Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “representation 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 1,801 records · Page 100Linked to original sources
PubMed2003

Extrinsic cues aid shape recognition from novel viewpoints.

It has been shown previously that the visual recognition of shape is susceptible to the mismatch between the retinal input and its representation in long-term memory, especially when this mismatch arises from rotations in depth. One possibility is that the visual recognition system deals with such mismatch by a transformation of the input or the representation thereby bringing both into alignment for comparison. In either case, knowing what transformation has taken place should facilitate recognition. In natural circumstances, objects do not disappear and appear in different orientations inexplicably and an observer usually knows what to expect according to the context. This context includes the environment, and the history of the observers' movements, which specify the transient relationship between the object, the environment and the observer. We used interactive computer graphics to study the effects of providing observers with either implicit or explicit indications of their view transformations in the recognition of a class of shape found previously to be highly view-dependent. Results show that these cues aid recognition to varying degrees but mostly for oblique views and primarily in terms of accuracy not response times. These results provide evidence for egocentric encoding of shape and suggest that knowing ones' transformation in view helps to reduce the problem space involved in matching a shape percept with a mental representation.

Adolescent↗
PubMed2005

Exploring motor cortical plasticity using transcranial magnetic stimulation in humans.

It is generally accepted that functional properties of the motor cortex in adult humans can be altered through behavioral or pharmacological manipulations, as well as in some pathological conditions. The ability and capacity of adult human cortex to undergo any adaptive or reorganizational changes is referred to as plasticity. Much of the evidence concerning motor cortical plasticity have been derived from studies using the non-invasive technique of transcranial magnetic stimulation (TMS). TMS has proven to be a suitable tool to explore representational plasticity and to interact with neuronal activity in settings of induction protocols either alone or coupled with altered sensory inputs. Furthermore, plastic changes induced by motor learning protocols have attracted particular interest because of their relevance in functional recovery. Recent studies support the view that learning in human motor cortex occurs through long-term potentiation (LTP)-like mechanisms. Purposeful modulation of motor cortical plastic changes by manipulative TMS protocols may offer useful rehabilitative strategies in patients with chronic motor deficits.

Adult↗
PubMed2006

The boundary vector cell model of place cell firing and spatial memory.

We review evidence for the boundary vector cell model of the environmental determinants of the firing of hippocampal place cells. Preliminary experimental results are presented concerning the effects of addition or removal of environmental boundaries on place cell firing and evidence that boundary vector cells may exist in the subiculum. We review and update computational simulations predicting the location of human search within a virtual environment of variable geometry, assuming that boundary vector cells provide one of the input representations of location used in mammalian spatial memory. Finally, we extend the model to include experience-dependent modification of connection strengths through a BCM-like learning rule - the size and sign of strength change is influenced by historic activity of the postsynaptic cell. Simulations are compared to experimental data on the firing of place cells under geometrical manipulations to their environment. The relationship between neurophysiological results in rats and spatial behaviour in humans is discussed.

Action Potentials↗
PubMed2001

Robustness of the two-dimensional structure of recognition of facial expression: evidence under different intensities of emotionality.

Many studies have indicated that recognition of facial expression is geometrically represented in terms of two underlying bipolar dimensions, pleasure-displeasure and the intensity of arousal. A psychological space defined by these two dimensions has been found under particular conditions using prototype faces: however, the question arises whether this two-dimensional psychological space is also an accurate representation for faces with lower emotional intensity. To test this we prepared four different conditions of emotional intensity (100%, 75%, 50%, and 25% of emotionality) for six basic facial expressions using morphing. Participants (29 for the 100%, 30 for 75%, 35 for 50%, and 34 for the 25% condition) in each condition rated the facial stimuli on a 6-point scale of "not at all" and "very much so". Multidimensional scaling for each condition clearly indicated a two-dimensional representation and multiple correlations showed significant similarity across conditions (p < .001). This shows that the two-dimensional representation of recognition of facial expression is robust.

Adult↗
PubMed2003

Reversible-figure perception: mechanisms of intentional control.

Observers can exert a degree of intentional control over the perception of reversible figures. Also, the portion of the stimulus that is selected for primary or enhanced processing (focal-feature processing) influences how observers perceive a reversible figure. Two experiments investigated whether voluntary control over perception of a Necker cube could be explained in terms of intentionally selecting appropriate focal features within the stimulus for primary processing. In Experiment 1, varying observers' intentions and the focus of primary processing produced additive effects on the percentage of time that one alternative was perceived. In Experiment 2, the effect of varying the focus of primary processing was eliminated by the use of a small cube, but the effect of intention was unaltered. The results indicate that intentional control over perception can be exerted independently of focal-feature processing, perhaps by top-down activation or priming of perceptual representations. The results also reveal the limits of intentional control.

Attention↗
PubMed2004

Is categorization performance really intact in amnesia? A meta-analysis.

Most published studies of category learning in amnesia have reported intact categorization performance. These results have been used to challenge single-system accounts of categorization and recognition, in which a single representational system mediates performance in these two tasks. Many of the published studies, however, have shown a numerical advantage for controls over amnesics and often have had low statistical power. A meta-analysis was conducted to assess whether this numerical advantage is significant when the data are pooled across studies. This analysis indicates that amnesic subjects do, in fact, show deficits in categorization tasks, which is consistent with single-system exemplar model predictions.

Amnesia↗
PubMed1996

Quantity-based interference and symbolic representations in chimpanzees (Pan troglodytes).

Five chimpanzees with training in counting and numerical skills selected between 2 arrays of different amounts of candy or 2 Arabic numerals. A reversed reinforcement contingency was in effect, in which the selected array was removed and the subject received the nonselected candies (or the number of candies represented by the nonselected Arabic numeral). Animals were unable to maximize reward by selecting the smaller array when candies were used as array elements. When Arabic numerals were substituted for the candy arrays, all animals showed an immediate shift to a more optimal response strategy of selecting the smaller numeral, thereby receiving the larger reward. Results suggest that a response disposition to the high-incentive candy stimuli introduced a powerful interference effect on performance, which was effectively overridden by the use of symbolic representations.

Animals↗
PubMed1998

Identification of divergent functions in homologous proteins by induction over conserved modules.

Homologous proteins do not necessarily exhibit identical biochemical function. Despite this fact, local or global sequence similarity is widely used as an indication of functional identity. Of the 1327 Enzyme Commission defined functional classes with more than one annotated example in the sequence databases, similarity scores alone are inadequate in 251 (19%) of the cases. We test the hypothesis that conserved domains, as defined in the ProDom database, can be used to discriminate between alternative functions for homologous proteins in these cases. Using machine learning methods, we were able to induce correct discriminators for more than half of these 251 challenging functional classes. These results show that the combination of modular representations of proteins with sequence similarity improves the ability to infer function from sequence over similarity scores alone.

Alcohol Dehydrogenase↗
PubMed2002

Integration of cardiac responses to serial stimuli after Pavlovian conditioning in rats.

The expression of cardiac responses to sequences of two sounds was studied in restrained rats following discriminative trace or delay conditioning. Stimuli paired with a tail shock 10 sec later (CS1) elicited conditioned bradycardia. Unpaired or neutral stimuli (CS0) elicited mostly tachycardia. Rats did not learn to suppress responding to nonreinforced sequences with an interval of 6 sec between sounds. Responses to the second stimulus were significantly augmented following a CS1 stimulus, but not following a CS0 stimulus. Real-time summation of simple responses provided a more complete and quantitative prediction of dual responses than did resetting or facilitation. These results extend the time range over which summation may be observed from less than 2 sec to at least 16 sec. They appear to be inconsistent with models involving competition between unitary representations of stimuli in short-term memory and suggest the existence of multiple stimulus traces with independent time courses.

Animals↗
PubMed2005

fMRI reveals a common neural substrate of illusory and real contours in V1 after perceptual learning.

Perceptual learning involves the specific and relatively permanent modification of perception following a sensory experience. In psychophysical experiments, the specificity of the learning effects to the trained stimulus attributes (e.g., visual field position or stimulus orientation) is often attributed to assumed neural modifications at an early cortical site within the visual processing hierarchy. We directly investigated a neural correlate of perceptual learning in the primary visual cortex using fMRI. Twenty volunteers practiced a curvature discrimination on Kanizsa-type illusory contours in the MR scanner. Practice-induced changes in the BOLD response to illusory contours were compared between the pretraining and the posttraining block in those areas of the primary visual cortex (V1) that, in the same session, had been identified to represent real contours at corresponding visual field locations. A retinotopically specific BOLD signal increase to illusory contours was observed as a consequence of the training, possibly signaling the formation of a contour representation, which is necessary for performing the curvature discrimination. The effects of perceptual training were maintained over a period of about 10 months, and they were specific to the trained visual field position. The behavioral specificity of the learning effects supports an involvement of V1 in perceptual learning, and not in unspecific attentional effects.

Adult↗
PubMed2002

Motor learning-dependent synaptogenesis is localized to functionally reorganized motor cortex.

The regional specificity and functional significance of learning-dependent synaptogenesis within physiologically defined regions of the adult motor cortex are described. In comparison to rats in a motor activity control group, rats trained on a skilled reaching task exhibited an areal expansion of wrist and digit movement representations within the motor cortex. No expansion of hindlimb representations was seen. This functional reorganization was restricted to the caudal forelimb area, as no differences in the topography of movement representations were observed within the rostral forelimb area. Paralleling the physiological changes, trained animals also had significantly more synapses per neuron than controls within layer V of the caudal forelimb area. No differences in the number of synapses per neuron were found in either the rostral forelimb or hindlimb areas. This is the first demonstration of the co-occurrence of functional and structural plasticity within the same cortical regions and provides strong evidence that synapse formation may play a role in supporting learning-dependent changes in cortical function.

Animals↗
PubMed2026

Systematic evaluation of one-dimensional-to-two-dimensional near-infrared spectroscopy transformations with deep learning for quantifying coconut sap adulteration.

Near-infrared (NIR) spectroscopy have limitations when combined with deep learning (DL) algorithms because they rely on low-dimensional datasets. Therefore, we investigated the potential of transforming one-dimensional (1D) NIR spectra into two-dimensional (2D) spectrograms using synchronous and asynchronous techniques and the continuous wavelet transform (CWT) and their effectiveness by integrating with DL for detecting adulteration in coconut sap. NIR spectra (12,500-4000&#xa0;cm-1) were collected from binary mixtures (0%-100%;w/w). The performance of all DL (convolutional neural networks-CNN, AlexNet and ResNet) models was compared with that of partial least squares (PLS). The models were ranked in the mentioned order based on their performances: 2D-CWT&#xa0;>&#xa0;2D-asynchronous > 2D-synchronous > 1D/2D-PLS. The important features of the best model can be explained and visualized using gradient-weighted-class-activation-mapping. The findings highlight that the 1D-to-2D NIR data transformation combined with DL is a highly robust approach because it addresses the feature representation gap in NIR data and effectively captures the spatial-spectral correlations.

Spectroscopy, Near-Infrared↗
PubMed2005

Age differences in mental scanning of locomotor maps.

PURPOSE: To investigate whether the metric quality of mental spatial representations of the environment changes with aging, by means of mental scanning of locomotor space. METHOD: Twenty young adults (age 21.1-30.8) and 20 old adults (age 60.4-76.4) were submitted to a battery of psychometric tests to assess the effects of aging on general cognitive functions and specific visuo-spatial abilities, and to determine which cognitive factors correlated with scanning performance. The mental scanning task followed. Participants firstly had to learn a path comprising six positions by walking and observing these; secondly, they had to use mental imagery to scan from a first to a second position, according to different distances. RESULTS: Aging reduces the validity of the metric information incorporated by mental spatial representations. Attentional and working memory resources, ability to manipulate spatial information and a general factor of abstract visuo-spatial reasoning strongly correlated with scanning performance. Finally, aging seemed to have a selective effect on cognitive functioning. CONCLUSIONS: Age differences emerge when demanding storage and processing of information are simultaneously required, as predicted by the Limited Resources view of aged cognition.

Adult↗
PubMed2004

Phonetic perceptual identification by native- and second-language speakers differentially activates brain regions involved with acoustic phonetic processing and those involved with articulatory-auditory/orosensory internal models.

This experiment investigates neural processes underlying perceptual identification of the same phonemes for native- and second-language speakers. A model is proposed implicating the use of articulatory-auditory and articulatory-orosensory mappings to facilitate perceptual identification under conditions in which the phonetic contrast is ambiguous, as in the case of second-language speakers. In contrast, native-language speakers are predicted to use auditory-based phonetic representations to a greater extent for perceptual identification than second-language speakers. The English /r-l/ phonetic contrast, although easy for native English speakers, is extremely difficult for native Japanese speakers who learned English as a second language after childhood. Twenty-two native English and twenty-two native Japanese speakers participated in this study. While undergoing event-related fMRI, subjects were aurally presented with syllables starting with a /r/, /l/, or a vowel and were required to rapidly identify the phoneme perceived by pushing one of three buttons with the left thumb. Consistent with the proposed model, the results show greater activity for second- over native-language speakers during perceptual identification of /r/ and /l/ relative to vowels in brain regions implicated with instantiating forward and inverse articulatory-auditory articulatory-orosensory models [Broca's area, anterior insula, anterior superior temporal sulcus/gyrus (STS/G), planum temporale (PT), superior temporal parietal area (Stp), SMG, and cerebellum]. The results further show that activity in brain regions implicated with instantiating these internal models is correlated with better /r/ and /l/ identification performance for second-language speakers. Greater activity found for native-language speakers especially in the anterior STG/S for /r/ and /l/ perceptual identification is consistent with the hypothesis that native-language speakers use auditory phonetic representations more extensively than second-language speakers.

Adult↗
PubMed1997

A 'Neural Sampling Theory (NST)' of learning and memory mechanisms.

The purpose of the Neural Sampling Theory (NST) is to propose a plausible neurobiological explanation for some general properties of learning and memory (LM) phenomena, based on the parallelism and redundancy of the nervous system organization; on the psychological side, the NST is inspired by the Stimulus Sampling and Encoding Variability theories. The sampling process which is its core, is not purely random; it depends on temporal and intensity factors. The NST may be implemented at different levels of the nervous system: synapse, neuron, assembly of neurons. Moreover, it may be incorporated in other formal models and improve their degree of neural realism. For instance it allows to give a more realistic representation of the connection weight in the connectionist models and of the noisy character of the nervous system.

Animals↗
PubMed2003

Differential pattern of cAMP response element-binding protein activation in the rat brain after conditioned aversion as a function of the associative process engaged: taste versus context association.

Ample data indicate that cAMP-response element-binding protein (CREB) is essential for the formation of long-term memory in various species and learning systems. This implies that activated CREB could delineate neuronal circuits that subserve items in memory, while leaving open the possibility that the specifics of CREB activation itself contribute to the specificity of the internal representation encoded by the relevant circuit. We describe here the differential activation of CREB in the rat brain as a function of two related yet distinct forms of aversive conditioning: conditioned taste aversion (CTA) and conditioned context aversion (CCA). We found that CTA induces strong CREB activation in the insular cortex (IC) and the lateral septum (LS), but not in the parietal cortex (PC) and the medial septum (MS). In contrast, CCA results in strong activation in the PC and MS, but not in the IC and LS. These findings are congruent with a model that links differential pattern of activity within the LS and the MS with the acquisition of elemental versus contextual conditioning and, more generally, with the notion that CREB activation delineates learning-dependent circuits as a function of the type of cognitive process engaged.

Animals↗
PubMed2001

Familiarisation with faces selectively enhances sensitivity to changes made to the eyes.

Two experiments were designed to isolate the facial information utilised in the learning of new faces. In experiments 1 and 2, two groups of subjects were each trained on different groups of faces by means of a dynamic video presentation. They were then shown both trained and novel faces in a same-different decision task, where 'different' trials included manipulations of internal and external facial features, and the task was to decide whether two images were identical or had a difference in one or more features. Both experiments showed that hair change was most easily detected in untrained (unfamiliar) faces. When faces had been trained (familiar), detection of eye changes was selectively enhanced and sensitivity to hair changes was maintained. While previous studies have suggested that familiar face representations are weighted towards their internal features, our experimental results show that this is due to selective enhancement of sensitivity to changes made to the eyes alone, with no reduction in the salience of the hair. Moreover, within the limits of the familiarisation used here, there was no enhancement of the representation of the other internal face features examined.

Analysis of Variance↗

Refine your search to explore more results.