PubMed Health⌕ Search

SEARCH · PubMed Health

Results for “Deep 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 487 records · Page 27Linked to original sources
PubMed1997

The index of deep burn injury: an analysis of 66 extremity sites in 15 children.

The group of burn injuries generally classified as 'fourth degree' vary enormously in extent and severity. This fact, and the absence of a widely accepted grading system for these relatively uncommon injuries, has hindered our ability to learn from the experiences of others as reported in the literature. Based on a series of cadaver dissections, the Index of Deep Burn Injury (IDBI) was developed to provide such a classification scheme. The initial experience with the IDBI in a group of locally destructive "fourth degree' wounds is presented. It is believed that this index has the potential to improve our ability to describe highly destructive wounds.

Adolescent↗
PubMed2006

A cerebellar model for predictive motor control tested in a brain-based device.

The cerebellum is known to be critical for accurate adaptive control and motor learning. We propose here a mechanism by which the cerebellum may replace reflex control with predictive control. This mechanism is embedded in a learning rule (the delayed eligibility trace rule) in which synapses onto a Purkinje cell or onto a cell in the deep cerebellar nuclei become eligible for plasticity only after a fixed delay from the onset of suprathreshold presynaptic activity. To investigate the proposal that the cerebellum is a general-purpose predictive controller guided by a delayed eligibility trace rule, a computer model based on the anatomy and dynamics of the cerebellum was constructed. It contained components simulating cerebellar cortex and deep cerebellar nuclei, and it received input from a middle temporal visual area and the inferior olive. The model was incorporated in a real-world brain-based device (BBD) built on a Segway robotic platform that learned to traverse curved paths. The BBD learned which visual motion cues predicted impending collisions and used this experience to avoid path boundaries. During learning, the BBD adapted its velocity and turning rate to successfully traverse various curved paths. By examining neuronal activity and synaptic changes during this behavior, we found that the cerebellar circuit selectively responded to motion cues in specific receptive fields of simulated middle temporal visual areas. The system described here prompts several hypotheses about the relationship between perception and motor control and may be useful in the development of general-purpose motor learning systems for machines.

Cerebellum↗
PubMed1998

Repair of dural defects in awkward areas-technical note.

Dural tears located at the base of the skull are difficult to repair due to the difficulties in the appropriate access and the fragility of the dura in such areas. In our experience the biggest problem when attempting to perform a dural repair in a deep narrow field is not to place the stitches, but rather to set the knots. A newly designed, easy-to-learn technique has been developed for dural closure in these situations. We present here a new technique for dural suturing of special interest when the space available is limited. In our hands it is possible to obtain a watertight dural closure in most microsurgical operations performed through a small hole and/or into a narrow, deep surgical field. These techniques can also be applied during a secondary procedure following development of a postoperative CSF leak. While simple and easy to learn, these techniques require practice in the laboratory setting before clinical application.

Cerebrospinal Fluid↗
PubMed2000

Cerebellar cortical degeneration disrupts discrimination learning but not delay or trace classical eyeblink conditioning.

The authors investigated classical eyeblink conditioning in a relatively rare patient, B.R., with extensive cerebellar cortical atrophy and marked sparing of the dentate nucleus. Patient B.R.'s ability to acquire and extinguish simple associations (delay and trace conditioning tasks) as well as her ability to acquire more complex associations (temporal and simple discrimination tasks) were examined. There are 2 primary findings from this study. First, B.R. showed normal acquisition and extinction in delay and trace conditioning. Second, she demonstrated a complete inability to learn associative discriminations, even in the case of a simple 2-tone discrimination within the context of a delay paradigm. The latter finding was unexpected because of the sparing of her deep cerebellar nuclei. These data suggest that the cerebellar cortex, or pathways traversing cerebellar cortex, play an important role in classical eyeblink discrimination learning.

Adult↗
PubMedDate not supplied

[Always learning; IDER's proposal].

Professional development should help a professional person look at life from the first line, with deep knowledge in his field, with his own opinion, with know-how, with a critical sense, with something to say and something to add. This is IDER's philosophy based on having students develop their scientific, technical and human capacities, relating their experiences with the knowledge provided by textbooks, by other professional by other students, by consultations and scientific work. And always at one's own rhythm because IDER has specialized in continuing professional development from afar.

Aged↗
PubMed1990

Autogenic drainage: efficacy of a simplified method.

A simplified and modified technique of the original autogenic drainage (AD) is described which is supplemented by breathing against an external flow resistance. The new method allows a better teaching and learning. It's efficacy is proven by means of a comparative trial of AD and PEP physiotherapy (i.e. expiration against a defined stenosis). The autogenic drainage (AD) was introduced by a Belgian working group (2). The basic idea was to support the elimination of mucus by deep breathing and by repressing the cough as long as possible. However, a rather sophisticated method impeded teaching and learning of AD (1). Thus, till now, there are only preliminary results which suggest the efficacy of AD (e.g. 3). Therefore, a simplified technique was developed and examined with regard to its sputum eliminating effect.

Adolescent↗
PubMed1997

Mechanisms underlying memory impairment in schizophrenia.

BACKGROUND: The purpose of this experiment was to investigate mechanisms underlying commonly-observed verbal memory impairments in schizophrenia, and especially the hypothesized encoding deficit. METHODS: A verbal memory task was administered to 38 patients with schizophrenia and 38 normal controls. Three functions involved in long-term memory-encoding, early phase of storage, retrieval-were investigated. First, non-organizable lists were compared to semantically-organizable lists in a free recall task, in order to vary encoding conditions. Superficial encoding (measured by a 'sequence' index) and deep encoding (measured by a categorization index) were assessed. Secondly, early storage was investigated by varying the delay between learning and recall. Lastly, cues were provided for organizable lists (semantic cues) and non-organizable lists (recognition sheet), in order to vary retrieval conditions. RESULTS: An analysis of variance revealed an interaction between type of list (organizable, non-organizable) and group, showing that patients used organization less than controls. A further analysis showed that deep encoding was impaired. Also, although the propensity to use superficial encoding was unimpaired, its efficiency was less. The analysis of variance revealed no interaction with delay or with either type of cue. A correlation was found between deep processing and memory performance in both groups. CONCLUSIONS: A major deficit in encoding appeared in the patient group, with a lesser use of deep encoding and a lesser efficiency of superficial encoding. On the other hand, the early phase of storage and the retrieval function seemed unaffected. Overall memory performance appeared to be related to the depth of encoding.

Adult↗
PubMed2005

Organization of optic lobes that support motion detection in a semiterrestrial crab.

There is a mismatch between the documentation of the visually guided behaviors and visual physiology of decapods (Malacostraca, Crustacea) and knowledge about the neural architecture of their visual systems. The present study provides a description of the neuroanatomical features of the four visual neuropils of the grapsid crab Chasmagnathus granulatus, which is currently used as a model for investigating the neurobiology of learning and memory. Visual memory in Chasmagnathus is thought to be driven from within deep retinotopic neuropil by large-field motion-sensitive neurons. Here we describe the neural architecture characterizing the Chasmagnathus lobula, in which such neurons are found. It is shown that, unlike the equivalent region of insects, the malacostracan lobula is densely packed with columns, the spacing of which is the same as that of retinotopic units of the lamina. The lobula comprises many levels of strata and columnar afferents that supply systems of tangential neurons. Two of these, which are known to respond to movement across the retina, have orthogonally arranged dendritic fields deep in the lobula. They also show evidence of dye coupling. We discuss the significance of commonalties across taxa with respect to the organization of the lamina and medulla and contrasts these with possible taxon-specific arrangements of deeper neuropils that support systems of matched filters.

Animals↗
PubMedDate not supplied

Medical expert systems based on causal probabilistic networks.

Causal probabilistic networks (CPNs) offer new methods by which you can build medical expert systems that can handle all types of medical reasoning within a uniform conceptual framework. Based on the experience from a commercially available system and a couple of large prototype systems, it appears that CPNs are now an attractive alternative to other methods. A CPN is an intensional model of a domain, and it is therefore conceptually much closer to qualitative reasoning systems and to simulation systems than to rule-based or logic-based systems. Recent progress in Bayesian inference in networks has yielded computationally efficient methods. The inference method used follows the fundamental axioms of probability theory, and gives a sound framework for causal and diagnostic (deductive and abductive) reasoning under uncertainty. Experience with the prototypes indicates that it may be possible to use decision theory as a rational approach to test planning and therapy planning. The way in which knowledge is acquired and represented in CPNs makes it easy to express 'deep knowledge' for example in the form of physiological models, and the facilities for learning make it possible to make a smooth transition from expert opinion to statistics based on empirical data.

Artificial Intelligence↗
PubMed1999

Deep vein thrombosis during pregnancy and the puerperium: a meta-analysis of the period of risk and the leg of presentation.

UNLABELLED: We performed a meta-analysis of all published studies of deep vein thrombosis during pregnancy and the puerperium using MEDLINE between 1966 and May 1998. Data were pooled using a random effects model. Fourteen studies included relevant information on deep vein thrombosis in pregnancy or the puerperium, and used objective testing to diagnose deep vein thrombosis. The pooled event rate for left sided or bilateral deep vein thrombosis was 82.2 percent (95 percent CI 75.1 to 87.5). There was no statistical evidence of heterogeneity for this figure (P = .08). Twelve studies reported on the trimester in which deep vein thrombosis was diagnosed. The deep vein thrombosis event rate during the first trimester was 21.9 percent (95 percent CI 17.4 to 27.3), 33.7 percent (95 percent CI 28.1 to 39.8) during the second trimester, and 47.6 percent (95 percent CI 39.2 to 56.2) for the third trimester. Heterogeneity testing was not significant. Nine studies compared deep vein thrombosis events between the antepartum and puerperium periods, with 65.5 percent (95 percent CI 58.1 to 72.1) arising during pregnancy, and 34.5 percent (95 percent CI 27.9 to 41.9) postpartum (P = .08, not heterogeneous). Using these figures, the estimated relative distribution of 100 deep vein thrombosis events during pregnancy and the puerperium would be 0.23 per day during pregnancy, and 0.82 per day in the postpartum period. During pregnancy and the puerperium, deep vein thrombosis is more likely to arise in the left leg. More than half of all deep vein thromboses during pregnancy occur during the first and second trimesters. Furthermore, during the puerperium, the risk of developing deep vein thrombosis is significantly higher than antepartum. TARGET AUDIENCE: Obstetricians & Gynecologists, Family Physicians. LEARNING OBJECTIVES: After completion of this article, the reader will be able to understand the incidence and distribution of DVTs during pregnancy and to be able to appreciate the differing risk of DVT during the various time periods in a pregnancy.

Female↗
PubMed2000

Right atrial angiographic evaluation of the posterior isthmus: relevance for ablation of typical atrial flutter.

BACKGROUND: Gaining anatomic information about the posterior isthmus is not generally part of flutter ablation procedures. We postulated that right atrial (RA) angiography could rationalize the ablation approach by revealing the conformation of the isthmus. METHODS AND RESULTS: In 100 consecutive patients, biplane RA angiography was performed before ablation to guide catheter contact with the isthmus along its length. Angiography showed a wide variation in the width of the isthmus (17 to 54 mm; 31.3+/-7.9), its angle with the inferior vena cava in the right anterior oblique projection (68 degrees to 114 degrees; 90.3+/-9.0 degrees ), and its lateral position relative to the inferior vena cava in the left anterior oblique projection. A deep sub-Eustachian recess was revealed in 47%, with a mean depth of 4.3+/-2.1 mm (1.5 to 9.4). A Eustachian valve was visualized in 24%. Ablation resulted in bidirectional conduction block (which could be transient) in all, with a median of 2 dragging radiofrequency (RF) applications (2.3+/-2.5 RF applications; 57 degrees C, < or =99 seconds each). Permanent block was achieved in 99%, with a median of 3 RF applications (3.4+/-3.0). The presence of a Eustachian valve or concave isthmus was associated with statistically more RF applications; the same trend was seen for patients with deep pouches. The number of RF applications decreased statistically throughout the study, indicating a learning curve. No patient had a recurrence after a follow-up of 13+/-11 months. CONCLUSIONS: Right atrial angiography reveals a highly variable isthmus anatomy, often showing particular configurations that can make ablation more laborious. Rational adaptation of the ablation approach to these anatomic findings may contribute to successful ablation.

Adolescent↗
PubMed2000

Retention of concurrent taste aversion learning after electrolytic lesioning of the interpositus-dentate region of the cerebellum.

Lesions in the interpositus-dentate region of the cerebellum impair short-term, or concurrent, TAL. In this type of learning, animals must discriminate between two flavor stimuli presented at the same time, one of which is associated with an aversive product. The task is learned by the control animals, and within this group the animals that acquire it adequately enough (15/22, 70% criterion) retain the learned taste discrimination when they are subjected to it again after being lesioned in the interpositus-dentate region. These results suggest that the deep nuclei are essential in the concurrent TAL acquisition process, but not in its retention.

Animals↗
PubMed2001

MRI-assessed volume of cerebellum correlates with associative learning.

Richard F. Thompson's cerebellar model of classical eyeblink conditioning highlights Purkinje cells in cerebellar cortex and principal cells in the deep cerebellar nucleus as the integrating cells for acquisition of conditioned responses (CRs). CR acquisition is significantly slower in rabbits with lesions to cerebellar cortex and in Purkinje cell-deficient mice that lose all cerebellar cortical Purkinje cells. Purkinje cells are the largest neurons in the cerebellum and contribute significantly to cerebellar volume. Magnetic resonance imaging (MRI) was used to assess cerebellar volume in humans. Cerebellar volume was related to eyeblink conditioning (400-ms delay procedure) in 8 adults (21-35 years) and compared to 8 older adults (77-95 years) tested previously (Woodruff-Pak, Goldenberg, Downey-Lamb, Boyko, & Lemieux, 2000). In the young adult sample, there was a high correlation between percentage of CRs in a session and cerebellar volume (corrected for total intracranial volume [TIV], r =.58, p =.066). There were statistically significant age differences in cerebellar volume, t(14) = 8.96, p <.001, and percentage of CRs, t(14) = 3.85, p <.002, but no age difference in TIV. Combining the young and older adult sample, the correlation between percentage of CRs and cerebellar volume (corrected for TIV) was.832 (p <.001). Cerebellar volume showed age-related deficits likely due to Purkinje cell loss. Individual differences in classical eyeblink conditioning are associated with differences in cerebellar volume, supporting Thompson's model of a cerebellar cortical role in facilitating this form of associative learning.

Adult↗
PubMed1998

Declarative and procedural learning, quantitative measures of the hippocampus, and subcortical white alterations in Alzheimer's disease and ischaemic vascular dementia.

This research investigated whether subjects with Alzheimer's disease (AD) and ischaemic vascular dementia (IVD) associated with periventricular and deep white matter alterations can be dissociated on tests of declarative and procedural memory, as well as on MRI indices of white matter alterations and the size of the hippocampal formation. The California Verbal Learning Test (CVLT) and the Pursuit Rotor Learning Tests (PRLT) were used to measure declarative and procedural memory, respectively. Subjects with IVD obtained a higher score on the CVLT recognition discriminability index; however, on the PRLT total time on target, carry-over between trial blocks, and slope calculated for all test trials was low. Subjects with AD exhibited the opposite profile. MRI studies indicated that subjects with IVD had considerably greater white matter alterations, but larger hippocampal formations than subjects with AD. Higher scores on the CVLT recognition discriminability index were correlated with increased size of the body of the hippocampus and parahippocampal gyrus. By contrast, as the severity of white matter alterations increased the slope on the PRLT declined. In sum, subjects with AD and IVD can be dissociated on the basis of differing patterns of impairment on tests of declarative and procedural memory, and MRI indices of white matter alteration and the integrity of the hippocampal formation.

Aged↗
PubMed1998

Polarity of long-term synaptic gain change is related to postsynaptic spike firing at a cerebellar inhibitory synapse.

Long-term potentiation and depression (LTP and LTD) in excitatory synapses can coexist, the former being triggered by stimuli that produce strong postsynaptic excitation and the latter by stimuli that produce weaker postsynaptic excitation. It has not been determined whether these properties also apply to LTP and LTD in the inhibitory synapses between Purkinje neurons and the neurons of the deep cerebellar nuclei (DCN), a site that has been implicated in certain types of motor learning. DCN cells exhibit a prominent rebound depolarization (RD) and associated spike burst upon release from hyperpolarization. In these cells, LTP can be elicited by short, high-frequency trains of inhibitory postsynaptic potentials (IPSPs), which reliably evoke an RD. LTD is induced if the same protocol is applied with conditions where the amount of postsynaptic excitation is reduced. The polarity of the change in synaptic strength is correlated with the amount of RD-evoked spike firing during the induction protocol. Thus, some important computational principles that govern the induction of use-dependent change in excitatory synaptic efficacy also apply to inhibitory synapses.

Action Potentials↗
PubMed1998

Segmentation and interpretation of MR brain images: an improved active shape model.

This paper reports a novel method for fully automated segmentation that is based on description of shape and its variation using point distribution models (PDM's). An improvement of the active shape procedure introduced by Cootes and Taylor to find new examples of previously learned shapes using PDM's is presented. The new method for segmentation and interpretation of deep neuroanatomic structures such as thalamus, putamen, ventricular system, etc. incorporates a priori knowledge about shapes of the neuroanatomic structures to provide their robust segmentation and labeling in magnetic resonance (MR) brain images. The method was trained in eight MR brain images and tested in 19 brain images by comparison to observer-defined independent standards. Neuroanatomic structures in all testing images were successfully identified. Computer-identified and observer-defined neuroanatomic structures agreed well. The average labeling error was 7%+/-3%. Border positioning errors were quite small, with the average border positioning error of 0.8+/-0.1 pixels in 256 x 256 MR images. The presented method was specifically developed for segmentation of neuroanatomic structures in MR brain images. However, it is generally applicable to virtually any task involving deformable shape analysis.

Algorithms↗
PubMed2006

Long-term depression at the mossy fiber-deep cerebellar nucleus synapse.

Several lines of evidence have indicated that the deep cerebellar nuclei (DCN) are a site of memory storage for certain forms of motor learning, most notably associative eyelid conditioning. In particular, these experiments, together with network models, have implicated the excitatory glutamatergic synapse between mossy fibers and DCN neurons in this memory trace. However, to date, evidence for persistent use-dependent change in the strength of this synapse has been almost entirely absent. Here, we report that high-frequency burst stimulation of mossy fibers, either alone or paired with postsynaptic depolarization, gives rise to long-term depression (LTD) of the mossy fiber-DCN synapse. This form of LTD is not associated with changes in the paired-pulse ratio and is blocked by loading with a postsynaptic Ca2+ chelator but not by bath application of an NMDA receptor antagonist. Mossy fiber-DCN LTD requires activation of a group I metabotropic glutamate receptor (mGluR) and protein translation. Unlike mGluR/translation-dependent LTD in other brain regions, this form of LTD requires mGluR1 and is mGluR5 independent.

Animals↗