SEARCH · PubMed Health

Results for “Contrastive learning”

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 721 records · Page 40Linked to original sources

PubMed1995

Attention and memory trials during neuronal recording from the primate pulvinar and posterior parietal cortex (area PG).

Cynomolgus monkeys were trained on a delayed match-to-sample (DMS) task. Several controls were performed for the separation of memory effects from other factors such as attention, stimulus attributes and eye movements. One of the controls employed a standardised sequence of events ('window trial') during which a window in a black screen was opened and a face, an object or a picture was presented. Unit cell activity was recorded from 209 sites in the posterior parietal cortex (area PG) and 186 sites in the pulvinar. Some units responded during the presentation of the DMS stimuli, mostly those in that part of area PG located in the superior temporal sulcus. They often appeared to be related to the task-related state of attention. There was no indication of pure memory-related changes in activity such as sustained responses during the inter-stimulus interval or specific responses to the second stimulus that would indicate a dependence on the kind of the preceding stimulus. In the window trial, some parietal units, mostly in that part of area PG located in the intraparietal sulcus (a region termed area LIP), responded in a relatively specific manner during and sometimes even after, the presentation of a human face. The responses often seemed to be related to the animal's state of attention. Almost no pulvinar unit responded in this paradigm. It is concluded that area PG and, to a lesser extent, the pulvinar are involved in the modulation of attention in relation to behaviourally relevant changes in the environment. If these structures have a role in memory function, it must be secondary to a role in controlling or regulating attention.

Animals↗

PubMed1994

Further investigation of preoperative overtraining, visual cortex lesions and black-white discrimination by the rat.

Preoperative overtraining can improve retention after brain lesions. However, studies of effects of overtraining on relearning of a black-white discrimination after visual cortex lesions in the rat have obtained conflicting results. Another experiment was done on this problem. The postoperative performance of overtrained rats was found to be slightly better than that of non-overtrained rats, but the difference was not significant. Nevertheless, when saving scores were computed, differences were significant. These saving score results seem to demonstrate a preoperative overtraining effect. It is suggested that overtraining before surgery produces a shift of function in the brain, and this shift is related to the development of long-term memory.

Animals↗

PubMed2005

Forward and backward visual masking in unaffected siblings of schizophrenic patients.

BACKGROUND: Visual masking tasks assess the earliest stages of visual processing. This study examined visual masking performance for forward and backward masking tasks in siblings of schizophrenic patients and healthy comparison subjects. METHODS: A staircase method was used to ensure that unmasked target identification was equivalent across subjects to eliminate differences due to discrimination of simple perceptual inputs. Four computerized visual masking tasks were administered to 43 siblings of patients and 42 normal comparison subjects. The tasks included: 1) locating a target; 2) identifying a target with a high-energy mask; 3) identifying a target with a low-energy mask; and 4) a paracontrast/metacontrast procedure with nonoverlapping target and mask. RESULTS: Across masking conditions, there was a significant group by forward/backward interaction, meaning that siblings showed a larger difference from control subjects in backward versus forward masking. This group difference was more pronounced in the location condition. CONCLUSIONS: These results support the theory that visual masking procedures may be indicators of vulnerability to schizophrenia. The pattern of findings in this report (larger group differences on backward versus forward masking and on the location condition) suggests that the activity of transient visual channels may be particularly linked to vulnerability.

Adult↗

PubMed2004

Investigation of color constancy with a neural network.

The perceptual stability of an object's color under different illuminants is called color constancy. We created a neural network to investigate this phenomenon. The net consisted of one input channel for the background and one for the test object. Each channel had a set of three (L, M, and S) receptors that were transmitting to three opponent neurons. The signals from the opponent neurons were passed to hidden neurons, which were connected to the output neurons. The output signal was generated from the three components of a color vector. The neural net was trained to identify the color of Munsell samples under various illuminants using the back-propagation algorithm. Our study investigated the properties of a successfully trained neural network. Based on the cross-neuron weight analysis, we report that the successfully trained neural net calculates color differences between the test object and the background. By comparing the human visual system to the neural net, we conclude that to satisfy the color constancy phenomenon, the human visual system has to contain two separate components: one to approximate the background color and the other to estimate the color difference between the object and the background.

Color↗

PubMed2006

The fusiform face area is tuned for curvilinear patterns with more high-contrasted elements in the upper part.

The ability to identify conspecifics from the face is of primary interest for human social behavior. Newborns' visual preference for schematic face-like stimuli has been recently related to a sensitivity for visual patterns with a greater number of elements in the upper compared to the lower part. At the adult level, neuroimaging studies have identified a network of cortical areas devoted to the detection and identification of faces. However, whether and how low-level structural properties of face stimuli contribute to the preferential response to faces in these areas remain to be clarified. Using functional magnetic resonance imaging (fMRI), here we investigated whether the adults' face-sensitive cortical areas show a preference for top-heavy patterns, similarly to newborns' preference. Twelve participants were presented with head-shaped and square patterns with either more elements in the upper or the lower vertical part. In the right fusiform gyrus ('fusiform face area', FFA), an area showing a preference for faces over other visual object categories, there was a larger activation for curvilinear patterns with more high-contrast elements in the upper part, even though these patterns were not perceived as face stimuli. These findings provide direct evidence that the FFA is tuned for geometrical properties fitting best with the structure of faces, a computational mechanism that might drive the automatic detection of faces in the visual world.

Adult↗

PubMed2004

Chromatic edges, surfaces and constancies in cerebral achromatopsia.

We tested achromatopsic observer, MS, on a number of tasks to establish the extent to which he can process chromatic contour. Stimuli, specified in terms of cone-contrast, were presented in a three-choice oddity paradigm. First we show that MS is able to discriminate the magnitude of chromatic and luminance contrast, but performance is inferior to that of normal observers. Moreover, MS can discriminate isoluminant borders of different chromatic composition. These abilities are not the result of unintended luminance differences and are abolished when chromatic borders are masked by sharp luminance change. In simple displays, local cone-contrast signals can make a significant contribution to surface colour appearance in normal observers. In more complex displays, the perception of a surface's colour becomes largely independent of the local contrast to its background, via processes presumed to be similar to the edge integration and anchoring stages of Land's Retinex algorithm. We show that in simple displays the percepts of both MS and normal observers are dominated by local chromatic-contrast. But, although the percepts of normal observers change in line with the predictions of retinex theory in more complex displays, those of MS do not, remaining dominated by local contrast signals. We conclude that MS has lost the ability to perform edge integration and that this loss is closely related to his absence of colour experience.

Adult↗

PubMed2006

Aging and visual motion discrimination in normal adults and schizophrenia patients.

Motion perception is impaired in many neuropathological conditions, including schizophrenia. Motion perception also declines in the course of normal aging. In this study, we ask whether aging is an additive factor in the motion-discrimination deficits of schizophrenia patients. We examined motion perception in schizophrenia patients (n=44) and non-psychiatric controls (n=40) whose ages ranged from 18 to 55. The tasks included velocity discrimination and contrast detection. Thresholds for each of the two tasks were determined for each subject using psychophysical methods. Schizophrenia patients showed significantly increased thresholds (degraded performance) for velocity discrimination compared with the controls. Degraded performance in patients was not related to age. In controls, however, velocity discrimination thresholds were significantly increased beginning by age 45. Performance on a contrast-detection task, which does not require precise discrimination of motion signals, was not significantly affected by age in either group. Aging, even in its early stages, degrades motion discrimination in normal adults. Aging, however, does not adversely affect motion-discrimination deficits in schizophrenia patients through age 55. A similar motion-discrimination deficit in schizophrenia patients and aging normal adults suggests that the mechanisms underlying motion processing in schizophrenia and normal aging may be associated.

Acceleration↗

PubMed2005

Implicit exploitation of regularities: novel correlations in images quickly alter visual perception.

Viewing a set of two thousand images in which the contrasts of a central and a surrounding pattern were highly correlated increased the suppressive influence of the surround on the perceived contrast of the central pattern. The apparent increase in inhibition supports the operation of an anti-Hebbian mechanism between the two groups of cells excited by the patterns (one group by the central pattern and the other by the surround). According to this mechanism, inhibitory connections between nearby cells increase in efficacy according to a simple function of the correlations between the cells' activities.

Adaptation, Psychological↗

PubMed2006

Synesthetic colors determined by having colored refrigerator magnets in childhood.

Synesthesia is a condition in which percepts in one modality reliably elicit secondary perceptions in the same or a different modality that are not in the stimulus. In a common manifestation, synesthetes see colors in response to spoken or written letters, words and numbers. In this paper we demonstrate that the particular colors seen by a grapheme-color synesthete AED were learned from a set of refrigerator magnets and that the synesthesia later transferred to Cyrillic in a systematic way, with the colors induced by the Cyrillic letters determined by their visual or phonetic similarity to English letters. Closer examination of the data reveals that letters of either language that are more visually similar to the English capitals in the magnet set are also more saturated. In order to differentiate AED's synesthesia from ordinary memory, we use a novel psychophysical method to show that AED's synesthetic colors are subject to ordinary lightness constancy mechanisms. This suggests that the level of representation at which her synesthesia arises is early in the stream of visual processing.

Adult↗

PubMed2002

A brain adaptation view of plasticity: is synaptic plasticity an overly limited concept?

A view that is emerging is that the brain has multiple forms of plasticity that must be governed, at least in part, by independent mechanisms. This view is illustrated by: (1) the apparent separate governance of some non-neural changes by activity, in contrast to synaptic changes driven by learning; (2) the apparent independence of different kinds of synaptic changes that occur in response to the learning aspects of training; (3) the occurrence of separate patterns of synaptic plasticity in the same system in response to different task demands; and (4) apparent dissociations between behaviorally induced synaptogenesis and LTP. The historical focus of research and theory in areas ranging from learning and memory to experiential modulation of brain development has been heavily upon synaptic plasticity since shortly after the discovery of the synapse. Based upon available data, it could be argued that: (1) synaptic, and even neuronal, plasticity is but a small fraction of the range of changes that occur in response to experience; and (2) we are just beginning to understand the importance of these other forms of brain plasticity. Appreciation of this aspect of the brain's adaptive process may allow us to better understand the capacity of the brain to tailor a particular set of changes to the demands of the specific experiences that generated them.

Adaptation, Physiological↗

PubMed2003

Non-elemental processing in olfactory discrimination tasks needs bilateral input in honeybees.

In patterning discriminations, animals have to differentiate a compound stimulus AB from each of its elements A and B. In positive patterning (PP), the compound is reinforced whilst the single elements are non-reinforced. In negative patterning (NP), single elements are reinforced whilst the compound is non-reinforced. Using olfactory conditioning of the proboscis extension response (PER), we asked whether honeybees (Apis mellifera) can solve these patterning problems when odorants are given unilaterally as well as bilaterally to the antennae. Separating the olfactory input space of bees into two independent zones using plastic walls placed between the antennae, we conditioned bees in PP and NP procedures, with input on one side, on both sides, or in an ambiguous problem where bees had to solve PP on one side and NP on the other side. We found that bees with simultaneous bilateral input solve both patterning tasks efficiently. In contrast, PP but not NP was learned by bees receiving unilateral olfactory input. Bees subjected to the ambiguous NP/PP problem only solved PP. As PP can be solved through mere elemental processes, but NP is critically dependent on the use of non-elemental learning processes, our results suggest that bilateral olfactory input is necessary for non-elemental processing to take place in the bee brain.

Animals↗

PubMed2006

Lightness compression and hue changes.

Two experiments were performed to relate the Bezold-Brücke (B-B) and lightness compression effects. The first used a calibrated screen to present an achromatic luminance staircase. In addition, it reproduced, the methodology and the essential aspects the lightness compression effect discovered by Cataliotti and Gilchrist (1995). That is, observers perceived a truncated grey scale (from white to medium grey) when the staircase was the only stimulation in the near background (Gelb condition), but not when presented on a Mondrian background, because of the high articulation level provided by this background. Experiment 1 design also included two other backgrounds that produced a partial compression effect. In Experiment 2, two chromatic staircases were used. Employing a naming task, changes in hue perception were only observed for the susceptible staircase. The observed changes were of two types. First, for the full staircase presentations, a Gelb background produced maximum lightness compression (more similarity in the lightness of the staircase stimuli) and, also, a minimum B-B effect (fewer differences in hue). Second, only for the Gelb condition, there were changes in the hue of the lowest luminance staircase stimuli depending on the staircase extension. Results are discussed in the framework of the anchoring theory of lightness perception.

Adult↗

PubMed1994

Stroop process dissociations: the relationship between facilitation and interference.

L.L. Jacoby's (1991) "process dissociation procedure" was used to quantitatively estimate the contributions of color-naming and word-reading processes to responding on the Stroop task. The results show that color naming and word reading can operate independently to determine responses. Degrading stimulus colors eliminated the typical asymmetry between Stroop facilitation and interference, as predicted by the equations (Experiments 1 and 2). Degrading stimulus colors reduced the estimated contribution of color naming to responding but had no effect on the estimated contribution of word reading (Experiment 2). In contrast, increasing the proportion of incongruent items reduced the estimated contribution of word reading but had no effect on the estimated contribution of color naming (Experiments 3 and 4). The results indicate that the facilitating and interfering effects of automatic processes cannot be accurately measured in terms of differences from baseline.

Adult↗

PubMed1996

Texture segmentation along the horizontal meridian: nonmonotonic changes in performance with eccentricity.

In 3 experiments, subjects were required to detect the presence of a small region of disparate texture embedded in a larger background at a range of eccentricities. Detection performance always peaked several degrees from fixation. Experiment 1 showed that the location of the peak was not retinally specific; scaling the display changed the location of the performance peak. Experiment 2 showed that poor foveal performance could not be explained by cross-frequency interference; filtering out high spatial frequencies did not lead to improved foveal performance. Experiment 3 showed that the effect is not unique to textures comprising left and right oblique line segments. A parsimonious account of these data is that, at the fovea, there is a mismatch between the scale of the texture and the scale of the mechanisms responsible for encoding texture differences. This mismatch diminishes as the textures are moved further into the periphery.

Attention↗

PubMed1998

Multiple-location access in vision: evidence from illusory line motion.

Four experiments with undergraduates used illusory line motion (ILM) to contrast Z. W. Pylyshyn's (1989) FINST theory of spatial indexing with predictions made by unitary attention models. Multiple-onset stimuli were able to cause ILM at disparate, noncontiguous spatial locations. Consistent with gradient explanations of ILM and with FINST theory predictions, varying line-drawing speed and the number of stimuli revealed a decrease in ILM and a capacity limitation, respectively. Modeling analyses suggested a limit in the number of locations (5-7) that could elicit the illusion. Requiring participants to report the locations of all stimuli exhibiting illusory motion in a specified direction suggested parallel access to between 2 and 5 display locations simultaneously. The results of all 4 experiments were predicted by FINST theory but not by a broad class of unitary attention hypotheses.

Acceleration↗

PubMed1999

Perceiving spatially inseparable objects: evidence for feature-based object selection not mediated by location.

In 4 experiments, stimulus elements were arranged into an LED-like array, and letters were defined within the array by feature similarity between the elements with respect to color and form. These stimuli allowed the display of a target and a distractor letter simultaneously at the same location. They were spatially inseparable but could be separated in feature space. Participants had to identify the letter on a prespecified feature dimension (color or form). As a result, the distractors produced specific compatibility effects. This showed that nontarget features could not be ignored at an early stage (i.e., that color and form were processed automatically and in parallel up to a high stage). The target was selected from the resulting objects according to the prespecified feature dimension. Results demonstrate that object selection is possible without selecting absolute spatial arrays.

Adult↗

PubMed1999

Do equiluminant object onsets capture visual attention?

S. Yantis and A. P. Hillstrom (1994) have claimed that abrupt onset of a new visual object captures attention even when the new object is equiluminant with its background, implying that attention is captured at the level of object descriptions rather than at the level of luminance change detection. S. Yantis and A. P. Hillstrom's experiments contained potential flaws that call their conclusion into question. The present article reports 5 experiments investigating these and related issues. The results suggest that for abruptly onsetting visual objects to capture attention, detection of change in a dimension of sensory stimulation is a necessary, but perhaps not a sufficient, requirement. Evidence is also presented against the view that attentional priority for new objects arises as a result of visual masking of old objects in the same displays.

Adult↗

PubMed2000

Contingent capture for onsets and offsets: attentional set for perceptual transients.

Four experiments were conducted to examine whether attentional set affects the ability of visual transients (onsets and offsets) to capture attention. In the experiments, visual search for an identity-defined target was conducted. In the first 3 experiments, the target display either onset entirely or was revealed by offsetting camouflaging line segments to reveal letters. Prior to the target display, there was a noninformative cue, either an onset or an offset, at one of the potential target locations. Cues that shared the same transient feature as the target display captured attention. The lack of predictable target transients led to attentional capture by all forms of transients. The final experiments with luminance changes without offsets or onsets showed attentional capture when the luminance changes were large. The results suggest that attentional set can be broadly or narrowly tuned to detect changes in luminance.

Adolescent↗