1-3. CCR5 expression was restricted to Vdelta2 gammadelta T cells, while the minor subset of Vdelta1 gammadelta T cells preferentially expressed CXCR1. Stimulation with heat-killed extracts of Mycobacterium tuberculosis down-modulated cell surface expression of CCR5 on gammadelta T cells in a macrophage-dependent manner, while synthetic phosphoantigen isopentenyl pyrophosphate and CCR5 ligands directly triggered CCR5 down-modulation on gammadelta T cells. The functionality of chemokine receptors CCR5 and CXCR3 on gammadelta T cells was demonstrated by Ca(2+) mobilization and chemotactic response to the respective chemokines. Our results identify high level expression of CCR5 as a characteristic and selective feature of circulating Vdelta2 gammadelta T cells, which is in line with their suspected function as Th1 effector T cells.

Adult↗

PubMed2002

Feature selection for the classification of movements from single movement-related potentials.

Classification of movement-related potentials recorded from the scalp to their corresponding limb is a crucial task in brain-computer interfaces based on such potentials. Many features can be extracted from raw electroencephalographic signals to be used for classification, but the utilization of irrelevant or superfluous features is detrimental to the performance of classification algorithms. It is, therefore, necessary to select a small number of relevant features for the classification task. This paper demonstrates the use of two feature selection methods to choose a small number (10-20) of relevant features from a bank containing upward of 1000 features. One method is based on information theory and the other on the use of genetic algorithms. We show that the former is poorly suited for the aforementioned classification task and discuss the probable reasons for this. However, using a genetic algorithm on data recorded from five subjects we demonstrate that it is possible to differentiate between the movements of two limbs with a classification accuracy of 87% using as little as 10 features without subject training. With the addition of a simple coding scheme, this method can be applied to multiple limb classification and a 63% classification accuracy rate can be reached when attempting to distinguish between three limbs.

Adult↗

PubMed2005

Feature selection in Bayesian classifiers for the prognosis of survival of cirrhotic patients treated with TIPS.

The transjugular intrahepatic portosystemic shunt (TIPS) is a treatment for cirrhotic patients with portal hypertension. A subgroup of patients dies in the first 6 months and another subgroup lives a long period of time. Nowadays, no risk factors have been identified in order to determine how long a patient will survive. An empirical study for predicting the survival rate within the first 6 months after TIPS placement is conducted using a clinical database with 107 cases and 77 variables. Applications of Bayesian classification models, based on Bayesian networks, to medical problems have become popular in the last years. Feature subset selection is useful due to the heterogeneity of the medical databases where not all the variables are required to perform the classification. In this paper, filter and wrapper approaches based on the feature subset selection are adapted to induce Bayesian classifiers (naive Bayes, selective naive Bayes, semi naive Bayes, tree augmented naive Bayes, and k-dependence Bayesian classifier) and are applied to distinguish between the two subgroups of cirrhotic patients. The estimated accuracies obtained tally with the results of previous studies. Moreover, the medical significance of the subset of variables selected by the classifiers along with the comprehensibility of Bayesian models is greatly appreciated by physicians.

Bayes Theorem↗

PubMed2005

The contribution of spike threshold to acoustic feature selectivity, spike information content, and information throughput.

Hypotheses of sensory coding range from the notion of nonlinear "feature detectors" to linear rate coding strategies. Here, we report that auditory neurons exhibit a novel trade-off in the relationship between sound selectivity and the information that can be communicated to a postsynaptic cell. Recordings from the cat inferior colliculus show that neurons with the lowest spike rates reliably signal the occurrence of stereotyped stimulus features, whereas those with high response rates exhibit lower selectivity. The highest information conveyed by individual action potentials comes from neurons with low spike rate and high selectivity. Surprisingly, spike information is inversely related to spike rates, following a trend similar to that of feature selectivity. Information per time interval, however, was proportional to measured spike rates. A neuronal model based on the spike threshold of the synaptic drive accurately accounts for this trade-off: higher thresholds enhance the spiking fidelity at the expense of limiting the total communicated information. Such a constraint on the specificity and throughput creates a continuum in the neural code with two extreme forms of information transfer that likely serve complementary roles in the representation of the auditory environment.

Acoustic Stimulation↗

PubMedDate not supplied

Image features selected by neurons of the cat primary visual cortex.

The sensitivity of neurons in field 17 of the visual cortex in cats to cross-shaped, Y-shaped, and star-shaped figures flashing in the receptive field was studied. About 40% of the neurons studied (114 of 289) were found to generate large responses (with an average response factor of 3.06 +/- 0.32) to one of the figures flashing in the center of the receptive field, as compared with the responses produced to a single bar in the optimal orientation. Most of these neurons (72%) were selectively sensitive to the shape and orientation of figures; the remainder demonstrated some degree of tuning invariance to these properties. The latent periods of responses to figures were usually shorter than those of responses to bars. Tuning parameters for bars and figures were generally related: neurons with acute orientational tuning to a bar were usually highly selective to both the configuration and the orientation people figures. Separate or combined stimulation with crosses in the center and near periphery of the receptive fields demonstrated summation, antagonism, or the lack of any interaction between these zones in producing sensitivity to crosses. Local blockade of intracortical GABAergic inhibition by microiontophoretic application of bicuculline showed that in one third of the neurons studied, sensitivity to figures was generated or enhanced by inhibition in normal conditions, while one third of cells showed suppression by inhibition, and sensitivity in the remainder was independent of inhibition. These data show that reconsideration of existing concepts of the role of field 17 in selecting only first-order shape features of images (i.e., the orientations of single lines) is needed, since almost half the neurons in the cat primary visual cortex can efficiently detect second-order features (angles and line intersections).

Action Potentials↗

PubMed2006

Feature selectivity in area 21a of the cat.

We compared the feature tuning of neuronal activity in area 21a with the tuning in areas 17/18. Local field potentials and multi-unit activity recorded in alert animals showed similar selectivity to orientation in both areas, higher selectivity to spatial frequencies in areas 17/18 and higher selectivity and tuning significance to temporal frequencies in areas 17/18. In addition, only at sites in areas 17/18 did the local field potential exhibit locking to a horizontal motion pattern extracted from a natural movie. These results suggest that area 21a is concerned with the analysis of spatial features but lacks a faithful representation of temporal features. Hence, they foster the hypothesis that cortical area 21a is part of a ventral form pathway.

Action Potentials↗

PubMed2026

engGNN: a dual-graph neural network for omics-based disease classification and feature selection.

Omics data, such as transcriptomics, proteomics, and metabolomics, provide critical insights into disease mechanisms and clinical outcomes. However, their high dimensionality, small sample sizes, and intricate biological networks pose major challenges for reliable prediction and meaningful interpretation. Graph neural networks offer a promising way to integrate prior knowledge by encoding feature relationships as graphs. Yet, existing methods typically rely solely on either an externally curated feature graph or a data-driven generated graph, which limits their ability to capture complementary information. To address this, we propose the external and generated Graph Neural Network (engGNN), a dual-graph framework that jointly leverages both external biological networks and data-driven generated graphs. Specifically, engGNN constructs a biologically informed undirected feature graph from established network databases and complements it with a directed feature graph derived from tree-ensemble models. This dual-graph design produces more comprehensive representations, thereby improving predictive performance and interpretability. Through extensive simulation studies and real-world applications to three independent gene expression datasets, engGNN consistently demonstrates strong classification performance compared with competitive baselines. Beyond classification, engGNN provides feature- and source-level interpretability, enabling biologically meaningful analyses such as pathway enrichment analysis. Taken together, these results highlight engGNN as a robust, flexible, and interpretable framework for disease classification and biomarker discovery in high-dimensional omics contexts.

Graph Neural Networks↗

PubMed2003

Learning distinct and complementary feature selectivities from natural colour videos.

Many biological and artificial neural networks require the parallel extraction of multiple features, and meet this requirement with distinct populations of neurons that are selective to one property of the stimulus while being non-selective to another property. In this way, several populations can resolve a set of features independently of each other, and thus achieve a parallel mode of processing. This raises the question how an initially homogeneous population of neurons segregates into groups with distinct and complementary response properties. Using a colour image sequence recorded from a camera mounted on the head of a freely behaving cat, we train a network of neurons to achieve optimally stable responses, that is, responses that change minimally over time. This objective leads to the development of colour-selective neurons. Adding a second objective, decorrelating activity within the network, a subpopulation of neurons develops with achromatic response properties. Colour selective neurons tend to be non-oriented while achromatic neurons are orientation-tuned. The proposed objective thus successfully leads to the segregation of neurons into complementary populations that are either selective for colour or orientation.

Animals↗

PubMed2005

Sensorimotor rhythm-based brain-computer interface (BCI): feature selection by regression improves performance.

People can learn to control electroencephalogram (EEG) features consisting of sensorimotor rhythm amplitudes and can use this control to move a cursor in one or two dimensions to a target on a screen. In the standard one-dimensional application, the cursor moves horizontally from left to right at a fixed rate while vertical cursor movement is continuously controlled by sensorimotor rhythm amplitude. The right edge of the screen is divided among 2-6 targets, and the user's goal is to control vertical cursor movement so that the cursor hits the correct target when it reaches the right edge. Up to the present, vertical cursor movement has been a linear function of amplitude in a specific frequency band [i.e., 8-12 Hz (mu) or 18-26 Hz (beta)] over left and/or right sensorimotor cortex. The present study evaluated the effect of controlling cursor movement with a weighted combination of these amplitudes in which the weights were determined by an regression algorithm on the basis of the user's past performance. Analyses of data obtained from a representative set of trained users indicated that weighted combinations of sensorimotor rhythm amplitudes could support cursor control significantly superior to that provided by a single feature. Inclusion of an interaction term further improved performance. Subsequent online testing of the regression algorithm confirmed the improved performance predicted by the offline analyses. The results demonstrate the substantial value for brain-computer interface applications of simple multivariate linear algorithms. In contrast to many classification algorithms, such linear algorithms can easily incorporate multiple signal features, can readily adapt to changes in the user's control of these features, and can accommodate additional targets without major modifications.

Adult↗

PubMed2000

Evidence for suppressive mechanisms in attentional selection: feature singletons produce inhibitory surrounds.

Four experiments explored inhibitory mechanisms related to attentional selection. Observers viewed multielement displays and performed a form discrimination task involving a probe element. Also present in the stimulus display was a singleton element (possessing a unique color or orientation). In Experiments 1-3, probe discrimination performance was measured as a function of the distance between the probe and the singleton. Experiment 1 revealed that probe discriminations suffered when the probe was adjacent to the singleton, but improved as the spatial separation between the probe and attentionally salient singleton increased. Experiment 2 added a control condition, revealing that probe discriminations were inhibited near the singleton, but returned to control level performance with increased separation. Further, the amount of inhibition increased with larger stimulus onset asynchronies between the singleton and probe. Experiment 3 demonstrated that the extent of the inhibitory region is spatially mediated. In Experiment 4, the task was modified to one of probe detection. No inhibition was observed in the detection task, indicating that the decrease in probe discrimination performance observed in Experiments 1-3 was not due to observers' inability to detect the probe element.

Attention↗

PubMed2003

[Diagnosis of calvarial lesions. Feature selection by neural network and logistic regression].

OBJECTIVES: To establish the minimun set of features needed in the diagnosis of calvarial lesions using computed tomography (CT) and to assess the accuracy of logistic regression (LR) and artificial neural networks (NN) for their diagnosis. MATERIAL AND METHODS: 167 patients with calvarial lesions as the only known disease were enrolled. The clinical and CT data were used for LR and NN models. Both models were tested with the jacknife method. The final results of each model were compared using the area under ROC curves (A 2 ). RESULTS: The lesions were 73.1 % benign and 26.9% malignant. There was no statistically significant difference between LR and NN in differentiating malignancy. In characterizing the histologic diagnoses, NN was statistically superior to LR. Important NN features needed for malignancy classification were age and edge definition, and for the histologic diagnoses matrix, marginal sclerosis and age. CONCLUSIONS: A minimum four features is needed to diagnose these lesions, not being important patients' symptoms. NNs offer wide possibilities over statistics for the calvarial lesions study besides a superior diagnostic performance.

Adolescent↗

PubMed1998

Calvarial eosinophilic granuloma: diagnostic models and image feature selection with a neural network.

RATIONALE AND OBJECTIVES: The authors analyzed the accuracy of diagnostic features used by an artificial neural network compared with logistic-regression analysis in the diagnosis with computed tomography (CT) of calvarial eosinophilic granuloma. MATERIALS AND METHODS: Thirty-one of 167 patients with calvarial lesions were found to have eosinophilic granuloma. Clinical and CT data were used for logistic-regression and neural network models. Both models were tested by using the leave-one-out method. The final results of each model were compared by means of the area under the receiver operating characteristic curve (Az). RESULTS: Identification of eosinophilic granuloma was significantly more accurate with the neural network than with logistic regression (Az = 0.9846 +/- 0.0157 [standard deviation] vs 0.9117 +/- 0.0373) (P = .001). The most important diagnostic features identified with the neural network were patient age and marginal sclerosis. For logistic regression, the most important features were age, shape, and lobularity. CONCLUSION: The neural network is a useful tool for analyzing the features of calvarial eosinophilic granuloma. Age and marginal sclerosis are important diagnostic features.

Adolescent↗

PubMed2006

Feature-selective attention enhances color signals in early visual areas of the human brain.

We used an electrophysiological measure of selective stimulus processing (the steady-state visual evoked potential, SSVEP) to investigate feature-specific attention to color cues. Subjects viewed a display consisting of spatially intermingled red and blue dots that continually shifted their positions at random. The red and blue dots flickered at different frequencies and thereby elicited distinguishable SSVEP signals in the visual cortex. Paying attention selectively to either the red or blue dot population produced an enhanced amplitude of its frequency-tagged SSVEP, which was localized by source modeling to early levels of the visual cortex. A control experiment showed that this selection was based on color rather than flicker frequency cues. This signal amplification of attended color items provides an empirical basis for the rapid identification of feature conjunctions during visual search, as proposed by "guided search" models.

Adult↗

PubMed2000

Computerized analysis of mammographic parenchymal patterns for breast cancer risk assessment: feature selection.

Our purpose in this study was to identify computer-extracted, mammographic parenchymal patterns that are associated with breast cancer risk. We extracted 14 features from the central breast region on digitized mammograms to characterize the mammographic parenchymal patterns of women at different risk levels. Two different approaches were employed to relate these mammographic features to breast cancer risk. In one approach, the features were used to distinguish mammographic patterns seen in low-risk women from those who inherited a mutated form of the BRCA1/BRCA2 gene, which confers a very high risk of developing breast cancer. In another approach, the features were related to risk as determined from existing clinical models (Gail and Claus models), which use well-known epidemiological factors such as a woman's age, her family history of breast cancer, reproductive history, etc. Stepwise linear discriminant analysis was employed to identify features that were useful in differentiating between "low-risk" women and BRCA1/BRCA2-mutation carriers. Stepwise linear regression analysis was employed to identify useful features in predicting the risk, as estimated from the Gail and Claus models. Similar computer-extracted mammographic features were identified in the two approaches. Results show that women at high risk tend to have dense breasts and their mammographic patterns tend to be coarse and low in contrast.

BRCA2 Protein↗

PubMed2000

Spectrophotometric assessment of pigmented skin lesions: methods and feature selection for evaluation of diagnostic performance.

This study documents the optical reflectance characteristics of pigmented skin lesions and evaluates their potential for improving the differential diagnosis of malignant melanoma from benign pigmented skin lesions. Optical reflectance spectra in the wavelength range 320-1100 nm were obtained from 121 lesions already selected by expert dermatologists as suspicious of malignancy. Characteristic differences in spectra from benign and malignant lesions were studied. Feature extraction showed significant differences between lesion groups classified by histology. Seven of the most relevant features were used in the discriminant analysis of reflectance spectra from 15 melanoma and 32 compound naevi which resulted in a sensitivity of 100% and specificity of 84.4% when compared with histology. This simple objective technique appears to perform as well as the expert dermatologist and may improve the diagnostic accuracy of non-specialists such as trainees and GPs. Further prospective clinical study of reflectance spectrophotometry in a larger patient group is now required.

Adolescent↗

PubMed1993

Selected features of marsupial genetics.

As a consequence of the ancient separation of the marsupial and eutherian lineages, comparative genetical studies of these two mammalian taxa can be particularly informative. The potential for marsupial genetical research has been enhanced by the development of laboratory colonies of three 'model' species--Macropus eugenii, Monodelphis domestica and Sminthopsis crassicaudata. In this paper two selected aspects of marsupial genetics are reviewed, one involving cytogenetics and the other linkage. Marsupials provide a spectacular example of karyotypic conservation. The so-called 'basic karyotype' (2n = 14) is probably ancestral in all extant marsupials. Karyotypes that do not conform to this basic arrangement are thought to have been derived from it. A notable feature of the basic karyotype is that it has been retained, possibly for as long as 150 million years, in morphologically, behaviourally and ecologically diverse species from at least five Australian and two American families; this suggests that selective forces, presently unknown, have acted to conserve the basic chromosome form and number in these species. With respect to genetic linkage, family studies in S. crassicaudata and more recently M. domestica have indicated extreme differences between the sexes with the recombination frequencies for linked loci being very much greater in males than in females, a situation that is strikingly different from that in eutherian mammals. These differences in linkage values are paralleled by differences in the number and distribution of chiasmata during male and female meiosis. Prospects for further research in marsupials, particularly research that builds upon the observations of karyotypic conservation and genetic linkage, are noted.

Animals↗

PubMed2004

Stimulus locking and feature selectivity prevail in complementary frequency ranges of V1 local field potentials.

The local field potential (LFP) is a population measure of neuronal activity complementary to spike trains. Whereas the response properties of the spiking activity in the visual cortex have been characterized extensively, the responses of the LFP have not been well explored. No coherent picture exists about which frequency ranges exhibit feature tuning or show stimulus locked activity. Addressing this, we recorded LFP in the primary visual cortex of alert cats and calculated the tuning indices for orientation, spatial and temporal frequency. Furthermore, we quantified the locking of the power in different LFP frequency bands to the velocity profile of artificial and natural stimuli. We found that the LFP in alert animals is well tuned with similar specificity to orientation, spatial frequency and temporal frequency. Tuning to these features is most prominent in two frequency bands (8-23 Hz and 39-109 Hz). In two complementary frequency bands (23-39 Hz and above 109 Hz) the dynamics of the LFP power is locked tightly to the temporal structure of the stimulus. This locking is furthermore independent of the spatial structure of the stimulus. Together these four frequency bands cover the whole frequency range investigated. In contrast to previous studies, which often reported correlates of visual processing only in a limited frequency range of the LFP, the present results suggest that the entire frequency range of the LFP can be assigned a role in visual processing.

Action Potentials↗

PubMed2001

Feature selection for examining behavior by pathology laboratories.

Australia has a universal health insurance scheme called Medicare, which is managed by Australia's Health Insurance Commission. Medicare payments for pathology services generate voluminous transaction data on patients, doctors and pathology laboratories. The Health Insurance Commission (HIC) currently uses predictive models to monitor compliance with regulatory requirements. The HIC commissioned a project to investigate the generation of new features from the data. Feature generation has not appeared as an important step in the knowledge discovery in databases (KDD) literature. New interesting features for use in predictive modeling are generated. These features were summarized, visualized and used as inputs for clustering and outlier detection methods. Data organization and data transformation methods are described for the efficient access and manipulation of these new features.

Artificial Intelligence↗