Spectral pruning of fully connected layers.
Ontology highlight
ABSTRACT:
SUBMITTER: Buffoni L
PROVIDER: S-EPMC9249877 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:Medical imaging data with thousands of spatially correlated data points are common in many fields. Methods that account for spatial correlation often require cumbersome matrix evaluations which are prohibitive for data of this size, and thus current work has either used low-rank approximations or analyzed data in blocks. We propose a method that accounts for nonstationarity, functional connectivity of distant regions of interest, and local signals, and can be applied to large multi-subject datasets using spectral methods combined with Markov Chain Monte Carlo sampling. We illustrate using simulated data that properly accounting for spatial dependence improves precision of estimates and yields valid statistical inference. We apply the new approach to study associations between cortical thickness and Alzheimer's disease, and find several regions of the cortex where patients with Alzheimer's disease are thinner on average than healthy controls.
| S-EPMC5874158 | biostudies-literature
Project description:Many diseases cause significant changes to the concentrations of small molecules (a.k.a. metabolites) that appear in a person's biofluids, which means such diseases can often be readily detected from a person's "metabolic profile"-i.e., the list of concentrations of those metabolites. This information can be extracted from a biofluids Nuclear Magnetic Resonance (NMR) spectrum. However, due to its complexity, NMR spectral profiling has remained manual, resulting in slow, expensive and error-prone procedures that have hindered clinical and industrial adoption of metabolomics via NMR. This paper presents a system, BAYESIL, which can quickly, accurately, and autonomously produce a person's metabolic profile. Given a 1D 1H NMR spectrum of a complex biofluid (specifically serum or cerebrospinal fluid), BAYESIL can automatically determine the metabolic profile. This requires first performing several spectral processing steps, then matching the resulting spectrum against a reference compound library, which contains the "signatures" of each relevant metabolite. BAYESIL views spectral matching as an inference problem within a probabilistic graphical model that rapidly approximates the most probable metabolic profile. Our extensive studies on a diverse set of complex mixtures including real biological samples (serum and CSF), defined mixtures and realistic computer generated spectra; involving > 50 compounds, show that BAYESIL can autonomously find the concentration of NMR-detectable metabolites accurately (~ 90% correct identification and ~ 10% quantification error), in less than 5 minutes on a single CPU. These results demonstrate that BAYESIL is the first fully-automatic publicly-accessible system that provides quantitative NMR spectral profiling effectively-with an accuracy on these biofluids that meets or exceeds the performance of trained experts. We anticipate this tool will usher in high-throughput metabolomics and enable a wealth of new applications of NMR in clinical settings. BAYESIL is accessible at http://www.bayesil.ca.
| S-EPMC4446368 | biostudies-literature
Project description:Here we introduce the fully quantified spectral imaging (FSI) method as a new tool to probe the stoichiometry and stability of protein complexes in biological membranes. The FSI method yields two dimensional membrane concentrations and FRET efficiencies in native plasma membranes. It can be used to characterize the association of membrane proteins: to differentiate between monomers, dimers, or oligomers, to produce binding (association) curves, and to measure the free energies of association in the membrane. We use the FSI method to study the lateral interactions of Vascular Endothelial Growth Factor Receptor 2 (VEGFR2), a member of the receptor tyrosine kinase (RTK) superfamily, in plasma membranes, in vivo. The knowledge gained through the use of the new method challenges the current understanding of VEGFR2 signaling.
| S-EPMC4755892 | biostudies-literature
Project description:Engineering, controlling, and simulating quantum dynamics is a strenuous task. However, these techniques are crucial to develop quantum technologies, preserve quantum properties, and engineer decoherence. Earlier results have demonstrated reservoir engineering, construction of a quantum simulator for Markovian open systems, and controlled transition from Markovian to non-Markovian regime. Dephasing is an ubiquitous mechanism to degrade the performance of quantum computers. However, all-purpose quantum simulator for generic dephasing is still missing. Here, we demonstrate full experimental control of dephasing allowing us to implement arbitrary decoherence dynamics of a qubit. As examples, we use a photon to simulate the dynamics of a qubit coupled to an Ising chain in a transverse field and also demonstrate a simulation of nonpositive dynamical map. Our platform opens the possibility to simulate dephasing of any physical system and study fundamental questions on open quantum systems.
| S-EPMC6110829 | biostudies-literature
Project description:Fully-connected triads (FCTs), such as the Oct4-Sox2-Nanog triad, have been implicated as recurring transcriptional motifs embedded within the regulatory networks that specify and maintain cellular states. To explore the possible connections between FCT topologies and cell fate determinations, we employed computational network screening to search all possible FCT topologies for multistability, a dynamic property that allows the rise of alternate regulatory states from the same transcriptional network. The search yielded a hierarchy of FCTs with various potentials for multistability, including several topologies capable of reaching eight distinct stable states. Our analyses suggested that complete auto-activation is an effective indicator for multistability, and, when gene expression noise was incorporated into the model, the networks were able to transit multiple states spontaneously. Different levels of stochasticity were found to either induce or disrupt random state transitioning with some transitions requiring layovers at one or more intermediate states. Using this framework we simulated a simplified model of induced pluripotency by including constitutive overexpression terms. The corresponding FCT showed random state transitioning from a terminal state to the pluripotent state, with the temporal distribution of this transition matching published experimental data. This work establishes a potential theoretical framework for understanding cell fate determinations by connecting conserved regulatory modules with network dynamics. Our results could also be employed experimentally, using established developmental transcription factors as seeds, to locate cell lineage specification networks by using auto-activation as a cipher.
| S-EPMC4109943 | biostudies-literature
Project description:A longstanding challenge in epilepsy research and practice is the need to classify synchronization patterns hidden in multivariate electroencephalography (EEG) data that is routinely superimposed with intensive noise. It is essential to select a suitable feature extraction method to achieve high recognition performance. A typical approach is to extract the mutual information (MI) between pairs of channels. This calculation, which considers the differences between the sequence pairs to build a reasonable partition, can improve the classification performance. On this basis, however, it is even more difficult to adaptively classify the synchronization patterns hidden in multivariate EEG data under circumstances of insufficient a priori knowledge of domain dependency, such as denoising, feature extraction on a special patient, etc. To address these problems by (1) effectively calculating the MI matrix (synchronization pattern) and (2) accurately classifying the seizure or non-seizure state, this study first accurately measures the synchronization between channel pairs in terms of affinity propagation clustering partition MI (APCPMI). The global synchronization measurement is then obtained by organizing APCPMIs of all channel pairs into a correlation matrix. Finally, a cross-layer fully connected net is designed to characterize the synchronization dynamics correlation matrices adaptively and identify seizure or non-seizure states automatically. Experiments are performed using the CHB-MIT scalp EEG dataset to evaluate the proposed approach. Seizure states are identified with an accuracy, sensitivity, and specificity of 0.9793 ± 0.002, 0.9942 ± 0.0005, and 0.9676 ± 0.003, respectively; the resulting performance is superior to those achieved by most existing methods over the same dataset. Furthermore, the approach alleviates the necessity for strictly preprocessing (denoising, removing interferences and artifacts) the EEG data using prior knowledge, which is usually required by existing approaches.
| S-EPMC6176510 | biostudies-literature
Project description:PurposeQuantification of retinal layers using automated segmentation of optical coherence tomography (OCT) images allows for longitudinal studies of retinal and neurological disorders in mice. The purpose of this study was to compare the performance of automated retinal layer segmentation algorithms with data from manual segmentation in mice using the Spectralis OCT.MethodsSpectral domain OCT images from 55 mice from three different mouse strains were analyzed in total. The OCT scans from 22 C57Bl/6, 22 BALBc, and 11 C3A.Cg-Pde6b+Prph2Rd2 /J mice were automatically segmented using three commercially available automated retinal segmentation algorithms and compared to manual segmentation.ResultsFully automated segmentation performed well in mice and showed coefficients of variation (CV) of below 5% for the total retinal volume. However, all three automated segmentation algorithms yielded much thicker total retinal thickness values compared to manual segmentation data (P < 0.0001) due to segmentation errors in the basement membrane.ConclusionsWhereas the automated retinal segmentation algorithms performed well for the inner layers, the retinal pigmentation epithelium (RPE) was delineated within the sclera, leading to consistently thicker measurements of the photoreceptor layer and the total retina.Translational relevanceThe introduction of spectral domain OCT allows for accurate imaging of the mouse retina. Exact quantification of retinal layer thicknesses in mice is important to study layers of interest under various pathological conditions.
| S-EPMC4555843 | biostudies-literature
Project description:Spectral analysis of the arterial pressure waveform, using specialized hardware, has been used for the retrospective calculation of the 'Spectral Peak Ratio' (SPeR) of the respiratory and cardiac arterial spectral peaks. The metric can quantify the cardiovascular response to volume loading by analysing the effect of changing tidal volume (indexed to body weight) (VTI) on pulse pressure variability. In this pilot study, the feasibility of real-time SPeR calculation, using a mobile computer which was wirelessly connected to the patient monitor, was evaluated by examining the determinants of SPeR in 60 cardiac-surgical patients. In 30 patients undergoing aortic valve replacement (AVR), graded cyclical changes in ventricular loading were induced by increasing VTI over 2 min, while performing spectral analysis at 1 Hz, before and after AVR. A strong, linear correlation between SPeR and VTI was found and the slope of the regression line (β) changed significantly after AVR. The change in β correlated with the width of the preoperative vena contracta. In another 30 patients, SPeR at constant VTI was calculated at 1 Hz during passive leg raising. β fell significantly on leg raising. The mean arterial pressure change during the manoeuvre was linearly related to the change in β. Real-time spectral analysis of the arterial waveform was easily accomplished. The regression of SPeR on VTI was linear. β appeared to represent the slope of the cardiac response curve at the venous return curve equilibrium point. Measurements were possible at a significantly lower VTI than the equivalent time domain metrics.
| S-EPMC6314984 | biostudies-literature
Project description:Develop the first normative database of the thickness of every inner retinal layer in the macular area in a healthy, Caucasian population between 18 to 87 years old, using Spectralis Optical Coherence Tomography (OCT).On this transversal, observational study, 300 patients between 18 to 87 years old and without an ophthalmological condition were recruited. Macular OCT scans were performed on all patients (Spectralis OCT, Heidelberg Engineering). An axial length measurement, and keratometry were performed using an optical biometer. The volume and thickness of the different macular sectors of the inner retinal layers (retinal nerve fiber layer (RNFL), ganglion cells layer (CGL) and inner plexiform layer (IPL)) were analyzed with the Spectralis OCT segmentation software. An eye was randomly selected for each patient.297 patients (179 females and 118 males) were included in the study. The mean age was 56.07 years (range: 40.50-72). The multivariate analysis showed a positive correlation between the RNFL thickness and the axial length (p < 0.001). The mean central retinal thickness was 278.2 ?m (range: 266-291), the mean central RNFL thickness was 12.61 ?m (range: 11-14), the mean central CGL thickness was 17.63 ?m (range: 14-21) and the mean central IPL thickness was 22.02 ?m (range: 20-25). The multivariate analysis showed a negative correlation between age and CGL thickness and inner IPL thickness (p< 0.001).This study provides a normative database of the volume of each of the inner retinal layers on a Caucasian population.
| S-EPMC5498048 | biostudies-literature
Project description:Quantitative analyses of soil and sediment samples are often used to complement stratigraphic interpretations in archaeological and geoscientific research. The outcome of such analyses often is confined to small parts of the examined profiles as only a limited number of samples can be extracted and processed. Recent laboratory studies show that such selectively measured soil and sediment characteristics can be spatially extrapolated using spectral image data, resulting in reliable maps of a variety of parameters. However, on-site usage of this method has not been examined. We therefore explore, whether image data (RGB data and visible and near infrared hyperspectral data), acquired under regular fieldwork conditions during an archaeological excavation, in combination with a sampling strategy that is close to common practice, can be used to produce maps of soil organic matter, hematite, calcite, several weathering indices and grain size characteristics throughout complex archaeological profiles. We examine two profiles from an archaeological trench in Yeha (Tigray, Ethiopia). Our findings show a promising performance of RGB data and its derivative CIELAB as well as hyperspectral data for the prediction of parameters via random forest regression. By including two individual profiles we are able to assess the accuracy and reproducibility of our results, and illustrate the advantages and drawbacks of a higher spectral resolution and the necessary additional effort during fieldwork. The produced maps of the parameters examined allow us to critically reflect on the stratigraphic interpretation and offer a more objective basis for layer delineation in general. Our study therefore promotes more transparent and reproducible documentation for often destructive archaeological fieldwork.
| S-EPMC7485874 | biostudies-literature