Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Transfer 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 145 records · Page 8Linked to original sources
PubMed1994

Transfer of learning but not memory after unilateral cerebellar lesion in rabbits.

Unilateral lesion of the cerebellum in rabbits completely and permanently abolishes previous learning and prevents new learning ipsilateral to the lesion. However, when training continues on the contralateral side, there is substantial savings in that it takes few trials to learn. This observation may imply that the memory survives the lesion. Rabbits were classically conditioned for an eyelid response and then the ipsilateral interpositus nucleus of the cerebellum was lesioned. Then the rabbits were trained on the contralateral side. There is no savings on the contralateral side without first trying to train on the lesioned side. The authors conclude that the survival of a memory after the lesion probably does not account for the rapid transfer, but rather that the act of trying to train on the lesioned side in previous studies first induces a new memory on the contralateral side.

Animals↗
PubMed2002

Transfer of learning across durations and ears in auditory frequency discrimination.

Frequency-discrimination thresholds (FDTs) for 1-kHz tone pips with durations of 40, 100, and 200 msec were measured in the left and right ears of 10 normal-hearing listeners, before and after six 2-h frequency-discrimination training sessions involving, exclusively, the 200-msec duration and the right ear. In the trained ear, highly significant improvements in FDTs were observed at all durations. Further inspection of the data suggested complete generalization between 200 and 100 msec, but not at 40 msec. Posttraining FDTs were not found to differ between the two ears for the two untrained durations, but proved significantly smaller in the right (trained) than in the left (untrained) ear at the trained (200-msec) duration only. A control experiment involving 10 additional subjects allowed us to establish the absence of intrinsic differences in pretraining FDTs between the right and left ears. Overall, these findings indicate that frequency-discrimination learning generalizes widely across stimulus durations and across ears, but that part of the improvement is specific to the range of durations and to the ear used in training.

Adult↗
PubMed2001

Consolidation of extinction learning involves transfer from NMDA-independent to NMDA-dependent memory.

Extinction of conditioned fear to a tone paired with foot shock is thought to involve the formation of new memory. In support of this, previous studies have shown that extinction of conditioned fear depends on NMDA receptor-mediated plasticity. To further investigate the role of NMDA receptors in extinction, we examined the effects of the NMDA antagonist d(-)-3-(2-carboxypiperazine-4-yl)-propyl-1-phosphonic acid (CPP) on the extinction of conditioned freezing and suppression of bar pressing (conditioned emotional response). Rats extinguished normally during a 90 min session in the presence of systemic CPP (10 mg/kg), but were unable to recall extinction learning 24 hr later. This suggests that an NMDA-independent form of plasticity supports short-term extinction memory, but NMDA receptors are required for consolidation processes leading to long-term extinction memory. Surprisingly, extinction learned in the presence of CPP was recalled normally when tested 48 hr after training, suggesting a delayed consolidation process that was able to improve memory in the absence of further training. Delayed consolidation involves NMDA receptors because CPP injected on the rest day between training and test prevented 48 hr recall of extinction learned under CPP. Control experiments showed that the effect of CPP on memory consolidation was not caused by state-dependent learning or reduced expression of freezing under CPP. These findings demonstrate that NMDA receptor activation is critical for consolidation of extinction learning and that this process can be initiated after training has taken place. We suggest that consolidation of extinction involves off-line relearning that reinforces extinction memory through NMDA-mediated plasticity, perhaps in prefrontal-amygdala circuits.

Amygdala↗
PubMed2006

Transfer of learned perception of sensorimotor simultaneity.

Synchronizing a motor response to a predictable sensory stimulus, like a periodic flash or click, relies on feedback (somesthetic, auditory, visual, or other) from the motor response. Practically, this results in a small (<50 ms) asynchrony in which the motor response leads the sensory event. Here we show that the perceived simultaneity in a coincidence-anticipation task (line crossing) is affected by changing the perceived simultaneity in a different task (pacing). In the pace task, human subjects were instructed to press a key in perfect synchrony with a red square flashed every second. In training sessions, feedback was provided by flashing a blue square with each key press, below the red square. There were two types of training pace sessions: one in which the feedback was provided with no delay, the other (adapting), in which the feedback was progressively delayed (up to 100 ms). Subjects' asynchrony was unchanged in the first case, but it was significantly increased in the pace task with delay. In the coincidence-anticipation task, a horizontally moving vertical bar crossed a vertical line in the middle of a screen. Subjects were instructed to press a key exactly when the bar crossed the line. They were given no feedback on their performance. Asynchrony on the line-crossing task was tested after the training pace task with feedback. We found that this asynchrony to be significantly increased even though there never was any feedback on the coincidence-anticipation task itself. Subjects were not aware that their sensorimotor asynchrony had been lengthened (sometimes doubled). We conclude that perception of simultaneity in a sensorimotor task is learned. If this perception is caused by coincidence of signals in the brain, the timing of these signals depends on something-acquired by experience-more adaptable than physiological latencies.

Adolescent↗
PubMed1976

A behavioral manifestation of the development of interhemispheric transfer of learning in children.

Two hundred and fifty normal, right-handed children ranging from aged 5 to 15 were administered measures of grip strength, finger-tapping speed and the Tactual Performance Test. In order to assess the development of manual preference, analysis for each of the measures were performed on the left hand score as a percentage of the right hand score. Small age-related trends were found for the pure motor measures of speed and strength. In the case of the TPT, a clear age-related practice effect was found. The results were discussed in terms of the development of hemispheric cooperation.

Adolescent↗
PubMed1993

Development of a clinical protocol for hepatic gene transfer: lessons learned in preclinical studies.

Strategies for hepatic gene therapy have been proposed that involve isolation of primary hepatocytes and introduction of recombinant genes into these cells in culture, followed by autologous hepatocellular transplantation (HCT). Consideration of clinical applications requires data suggesting that HCT can be performed safely in human subjects in addition to data indicating that recombinant gene expression can reverse a disease process. This report describes preclinical studies that underlie a clinical trial of HCT in which hepatocytes would be labeled with a marker gene to facilitate assessment of engraftment in the recipient. Human hepatocytes were harvested from liver segments preserved in Belzar's solution and transduced with an amphotropic retroviral vector carrying a recombinant marker gene (neomycin phosphotransferase II). Human hepatocytes were recovered from monolayer culture, stained with the fluorescent dye 1,1'-dioctadecyl-3,3,3,3'-tetra-methylindo-carbocyanine perchlorate (DiI) and transplanted into severe combined immunodeficient mice by splenic injection. Engrafted hepatocytes were identified in the liver and spleen of severe combined immunodeficient mice but not immunocompetent controls. Two large animal models of HCT are described. In a dog model, neomycin phosphotransferase II-containing hepatocytes were identified in the liver 7 wk after transplantation. In a baboon model, autologous HCT with DiI-stained cells demonstrated that transplanted cells assume a normal morphology and constitute up to 5% of hepatocytes. These data demonstrate transduction and transplantation of human hepatocytes and the feasibility of HCT in large animals. On the basis of these studies, the proposed clinical trial for gene transfer and transplantation in human subjects has been approved by the National Institutes of Health and the Food and Drug Administration.(ABSTRACT TRUNCATED AT 250 WORDS)

Animals↗
PubMed2001

Two modes of transfer in artificial grammar learning.

Participants can transfer grammatical knowledge acquired implicitly in 1 vocabulary to new sequences instantiated in both the same and a novel vocabulary. Two principal theories have been advanced to account for these effects. One suggests that sequential dependencies form the basis for cross-domain transfer (e.g., Z. Dienes, G. T. M. Altmann, & S. J. Gao, 1999). Another argues that a form of episodic memory known as abstract analogy is sufficient (e.g., L. R. Brooks & J. R. Vokey, 1991). Three experiments reveal the contributions of the 2. In Experiment 1 sequential dependencies form the only basis for transfer. Experiment 2 demonstrates that this process is impaired by a change in the distributional properties of the language. Experiment 3 demonstrates that abstract analogy of repetition structure is relatively immune to such a change. These findings inform theories of artificial grammar learning and the transfer of grammatical knowledge.

Adult↗
PubMed2025

Deep learning-based annotation of plant abiotic stress resistance genes for crops.

The declining costs of DNA sequencing have expanded genomic data, crucial for understanding plant abiotic stress responses and crop improvement. However, accurate gene annotation remains challenging. To address this limitation, we propose the PASRGA, a deep learning approach that leverages transfer learning and contrastive learning to annotate genes related to drought, salt, cold, and UV resistance. PASRGA achieves high F1-scores, area under the receiver operating characteristic (AUROC), area under the precision-recall curve (AUPRC), and Matthews correlation coefficient (MCC) in annotating stress resistance genes, significantly outperforming the general protein annotation model CLEAN, the plant phosphatase gene annotation model PF-NET, the top-ranked model in the CAFA5 challenge NetGO 4.0, and four traditional machine learning methods. Its effectiveness was further validated with a salt stress treatment experiment in Eutrema salsugineum. To facilitate crop breeding practices, we utilized PASRGA to annotate the genomes of 17 major crops. To improve accessibility and utility, we incorporated both manually curated and PASRGA-predicted gene data, together with the PASRGA tool, into the PlantASRG database (https://bioinfor.nefu.edu.cn/PlantASRG/). This comprehensive resource aims to support crop breeding initiatives and ensure food security.

Crops, Agricultural↗
PubMed2005

Auditory frequency discrimination learning is affected by stimulus variability.

We explored the effects of training set variability on learning and generalization of pure-tone frequency discrimination (FD) in three groups of untrained, normally hearing adult listeners. Group 1 trained using a fixed standard tone at 1 kHz (fixed), Group 2 on slightly varying (roving) tones around 1 kHz, and Group 3 on widely varying standard frequencies (wide-roving). Initially, two thirds of all listeners had low FD thresholds (good listeners) and one third had intermediate to high thresholds (poor listeners). For good listeners, slight variations in the training set slowed learning but wide variations did not. Transfer to untrained frequencies (up to 4 kHz) and to the fixed condition was equivalent regardless of training condition, but Group 1 listeners did not fully transfer learning to the roving condition. For poor listeners, any variation in the training condition slowed learning and impaired transfer to other frequencies but did not affect transfer to untrained conditions. Thus, the effects of training set on progress and outcome depend on set variability and individual FD ability.

Adult↗