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 595 records · Page 33Linked to original sources
PubMed2001

[New functional cerebral cartography: studies of plasticity of the human brain].

The non-invasive brain scanning techniques, introduced a quarter of a century ago, have become crucial for diagnosis in clinical neurology. They have also been used to investigate brain function and have provided information about normal activity and pathogenesis. They have been used to investigate functional specialisation in the brain and how specialised areas communicate to generate complex integrated functions such as speech, memory, the emotions and so on. Brain plasticity is a generic term that has now come to be used to signify any changes in brain structure or function. The phenomenon is poorly understood, and yet clinical neurologists are aware that spontaneous recovery from brain lesions is not uncommon. An improved understanding of the mechanisms of recovery may generate new therapeutic strategies and indicate ways of modulating mechanisms that promote plastic compensation for loss of function. The main methods used to investigate these issues are positron emission tomography and magnetic resonance imaging (MRI). The techniques of functional brain mapping and computational morphometrics depend on high performance scanners and a validated set of analytic statistical procedures that generate reproducible data and meaningful inferences from them. The motor system presents a good paradigm to illustrate advances made by scanning towards an understanding of plasticity at the level of brain areas. The normal motor system is organised in a nested hierarchy. Recovery from paralysis caused by internal capsule strokes involves functional reorganisation manifesting as changed patterns of activity in the component brain areas of the normal motor system. The pattern of plastic modification depends in part on patterns of residual or disturbed connectivity after brain injury. Therapeutic manipulations in patients with Parkinson's disease using deep brain stimulation, dopaminergic agents or foetal mesencephalic transplantation provide a means to examine mechanisms underpinning plastic change. Other models of plastic change, such as normal visuospatial learning or re-establishing speech comprehension after cochlear implantation in the deaf illustrate how patterns of brain function adapt over time. Limitations of the scanning techniques and prospects for the future are discussed in relation to new developments in the neuroimaging field.

Brain↗
PubMedDate not supplied

Forthergillian Lecture. Imaging human brain function.

The non-invasive brain scanning techniques introduced a quarter of a century ago have become crucial for diagnosis in clinical neurology. They have also been used to investigate brain function and have provided information about normal activity and pathogenesis. They have been used to investigate functional specialization in the brain and how specialized areas communicate to generate complex integrated functions such as speech, memory, the emotions and so on. The phenomenon of brain plasticity is poorly understood and yet clinical neurologists are aware, from everyday observations, that spontaneous recovery from brain lesions is common. An improved understanding of the mechanisms of recovery may generate new therapeutic strategies and indicate ways of modulating mechanisms that promote plastic compensation for loss of function. The main methods used to investigate these issues are positron emission tomography and magnetic resonance imaging (M.R.I.). M.R.I. is also used to map brain structure. The techniques of functional brain mapping and computational morphometrics depend on high performance scanners and a validated set of analytic statistical procedures that generate reproducible data and meaningful inferences from brain scanning data. The motor system presents a good paradigm to illustrate advances made by scanning towards an understanding of plasticity at the level of brain areas. The normal motor system is organized in a nested hierarchy. Recovery from paralysis caused by internal capsule strokes involves functional reorganization manifesting itself as changed patterns of activity in the component brain areas of the normal motor system. The pattern of plastic modification depends in part on patterns of residual or disturbed connectivity after brain injury. Therapeutic manipulations in patients with Parkinson's disease using deep brain stimulation, dopaminergic agents or fetal mesencephalic transplantation provide a means to examine mechanisms underpinning plastic change. Other models of plastic change, such as normal visuospatial learning or re-establishing speech comprehension after cochlear implantation in the deaf illustrate how patterns of brain function adapt over time. Limitations of the scanning techniques and prospects for the future are discussed in relation to new developments in the neuroimaging field.

Brain↗
PubMed2000

Time window control: a model for cerebellar function based on synchronization, reverberation, and time slicing.

We present a new hypothesis of cerebellar function that is based on synchronization, delayed reverberation, and time windows for triggering spikes. Our model suggests that granule cells admit mossy fiber activity to the parallel fibers only if the Golgi cells are firing synchronously and if the mossy-fiber spikes arrive within short and well-defined time windows. The concept of time window control organizes neuronal activity in discrete 'time slices' that can be used to discern meaningful information from background noise. In particular, Purkinje cell activity can trigger rebound spikes in deep cerebellar nuclei cells, which project via brain stem nuclei and mossy fibers back to the cerebellar cortex. Using a detailed model of deep cerebellar nuclei cells, we demonstrate that the delayed firing of rebound spikes is a robust mechanism so as to ensure that the reverberated activity re-arrives in the mossy fibers just during the granule-cell time window. Large network simulations reveal that synaptic plasticity (LTD and LTP) at the parallel fiber/Purkinje cell synapses that relies on the timing of the parallel fiber and climbing fiber activities allows the system to learn, store, and recall spatiotemporal patterns of spike activity. Climbing fiber spikes function both as teacher and as synchronization signals. The temporal characteristics of the climbing fiber activity are due to intrinsic oscillatory properties of inferior olivary neurons and to reverberating projections between deep cerebellar nuclei, the mesodiencephalic junction, and the inferior olive. Thus, the reverberating loops of the mossy fiber system and climbing fiber system may interact directly with the time windows provided by the circuitry of the cerebellar cortex so as to generate the appropriate spatio-temporal firing patterns in the deep cerebellar nuclei neurons that control premotor systems. In future studies the model will be extended in that high frequency simple spike activities will be included and that their relevance for motor control will be addressed.

Animals↗
PubMed2006

Synchronizing activity of basal ganglia and pathophysiology of Parkinson's disease.

Early physiological studies emphasized changes in the discharge rate of basal ganglia in the pathophysiology of Parkinson's disease (PD), whereas recent studies stressed the role of the abnormal oscillatory activity and neuronal synchronization of pallidal cells. However, human observations cast doubt on the synchronization hypothesis since increased synchronization may be an epi-phenomenon of the tremor or of independent oscillators with similar frequency. Here, we show that modern actor/ critic models of the basal ganglia predict the emergence of synchronized activity in PD and that significant non-oscillatory and oscillatory correlations are found in MPTP primates. We conclude that the normal fluctuation of basal ganglia dopamine levels combined with local cortico-striatal learning rules lead to noncorrelated activity in the pallidum. Dopamine depletion, as in PD, results in correlated pallidal activity, and reduced information capacity. We therefore suggest that future deep brain stimulation (DBS) algorithms may be improved by desynchronizing pallidal activity.

Animals↗
PubMed2005

Long-term anti-kindling effects of desynchronizing brain stimulation: a theoretical study.

In a modeling study we show that desynchronization stimulation may have powerful anti-kindling effects. For this, we incorporate spike-timing-dependent plasticity into a generic network of coupled phase oscillators, which serves as a model network of synaptically interacting neurons. Two states may coexist under spontaneous conditions: a state of uncorrelated firing and a state of pathological synchrony. Appropriate stimulation protocols make the network learn or unlearn the pathological synaptic interactions, respectively. Low-frequency periodic pulse train stimulation causes a kindling. Permanent high-frequency stimulation, used as golden standard for deep brain stimulation in medically refractory movement disorders, basically freezes the synaptic weights. In contrast, desynchronization stimulation, e.g., by means of a multi-site coordinated reset, has powerful long-term anti-kindling effects and enables the network to unlearn pathologically strong synaptic interactions. We propose desynchronization stimulation for the therapy of movement disorders and epilepsies.

Brain↗
PubMed1999

Vermectomy enhances parvalbumin expression and improves motor performance in weaver mutant mice: an animal model for cerebellar ataxia.

In the Weaver mutant mouse (wv/wv), an animal model for hereditary cerebellar ataxia, electrophysiological experiments have revealed a disorganized output of cerebellar Purkinje cells (the latter using GABA as an inhibitory transmitter) which, by a cascade of mechanisms, was thought to be the cause of the poor motor abilities. In Purkinje cell degeneration mice (pcd/pcd) lacking nearly all Purkinje cells and displaying milder motor deficiencies than wv, in comparison to wild-type mice, a strong increase in parvalbumin- and (co-localized with parvalbumin) glycine-immunopositive somata in the deep cerebellar and vestibular nuclei has recently been found. It was therefore intriguing to investigate whether motor performance in weaver mutants could be ameliorated by applying cerebellar lesions to eliminate the faulty output and to look for a change in transmitter weighting, indicated by a strong increase in parvalbumin-positive somata in areas (the respective target areas) which were formerly devoid of it. Ten Weaver mutants were subjected to cerebellar lesions. After removal of the vermis a total abolition of tremor, a definite improvement in the balance of affected body parts, an increase in locomotor activity when tested in an open-field matrix, and a strong increase in parvalbumin expression in Weaver mutant deep cerebellar and vestibular nuclei in comparison to wild-types have indeed been found. Increase in motor activity (or explorative behaviour) has been placed in relation to learning mechanisms. The increase in parvalbumin expression and the observed improvement in motor abilities and mechanisms probably related to learning underline the hypothesis that any change in the physiological equilibrium of the brain function by removal of input or output related to an assembly of nerve cells leads to a cascade of changes at the transmitter and neuronal level in near or distant connected brain structures.

Animals↗
PubMedDate not supplied

[The pedunculopontine nucleus. A structure involved in motor and emotional processing].

INTRODUCTION: There is currently a growing interest for conducting studies into the electrical and neurochemical activity of the pedunculopontine nucleus (PPN) due to the privileged position occupied by this structure in the flow of information to and from the cortex. This nucleus acts as a relay, not only for the motor information that is processed in the basal ganglia but also for information of an emotional type, whose main centre is the nucleus accumbens. It is also strongly linked with the aspects that determine the mechanisms governing addiction to certain drugs. DEVELOPMENT: We conduct a detailed analysis of the main findings from studies of the role played by the PPN in the physiopathology of Parkinsonism, namely the study of metabolic activity, immunohistochemical studies with different tracers, electrophysiological studies that have confirmed the immunohistochemical observations, as well as deep electrical stimulation carried out in non human primates. Furthermore, we also examine the part played by this structure in the processing of emotional information associated with different learning tasks. CONCLUSIONS: Overall, the authors grant the PPN a privileged position in the physiopathology of the axial disorders related to Parkinson s disease; its most important afference, stemming from the subthalamic nucleus, appears to play a key role in the understanding of the part played by the PPN in Parkinsonism.

Animals↗
PubMed2006

Students learning from patients: let's get real in medical education.

Medical students must be prepared for working in inter-professional and multi-disciplinary clinical teams centred on a patient's care pathway. While there has been a good deal of rhetoric surrounding patient-centred medical education, there has been little attempt to conceptualise such a practice beyond the level of describing education of communication skills and empathy within a broad 'professionalism' framework. Paradoxically, while aiming to strengthen patient-student interactions, this approach tends to refocus on the role modelling of the physician, and opportunities for potentially deep collaborative working relationships between students and patients are missed. A radical overhaul of conventional doctor-led medical education may be necessary, that also challenges the orthodoxies of individualistic student-centred approaches, leading to an authentic patient-centred model that shifts the locus of learning from the relationship between doctor as educator and student to the relationship between patient and student, with expert doctor as resource. Drawing on contemporary poststructuralist theory of text and identity construction, and on innovative models of work-based learning, the potential quality of relationship between student and patient is articulated in terms of collaborative knowledge production, involving close reading with the patient as text, through dialogue. Here, a medical 'education' displaces traditional forms of medical 'training' that typically involve individual information reproduction. Students may, paradoxically, improve clinical acumen through consideration of silences, gaps, and contradictions in patients as texts, rather than treating communication as transparent. Such paradoxical effects have been systematically occluded or denied in traditional medical education.

Education, Medical↗
PubMed2003

A two-stage unsupervised learning algorithm reproduces multisensory enhancement in a neural network model of the corticotectal system.

Multisensory enhancement (MSE) is the augmentation of the response to sensory stimulation of one modality by stimulation of a different modality. It has been described for multisensory neurons in the deep superior colliculus (DSC) of mammals, which function to detect, and direct orienting movements toward, the sources of stimulation (targets). MSE would seem to improve the ability of DSC neurons to detect targets, but many mammalian DSC neurons are unimodal. MSE requires descending input to DSC from certain regions of parietal cortex. Paradoxically, the descending projections necessary for MSE originate from unimodal cortical neurons. MSE, and the puzzling findings associated with it, can be simulated using a model of the corticotectal system. In the model, a network of DSC units receives primary sensory input that can be augmented by modulatory cortical input. Connection weights from primary and modulatory inputs are trained in stages one (Hebb) and two (Hebb-anti-Hebb), respectively, of an unsupervised two-stage algorithm. Two-stage training causes DSC units to extract information concerning simulated targets from their inputs. It also causes the DSC to develop a mixture of unimodal and multisensory units. The percentage of DSC multisensory units is determined by the proportion of cross-modal targets and by primary input ambiguity. Multisensory DSC units develop MSE, which depends on unimodal modulatory connections. Removal of the modulatory influence greatly reduces MSE but has little effect on DSC unit responses to stimuli of a single modality. The correspondence between model and data suggests that two-stage training captures important features of self-organization in the real corticotectal system.

Algorithms↗
PubMed1996

Two approaches to learning a theatrical script.

Previous research has shown that actors recall text with word-for-word fidelity by using a type of meaning-based learning strategy that usually results only in retention of the gist (Noice, 1993). Six actors generated think-aloud protocols while studying a theatrical script. Analysis of these protocols indicated that the actors' application of their strategy required minute attention to the exact wording, punctuation, etc. in order to derive the deep meaning they require for interpretation and performance. A think-aloud protocol of the well-known mnemonist, Harry Lorayne, was collected for the same theatrical script. Although the strategies of the actors and the mnemonist were totally different, they appear to be dependent on the same underlying processes. An additional finding was that, although the mnemonist employed imagery-mediation (usually used for learning discrete items), he had devised an original variation capable of yielding verbatim retention. The reasons why mnemonic devices might be useful to actors in certain situations are discussed.

Association Learning↗
PubMed1998

Changing attitudes to infection management in primary care: a controlled trial of active versus passive guideline implementation strategies.

BACKGROUND AND OBJECTIVES: When attempting to implement evidence-based medicine, such as through clinical guidelines, we often rely on passive educational tactics, for example didactic lectures and bulletins. These methods involve the recipient in relatively superficial processing of information, and any consequent attitude changes can be expected to be short-lived. However, active methods, such as practice-based discussion, should involve recipients in deep processing, with more enduring attitude changes. In this experiment, the aim was to assess the efficacy of an active strategy at promoting deep processing and its effectiveness, relative to a typical passive method, at changing attitudes between groups of GPs over 12 months across an English Health District. METHODS: All 191 GPs operating from 69 practices in the Wirral Health District of Northwest England were assigned, with minimization of known confounding variables, to three experimental groups: active, passive and control. The groups were shown to have similar learning styles. The objective of the study was to impart knowledge of best management of infections as captured in a series of locally developed clinical guidelines. The passive group GPs were given a copy of the guidelines and were invited to an hour-long lecture event. The GPs in the deep group were given a copy of the guidelines and were invited to engage in an hour-long discussion about the guideline content at their own premises. The control group received neither the guidelines nor any educational contact regarding them. Three months before and 12 months after the interventions, all GPs were sent a postal questionnaire on their preferred empirical antibiotic for 10 common bacterial infections. The responses were compared in order to ascertain whether increased knowledge of best clinical practice was evident in each group. RESULTS: Seventy-five per cent (144/191) of GPs responded to the pre-intervention questionnaire, 62 % (119/191) post-intervention. Thirty-four per cent (22/64) of GPs in the passive group attended the lecture; 91% (60/66) of the active group engaged in discussion at meetings with the authors. A significantly higher proportion of the active group participants' speaking time, during a sample of four visits, was devoted to verbal indicators of active processing than the passive group lecture attenders (difference = 55%, Fisher's exact test P = 0.002, OR = 11.5, 95% CI 2.1-113.4). Inter-observer agreement on the classification of the verbal evidence was highly statistically significant for all classes (Pearson's product moment correlation, P < 0.0005, r = +0.893 to +0.999) except repetition (P > 0.05, r = +0.407). Median compliance of responses with the guidelines improved by 2.5% within the control group and 4% within the passive, but by 23% within the active. The difference between the changes in the active and control groups was highly statistically significant at 17.5% (Mann-Whitney test, P = 0.004, 95% CI 6-29%). However, for the 10 infections, the median difference between the changes in the passive and control groups was not significant at 3% (P = 0.75, 95% CI -8 to +12. The median difference between changes in the active and passive groups was significant at 17% (P = 0.015, 95% CI 7-24%) in favour of the active. DISCUSSION: An active educational strategy attracted more participation and was more effective at generating deep cognitive processing than a passive strategy. A large improvement, lasting for at least 12 months, in attitude-compliance with guidelines on the optimal treatment of infections was imparted by the active processing method. A typical passive method was much less popular and had an insignificant impact on attitudes. The findings suggest that initiatives aiming to implement evidence-based guidelines must employ active educational strategies if enduring changes in attitude are to result.

Attitude of Health Personnel↗
PubMed2002

Improved survival of patients undergoing palliation of hypoplastic left heart syndrome: lessons learned from 115 consecutive patients.

BACKGROUND: Outcome of stage 1 palliation (S1P) for hypoplastic left heart syndrome (HLHS) has improved coincident with application of treatment strategies including continuous superior vena cava oximetry (SvO2), phenoxybenzamine (POB), strategies to minimize the duration of deep hypothermic circulatory arrest (DHCA) and efforts to ameliorate the inflammatory response to cardiopulmonary bypass (CPB) using aprotinin and modified ultrafiltration. METHODS AND RESULTS: Analysis of a consecutive series of 115 patients undergoing S1P was done to identify the risk factors for mortality and the impact of new treatment strategies. For the current era, July 1996 to October 2001, hospital survival was 93% (75/81) compared with 53% (18/34) for the time period, January 1992 to June 1996, P<0.001. Survival to stage 2 palliation (S2P) was also significantly improved in the current era, 81% (66/81) versus 44% (15/34), P<0.01. Anti-inflammatory treatment strategies demonstrated improved survival by univariate analysis (P<0.001). Multivariate analysis identified continuous SvO2 monitoring as a factor favoring S1P survival (P=0.02) and use of POB as a factor favoring survival to S2P (P=0.003). In the current era shorter duration of DHCA was associated with improved survival to S2P (P=0.02). CONCLUSIONS: Improved survival following S1P can be achieved with strategies that allow for early identification of decreased systemic output and the use of afterload reduction to stabilize systemic vascular resistance and therefore the pulmonary to systemic flow ratio. Strategies to ameliorate the inflammatory response to CPB may decrease the degree and duration of postoperative support. Strategies to minimize duration of DHCA may improve intermediate survival and merit additional studies.

Anastomosis, Surgical↗
PubMed2003

How to use participatory action research in primary care.

OBJECTIVE: The aim of the article is to demonstrate the usefulness of participatory action research (PAR) in primary care. The author used PAR firstly to develop a deeper understanding of mutual participation in the doctor-patient encounter and secondly to apply this learning in a rural cross-cultural practice setting. METHOD: PAR was done with four patient groups. Four patients with terminal illnesses formed groups with their family members, neighbours and friends. Seven meetings were held with each group over a period of 6 months. The meetings were conducted in Tsonga, which is the local vernacular. All the meetings were audio-taped. The primary question for each meeting was how the group could work together to achieve the best possible health outcome for the patient. Additionally, the author, who facilitated the meetings, kept a reflective diary, including field notes over the research period. One member of each group kept a written record of each meeting. Three free attitude interviews were conducted with the author over the research period to elicit the development of his understanding about mutual participation in the doctor-patient encounter. The recorded meetings and interviews were transcribed and translated and themes subsequently identified using the transcripts. The reflective diary was analysed similarly. A model was constructed to depict the themes and their interrelatedness. The model was interpreted and conclusions were drawn. RESULTS: The PAR process had a positive effect on the doctor-patient encounter. PAR greatly resembles a mutual participatory doctor-patient encounter. The research facilitator had certain basic tenets in order to facilitate participation. The patients who participated actively benefited most. Basic interviewing techniques were used to facilitate the mutual participation in PAR. CONCLUSIONS: PAR is very applicable in primary care. The principles of PAR such as mutual collaboration, reciprocal respect, co-learning and acting on results from the enquiry are essential in the doctor-patient relationship. Self-awareness, the ability to self-critique and reflect in a deep manner using such tools as a reflective diary are essential for nurturing the development of effective primary health care workers and consequently care structures for the patients and their families.

Cross-Cultural Comparison↗
PubMed2000

The analysis system COGITAT for the study of cognitive deficiencies in rodents.

COGITAT is an automated hole board system that, following minimal experimental interventions, makes it possible to measure a variety of parameters associated with learning, memory, relearning, cognition, and cognitive shifts, but also changes in exploratory and sensorimotor performance in rodent models. The individual parameters--that is, overall exploratory activity, number of visits (deep in the hole) into or inspections of (at the upper surface) holes, number of baited, unbaited, or previously baited holes visited or inspected, reinspections of or revisits into any holes, number of pellets eaten, time to find pellets, serial order collection (without intermediate inspections or visits), and reference and working memory errors (visits, inspections, or total)--are obtained simultaneously, and the results are immediately available after the end of each experiment. The system appears to be well suited to neurophysiological, neuropharmacological, and gene-technological investigations in rodent models.

Animals↗
PubMed1998

Versatility of a multilingual and bi-directional approach for medical language processing.

At the dawn of the 21st century, we are experiencing an exponential growth of online information that is mostly textual, and that benefits from new electronic media, such as the World Wide Web (WWW), to be broadly diffused across borders. However, there is a gap to bridge between holding information and accessing in a relevant way the deep underlying knowledge. Multilingual natural language processing (NLP), once tuned, is certainly the best solution to cope with this era of textual information. This paper focuses on the lesson learned through the joint development of an analyzer and a generator of medical language, within a multilingual context. Concrete examples, derived from the efforts under way in the European GALEN-IN-USE project, illustrate the use of these linguistic tools for the handling of surgical procedures.

Multilingualism↗
PubMed1980

Cingulate cortical and anterior thalamic neuronal correlates of reversal learning in rabbits.

Unit activity of the cingulate cortex and the anteroventral (AV) nucleus of the thalamus in rabbits was recorded during reversal training following differential conditioning of a locomotory (wheel rotation) avoidance response. The positive and negative conditional stimuli (CS+ and CS-) were pure tones (1 kHz and 8 kHz), and the unconditional stimulus was footshock (1.5 mA). During the first session, activity appropriate to the original task (i.e., a greater neuronal response to the CS- than to the CS+) occurred in the deep cortical laminae (V and VI), but the superficial laminae (I-IV) manifested the original discrimination in certain records and the reverse discrimination in others. All discriminative effects diminished during subsequent sessions except the short-latency (25-msec) deep laminar original discrimination, which persisted throughout reversal training. Activity of the AV nucleus underwent gradual transition, in parallel with the behavior, from the original to the reverse discrimination. Arousal-producing footshocks given after behavioral reversal educed the original and reverse discriminations in the deep and superficial laminae, respectively. These data are interpreted to indicate the role of cingulate cortex and anterior thalamus for encoding the associative significance of the conditional stimuli.

Animals↗
PubMed1997

Salient responsiveness of parabrachial neurons to the conditioned stimulus after the acquisition of taste aversion learning in rats.

The pontine parabrachial nucleus is considered to be one of the most critical regions for the acquisition of conditioned taste aversion which is an associative learning of taste and illness. To further clarify the possible involvement of the parabrachial nucleus in conditioned taste aversion, we recorded neuronal responses to taste stimuli from the parabrachial nucleus of rats under deep urethane anaesthesia. Animals were separated into two groups: the conditioned taste aversion group that had acquired a taste aversion to 0.1 M NaCl (conditioned stimulus) after paired presentations of the taste stimulus with intraperitoneal injection of LiCl (unconditioned stimulus), and the control group that had received only the unconditioned stimulus before experiments. Taste-responsive neurons in the conditioned taste aversion group showed larger responses to NaCl at below 0.1 M, but similar responses to 0.3 M and 0.5 M NaCl when compared with those in the control group. Furthermore, hierarchical cluster analyses revealed a strong similarity among responses to sodium salts in neurons of the conditioned taste aversion group compared with the control group. These results suggest that the aversive conditioning to NaCl modified parabrachial units so that the sodium taste was more salient than other tastes. This modification may reflect a long-term plastic change persisting without support of forebrain structures, and would facilitate the gustatory discrimination of the conditioned stimulus, which is required by conditioned animals.

Animals↗