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 1,009 records · Page 56Linked to original sources
PubMed2006

Appetitive odor learning does not change olfactory coding in a subpopulation of honeybee antennal lobe neurons.

Odors elicit spatio-temporal patterns of activity in the olfactory bulb of vertebrates and the antennal lobe of insects. There have been several reports of changes in these patterns following olfactory learning. These studies pose a conundrum: how can an animal learn to efficiently respond to a particular odor with an adequate response, if its primary representation already changes during this process? In this study, we offer a possible solution for this problem. We measured odor-evoked calcium responses in a subpopulation of uniglomerular AL output neurons in honeybees. We show that their responses to odors are remarkably resistant to plasticity following a variety of appetitive olfactory learning paradigms. There was no significant difference in the changes of odor-evoked activity between single and multiple trial forward or backward conditioning, differential conditioning, or unrewarded successive odor stimulation. In a behavioral learning experiment we show that these neurons are necessary for conditioned odor responses. We conclude that these uniglomerular projection neurons are necessary for reliable odor coding and are not modified by learning in this paradigm. The role that other projection neurons play in olfactory learning remains to be investigated.

Analysis of Variance↗
PubMed1998

Association learning in the acute confusional state.

The usefulness of cognitive rehabilitative treatment in the acute stages after brain injury seems questionable because patients in severe acute confusional state early after coma clinically seem unable to learn and store new information. Therefore, the capability of patients in acute confusional state to learn and retain associative information was assessed. On two occasions pairs of simple nouns were presented to six patients in severe acute confusional state. Stimuli were presented repeatedly either in written form only or with additional pictorial representations. Immediate and 20 minutes delayed recall was measured. Patients in acute confusional state were able to learn progressively more word pairs across several presentations. They retained some information over an interval of 20 minutes. In addition, they learned and remembered pictorially supported associations better than pure verbal associations. Patients in severe acute confusional state may retain some explicit information and may profit from an imagery mnemonic aid. These results were not expected on the basis of clinical findings alone and they have potential implications for the care of patients in acute confusional state.

Adult↗
PubMed1983

On learning to speak.

Every language, spoken or signed, deploys a large lexicon, made possible by permutation and combination of a small set of linguistic elements. In speech, rapid interleaving of the gestures that form these elements (consonants and vowels) leads to a complex acoustic signal in which the boundaries between elements are lost. However, for the child learning to speak, the initial task is not to recover these elements, but simply to imitate the sound pattern that it hears. Studies of "lipreading" in adults and infants suggest that imitation is mediated by an amodal representation, closely related to the dynamics of articulation, and that a left-hemisphere perceptuo-motor mechanism specialized to make use of this representation develops during the first six months of life. By drawing on this specialized mechanism, the infant learns the recurrent patterns of acoustic structure and articulatory gesture from which linguistic segments must be presumed to emerge.

Dominance, Cerebral↗
PubMed2006

Simulating the 'other-race' effect with autoassociative neural networks: further evidence in favor of the face-space model.

Other-race (OR) faces are less accurately recognized than same-race (SR) faces, but faster classified by race. This phenomenon has often been reported as the 'other-race' effect (ORE). Valentine (1991 Quarterly Journal of Experimental Psychology A: Human Experimental Psychology 43 161-204) proposed a theoretical multidimensional face-space model that explained both of these results, in terms of variations in exemplar density between races. According to this model, SR faces are more widely distributed across the dimensions of the space than OR faces. However, this model does not quantify nor state the dimensions coded within this face space. The aim of the present study was to test the face-space explanation of the ORE with neural network simulations by quantifying its dimensions. We found the predicted density properties of Valentine's framework in the face-projection spaces of the autoassociative memories. This was supported by an interaction for exemplar density between the race of the learned face set and the race of the faces. In addition, the elaborated face representations showed optimal responses for SR but not for OR faces within SR face spaces when explored at the individual level, as gender errors occurred significantly more often in OR than in SR face-space representations. Altogether, our results add further evidence in favor of a statistical exemplar density explanation of the ORE as suggested by Valentine, and question the plausibility of such coding for faces in the framework of recent neuroimaging studies.

Asian People↗
PubMed2000

Collage work as a medium for guided reflection in the clinical supervision relationship.

This paper identifies the value of one form of art as a means of enhancing communication and developing self expression. The medium of magazine picture collage is described with explanation of its application in the clinical supervision setting where it is used as a trigger for reflective questioning to clarify issues, examine perceptions and facilitate learning from experience. The assembly of magazine pictures as a collage allows metaphorical representation of events and influences through which the supervisee can 'see' their situation from different perspectives. This medium is used as a starting point for discussion in the supervision setting where reflective questioning by the supervisor is framed around the collage images in order to clarify relationships of the issues presented and the personal meanings and values of these to the supervisee. This method is offered as a means of supporting existing systems of supervision not a replacement for these and is perceived to be of particular value in situations where facilitation of expression by means other than dialogue alone is felt to be required in order to move the supervision process forward and enhance personal learning from this.

Art↗
PubMed1993

Responses of monkey dopamine neurons to reward and conditioned stimuli during successive steps of learning a delayed response task.

The present investigation had two aims: (1) to study responses of dopamine neurons to stimuli with attentional and motivational significance during several steps of learning a behavioral task, and (2) to study the activity of dopamine neurons during the performance of cognitive tasks known to be impaired after lesions of these neurons. Monkeys that had previously learned a simple reaction time task were trained to perform a spatial delayed response task via two intermediate tasks. During the learning of each new task, a total of 25% of 76 dopamine neurons showed phasic responses to the delivery of primary liquid reward, whereas only 9% of 163 neurons responded to this event once task performance was established. This produced an average population response during but not after learning of each task. Reward responses during learning were significantly more numerous and pronounced in area A10, as compared to areas A8 and A9. Dopamine neurons also showed phasic responses to the two conditioned stimuli. These were the instruction cue, which was the first stimulus in each trial and indicated the target of the upcoming arm movement (58% of 76 neurons during and 44% of 163 neurons after learning), and the trigger stimulus, which was a conditioned incentive stimulus predicting reward and eliciting a saccadic eye movement and an arm reaching movement (38% of neurons during and 40% after learning). None of the dopamine neurons showed sustained activity in the delay between the instruction and trigger stimuli that would resemble the activity of neurons in dopamine terminal areas, such as the striatum and frontal cortex. Thus, dopamine neurons respond phasically to alerting external stimuli with behavioral significance whose detection is crucial for learning and performing delayed response tasks. The lack of sustained activity suggests that dopamine neurons do not encode representational processes, such as working memory, expectation of external stimuli or reward, or preparation of movement. Rather, dopamine neurons are involved with transient changes of impulse activity in basic attentional and motivational processes underlying learning and cognitive behavior.

Animals↗
PubMed1999

Learning through maps: functional significance of topographic organization in primary sensory cortex.

The presence of "maps" in sensory cortex is a hallmark of the mammalian nervous system, but the functional significance of topographic organization has been called into question by physiological studies claiming that patterns of neural behavioral activity transcend topographic boundaries. This paper discusses recent behavioral and physiological studies suggesting that, when animals or human subjects learn perceptual tasks, the neural modifications associated with the learning are distributed according to the spatial arrangement of the primary sensory cortical map. Topographical cortical representations of sensory events, therefore, appear to constitute a true structural framework for information processing and plasticity.

Animals↗
PubMed2006

Early-phase of learning enhances communication between brain hemispheres.

In the somatosensory system, inputs from one side of the body are only transmitted to the contralateral primary somatosensory cortex, but both sides of the body representation can interact via interhemispheric connections. These interactions depend on the behavioural requirements of the animal and its level of arousal. During the process of learning, alertness and attention may modify the responsiveness of neuronal pathways. We functionally mapped the brains of mice by using [14C]2-deoxyglucose (2DG) autoradiography during the first and the third session of a classical conditioning paradigm, involving whiskers stimulation on one side of the muzzle paired with an aversive or appetitive unconditioned stimulus. During the first pairing session, an increased 2DG uptake was seen in the barrel cortex of both hemispheres, independently of the type of applied unconditioned stimulus. In the third session of the sensory pairing, activation of the barrel cortex was solely contralateral, as expected after unilateral whisker stimulation. Thus, sensory stimulation directed to one cerebral hemisphere during the initial stages of Pavlovian conditioning activates the primary sensory area in both hemispheres. These results suggest that during the early phase of conditioning, when alertness is presumably strongest, the interhemispheric interactions are enhanced.

Animals↗
PubMed1997

Spatial learning by rats across visually disconnected environments.

Two spatial tasks were designed to test specific properties of spatial representation in rats. In the first task, rats were trained to locate an escape hole at a fixed position in a visually homogeneous arena. This arena was connected with a periphery where a full view of the room environment existed. Therefore, rats were dependent on their memory trace of the previous position in the periphery to discriminate a position within the central region. Under these experimental conditions, the test animals showed a significant discrimination of the training position without a specific local view. In the second task, rats were trained in a radial maze consisting of tunnels that were transparent at their distal ends only. Because the central part of the maze was non-transparent, rats had to plan and execute appropriate trajectories without specific visual feedback from the environment. This situation was intended to encourage the reliance on prospective memory of the non-visited arms in selecting the following move. Our results show that acquisition performance was only slightly decreased compared to that shown in a completely transparent maze and considerably higher than in a translucent maze or in darkness. These two series of experiments indicate (1) that rats can learn about the relative position of different places with no common visual panorama, and (2) that they are able to plan and execute a sequence of visits to several places without direct visual feed-back about their relative position.

Animals↗
PubMed2000

Structural priming as implicit learning: a comparison of models of sentence production.

Structural priming reflects a tendency to generalize recently spoken or heard syntactic structures to different utterances. We propose that it is a form of implicit learning. To explore this hypothesis, we developed and tested a connectionist model of language production that incorporated mechanisms previously used to simulate implicit learning. In the model, the mechanism that learned to produce structured sequences of phrases from messages also exhibited structural priming. The ability of the model to account for structural priming depended on representational assumptions about the nature of messages and the relationship between comprehension and production. Modeling experiments showed that comprehension-based representations were important for the model's generalizations in production and that nonatomic message representations allowed a better fit to existing data on structural priming than traditional thematic-role representations.

Cognition↗
PubMed2005

Bimanual adaptation: internal representations of bimanual rhythmic movements.

From tying your shoes and clipping your tie to the claps at the end of a fine seminar, bimanual coordination plays a major role in our daily activities. An important phenomenon in bimanual coordination is the predisposition toward mirror symmetry in the performance of bimanual rhythmic movements. Although learning and adaptation in bimanual coordination are phenomena that have been observed, they have not been studied in the context of adaptive control and internal representations-approaches that were successfully employed in the arena of reaching movements and adaptation to force perturbations. In this paper we examine the dynamics of the learning mechanisms involved when subjects are trained to perform a bimanual non-harmonic polyrhythm in a bimanual index finger tapping task. Subjects are trained in this task implicitly, using altered visual feedback, while their performance is continuously monitored throughout the experiment. Our experimental results indicate the existence of significant (p<<0.01) learning curves (i.e., error plots with significantly negative slopes) during training and aftereffects with a washout period after the visual feedback ceases to be altered. These results confirm the formation of internal representations in bimanual motor control. We present a simple, physiologically plausible, neural model that combines feedback and adaptation in the control process and which is able to reproduce key phenomena of bimanual coordination and adaptation.

Adaptation, Physiological↗
PubMed1986

Comment on "Memory storage and retrieval processes in category learning".

This article is based on the Estes (1986b) article that examined learning processes associated with categorization in relation to new-old recognition. The focus of the commentary is on alternative views of recognition/classification relations as well as the implications of the more detailed analyses of learning for exemplar-based classification models. The argument is made that strategies typically used by experimental participants and exemplar processing have some fundamental properties in common. This implies that a good fit to classification data by an exemplar model does not necessarily mean that performance is based on comparisons with remembered exemplars but suggests that abstract representations may not be different kinds of entities from the memory representation of a specific experience.

Classification↗
PubMed2005

It was the best of times, it was the worst of times: doctoral students' experiences of family therapy research training through alternative forms of data representation.

In this study, we interviewed 14 doctoral students from 10 COAMFTE-accredited doctoral programs to learn more about how they experienced their research training and what they might suggest to strengthen the research culture in their training programs. We solicited somewhat unconventional data--metaphors, poetry, free associations, critical experiences--to (a) tap into our participants' underlying thought processes, (b) capture the multifaceted nature of their doctoral research training, and (c) represent the richness of our participants' subjective experiences. The themes we identified reflect both positive and negative research training experiences and suggest several ways that family therapy program faculty might improve their programs' research training and culture.

Adult↗
PubMed1998

[O-15-butanol Pet activation study on the cerebral representation of declarative memory].

AIM: In this study, neuroanatomical correlates of encoding and retrieval in paired associate learning were evaluated with positron emission tomography using auditorily presented highly imaginable words. METHODS: Six right-handed normal male volunteers took part in the study. Each subject underwent six O-15-butanol PET scans. On each of the six trials the memory task began with the injection of a bolus of O-15-butanol. The subjects had to learn and retrieve twelve word pairs (highly imaginable words, not semantically related). The presentation of nonsense words served as reference condition. RESULTS: Recall accuracy after 2-4 presentations was high during the PET measurement. In both encoding and retrieval we found anterior cingulate activation. We show bilateral dorsolateral prefrontal activation during the encoding of auditorily presented word pair associates, whereas retrieval led to left frontal activation. Furthermore, we demonstrate the importance of the precuneus in the retrieval of highly imaginable word-pair associates. CONCLUSION: Our results support the hypothesis of the presence of distributed widespread brain structures subserving episodic declarative memory.

Adult↗
PubMed1992

Difference and ratio information in children's and adults' memory representations.

A study was made of the choice reaction times of 8-, 10-, 12-year-olds and adults in order to investigate whether length information stored in memory was coded to preserve numerical relations between stimuli. Subjects learned unique line length-color associations and were later presented with pairs of colors and were instructed to choose the color that had been associated with the longer line. For all age groups, choice reaction times were related to both differences and ratios. These results were interpreted as supporting the analog theory of representation.

Adolescent↗
PubMed1999

A blueprint for movement: functional and anatomical representations in the human motor system.

Despite a clear somatotopic organization of the motor cortex, a movement can be learned with one extremity and performed with another. This suggests that there exists a limb-independent coding for movements. To dissociate brain regions coding for movement parameters from those relevant to the chosen effector, subjects wrote their signature with their dominant index finger and ipsilateral big toe, and we determined those areas activated by both conditions using functional magnetic resonance imaging. The results show that movement parameters for this highly trained movement are stored in secondary sensorimotor cortices of the extremity with which it is usually performed, i.e., the dominant hand, including dorsal and ventral lateral premotor cortices. These areas can be accessed by the foot and are therefore functionally independent from the primary representation of the effector. Thus, somatotopy in secondary structures in the human motor system seems to be defined functionally, and not on the basis of anatomical representations.

Adult↗
PubMed2003

A challenge to chaotic itinerancy from brain dynamics.

Brain hermeneutics and chaotic itinerancy proposed by Tsuda are attractive characterizations of perceptual dynamics in the mammalian olfactory system. This theory proposes that perception occurs at the interface between itinerant neural representation and interaction with the environment. Quantifiable application of these dynamics has been hampered by the lack of definable history and action processes which characterize the changes induced by behavioral state, attention, and learning. Local field potentials measured from several brain areas were used to characterize dynamic activity patterns for their use as representations of history and action processes. The signals were recorded from olfactory areas (olfactory bulb, OB, and pyriform cortex) and hippocampal areas (entorhinal cortex and dentate gyrus, DG) in the brains of rats. During odor-guided behavior the system shows dynamics at three temporal scales. Short time-scale changes are system-wide and can occur in the space of a single sniff. They are predictable, associated with learned shifts in behavioral state and occur periodically on the scale of the intertrial interval. These changes occupy the theta (2-12 Hz), beta (15-30 Hz), and gamma (40-100 Hz) frequency bands within and between all areas. Medium time-scale changes occur relatively unpredictably, manifesting in these data as alterations in connection strength between the OB and DG. These changes are strongly correlated with performance in associated trial blocks (5-10 min) and may be due to fluctuations in attention, mood, or amount of reward received. Long time-scale changes are likely related to learning or decline due to aging or disease. These may be modeled as slow monotonic processes that occur within or across days or even weeks or years. The folding of different time scales is proposed as a mechanism for chaotic itinerancy, represented by dynamic processes instead of static connection strengths. Thus, the individual maintains continuity of experience within the stability of fast periodic and slow monotonic processes, while medium scale events alter experience and performance dramatically but temporarily. These processes together with as yet to be determined action effects from motor system feedback are proposed as an instantiation of brain hermeneutics and chaotic itinerancy.

Animals↗
PubMed2025

Boolean matrix logic programming for active learning of gene functions in genome-scale metabolic network models.

Reasoning about hypotheses and updating knowledge through empirical observations are central to scientific discovery. In this work, we applied logic-based machine learning methods to drive biological discovery by guiding experimentation. Genome-scale metabolic network models (GEMs) - comprehensive representations of metabolic genes and reactions - are widely used to evaluate genetic engineering of biological systems. However, GEMs often fail to accurately predict the behaviour of genetically engineered cells, primarily due to incomplete annotations of gene interactions. The task of learning the intricate genetic interactions within GEMs presents computational and empirical challenges. To efficiently predict using GEM, we describe a novel approach called Boolean Matrix Logic Programming (BMLP) by leveraging Boolean matrices to evaluate large logic programs. We developed a new system, [Formula: see text], which guides cost-effective experimentation and uses interpretable logic programs to encode a state-of-the-art GEM of a model bacterial organism. Notably, [Formula: see text] successfully learned the interaction between a gene pair with fewer training examples than random experimentation, overcoming the increase in experimental design space. [Formula: see text] enables rapid optimisation of metabolic models to reliably engineer biological systems for producing useful compounds. It offers a realistic approach to creating a self-driving lab for biological discovery, which would then facilitate microbial engineering for practical applications.

Active learning↗