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 1,783 records · Page 99Linked to original sources
PubMed1991

Limited retention of information in the graphemic buffer.

This study examined a patient who mainly made letter deletion errors in spelling. It was hypothesized that his errors were due primarily to limited ability to retain information in the graphemic buffer, a structure that holds sequences of abstract letter identities for output. Consistent with this hypothesis, the patient's spelling accuracy declined on long words, but the number of letters he wrote per response was not related to word length. Moreover, by having him write words forward and backward, it was shown that his accuracy within a word depended on which part of the word he tried to output first. These results also ruled out alternative accounts of the patient's spelling deficit based on neglect or damage to lexical representations.

Anomia↗
PubMed1994

Role of chaotic dynamics in neural plasticity.

Mathematical models are essential for the study of complex neural systems at all levels of the hierarchy from macromolecules through neurons to brain systems. ANN are readily available, but most of them are inappropriate for modeling brain function in normal behavior, because they stem from studies of neural systems in anesthetized or paralyzed animals which are capable only of reflux output. That class of models lacks the goal-directed, self-organizing properties of neural systems in behaving animals. In contrast to the stability of ANN and their reliance on asymptotic convergence to steady states (point attractors) and periodic oscillations (limit cycle attractors), BNN are intrinsically unstable. They continuously generate 'spontaneous' aperiodic activity that manifests the operation of chaotic dynamics undergoing repeated state transitions. Observations on the activity patterns of sensory cortex reveal that the perceptual outputs of BNN are by construction of spatial patterns and dynamic trajectories and not by computation using symbolic representations. Chaotic dynamics plays essential roles both in the construction of perceptions and in the continuing update of cortical populations, which requires selective synaptic modification during associative learning and habituation. Simultaneous multichannel recording from the olfactory bulb and cortex has given the following experimental results. (1) The cortical activity that relates to the perception of a sensory stimulus is carried macroscopically by a smaller number of single neurons e.g. 'units', 'feature detectors'. (2) The macroscopic activity reflects the meaning and significance of the stimulus for the experimental subject and not the stimulus as it is known to the observer. (3) The activity carries the meaning in spatial patterns, not in time series (the difference between a phonograph or radio and movie or TV). (4) The spatial patterns of activity that accompany previously learned stimuli or responses are changed by the introduction of new stimuli and also by modifications in reinforcement contingencies. There is no invariance in the memory store within the populations. (5) The patterns of activity are created by dynamic neural interactions in sensory cortex, not by registration or filtering of stimuli. There is no evidence for storage, retrieval, cross-correlation or logical tree search. (6) The dynamics is chaotic, not merely noisy, so that each act of perception involves a new construction by the cortex not by mere information processing. From these findings we infer that chaotic dynamics plays a crucial role in the construction of the associational contexts comprising the memory systems of experimental subjects.

Animals↗
PubMed2002

Decision trees: an overview and their use in medicine.

In medical decision making (classification, diagnosing, etc.) there are many situations where decision must be made effectively and reliably. Conceptual simple decision making models with the possibility of automatic learning are the most appropriate for performing such tasks. Decision trees are a reliable and effective decision making technique that provide high classification accuracy with a simple representation of gathered knowledge and they have been used in different areas of medical decision making. In the paper we present the basic characteristics of decision trees and the successful alternatives to the traditional induction approach with the emphasis on existing and possible future applications in medicine.

Clinical Medicine↗
PubMed2000

The extralist-feature effect: evidence against item matching in short-term recognition memory.

The authors varied the similarity between negative probes and study items in a short-term item-recognition task. Current models treat similarity as a function of the number of occurrences of the probe's features in the study set, a factor that is often confounded with the number of the probe's features occurring in the study set. Unconfounded comparisons showed that performance reflected only the latter factor, with response time a linear function of the number of probe features in the study set. The effect was obtained for both stimuli with manipulated features (colored shapes) and words. Number of presented features is a global property of the study list, but existing global models calculate familiarity by averaging across item matches and cannot readily accommodate the data. The authors proposed that the probe's features are compared with a global representation of the study set's features.

Adult↗
PubMed1991

Differential effectiveness of various prior-cuing treatments in the reactivation and maintenance of memory.

In 3 experiments, changes were examined in the characteristics of newly acquired and reinstated memories over time in preweanling rats. Experiment 1 indicated that forgetting after conditioning was monotonic, with the upper limit of retention at approximately 120 min posttraining. In Experiment 2, Ss were exposed to various elements of the training episode before testing, after either a 3- or a 24-hr retention interval. The results indicated that the prior-cuing treatments were differentially effective and that the effectiveness of a reactivation treatment may change as a function of the retention interval. Experiment 3 indicated that Ss expressed a conditioned aversion at much longer intervals following reactivation treatments than after initial conditioning. Furthermore the susceptibility of the reinstated memory to forgetting was dependent on the prior-cuing treatment used. The results suggest a change in the memorial representation of the conditioning episode over time.

Animals↗
PubMed1992

Priming and recognition of transformed three-dimensional objects: effects of size and reflection.

In 2 experiments exploring memory for unfamiliar 3-dimensional objects, Ss studied drawings under conditions that encouraged encoding of global object structure. Implicit memory for objects was assessed by a judgment of structural possibility; explicit memory was assessed by recognition. The principal manipulation was the relationship between the sizes or the left-right parities of the studied and tested objects. Priming was observed on the possible-impossible object decision task despite transformations of size or reflection. Recognition, by contrast, was significantly impaired by the transformations. These results suggest that a structural description system constructs representations of objects invariant over size and reflection, whereas a separable episodic system encodes these transformations as properties of an object's distinctive representation in memory.

Adult↗
PubMed1992

Dynamic binding in a neural network for shape recognition.

Given a single view of an object, humans can readily recognize that object from other views that preserve the parts in the original view. Empirical evidence suggests that this capacity reflects the activation of a viewpoint-invariant structural description specifying the object's parts and the relations among them. This article presents a neural network that generates such a description. Structural description is made possible through a solution to the dynamic binding problem: Temporary conjunctions of attributes (parts and relations) are represented by synchronized oscillatory activity among independent units representing those attributes. Specifically, the model uses synchrony (a) to parse images into their constituent parts, (b) to bind together the attributes of a part, and (c) to bind the relations to the parts to which they apply. Because it conjoins independent units temporarily, dynamic binding allows tremendous economy of representation and permits the representation to reflect the attribute structure of the shapes represented.

Depth Perception↗
PubMed1991

Conflicting figure--ground and depth information reduces moving phantom visibility.

Moving phantom visibility was measured in two experiments where the global figure-ground and depth relations within phantom-inducing patterns were manipulated. The local inducing environment where the illusion occurred was identical for all patterns. Phantom visibility was significantly reduced when occlusion cues specified the phantom-inducing parts of a pattern as ground. These results suggest conflicting figure-ground and depth information interferes with the representation and perception of phantoms as figure regions.

Adult↗
PubMed1990

Representation of letters when mislocation errors occur.

Errors in reporting a cued target letter appearing among a string of letters more often reflect the mislocation of a letter appearing elsewhere in the string than the intrusion of one not in the string. The current experiment was conducted to determine the representation of letters at the stage at which errors occur. Four letters (from a set of 12 chosen to contain counterpart pairs that were similar physically, phonemically, or both physically and phonemically) appeared in each exposure, with the target letter indicated by a cue in the postexposure mask. Letter strings presented to one group of subjects were flanked on either side by a pound sign (#) to assess the effect of lateral masking on the terminal letter in the string. Physical similarity dictated the pattern of mislocations between counterparts, suggesting a physical rather than phonemic or abstract representation. Lateral masking played no significant role in the difference between intrusions and mislocations.

Attention↗
PubMed2002

Alcohol seeking by rats: action or habit?

In two experiments, we examined the relative susceptibility to outcome devaluation of lever pressing by rats for either a 10% ethanol solution or food pellets. The rats were trained to press different levers for these two reinforcers using a sucrose-substitution procedure. An aversion was then conditioned from either the ethanol solution or the food pellets by pairing consumption with illness induced by lithium chloride. When instrumental performance was subsequently tested in extinction, the rats pressed less on the pellet lever if the pellets, rather than the ethanol, had been devalued by aversion conditioning. By contrast, performance on the ethanol lever was unaffected by whether the ethanol or pellets were devalued. Moreover, noncontingent presentations of the devalued reinforcer had no impact on test performance. The differential resistance to outcome devaluation suggests that, in contrast to food seeking, alcohol seeking is a stimulus-response habit rather than a goal-directed action mediated by a representation of the action-outcome contingency.

Alcohol Drinking↗
PubMedDate not supplied

Fixational instability and natural image statistics: implications for early visual representations.

Under natural viewing conditions, small movements of the eye, head and body prevent the maintenance of a steady direction of gaze. It is known that stimuli tend to fade when they are stabilized on the retina for several seconds. However, it is unclear whether the physiological motion of the retinal image serves a visual purpose during the brief periods of natural visual fixation. This study examines the impact of fixational instability on the statistics of the visual input to the retina and on the structure of neural activity in the early visual system. We show that fixational instability introduces a component in the retinal input signals that, in the presence of natural images, lacks spatial correlations. This component strongly influences neural activity in a model of the LGN. It decorrelates cell responses even if the contrast sensitivity functions of simulated cells are not perfectly tuned to counter-balance the power-law spectrum of natural images. A decorrelation of neural activity at the early stages of the visual system has been proposed to be beneficial for discarding statistical redundancies in the input signals. The results of this study suggest that fixational instability might contribute to the establishment of efficient representations of natural stimuli.

Animals↗
PubMed2003

Interference and integration: the fan effect in children and adults.

In three experiments, we explored (1) possible developmental changes in interference and (2) the use of integration processes to eliminate them. Two versions of the fan effect procedure were used. Results of Experiment 1 with unrelated sentences showed similar fan effects for 12-13-year-old children and adults. In Experiment 2, the sentences were thematically related so that it was possible to use elaboration processes to construct integrated representations. Under these conditions only adult participants eliminated the fan effect. In Experiment 3 a new procedure was used where children and adults studied highly familiar material in an organised manner. The results showed that in these conditions children integrated information and reduced interference. Thus, prior knowledge and organisation seem to be important factors to create meaningful integrated representations in memory.

Adolescent↗
PubMed1991

Distinguishing representation deficits and processing deficits in a case of acquired dysgraphia.

The same acquired disorder of spelling may be due to deficits affecting lexical representations of word spelling or deficits affecting the mechanisms that process those representations. This study sought to distinguish these possibilities in a dysgraphic patient. The integrity of the patient's lexical orthographic representations was assessed by having him decide whether or not pairs of words presented auditorily rhymed. Although the patient was impaired on a variety of spelling tasks and with all types of stimulus material, he showed a normal effect of spelling on the rhyme task. Like normal subjects, he was faster at deciding that words rhymed when they were spelled similarly (e.g. tool-cool) than when they were spelled dissimilarly (e.g. rule-cool) and slower at deciding that words did not rhyme when they were spelled similarly (e.g. toad-broad) than when they were spelled dissimilarly (e.g. code-broad). Therefore, as the patient's lexical representations of word spelling seemed to be generally intact, his spelling problems were probably due to difficulty in processing those representations.

Agraphia↗
PubMedDate not supplied

Memory for items and memory for relations in the procedural/declarative memory framework.

A major area of research in memory and amnesia concerns the item specificity of implicit memory. In this paper we address several issues about the nature of implicit memory phenomena and about what constitutes an "item", using the procedural/declarative memory theory to guide us. We consider the nature of memory for items and of memory for relations among items, within the context of the procedural/declarative framework, providing us with the foundation necessary to analyse the basis for item-specific implicit memory phenomena. We review recent work from our laboratories demonstrating the fundamentally relational and flexible nature of declarative memory representation, in both humans and animals, and the essential role of the hippocampal system in relational memory processing. We show, further, that the memory representations supporting implicit memory phenomena are inflexible and nonrelational, and are tied to specific processing modules. Finally, we introduce empirical approaches that blur the distinction between skill learning and repetition priming, and show computational modelling results that demonstrate how these two implicit memory phenomena can be mediated by a single incremental learning mechanism, in accord with the claims of the procedural-declarative theory. Taken together, these various analyses of memory for items and memory for relations help to illuminate the nature of the functional deficit in amnesia and the memory systems of the brain.

Amnesia↗
PubMed2003

Global transsaccadic change blindness during scene perception.

Each time the eyes are spatially reoriented via a saccadic eye movement, the image falling on the retina changes. How visually specific are the representations that are functional across saccades during active scene perception? This question was investigated with a saccade-contingent display-change paradigm in which pictures of complex real-world scenes were globally changed in real time during eye movements. The global changes were effected by presenting each scene as an alternating set of scene strips and occluding gray bars, and by reversing the strips and bars during specific saccades. The results from two experiments demonstrated a global transsaccadic change-blindness effect, suggesting that point-by-point visual representations are not functional across saccades during complex scene perception.

Adult↗
PubMed2005

Auditory-visual intermodal matching of small numerosities in 6-month-old infants.

Recent studies have reported that preverbal infants are able to discriminate between numerosities of sets presented within a particular modality. There is still debate, however, over whether they are able to perform intermodal numerosity matching, i.e. to relate numerosities of sets presented with different sensory modalities. The present study investigated auditory-visual intermodal matching of small numerosities in infancy by using a violation-of-expectation paradigm. After being familiarized with events of a few objects impacting a surface successively, 6-month-old infants were alternatively presented with two and three tones while the movement of each object remained hidden behind an opaque screen. The screen was then removed to reveal either two or three objects. Results showed that the infants looked significantly longer at the numerically nonequivalent events (the three-tone/two-object and the two-tone/three-object events) than at the numerically equivalent events (the two-tone/two-object and the three-tone/three-object events) irrespective of the rate or duration of auditory tones presented. These findings suggest that infants are capable of performing intermodal matching of small numerosities and that they might possess abstract representations of numerosity beyond sensory modalities.

Auditory Perception↗
PubMed2001

Sensory input directs spatial and temporal plasticity in primary auditory cortex.

The cortical representation of the sensory environment is continuously modified by experience. Changes in spatial (receptive field) and temporal response properties of cortical neurons underlie many forms of natural learning. The scale and direction of these changes appear to be determined by specific features of the behavioral tasks that evoke cortical plasticity. The neural mechanisms responsible for this differential plasticity remain unclear partly because important sensory and cognitive parameters differ among these tasks. In this report, we demonstrate that differential sensory experience directs differential plasticity using a single paradigm that eliminates the task-specific variables that have confounded direct comparison of previous studies. Electrical activation of the basal forebrain (BF) was used to gate cortical plasticity mechanisms. The auditory stimulus paired with BF stimulation was systematically varied to determine how several basic features of the sensory input direct plasticity in primary auditory cortex (A1) of adult rats. The distributed cortical response was reconstructed from a dense sampling of A1 neurons after 4 wk of BF-sound pairing. We have previously used this method to show that when a tone is paired with BF activation, the region of the cortical map responding to that tone frequency is specifically expanded. In this report, we demonstrate that receptive-field size is determined by features of the stimulus paired with BF activation. Specifically, receptive fields were narrowed or broadened as a systematic function of both carrier-frequency variability and the temporal modulation rate of paired acoustic stimuli. For example, the mean bandwidth of A1 neurons was increased (+60%) after pairing BF stimulation with a rapid train of tones and decreased (-25%) after pairing unmodulated tones of different frequencies. These effects are consistent with previous reports of receptive-field plasticity evoked by natural learning. The maximum cortical following rate and minimum response latency were also modified as a function of stimulus modulation rate and carrier-frequency variability. The cortical response to a rapid train of tones was nearly doubled if BF stimulation was paired with rapid trains of random carrier frequency, while no following rate plasticity was observed if a single carrier frequency was used. Finally, we observed significant increases in response strength and total area of functionally defined A1 following BF activation paired with certain classes of stimuli and not others. These results indicate that the degree and direction of cortical plasticity of temporal and receptive-field selectivity are specified by the structure and schedule of inputs that co-occur with basal forebrain activation and suggest that the rules of cortical plasticity do not operate on each elemental stimulus feature independently of others.

Acoustic Stimulation↗
PubMed2003

Competition and selection during visual processing of natural scenes and objects.

When a visual scene, containing many discrete objects, is presented to our retinae, only a subset of these objects will be explicitly represented in visual awareness. The number of objects accessing short-term visual memory might be even smaller. Finally, it is not known to what extent "ignored" objects (those that do not enter visual awareness) will be processed--or recognized. By combining free recall, forced-choice recognition and visual priming paradigms for the same natural visual scenes and subjects, we were able to estimate these numbers, and provide insights as to the fate of objects that are not explicitly recognized in a single fixation. When presented for 250 ms with a scene containing 10 distinct objects, human observers can remember up to 4 objects with full confidence, and between 2 and 3 more when forced to guess. Importantly, the objects that the subjects consistently failed to report elicited a significant negative priming effect when presented in a subsequent task, suggesting that their identity was represented in high-level cortical areas of the visual system, before the corresponding neural activity was suppressed during attentional selection. These results shed light on neural mechanisms of attentional competition, and representational capacity at different levels of the human visual system.

Attention↗