Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Ensemble learning”

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 235 records · Page 13Linked to original sources
PubMed2004

Dopamine-induced dispersion of correlations between action potentials in networks of cortical neurons.

The involvement of dopamine in the process of learning, at the cellular and behavioral levels, has been studied extensively. Evidently, dopamine is released from midbrain nuclei neurons on exposure to salient unpredicted stimuli and binds to neurons of cortical and subcortical structures, where its neuromodulatory effects are exerted. The neuromodulatory effects of dopamine at the synaptic and cellular levels are very rich, but it is difficult to extrapolate from these elementary levels what their effect might be at the behaviorally relevant level of neuronal ensembles. Using multi-site recordings from networks of cortical neurons developing ex vivo, we studied the effects of dopamine on connectivity within neuronal ensembles. We found that dopamine disperses correlations between individual neuronal activities while preserving the global distribution of correlations at the network level. Using selective D(1) and D(2) modulators, we show that both receptor types are contributing to dopamine-induced dispersion. Our results indicate that, at the neuronal ensemble level, dopamine acts to enhance changes in network connectivity rather than stabilize such connections.

Action Potentials↗
PubMed2006

An ensemble of K-local hyperplanes for predicting protein-protein interactions.

Prediction of protein-protein interaction is a difficult and important problem in biology. In this paper, we propose a new method based on an ensemble of K-local hyperplane distance nearest neighbor (HKNN) classifiers, where each HKNN is trained using a different physicochemical property of the amino acids. Moreover, we propose a new encoding technique that combines the amino acid indices together with the 2-Grams amino acid composition. A fusion of HKNN classifiers combined with the 'Sum rule' enables us to obtain an improvement over other state-of-the-art methods. The approach is demonstrated by building a learning system based on experimentally validated protein-protein interactions in human gastric bacterium Helicobacter pylori and in Human dataset.

Algorithms↗
PubMed2002

Schisandrin B protects against tacrine- and bis(7)-tacrine-induced hepatotoxicity and enhances cognitive function in mice.

Intragastric administration (100-200 micromol/kg) of tacrine (THA) or bis(7)-THA could cause an acute and dose-dependent increase in plasma alanine aminotransferases activity in mice at 6 h after the drug administration. The increase in plasma enzyme activity was associated with an increase in hepatic malondialdehyde level, an indirect index of oxidative tissue damage. Pretreating mice with schisandrin B (Sch B), an active dibenzocyclooctadiene derivative isolated from the fruit of Schisandra chinensis, at a daily dose of 0.125-0.5 mmol/kg for 3 days protected against the THA/bis(7)-THA induced hepatic oxidative damage in a dose-dependent manner. Sch B treatment (0.025-0.5 mmol/kg/day x 5) also enhanced the passive avoidance-response in mice as assessed by the step-through task experiment. The ensemble of results suggests that Sch B may be useful for reducing the potential hepatotoxicity of THA/bis(7)-THA in anti-Alzheimer's therapy.

Alanine Transaminase↗
PubMed1991

Specific respiratory patterns distinguish among human basic emotions.

Prototypical respiratory-facial-postural actions ('emotional effector patterns') related to six basic emotions had been extracted from an ensemble of physiological reactions present in subjects reliving intense emotional situations (Bloch & Santibañez, 1972). Subjects reproducing these actions could evoke the corresponding subjective experience, which suggested their use as an experimental model for generating controlled emotional states. The aim of the present study was to quantify the respiratory parameters which characterize the emotions of joy-laughter, sadness-crying, fear-anxiety, anger, erotic love and tenderness. Respiratory movements and facial/postural expressions were recorded from 36 young actors who had learned in previous workshops to express these emotions by reproducing the corresponding prototypical actions. A qualitative analysis of the recordings showed that as the emotional reproduction went along, both breathing and expression evolved from an initial 'robot-like' phase to a more natural stage in which spontaneous vocalizations and gestures appeared. This suggested a partial activation of the emotional network. The quantitative analysis of the respiratory movements for the fundamental cycles showed that for anger, erotic love and tenderness significant changes in amplitude, rate and duration of the 'expiratory pause' were the major elements of differentiation, while for sadness, joy and fear inspiratory over expiratory time ratios were the elements of differentiation. These last three emotions were further characterized by small amplitude/high rate saccadic respiratory movements superposed to different phases of the fundamental cycles. It is concluded that quantitatively well differentiated sets of respiratory changes characterize each of six basic emotions. The bottom-up experimental model for generating such emotions based on the joint activation of the respiratory-facial-postural systems and its relation to corresponding 'real-life' emotions is discussed.

Adult↗
PubMed2005

Trial-by-trial discrimination of three enantiomer pairs by neural ensembles in mammalian olfactory bulb.

Populations of output neurons in the mammalian olfactory bulb (OB) exhibit distinct, widespread spatial and temporal activation patterns when stimulated with odorants. However, questions remain as to how ensembles of mitral/tufted (M/T) neurons in the mammalian OB represent odorant information. In this report, the single-trial encoding limits of random ensembles of putative single- and multiunit M/T cells in the anesthetized rat OB during presentations of enantiomers of limonene, carvone, and 2-butanol are investigated using simultaneous multielectrode recording techniques. The results of these experiments are: the individual constituents of our recorded ensembles broadly represent information about the presented odorants, the ensemble single-trial response of small spatially distributed populations of M/T neurons can readily discriminate between six different odorants, and the most consistent odorant discrimination is attained when the ensemble consists of all available units and their responses are integrated over an entire breathing cycle. These results suggest that small differences in spike counts among the ensemble members become significant when taken within the context of the entire ensemble. This may explain how ensembles of broadly tuned OB neurons contribute to olfactory perception and may explain how small numbers of individual units receiving input from distinct olfactory receptor neurons can be combined to form a robust representation of odorants.

Action Potentials↗
PubMedDate not supplied

Selection and combination of machine learning classifiers for prediction of linear B-cell epitopes on proteins.

Recently, new machine learning classifiers for the prediction of linear B-cell epitopes were presented. Here we show the application of Receiver Operator Characteristics (ROC) convex hulls to select optimal classifiers as well as possibilities to improve the post test probability (PTP) to meet real world requirements such as high throughput epitope screening of whole proteomes. The major finding is that ROC convex hulls present an easy to use way to rank classifiers based on their prediction conservativity as well as to select candidates for ensemble classifiers when validating against the antigenicity profile of 10 HIV-1 proteins. We also show that linear models are at least equally efficient to model the available data when compared to multi-layer feed-forward neural networks.

Algorithms↗
PubMedDate not supplied

Using genomic databases for sequence-based biological discovery.

The inherent potential underlying the sequence data produced by the International Human Genome Sequencing Consortium and other systematic sequencing projects is, obviously, tremendous. As such, it becomes increasingly important that all biologists have the ability to navigate through and cull important information from key publicly available databases. The continued rapid rise in available sequence information, particularly as model organism data is generated at breakneck speed, also underscores the necessity for all biologists to learn how to effectively make their way through the expanding "sequence information space." This review discusses some of the more commonly used tools for sequence discovery; tools have been developed for the effective and efficient mining of sequence information. These include LocusLink, which provides a gene-centric view of sequence-based information, as well as the 3 major genome browsers: the National Center for Biotechnology Information Map Viewer, the University of California Santa Cruz Genome Browser, and the European Bioinformatics Institute's Ensembl system. An overview of the types of information available through each of these front-ends is given, as well as information on tutorials and other documentation intended to increase the reader's familiarity with these tools.

Animals↗
PubMed2004

From plasticity to complexity: a new diagnostic method for psychiatry.

There is growing dissatisfaction regarding the available diagnostic systems for psychiatric disorders (DSM, ICD). Psychiatrists acknowledge that though mental disease reflects brain disorders, the descriptive and symptom based nature of psychiatric diagnosis bears no relation to brain functions. According to Helmut's article published in the October 2003 issue of Science, in the coming decade researchers and psychiatrists will be called upon to propose a basis for the psychiatric diagnostic system of the future. I propose a new etiology-oriented diagnostic system for psychiatry by integrating two recently emerging bodies of knowledge, one regarding plasticity and the other involving complex systems. Plasticity refers to all brain processes involved in dynamic alterations within communicating neuronal ensembles or networks, in the brain. Complexity refers to certain formulations from system theories relevant to brain dynamics and plasticity. It is proposed to divide plasticity processes into three types based on time domains: (1) "developmental plasticity", (2) "tuning plasticity" and (3) "fast stabilizing plasticity". Each type of plasticity is related to different complexity models achieved by the brain, developmental plasticity is life-long brain organization, it is related to state-space configurations molded into brain representations internalized via processes such as Hebbian learning. Tuning plasticity is related to "matching complexity" a measure of adaptability between internal configurations in the brain-system and externally originating event stimuli. Fast stabilizing plasticity is related to "neural complexity" a measure of neural network integration in the brain. Neural complexity meets the mental requirement to extract important features from different sensory inputs and to simultaneously generate coherent perceptual and cognitive states, thus balancing specialized segregated brain processes with coherent globally integrated whole brain activity. Mental disorders can be reconceptualized as disorders of plasticity resulting in disturbances of state-space brain configurations, matching and neural complexities. Personality disorders result from altered internal representations of the psychosocial environment. Depression and anxiety have been recently linked to alterations of adaptive neuronal plasticity thus reconceptualized as disorders of matching complexity. Finally, psychoses, including schizophrenia spectrum disorders, are reconceptualized as disturbances of neural complexity resulting in altered fast stabilizing plasticity. The new diagnostic system generates testable predictions regarding diagnosis and treatments of mental disorders which may be the future of psychiatry.

Cognition↗
PubMed2005

Orbitofrontal ensemble activity monitors licking and distinguishes among natural rewards.

The classification of rhythmic licking into clusters has proved to be useful for characterizing brain mechanisms that modulate the ingestion of natural rewards (sucrose and water). One cortical area that is responsive to rewarding stimuli is the orbitofrontal cortex (OFC). However, it is not presently known how OFC neurons respond while rodents freely lick for natural rewards and whether these responses are related to the structure of licking clusters. We addressed these issues by showing that temporary inactivation of the OFC decreases the duration and increases the number of clusters and that the activity of OFC neurons changed at precise times before, during, and after the cluster terminates. Furthermore, analysis of the activity of OFC neuronal ensembles showed that they could discriminate cluster onset from termination, predict when a behaving animal will begin a cluster, and distinguish and anticipate between natural rewards. These results provide a new role for the OFC in influencing licking clusters and anticipating specific rewards.

Animals↗
PubMed2026

Predicting ACL injury risk in athletes: A systematic review of machine learning-based models.

BACKGROUND: Early ACL injury risk identification in athletes is essential. This systematic review examines machine learning (ML) models for predicting ACL injuries, evaluating their methodological quality, performance, and reliability. METHOD: A comprehensive electronic search was conducted across PubMed, Scopus, Web of Science, and IEEE Xplore databases, supplemented by Google Scholar for grey literature, covering articles published between January 1, 2015, and August 30, 2025. Eligible studies were appraised using the Prediction Model Study Risk of Bias Assessment Tool (PROBAST) for methodological quality and risk of bias, and the Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis (TRIPOD) guidelines for quality of evidence. RESULTS: Ten studies were included. PROBAST showed eight studies had moderate risk of bias and two low risk. TRIPOD found only two studies met quality criteria. ML models included logistic regression (n = 5), support vector machines (n = 4), k-nearest neighbor (n = 3), decision trees (n = 3), random forests (n = 5), neural networks (n = 2), linear discriminant analysis (n = 1), and pre-trained CNNs (n = 1). AUC ranged from 0.63 to 0.98. Accuracy (reported in six studies) ranged from 26% to 95%; however, these values should be interpreted with caution due to the absence of confidence intervals, lack of class imbalance handling, and limited external validation across studies. Tree-based ensemble methods such as random forest achieved competitive accuracy (74-86%), while SVM, a non-ensemble classifier, reported accuracy ranging from 71% to 95%; however, the highest values were obtained in studies with notably small sample sizes (n = 12 to n = 39), raising concerns about overfitting and generalizability. CONCLUSION: Current ML algorithms show promise for identifying athletes at high ACL injury risk and detecting relevant risk factors. Although study quality was generally satisfactory, future research should prioritize external validation and model interpretability to support clinical translation.

Humans↗
PubMed1990

Development of a spin-glass model of prebiotic evolution: environmental effects on ensembles of genetic polymers.

We develop in two ways an existing spin-glass model of prebiotic polymer evolution. First, by choosing the environment J in a prescribed manner, similar to neural network presciptions, we may create an environment which favors a linearly independent set of evolutionary niches (Ea). That is, we may control which polymer "species" will evolve in our system. Computer simulations confirm this result. We obtain a quantitative value for the sharpness of a niche. Second, we extend the model by allowing a surviving polymer to act upon--to "remold"--its environment; the nature of the environmental action is governed by the "molding" matrix M. When the mold M is the identity matrix, the feedback algorithm reduces to a Hebb learning algorithm form, and a surviving polymer acts to enhance its own survival prospects. Molds having a structure analogous to (temporal) associative memories in neural networks can generate autocatalytic species or can exhibit symbiotic interspecies relationships.

Biological Evolution↗
PubMed1997

Configural processing in memory retrieval: multiple cues and ensemble representations.

When retrieving information from memory, a number of contextual cues may interact to determine which ideas will be easily accessible. Even the simplest case in which joint cue action obtains (two cues) is very revealing of the principles of memory access and representation of compounds. Mechanisms by which dual cues interact to constrain retrieval from episodic memory are considered. A holistic mechanism of cue integration is contrasted with two nonholistic mechanisms: a multiplicative or intersection mechanism and an independent-contributions mechanism. Holistic- and intersection-cuing mechanisms are consistent with different variants of compound cue models of priming. The independent cuing mechanism is consistent with spreading activation models of priming. Data from four experiments which examined dual-cued recognition of items from (newly learned) triples demonstrated strongly configural, holistic, action of dual cues. The two cues and test item must form an encoded compound to yield cuing advantages. Two independent cues to the test item are ineffective if the two cues and test were not learned together as a triple; one valid and one invalid cue are also ineffective. This is so despite the availability of pairwise information for each cue-test relation, and despite the fact that these cues are effective when operating alone. A compound cue model which predicts precisely this surprising pattern of priming is developed. The compound cue model also predicts previously obtained configural priming of associative judgments, as well as the bias priming generally observed in item recognition and similar paradigms.

Analysis of Variance↗
PubMed2003

Place representation within hippocampal networks is modified by long-term potentiation.

In the brain, information is encoded by the firing patterns of neuronal ensembles and the strength of synaptic connections between individual neurons. We report here that representation of the environment by "place" cells is altered by changing synaptic weights within hippocampal networks. Long-term potentiation (LTP) of intrinsic hippocampal pathways abolished existing place fields, created new place fields, and rearranged the temporal relationship within the affected population. The effect of LTP on neuron discharge was rate and context dependent. The LTP-induced "remapping" occurred without affecting the global firing rate of the network. The findings support the view that learned place representation can be accomplished by LTP-like synaptic plasticity within intrahippocampal networks.

Animals↗
PubMed1998

Dopamine and the mechanisms of cognition: Part I. A neural network model predicting dopamine effects on selective attention.

BACKGROUND: Dopamine affects neural information processing, cognition, and behavior; however, the mechanisms through which these three levels of function are affected have remained unspecified. We present a parallel-distributed processing model of dopamine effects on neural ensembles that accounts for effects on human performance in a selective attention task. METHODS: Task performance is stimulated using principles and mechanisms that capture salient aspects of information processing in neural ensembles. Dopamine effects are simulated as a change in gain of neural assemblies in the area of release. RESULTS: The model leads to different predictions as a function of the hypothesized location of dopamine effects. Motor system effects are simulated as a change in gain over the response layer of the model. This induces speeding of reaction times but an impairment of accuracy. Cognitive attentional effects are simulated as a change in gain over the attention layer. This induces a speeding of reaction times and an improvement of accuracy, especially at very fast reaction times and when processing of the stimulus requires selective attention. CONCLUSIONS: A computer simulation using widely accepted principles of processing in neural ensembles can account for reaction time distributions and time-accuracy curves in a selective attention task. The simulation can be used to generate predictions about the effects of dopamine agonists on performance. An empirical study evaluating these predictions is described in a companion paper.

Attention↗
PubMed2001

Synchronous activity in the hippocampus and nucleus accumbens in vivo.

The hippocampus is one of the brain regions involved in cognitive functions, including learning and memory. Extensive studies have unveiled how information is processed within this system. However, the mechanisms by which hippocampal activity is translated into action remain unsolved. One important target of hippocampal projections is the nucleus accumbens, which has been described as the motivation-to-action interface. Previous experiments indicate that these projections can control information processing in this region by setting neurons into a depolarized state. Here, we report that membrane potential transitions in nucleus accumbens neurons are correlated with electrical activity in the ventral hippocampus, suggesting that hippocampal neural activity can determine ensembles of active accumbens neurons.

Animals↗
PubMed1989

The effect of learning on the face selective responses of neurons in the cortex in the superior temporal sulcus of the monkey.

Neurophysiological studies have shown that some neurons in the cortex in the superior temporal sulcus and the inferior temporal gyrus of macaque monkeys respond to faces. These neurons provided a consistently identifiable substrate with which studies of the storage of visual information were performed. To determine whether face responsive neurons change how much they respond to different novel faces as they become familiar, neurons were tested with two experimental designs. In the first experiment, 22 neurons were tested on their responsiveness to the different members of a large set of novel faces as the set was presented repeatedly until the faces became familiar. 6 neurons altered the relative degree to which they responded to the different members of the set between the first two presentations and subsequent presentations. In a control condition, only 1 out of 17 neurons showed a significant response difference between the first two presentations and subsequent presentations when the experiment started with faces which were already familiar to the monkey. In the second experiment, 26 neurons were tested on their responsiveness to the different members of a set of familiar faces before and after the addition of a novel face to the set. 5 neurons altered the relative degree in which they responded to the different members of the set of familiar faces after addition of a novel face. It is suggested that these changes in neuronal responsiveness to different stimuli reflect the setting up of an ensemble encoded representation of face stimuli. This alteration of neuronal responsiveness as novel faces become familiar suggests that face responsive neurons may store information useful in visual recognition. In addition to this relatively long-term alteration of relative neuronal responsiveness to different stimuli, it was found that a large number of cells showed a higher mean response to the first presentation of a set of novel faces than to subsequent presentations of the faces. However, the response to the first presentation of a set of familiar faces was also higher than to subsequent presentations in that sequence. This pattern indicates a short term recency effect in the response of these neurons to visual stimuli which is similar to that previously reported (Baylis and Rolls 1987).

Animals↗
PubMed2003

Interval timing and the encoding of signal duration by ensembles of cortical and striatal neurons.

This study investigated the firing patterns of striatal and cortical neurons in rats in a temporal generalization task. Striatal and cortical ensembles were recorded in rats trained to lever press at 2 possible criterion durations (10 s or 40 s from tone onset). Twenty-two percent of striatal and 15% of cortical cells had temporally specific modulations in their firing rate, firing at a significantly different rate around 10 s compared with 40 s. On 80% of trials, a post hoc analysis of the trial-by-trial consistency of the firing rates of an ensemble of neurons predicted whether a spike train came from a time window around 10 s versus around 40 s. Results suggest that striatal and cortical neurons encode specific durations in their firing rate and thereby serve as components of a neural circuit used to represent duration.

Animals↗
PubMed2006

Cortical memory mechanisms and language origins.

We have previously proposed that cortical auditory-vocal networks of the monkey brain can be partly homologized with language networks that participate in the phonological loop. In this paper, we suggest that other linguistic phenomena like semantic and syntactic processing also rely on the activation of transient memory networks, which can be compared to active memory networks in the primate. Consequently, short-term cortical memory ensembles that participate in language processing can be phylogenetically tracked to more simple networks present in the primate brain, which became increasingly complex in hominid evolution. This perspective is discussed in the context of two current interpretations of language origins, the "mirror-system hypothesis" and generativist grammar.

Animals↗