Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Deep Neural Network”

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 91 records · Page 5Linked to original sources
PubMed2026

Accurately Deciphering Tissue Heterogeneity From Spatial Multi-Modal and Multi-Omics With STransformer.

Advances in spatially resolved technologies enable the simultaneous acquisition of diverse data modalities within a tissue slice while preserving critical spatial context, which presents unprecedented opportunities to decipher intricate tissue heterogeneity. However, existing computational approaches lack the intrinsic flexibility to universally process both spatial multi-modal and multi-omics data. Here, we introduce STransformer, a unified deep learning framework designed to seamlessly accommodate a comprehensive landscape of spatial data. By simultaneously capturing short-range cellular interactions and tissue-wide semantic patterns, it extracts robust representations to accurately dissect complex tissue heterogeneity. Systematic evaluations across diverse species, tissue types, and data modalities highlight its profound versatility. For spatial multi-modal data, STransformer delineates intricate anatomical structures in the human cortex, uncovers pathological mechanisms in Alzheimer's disease, and characterizes dynamic spatiotemporal developmental trajectories during chicken cardiogenesis. Scaling to spatial multi-omics data, STransformer synergizes spatial transcriptomic and proteomic profiles to decipher intricate immune microenvironments within the human tonsil, and jointly analyzes spatial epigenomic and transcriptomic data to infer regulatory mechanisms in the mouse embryonic brain. Consequently, STransformer serves as a highly versatile and robust analytical framework for advancing our understanding of tissue heterogeneity and disease pathogenesis.

Multiomics↗
PubMed2025

A neural network model enables worm tracking in challenging conditions and increases signal-to-noise ratio in phenotypic screens.

High-resolution posture tracking of C. elegans has applications in genetics, neuroscience, and drug screening. While classic methods can reliably track isolated worms on uniform backgrounds, they fail when worms overlap, coil, or move in complex environments. Model-based tracking and deep learning approaches have addressed these issues to an extent, but there is still significant room for improvement in tracking crawling worms. Here we train a version of the DeepTangle algorithm developed for swimming worms using a combination of data derived from Tierpsy tracker and hand-annotated data for more difficult cases. DeepTangleCrawl (DTC) outperforms existing methods, reducing failure rates and producing more continuous, gap-free worm trajectories that are less likely to be interrupted by collisions between worms or self-intersecting postures (coils). We show that DTC enables the analysis of previously inaccessible behaviours and increases the signal-to-noise ratio in phenotypic screens, even for data that was specifically collected to be compatible with legacy trackers including low worm density and thin bacterial lawns. DTC broadens the applicability of high-throughput worm imaging to more complex behaviours that involve worm-worm interactions and more naturalistic environments including thicker bacterial lawns.

Caenorhabditis elegans↗
PubMed2002

A model of prefrontal cortex dopaminergic modulation during the delayed alternation task.

Working memory performance is modulated by the level of dopamine (DA) D1 receptors stimulation in the prefrontal cortex (PFC). This modulation is exerted at different time scales. Injection of D1 agonists/antagonists exerts a long-lasting influence (several minutes or hours) on PFC pyramidal neurons. In contrast, during performance of a cognitive task, the duration of the postsynaptic effect of phasic DA release is short lasting. The functional relationships of these two time scales of DA modulation remain poorly understood. Here we propose a model that combines these two time scales of DA modulation on a prefrontal neural network. The model links the cellular and behavioral levels during performance of the delayed alternation task. The network, which represents the activity of deep-layer pyramidal neurons with intrinsic neuronal properties, exhibits two stable states of activity that can be switched on and off by excitatory inputs from long-distance cortical areas arriving in superficial layers. These stable states allow PFC neurons to maintain representations during the delay period. The role of an increase of DA receptors stimulation is to restrict inputs arriving on the prefrontal network. The model explains how the level of working memory performance follows an inverted U-shape with an increased stimulation of DA D1 receptors. The model predicts that (1) D1 receptor agonists increase perseverations, (2) D1 antagonists increase distractability, and (3) the duration of the postsynaptic effect of phasic DA release in the PFC is adjusted to the delay period of the task. These results show how the precise duration of the postsynaptic effect of phasic DA release influences behavioral performance during a simple cognitive task.

Dopamine↗
PubMed1995

A system of IAC neural networks as the basis for self-organization in a sociological dynamical system simulation.

This sociological simulation uses the ideas of semiotics and symbolic interactionism to demonstrate how an appropriately developed associative memory in the minds of individuals on the microlevel can self-organize into macrolevel dissipative structures of societies such as racial cultural/economic classes, status symbols and fads. The associative memory used is based on an extension of the IAC neural network (the Interactive Activation and Competition network). Several IAC networks act together to form a society by virtue of their human-like properties of intuition and creativity. These properties give them the ability to create and understand signs, which lead to the macrolevel structures of society. This system is implemented in hierarchical object oriented container classes which facilitate change in deep structure. Graphs of general trends and an historical account of a simulation run of this dynamical system are presented.

Black or African American↗
PubMed1998

Principal component ANN for modelling and control of baker's yeast production.

Modelling of baker's yeast production by the principal component based artificial neural networks (ANN) is presented. The models are derived for their application in adaptive control of fermentation by the internal model control (IMC) method. Modelling data are from industrial production in 40 m3 deep jet bioreactor and from computer simulations. The modelling effort is focused on selection of ANN structure and model verification. Principal component analysis of process variables results in projection of patterns to a space of low dimension, which enables determination of ANN structure, removes data colinearity and random components of measurement signals, and model degradation by over-training is eliminated. In view of IMC application, the models for prediction of the controlled variable (ethanol partial pressure) and the inverse model for manipulative variable (molasses feed rate) are determined. The models are tested for their predictability in the time horizon from 1 to 20 min. ANN models are derived with average relative errors for untrained patterns in the range from 1 to 10%.

Bioreactors↗
PubMed2005

Divergent evolution within protein superfolds inferred from profile-based phylogenetics.

Many dissimilar protein sequences fold into similar structures. A central and persistent challenge facing protein structural analysis is the discrimination between homology and convergence for structurally similar domains that lack significant sequence similarity. Classic examples are the OB-fold and SH3 domains, both small, modular beta-barrel protein superfolds. The similarities among these domains have variously been attributed to common descent or to convergent evolution. Using a sequence profile-based phylogenetic technique, we analyzed all structurally characterized OB-fold, SH3, and PDZ domains with less than 40% mutual sequence identity. An all-against-all, profile-versus-profile analysis of these domains revealed many previously undetectable significant interrelationships. The matrices of scores were used to infer phylogenies based on our derivation of the relationships between sequence similarity E-values and evolutionary distances. The resulting clades of domains correlate remarkably well with biological function, as opposed to structural similarity, indicating that the functionally distinct sub-families within these superfolds are homologous. This method extends phylogenetics into the challenging "twilight zone" of sequence similarity, providing the first objective resolution of deep evolutionary relationships among distant protein families.

Evolution, Molecular↗
PubMed1999

Towards the networks of the brain: from brain imaging to consciousness.

The manner in which the brain computes in various tasks is being probed at a deep level by modern brain imaging techniques, with an increasing appreciation of the different networks being used to solve these tasks. There is simultaneously developing a neural modelling technology, which attempts to explain the underlying computations being performed by this set of networks. This paper describes results from brain imaging and how they may be related to the underlying neural networks by means of structural modelling. It thereby attempts to give an initial glimpse of the emerging picture of the functionality of brain networks. It concludes with a discussion of the role of consciousness in global processing, and how particular styles of neural processing can attain this.

Journal Article↗
PubMed2025

Efficient Detection and Characterization of Targets of Natural Selection Using Transfer Learning.

Natural selection leaves detectable patterns of altered spatial diversity within genomes, and identifying affected regions is crucial for understanding species evolution. Recently, machine learning approaches applied to raw population genomic data have been developed to uncover these adaptive signatures. Convolutional neural networks (CNNs) are particularly effective for this task, as they handle large data arrays while maintaining element correlations. However, shallow CNNs may miss complex patterns due to their limited capacity, while deep CNNs can capture these patterns but require extensive data and computational power. Transfer learning addresses these challenges by utilizing a deep CNN pretrained on a large dataset as a feature extraction tool for downstream classification and evolutionary parameter prediction. This approach reduces extensive training data generation requirements and computational needs while maintaining high performance. In this study, we developed TrIdent, a tool that uses transfer learning to enhance detection of adaptive genomic regions from image representations of multilocus variation. We evaluated TrIdent across various genetic, demographic, and adaptive settings, in addition to unphased data and other confounding factors. TrIdent demonstrated improved detection of adaptive regions compared to recent methods using similar data representations. We further explored model interpretability through class activation maps and adapted TrIdent to infer selection parameters for identified adaptive candidates. Using whole-genome haplotype data from European and African populations, TrIdent effectively recapitulated known sweep candidates and identified novel cancer, and other disease-associated genes as potential sweeps.

Selection, Genetic↗
PubMed2000

Interrelated modification of excitatory and inhibitory synapses in three-layer olivary-cerebellar neural network.

The model of three-layer olivary-cerebellar neural network with modifiable excitatory and inhibitory connections between diverse elements is suggested. The same Hebbian modification rules are proposed for Purkinje cells, granule (input) cells, and deep cerebellar nuclei (output) cells. The inverse calcium-dependent modification rules for these cells and hippocampal/neocortical neurones or Golgi cells are conceivably the result of the involvement of cGMP and cAMP in postsynaptic processes. The sign of simultaneous modification of excitatory and inhibitory inputs to a cell is opposite and determined by the variations in pre- and/or postsynaptic cell activity. Modification of excitatory transmission between parallel fibers and Purkinje cells, mossy fibers and granule cells, and mossy fibers and deep cerebellar nuclei cells essentially depends on inhibition effected by stellate/basket cells, Golgi cells and Purkinje cells, respectively. The character of interrelated modifications of diverse synapses in all three layers of the network is influenced by olivary cell activity. In the absence (presence) of a signal from inferior olive, the long-term potentiation (depression) in the efficacy of a synapse between input mossy fiber and output cell can be induced. The results of the suggested model are in accordance with known experimental data.

Animals↗
PubMed1992

Projections from cortical visual areas of the superior temporal sulcus to the superior colliculus, in macaque monkeys.

1. The distribution of tectal projections of two visual areas of the superior temporal sulcus (MT and MST areas) has been studied, in five Macaca fascicularis, by means of the autoradiographic method tracing the anterograde transport of tritiated aminoacids intracortically injected. 2. In all cases the ipsilateral superior colliculi (SC) were found labelled, whereas the contralateral ones were devoid of label. 3. The three brains injected in the MT area resulted in SC labels that involved the superficial gray layer (SGS), the stratum opticum (SO) and the intermediate gray layer (SGI), sparing the layers below SGI. 4. The collicular labels found after injections within the MST area exhibited their distribution over the deep SC subdivision, whereas they spared all the superficial layers but the deep part of the SO. 5. In two animals with large uptake zones, one in MT and the other in MST, the labelling within the SGI showed a cluster-like pattern. 6. The distinct found bulk of projections of MT and MST respectively to the superficial and deep subdivisions of the SC, along with a number of peculiar connections of the MST area as mentioned in the text, contribute to depict an overall neural network in which MST appears to be more strongly involved than MT in linking sensory visual with oculomotor attentive functions.

Afferent Pathways↗
PubMed1995

A functional neuroanatomy of hallucinations in schizophrenia.

Hallucinations, perceptions in the absence of external stimuli, are prominent among the core symptoms of schizophrenia. The neural correlates of these brief, involuntary experiences are not well understood, and have not been imaged selectively. We have used new positron emission tomography (PET) methods to study the brain state associated with the occurrence of hallucinations in six schizophrenic patients. Here we present a group study of five patients with classic auditory verbal hallucinations despite medication, demonstrating activations in subcortical nuclei (thalamic, striatal), limbic structures (especially hippocampus), and paralimbic regions (parahippocampal and cingulate gyri, as well as orbitofrontal cortex). We also present a case study of a unique, drug-naive patient with visual as well as auditory verbal hallucinations, demonstrating activations in visual and auditory/linguistic association cortices as part of a distributed cortical-subcortical network. Activity in deep brain structures, identified with group analysis, may generate or modulate hallucinations, and the particular neocortical regions entrained in individual patients may affect their specific perceptual content. The interaction of these distributed neural systems provides a biological basis for the bizarre reports of schizophrenic patients.

Adult↗
PubMed2006

Deep brain stimulation of the substantia nigra pars reticulata exerts long lasting suppression of amygdala-kindled seizures.

Deep brain stimulation (DBS) has been used to treat a variety of neurological disorders including epilepsy. However, we have limited knowledge about effective target areas, optimal stimulation parameters, and long-term effect of DBS on epileptic seizures. Here we examined the effects of DBS of the substantia nigra pars reticulata (SNr) on amygdala-kindled seizures. Microwire electrodes were implanted into the SNr and amygdala of adult male rats. When stage 5-kindled seizures were achieved by daily amygdala kindling, high frequency stimulation was delivered to the SNr bilaterally 1 s after cessation of kindling. Our DBS protocol completely blocked kindled seizures in 10 out of 23 (43.5%) rats studied. Furthermore, when the same amygdala kindling procedure was performed 24 h later without DBS, the kindling failed to elicit any seizure signs in 6 of these 10 rats. Some of the post-DBS period of seizure suppression lasted for up to 4 days. In other 3 rats, only mild stage 1 to 2 seizures appeared following amygdala kindling. Only 1 of the 10 rats for which DBS had blocked kindled seizures exhibited full-scale 5 stage-kindled seizures 24 h after DBS. These results suggest that highly plastic neural networks are involved in amygdala-kindled seizures and that DBS, if well timed with the onset of amygdala kindling, may exert long lasting effects on the networks that may prevent the recurrence of kindled seizures.

Amygdala↗
PubMed2005

Dynamic properties of network motifs contribute to biological network organization.

Biological networks, such as those describing gene regulation, signal transduction, and neural synapses, are representations of large-scale dynamic systems. Discovery of organizing principles of biological networks can be enhanced by embracing the notion that there is a deep interplay between network structure and system dynamics. Recently, many structural characteristics of these non-random networks have been identified, but dynamical implications of the features have not been explored comprehensively. We demonstrate by exhaustive computational analysis that a dynamical property--stability or robustness to small perturbations--is highly correlated with the relative abundance of small subnetworks (network motifs) in several previously determined biological networks. We propose that robust dynamical stability is an influential property that can determine the non-random structure of biological networks.

Animals↗
PubMed2003

Age-related changes in the neural correlates of motor performance.

Age-related neurodegenerative and neurochemical changes are thought to underlie decline in motor and cognitive functions, but compensatory processes in cortical and subcortical function may allow maintenance of performance level in some people. Our objective was to investigate age-related changes in the motor system of the human brain using functional MRI. Twenty six right handed volunteers were scanned whilst performing an isometric, dynamic, visually paced hand grip task, using dominant (right) and non-dominant (left) hands in separate sessions. Hand grip with visual feedback activated a network of cortical and subcortical regions known to be involved in the generation of simple motor acts. In addition, activation was seen in a putative human 'grasping circuit', involving rostral ventral premotor cortex (Brodmann area 44) and intraparietal sulcus. Within this network, a number of regions were more likely to be activated the older the subject. In particular, age-related changes in task- specific activations were demonstrated in left deep anterior central sulcus when using the dominant or non-dominant hand. Additional age-related increases were seen in caudal dorsal premotor cortex, caudal cingulate sulcus, intraparietal sulcus, insula, frontal operculum and cerebellar vermis. We have demonstrated a clear age-related effect in the neural correlates of motor performance, and furthermore suggest that these changes are non-linear. These results support the notion that an adaptable and plastic motor network is able to respond to age-related degenerative changes in order to maintain performance levels.

Adult↗
PubMed2025

Efficient detection and characterization of targets of natural selection using transfer learning.

Natural selection leaves detectable patterns of altered spatial diversity within genomes, and identifying affected regions is crucial for understanding species evolution. Recently, machine learning approaches applied to raw population genomic data have been developed to uncover these adaptive signatures. Convolutional neural networks (CNNs) are particularly effective for this task, as they handle large data arrays while maintaining element correlations. However, shallow CNNs may miss complex patterns due to their limited capacity, while deep CNNs can capture these patterns but require extensive data and computational power. Transfer learning addresses these challenges by utilizing a deep CNN pre-trained on a large dataset as a feature extraction tool for downstream classification and evolutionary parameter prediction. This approach reduces extensive training data generation requirements and computational needs while maintaining high performance. In this study, we developed TrIdent, a tool that uses transfer learning to enhance detection of adaptive genomic regions from image representations of multilocus variation. We evaluated TrIdent across various genetic, demographic, and adaptive settings, in addition to unphased data and other confounding factors. TrIdent demonstrated improved detection of adaptive regions compared to recent methods using similar data representations. We further explored model interpretability through class activation maps and adapted TrIdent to infer selection parameters for identified adaptive candidates. Using whole-genome haplotype data from European and African populations, TrIdent effectively recapitulated known sweep candidates and identified novel cancer, and other disease-associated genes as potential sweeps.

Journal Article↗
PubMed1999

Control systems of gastrointestinal motility are immature at birth in dogs.

Networks of interstitial cells of Cajal (ICC) in the myenteric plexus (Myp) or circular muscle (CM) function as pacemakers for gastrointestinal slow waves. ICC in contact with muscle and closely associated with nerves in the CM may mediate inhibitory neurotransmission. We wondered if ICC in Myp and CM and their connections are immature at birth and mature first in the proximal gut in association with nerves. Tissues from lower esophageal sphincter (LES), pylorus (PYL), small intestine (SI) and colon (CO) of 18 term fetal dogs taken from six females were fixed and prepared for ultrastructural examination and studied. Ganglia were present where expected in the Myp and submucous plexus (SMP). ICC cells were present in the Myp of PYL, SI and CO and appeared to have normal relationships to the outer border of CM as in adults. ICC in CM were found associated with nerves in the LES and in PYL, but not in SI or CO. However, axons in CM were everywhere usually free of glial covering, indicating ongoing migration or development. No organized deep muscular plexus (DMP) in SI or submuscular plexus (SP) in colon was present. Visible gap junctions were absent everywhere except for very rare ones between circular muscle cells. We conclude that at birth the neural and ICC networks of CM are more immature in intestine and colon than in oesophagus and stomach. Development of nerve and ICC of CM in oesophagus and stomach apparently precedes that in the remaining gut. However networks in these regions have not achieved adult organization and ICC and smooth muscle cells are anatomically poorly coupled. These findings suggest the reasons that gut motility at birth will not be adult in pattern are because ICC, nerve and muscle control systems are not fully differentiated. Further developmental delays in ICC and nerve maturation could have serious consequences for feeding of infant animals.

Animals↗
PubMed2025

Hepatocyte proteome destabilization and novel targets for PFASs unveiled through combined thermal proteome profiling and deep transfer learning.

Identifying protein targets for per- and polyfluoroalkyl substances (PFASs) is essential to understand their toxicity and health risks. However, knowledge about their interacting proteins is limited since reliable identification methods are lacking. We developed an integrated approach combining thermal proteome profiling (TPP) and deep transfer learning (DTL) modeling to efficiently identify cellular targets of PFAS. TPP measured PFAS binding proteins and the affinities by nanospray liquid chromatography tandem mass spectrometry, while DTL models were constructed to predict PFAS-protein affinities using neural network algorithms. TPP results revealed that PFASs uniquely destabilized the proteome of HepG2 cells, unlike the stabilizing effects by other xenobiotics. Key protein targets for three representative PFASs (PFOA, GenX and Novec 649) were identified, which exhibited weak binding affinities (median EC50 ≈ 30 μM). The number of protein targets increased with molecular weights among the three PFASs. The DTL model achieved a higher Pearson correlation coefficient of 0.89, and reduced mean squared errors by 54 % over previous models for drug-protein interactions. Notably, TPP and DTL jointly pinpointed ribosomal proteins as novel targets of GenX, potentially linking it to cell apoptosis through disrupted protein synthesis. Biolayer interferometry validated GenX binding to RPL4 protein, driven by electrostatic interactions and halogen bonds. This integrated approach effectively uncovers novel PFASs targets, advancing insights into their adverse health effects.

Humans↗
PubMed2026

Artificial intelligence in treatment prediction for skeletal Class III malocclusion: A systematic review.

In skeletal Class III patients, treatment options range from orthodontics to orthognathic surgery. Choosing the optimal approach requires a comprehensive clinical evaluation, which may be supported by AI tools. The aim of this study was to assess the performance of AI models in predicting the need for orthognathic surgery and in identifying predictors influencing treatment decisions. A PRISMA-guided electronic database search (PubMed, Web of Science; 2009-2024; English/French) was performed to identify studies using machine learning (ML) or deep learning (DL) on cephalometric and clinical data. After screening and assessment for eligibility, 15 studies were critically appraised. Model performance was summarized using accuracy, sensitivity, specificity, and the area under the curve (AUC). ML algorithms (particularly Random Forest and XGBoost) and DL models (ResNet-based convolutional neural networks (CNNs)) achieved high accuracy for predicting surgical need. Frequently selected predictors included Wits appraisal, ANB angle, the maxillomandibular ratio (Mx/Md), overjet, and the divergence of the lower gonial angle. AI methods show promise for assisting treatment decisions in Class III malocclusion, with Random Forest and XGBoost performing well on tabular cephalometric data and CNNs on imaging. Larger, multicentre datasets and external validation are needed to improve reliability, address bias, and support clinical implementation.

Humans↗