PubMed HealthSearch

SEARCH · PubMed Health

Results for “Deep generative model”

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 127 records · Page 7Linked to original sources

[Microwave hyperthermia: influence of blood flow and thermoregulation processes (author's transl)].

Various problems are encountered during production of local, deep-seated microwave hyperthermia, involving the technology of the irradiating system, and the complexity of heat transfer in living tissues. Before starting investigations on patients, preliminary studies were conducted on different models. Taking the influence of the blood flow into account, the thermal effects of microwaves were simulated on a numerical model, and a perfused phantom. These studies were completed by investigations in animals. The analysis of findings demonstrates that, in given conditions of irradiation, the temperature distribution is strongly dependent on blood flow. This means that the phantom models are only useful to evaluate the influence of the irradiation parameters and to develop and compare the generator-applicator systems, and that accurate planning of therapeutic trials requires in vivo studies on animals as well as on patients.

Animals

Spatially-localized NMR spectroscopy employing an inhomogeneous surface-spoiling magnetic field gradient. 2. Surface coil experiments with multicompartment phantom and rat in vivo.

The use of inhomogeneous surface-spoiling magnetic field gradients for elimination of signal from surface lying regions of a sample was theoretically examined in the companion article (W. Chen and J.J.H. Ackerman, NMR Biomed. 3, 147-157 (1990)). Using the spoiling gradient coil design described therein, this article presents experimental verification of the feasibility of such an approach to enhanced spatial localization. Single coil mode 31P NMR surface coil interrogation of both a two compartment phantom and rat in vivo are shown to provide excellent suppression of surface lying regions with minimal degradation of signal from the deep lying region of interest. Both pulse-and-collect and spin echo sequences were highly efficient in concert with spoiling gradient periods of 0.5-2 ms and driving currents of 0.5-2 A. The use of a current-generated surface spoiling gradient offers a robust means to remove surface tissue signal contributions and can be implemented with a wide range of localizing pulse sequences and imaging protocols.

Animals

[Formation of a generator of excitation in the gigantocellular nucleus of the medulla oblongata during disruption of inhibitory processes].

Neuronal activity in the gigantocellular nucleus after injection of tetanus toxin was studied on decerebrated cats. The toxin was used as a substance producing a deep and continuous suppression of inhibitory processes. The increase in the amplitude and rate of neuronal discharges, in the integral background and evoked activity as well as in the number of active neurons and that of neurons with burst activity was recorded in the "poisoned" nucleus. The enhanced activity in the investigated regions of the poisoned nucleus might be temporarily suppressed by a strong direct electrical shock and by glycin administrations to those regions. The obtained data indicate that a pool of neurons with disturbed inhibitory processes forms a generator of enhanced excitation. The mechanisms and characteristic features of the activity of such generators are discussed. The possibility of modelling neurological syndromes by production of similar generators in various parts of the central nervous system and their relation to the earlier described phenomenon of "dispatch station" are considered.

Animals

Numerical phase algorithm for decompression computers and application.

Present generation decompression computers employ a simplified algorithm, limiting dissolved gas build-up in tissue and blood according to a method proposed by Haldane 80 years ago. Such a model works well for single dives, but is usually liberal and theoretically incomplete for multiple exposures within 24 hr spans. Using the critical phase hypothesis in a bubble model, we have extended the classical model of Haldane to multi-exposures. This model is discussed, and a decomputer algorithm described for multi-diving. The focus is permissible bubble excess, not just dissolved gas per se, with phase constraints affecting all tissues, fast and slow, and requiring a systematic lowering of repetitive tissue tensions. Deep repetitive and shallow multi-day exposures are impacted most by the procedure. Within nucleation theory deeper-than-first dives are also treated. A set of multi-diving fractions, xi, accounting for micronuclei excitation and regeneration, reduced bubble elimination in repetitive activity, and coupled effects on tissue tension, are proposed, with xi representing a set of multiplicative factors (less than one) applied to critical tissue tensions for multi-exposures. These factors affect repetitive activity over short time spans, deeper-than-previous and continuous multi-day activities, compared to standard computer software, and are easily encoded into existing decompression meters, potentially extending their range and flexibility over exposure regimes.

Algorithms

A prototype epithermal neutron beam for boron neutron capture therapy.

An epithermal neutron beam has been designed and tested at the Georgia Institute of Technology's 5-MW Research Reactor. The prototype facility consists of aluminum and sulfur disks in a tangential beam port for fast neutron filtration. A cadmium sheet at the port exit removes the thermal neutrons from the transmitted beam, leaving an intensely epithermal neutron beam spanning five energy decades, each contributing to the flux demanded by boron neutron capture therapy. The thermal neutron flux generated by the incident epithermal neutrons in a polyethylene head phantom peaks at a depth of 3 cm and remains above the incident thermal flux to a 7-cm depth. The beam thus provides the penetration required for treating deep-seated gliomas. Photon contamination in the prototype facility is high, and a number of basic modifications are proposed for reducing it to safer levels.

Boron

Three-dimensional reconstruction of microleakage pattern using a sequential grinding technique.

Dye penetration tests are very commonly used to detect the absence of a fluid seal at the tooth-restoration interface. Airlocks in the marginal gap, leaching of water-soluble tracers during processing, and the failure of only a few sections to allow interpretation of the full pattern, limit these tests to low reproducibility and precision. The purpose of this present study was to generate high-resolution three-dimensional images of waterfast tracer patterns. Cylindrical class V (3 mm diameter, 2 mm deep) dentine-bonded resin composite restorations in buccal coronal dentine were thermally cycled (1000 x, 8 degrees C, 55 degrees C, 30 s dwell at each temperature) and then silver stained using an initial vacuum (100 mmHg pressure). Each restoration was sequentially abraded from the free surface on wet 180 grit silicon carbide paper, producing up to 30 parallel surfaces at approximately 0.15 mm separation through the restoration down to the pulp. Images of the ground surfaces were captured, and assembled by a computer image analyser program to give a three-dimensional model of the tracer pattern. The maximum depths of tracer penetration below the reference surfaces were 3.00 mm, 2.09 mm, 3.16 mm and > 2.29 mm for the four specimens. Projections of the models were viewed from several directions with sections in various locations to allow investigation of the full tracer pattern. This method allows the creation of high-resolution three-dimensional tracer patterns.

Composite Resins

A molecular scheme for the reaction between gamma-aminobutyric acid and the most abundant chloride channel on crayfish deep extensor abdominal muscle.

Single-channel measurements were performed with the aim of constructing a detailed molecular scheme for the reaction between gamma-aminobutyric acid (GABA) and a chloride channel of crayfish deep extensor abdominal muscle (DEAM). GABA was applied in pulses to outside-out patches of muscle membrane, and, based on the dose-response of the peak currents and of their rise times, a linear model with five binding steps has been proposed. Evaluation of the single-channel kinetics indicated at least three open states. Two of them originate most probably from the fully liganded receptor state and are grouped in mixed bursts due to their different life times. The third one appears independently, outside the bursts, and originates from a lower liganded receptor state. Simulations of the dose-responses and the open time distributions with this model led to a set of rate constants which generated relatively optimal fits.

Abdominal Muscles

Diagnostic performance of machine learning models versus established risk stratification for intracranial aneurysm rupture: a systematic review and bivariate meta-analysis.

BACKGROUND: Machine learning (ML) models have been proposed to improve the discrimination of intracranial aneurysm rupture status beyond established clinical risk stratification tools. However, reported performance is heterogeneous and the relative contribution of model architecture and feature dominance remains unclear. METHODS: We performed a Preferred Reporting Items for Systematic Reviews and Meta-Analyses-diagnostic test accuracy systematic review and diagnostic meta-analysis of studies evaluating ML models for intracranial aneurysm rupture discrimination. PubMed, Embase and CENTRAL were searched to February 2026. Sensitivity and specificity were pooled using a bivariate random-effects model, with summary receiver operating characteristic curves generated across training, internal testing and external validation datasets. Models were compared with regression-based approaches and Population, Hypertension, Age, Size of aneurysm, Earlier subarachnoid haemorrhage, Site of aneurysm (PHASES) scores. Subgroup and meta-regression analyses explored associations between algorithm family and feature domain. RESULTS: Sixty-two retrospective cohorts (29 709 patients 209 models) met the inclusion criteria. In training datasets, pooled sensitivity and specificity for ML were 0.81 (95% CI 0.75 to 0.85) and 0.83 (0.80-0.86), with an area under the curve (AUC) of 0.878, exceeding PHASES (AUC 0.667). In testing datasets, ML retained higher discrimination (AUC 0.837) than regression models (0.806) and PHASES (0.646). In external validation, sensitivity was preserved (0.82), but specificity declined (0.66). Deep learning demonstrated the highest AUCs (training and testing). Incorporation of haemodynamic or radiomic features improved pooled discrimination relative to morphology alone. Evidence of small-study effects and mostly unclear Prediction Model Risk Of Bias Assessment Tool ratings were observed. CONCLUSIONS: ML approaches demonstrate higher pooled discrimination for aneurysm rupture status than conventional risk scores in retrospective datasets, but reduced external validation specificity and heterogeneity limit confidence for clinical translation. Prospective, externally validated, calibrated models are required before integration into routine cerebrovascular risk stratification.

Humans

[OMRON RF hyperthermia treatment system HEH-500 C].

The RF capacitive type hyperthermia system "HEH-500 C" enables regional heating for superficial and deep seated tumors. It consists of a high frequency generator (frequency; 13.56 MHz, out put power; 500 watts), a control unit, a cooling unit for the applicators, an I/F unit (communicate the thermometer to RF generator and plotter printer), a thermometer (thermocouple thermometer or fiber fluorothermometer) and a plotter printer. Heating profile of HEH-500 C was presented with thermographic picture on TX-150 muscle equivalent phantom. Effects of bone and fat layer on heating profile of phantom were also examined. Points to be solved on thermometry, electromagnetic field environment and clinical use were also discussed.

Humans

Head phantom experiment and calculation for boron neutron capture therapy.

Head phantom experiments with various neutron beams and calculations were carried out in order to provide useful information for boron neutron capture therapy (BNCT). Thermal neutron beams for thermal neutron capture therapy were used for phantom experiments with various neutron collimator aperture sizes. The filtered beam neutrons of 24 and 144 keV generated with iron and silicon filters were also used to investigate the possible application of BNCT in the treatment of deep-seated cancers. Thermal neutron fluence and induced capture gamma dose distributions within the phantom were calculated with a transport code DOT 3.5 and compared with the experimental results. The results showed that the calculation used was consistent with the experimental results and provided useful information on BNCT. The filtered beam neutron may be very useful for the treatment of deep or widespread cancer, if there were a high power research reactor constructed for this purpose.

Brain Neoplasms

Human air space shapes, sizes, areas, and volumes.

The geometry of an enlarged reconstructed human acinus (i.e., a terminal bronchiole and distal airways and air spaces) was studied. Alveoli were categorized in six shapes: three-fourths of a spheroid, a slightly truncated cone, one-fourth of a spheroid, a cylindroid with a hemispherical bottom, a shallow cylindroid with a flat bottom, and a truncated deep ellipsoid. Sacs were usually either hemispheroids or shallow truncated cones. Ducts of eight generations were spheroid and gradually decreased in diameter (D) and length (L) as the generation number (z) increased. Considering the terminal bronchiole as the 15th generation and using Weibel's data for the first 10 generations, the dimensions, in mm, for z of 1-10 and 10-26 were reasonably described by D-z = 12e-(0.27-0.005z)z and L-z = 25e-0.187z. The predicted volume of the acinus at three-fourths total lung volume was 182.8 mm3, a volume equivalent to that of a sphere 7.04 mm in diameter. The reconstruction demonstrated a great increase in respiratory bronchiolar and ductal cross-sectional area and alveolar surface area, considerably more rapid than predicted by Weibel's model A.

Anthropometry

Modeling Alternative Conformational States in CASP16.

The CASP16 Ensemble Prediction experiment assessed advances in methods for modeling proteins, nucleic acids, and their complexes in multiple conformational states. Targets included systems with experimental structures determined in two or three states, evaluated by direct comparison to experimental coordinates, as well as domain-linker-domain (D-L-D) targets assessed against statistical models from NMR and SAXS data. This paper focuses on the former class of multi-state targets. Ten ensembles were released as community challenges, including ligand-induced conformational changes, protein-DNA complexes, a trimeric protein, a stem-loop RNA, and multiple oligomeric states of a single RNA. For five targets, some groups produced reasonably accurate models of both reference states (best TM-score >0.75). However, with the exception of one protein-ligand complex (T1214), where an apo structure was available as a template, predictors generally failed to capture key structural details distinguishing the states. Overall, accuracy was significantly lower than for single-state targets in other CASP experiments. The most successful approaches generated multiple AlphaFold2 models using enhanced multiple sequence alignments and sampling protocols, followed by model quality based selection. While the AlphaFold3 server performed well on several targets, individual groups outperformed it in specific cases. By contrast, predictions for one protein-DNA complex, three RNA targets, and multiple oligomeric RNA states consistently fell short (TM-score <0.75). These results highlight both progress and persistent challenges in multi-state prediction. Despite recent advances, accurate modeling of conformational ensembles, particularly RNA and large multimeric assemblies, remains a critical frontier for structural biology.

AlphaFold2

Pulse duration and peak intensity during focused ultrasound surgery: theoretical and experimental effects in rabbit brain in vivo.

The goal of this study was to establish the exposure parameters that will generate predictable thermally induced lesions in brain. In addition, the accuracy of a theoretical model for prediction of the lesion size was tested. To do this, 160 adult rabbits were sonicated (frequency 0.936 and 1.72 MHz) and then sacrificed at various intervals after the sonications. The results showed that predictable thermal lesions could be induced if the exposure durations were between 0.5 and 2 s. Dimensions of the necrosed tissue volume were roughly predictable by the theoretical calculations based on purely thermal effects. Shorter sonications required higher intensities (above 3700 W cm-2 at 1.72 MHz) resulting in mechanical effects with extensive vascular damage. Lesion size varied more at longer exposures (5 and 10 s), perhaps due to the increased effect of tissue perfusion. As a conclusion, focused ultrasound can be used for destruction of tissues deep in brain without causing undesirable mechanical effects, if the exposure parameters are selected properly.

Animals

Ultrastructural pathology of experimental autoimmune uveitis. Quantitative evidence of activation and possible high endothelial venule-like changes in retinal vascular endothelium.

BACKGROUND: Experimental autoimmune uveitis (EAU) is a highly organ-specific autoimmune disease in which the target is the retinal photoreceptors. It is well recognized as a model of uveoretinitis in humans. The mechanisms that control the homing of sensitized lymphocytes and other leukocytes to the retina is unknown. The aim of this study was to investigate changes in the retinal vasculature that may be involved with aiding leukocyte-endothelial cell interactions and subsequent extravasation of leukocytes into the retina. EXPERIMENTAL DESIGN: Lewis rats immunized with S-antigen were used to produce EAU. The retinal vasculature was assessed by morphologic (light and electron microscopy) and morphometric techniques at various stages in the generation and course of the disease (days 3, 7, 11, 14, 21, 28 and 49 postimmunization) for evidence of endothelial cell (EC) activation and leukocyte-EC interaction. Image analysis of the retinal vessels at the electron microscopic level was performed to detect alterations in the thickness and irregularity of the EC surface, both considered to be important in lymphocyte homing in the high endothelial venules (HEVs) of lymphoid tissues. Control values were obtained from normal eyes, pertussis-only treated animals, and normal lymph node HEVs. RESULTS: The clinical and histopathologic changes in the eyes were consistent with previous descriptions of EAU and included perivasculitis, focal mononuclear infiltrate in the outer retina, and choroid with destruction of the photoreceptor outer segments and eventually loss of large portions of the outer retina. During the course of EAU, a significant proportion of retinal venules underwent both qualitative and quantitative morphologic changes including EC activation evident as increased cytoplasmic organelles, a 230% average increase in mean EC thickness, and a concomitant 4-fold increase in irregularity of the EC, that produced plump irregular EC with deep intercellular clefts. These alterations were maximal at day 21, however from day 11 onward, large numbers of lymphocytes and monocytes were observed adhering to or lodged in the clefts of plump EC, migrating through the EC cytoplasm, or lying beneath the EC. CONCLUSIONS: The characteristics acquired by the retinal venules during EAU are reminiscent of HEVs. This study suggests that tissue-specific changes in the endothelial cells of retinal venules may be responsible for the homing of S-antigen specific autoreactive lymphocytes to the target organ in this model of retinal autoimmunity.

Animals

[Thyroid gland function and intrapulmonary temperature in tuberculosis].

The experiment on 140 albino mice examined the relationship between the pulmonary thermogenesis and the functional activity of the hypophyseal-thyroid system at different stages of pulmonary tuberculous inflammation development, during Staphylococcus-induced pneumonia, aseptic inflammation in lung tissue. Deep abnormalities of the heat-generating function of the lung were revealed just at the early periods of specific inflammation. The degree of hypothermal reactions of lung tissue correlated with the inhibition of hypophyseal-thyroidal function at all developmental stages of a tuberculous process. The changes were rather pronounced, stable and phasic. Spontaneous regression of the tuberculous process was not accompanied by recovery in the activity of the hypophyseal-thyroidal system, despite the fact that there was a clear-cut trend to normalization of pulmonary thermogenesis.

Animals

A common protein fold and similar active site in two distinct families of beta-glycanases.

The structure of Clostridium thermocellum endoglucanase CelC, a member of the largest cellulase family (family A), has been determined at 2.15 A resolution. The protein folds into an (alpha/beta)8 barrel, with a deep active-site cleft generated by the insertion of a helical subdomain. The structure of the catalytic core of xylanase XynZ, which belongs to xylanase family F, has been determined at 1.4 A resolution. In spite of significant differences in substrate specificity and structure (including the absence of the helical subdomain), the general polypeptide folding pattern, architecture of the active site and catalytic mechanism of XynZ and CelC are similar, suggesting a common evolutionary origin.

Amino Acid Sequence

The crystal structure of d(G-G-G-G-C-C-C-C). A model for poly(dG).poly(dC).

The structure of the DNA oligomer d(G-G-G-G-C-C-C-C) has been determined at a resolution of 2.5 A by single-crystal X-ray methods. There are two strands in the asymmetric unit, and these coil about each other to form a right-handed double-helix of the A-type with Watson-Crick hydrogen bonds between base-pairs. The helix has a shallow minor groove and a deep, water-filled major groove; almost all exposed functional groups on the DNA are hydrated, and 106 ordered solvent molecules have been found. The two d(G-G-G-G).d(C-C-C-C) segments in the octamer exhibit similar and uniform structures, but there is a slight discontinuity at the GpC step between them. A recurring feature of the structure is the overlap of adjacent guanine bases in each GpG step, with the five-membered ring of one guanine stacking on the six-membered ring of its neighbour. There is little or no overlap between adjacent cytosine rings. Conformational parameters for these GpG steps are compared with those from other single-crystal X-ray analyses. In general, GpG steps exhibit high slide, low roll and variable twist. Models for poly(dG).poly(dC) were generated by applying a simple rotation and translation to each of the unmodified d(G-G-G-G).d(C-C-C-C) units. Detailed features of these models are shown to be compatible with various assays of poly(dG).poly(dC) in solution, and are useful in understanding the polymorphic behaviour of this sequence under a variety of experimental conditions.

Chemical Phenomena

Upcycling Vegetable Waste Into Functional Food Ingredients via Synergistic Microbial Engineering and Artificial Intelligence.

The escalating generation of global vegetable waste represents a critical loss of bioactive resources, necessitating a paradigm shift from passive disposal to active nutrient upcycling. However, the industrial conversion of this heterogeneous biomass into standardized functional food ingredients is currently impeded by significant techno-economic barriers, primarily structural recalcitrance, compositional inconsistency, and the presence of toxic fermentation inhibitors. This review provides a comprehensive analysis of the synergistic application of microbial engineering and artificial intelligence (AI) to resolve these bioprocessing bottlenecks within a food-to-food closed-loop framework (as shown in the graphical abstract). We evaluate recent advances in engineering food-grade microbial chassis (e.g., Saccharomyces cerevisiae and Escherichia coli) to enhance lignocellulose degradation and stress tolerance. Concurrently, we examine the integration of AI across the entire value chain, covering deep learning-based rational enzyme design, genome-scale metabolic modeling, and intelligent process control for precision fermentation. Current evidence demonstrates that the hardware-software coupling of engineered strains and AI algorithms significantly enhances conversion efficiency and process robustness. Key findings highlight that AI-driven Design-Build-Test-Learn cycles facilitate the de novo creation of enzymes with superior kinetics and strains with adaptive stress response capabilities against toxins. Moreover, dynamic digital twin models effectively mitigate the impact of substrate variability, ensuring the batch-to-batch consistency required for food applications. We conclude that this data-driven synergistic paradigm is pivotal for establishing a resilient circular bioeconomy, enabling the reliable bioconversion of waste into high-value single-cell proteins, natural flavor additives, and sustainable packaging materials.

Artificial Intelligence