9682 research outputs found
Sort by
MHDeep: Mental Health Disorder Detection System based on Body-Area and Deep Neural Networks
No full textMental health problems impact quality of life of millions of people around the world. However, diagnosis of mental health disorders is a challenging problem that often relies on self-reporting by patients about their behavioral patterns and social interactions. Therefore, there is a need for new strategies for diagnosis and daily monitoring of mental health conditions. The recent introduction of body-area networks consisting of a plethora of accurate sensors embedded in smartwatches and smartphones and edge-compatible deep neural networks (DNNs) points towards a possible solution. Such wearable medical sensors (WMSs) enable continuous monitoring of physiological signals in a passive and non-invasive manner. However, disease diagnosis based on WMSs and DNNs, and their deployment on edge devices, such as smartphones, remains a challenging problem. To this end, we propose a framework called MHDeep that utilizes commercially available WMSs and efficient DNN models to diagnose three important mental health disorders: schizoaffective, major depressive, and bipolar. MHDeep uses eight different categories of data obtained from sensors integrated in a smartwatch and smartphone. These categories include various physiological signals and additional information on motion patterns and environmental variables related to the wearer. MHDeep eliminates the need for manual feature engineering by directly operating on the
data streams obtained from participants. Since the amount of the data is limited, MHDeep uses a synthetic data generation module to augment real data with synthetic data drawn from the same probability distribution. We use the synthetic dataset to pre-train the weights of the DNN models, thus imposing a prior on the weights. We use a grow-and-prune DNN synthesis approach to learn both the architecture and weights during the training process. We use three different data partitions to evaluate the MHDeep models trained with data collected from 74 individuals. We conduct two types of evaluations: at the data instance level and at the patient level. MHDeep achieves an average test accuracy, across the three data partitions, of 90.4%, 87.3%, and 82.4%, respectively, for classifications between healthy and schizoaffective disorder instances, healthy and major depressive disorder instances, and healthy and bipolar disorder instances. At the patient level, MHDeep DNNs achieve an accuracy of 100%, 100%, and 90.0% for the three mental health disorders, respectively, based on inference that uses 40, 16, and 22 minutes of data from each patient
Molecular dynamics simulation of organic contaminant adsorption on organic‐coated smectite clay
No full textMolecular dynamics simulations are used to examine the adsorption and aggregation of tyrosine and glutamate molecules on a stack of smectite particles at three different organic loadings. The results reveal a strong affinity of the smectite surface for the organic molecules with the zwitterionic tyrosine molecules coating the exterior surfaces (and entering the interlayer region to a lesser extent), whereas the negatively charged glutamate molecules are generally found near the clay edge sites and in coordination with aqueous and coordinating Ca ions. Additional simulations examine the effects of the tyrosine and glutamate organic coatings on the adsorption of two organic contaminants, dimethyl phthalate (DMP) and perfluorobutanesulfonic
acid (PFBS). The addition of these coatings did not prevent the DMP and PFBS molecules from accessing previously identified favorable adsorption domains. An analysis of the adsorption energetics shows an initial decrease in contaminant adsorption relative to pure mineral surfaces as tyrosine and glutamate are introduced to the system, followed by increasing adsorption with increasing organic loadings. Overall, this research advances the mechanistic understanding of the interplay between smectite surfaces, organic coatings, and organic contaminants
Lead federated neuromorphic learning for wireless edge artificial intelligence
No full textIn order to realize the full potential of wireless edge artificial intelligence (AI), very large and diverse datasets will often be required for energy-demanding model training on resource-constrained edge devices. This paper proposes a lead federated neuromorphic learning (LFNL) technique, which is a decentralized energy-efficient brain-inspired computing method based on spiking neural networks. The proposed technique will enable edge devices to exploit brain-like biophysiological structure to collaboratively train a global model while helping preserve privacy. Experimental results show that, under the situation of uneven dataset distribution among edge devices, LFNL achieves a comparable recognition accuracy to existing edge AI techniques, while substantially reducing data traffic by >3.5× and computational latency by >2.0×. Furthermore, LFNL significantly reduces energy consumption by >4.5× compared to standard federated learning with a slight accuracy loss up to 1.5%. Therefore, the proposed LFNL can facilitate the development of brain-inspired computing and edge AI
Humanized mice reveal a macrophage-enriched gene signature defining human lung tissue protection during SARS-CoV-2 infection
No full textThe human immunological mechanisms defining the clinical outcome of SARS-CoV-2 infection remain
elusive. This knowledge gap is mostly driven by the lack of appropriate experimental platforms recapitulating human immune responses in a controlled human lung environment. Here, we report a mouse model (i.e., HNFL mice) co-engrafted with human fetal lung xenografts (fLX) and a myeloid-enhanced human immune system to identify cellular and molecular correlates of lung protection during SARS-CoV-2 infection. Unlike mice solely engrafted with human fLX, HNFL mice are protected against infection, severe inflammation, and histopathological phenotypes. Lung tissue protection from infection and severe histopathology associates with macrophage infiltration and differentiation and the upregulation of a macrophage-enriched signature composed of 11 specific genes mainly associated with the type I interferon signaling pathway. Our work highlights the HNFL model as a transformative platform to investigate, in controlled experimental settings, human myeloid immune mechanisms governing lung tissue protection during SARS-CoV-2 infection
GAGA factor: a multifunctional pioneering chromatin protein
No full textThe Drosophila GAGA factor (GAF) is a multifunctional protein implicated in nucleosome
organization and remodeling, activation and repression of gene expression, long distance
enhancer-promoter communication, higher order chromosome structure and mitosis. This broad
range of activities poses questions about how a single protein can perform so many seemingly
different and unrelated functions. Current studies argue that GAF acts as a “pioneer” factor,
generating nucleosome free regions of chromatin for different classes of regulatory elements. The
removal of nucleosomes from regulatory elements in turn enables other factors to bind to these
elements and carry out their specialized functions. Consistent with this view, GAF associates
with a collection of chromatin remodelers and also interacts with proteins implicated in different
regulatory functions. In this review, we summarize the known activities of GAF and the functions
of its protein partners
Preformation and epigenesis converge to specify primordial germ cell fate in the early Drosophila embryo
No full textA critical step in animal development is the specification of primordial germ cells (PGCs), the precursors of the germline. Two seemingly mutually exclusive mechanisms are implemented across the animal kingdom: epigenesis and preformation. In epigenesis, PGC specification is non-autonomous and depends on extrinsic signaling pathways. The BMP pathway provides the key PGC specification signals in mammals. Preformation is autonomous and mediated by determinants localized within PGCs. In Drosophila, a classic example of preformation, constituents of the germ plasm localized at the embryonic posterior are thought to be both necessary and sufficient for proper determination of PGCs. Contrary to this longstanding model, here we show that these localized determinants are insufficient by themselves to direct PGC specification in blastoderm stage embryos. Instead, we find that the BMP signaling pathway is required at multiple steps during the specification process and functions in conjunction with components of the germ plasm to orchestrate PGC fate
The Atacama Cosmology Telescope: measurement and analysis of 1D beams for DR4
No full textWe describe the measurement and treatment of the telescope beams for the Atacama Cosmology Telescope's fourth data release, DR4. Observations of Uranus are used to measure the central portion (<12') of the beams to roughly -40 dB of the peak. Such planet maps in intensity are used to construct azimuthally averaged beam profiles, which are fit with a physically motivated model before being transformed into Fourier space. We investigate and quantify a number of percent-level corrections to the beams, all of which are important for precision cosmology. Uranus maps in polarization are used to measure the temperature-to-polarization leakage in the main part of the beams, which is ≲ 1% (2.5%) at 150 GHz (98 GHz). The beams also have polarized sidelobes, which are measured with observations of Saturn and deprojected from the ACT time-ordered data. Notable changes relative to past ACT beam analyses include an improved subtraction of the atmospheric effects from Uranus calibration maps, incorporation of a scattering term in the beam profile model, and refinements to the beam model uncertainties and the main temperature-to-polarization leakage terms in the ACT power spectrum analysis
Bacteriophage and endolysin engineering for biocontrol of food pathogens/pathogens in the food: recent advances and future trends
No full textDespite advances in modern technologies, various foodborne outbreaks have continuously threatened the food safety. The overuse of and abuse/misuse of antibiotics have escalated this threat due to the prevalence of multidrug-resistant (MDR) pathogens. Therefore, the development of new methodologies for controlling microbial contamination is extremely important to ensure the food safety. As an alternative to antibiotics, bacteriophages(phages) and derived endolysins have been proposed as novel, effective, and safe antimicrobial agents and applied for the prevention and/or eradication of bacterial contaminants even in foods and food processing facilities. In this review, we describe recent genetic and protein engineering tools for phages and endolysins. The major aim of engineering is to overcome limitations such as a narrow host range, low antimicrobial activity, and low stability of phages and endolysins. Phage engineering also aims to deter the emergence of phage resistance. In the case of endolysin engineering, enhanced antibacterial ability against Gram-negative and Gram-positive bacteria is another important goal. Here, we summarize the successful studies of phages and endolysins treatment in different types of food. Moreover, this review highlights the recent advances in engineering techniques for phages and endolysins, discusses existing challenges, and suggests technical opportunities for further development, especially in terms of antimicrobial agents in the food industry
The non-centrosymmetric layered compounds IrTe2I and RhTe2I
No full textThe previously unreported layered compounds IrTe2I and RhTe2I were prepared by a high-pressure synthesis method. Single crystal X-ray and powder X-ray diffraction studies find that the compounds are isostructural, crystallizing in a layered orthorhombic structure in the non-centrosymmetric, non-symmorphic space group Pca21 (#29). Characterization reveals diamagnetic, high resistivity, semiconducting behavior for both compounds, consistent with the +3 chemical valence and d6 electronic configurations for both iridium and rhodium and the Te–Te dimers seen in the structural study. Electronic band structures are calculated for both compounds, showing good agreement with the experimental results
Inexpensive Multipatient Respiratory Monitoring System for Helmet Ventilation During COVID-19 Pandemic
No full textHelmet continuous positive applied pressure is a form of noninvasive ventilation (NIV) that has been used to provide respiratory support to COVID-19 patients. Helmet NIV is low-cost, readily available, provides viral filters between the patient and clinician, and may reduce the need for invasive ventilation. Its widespread adoption has been limited, however, by the lack of a respiratory monitoring system needed to address known safety vulnerabilities and to monitor patients. To address these safety and clinical needs, we developed an inexpensive respiratory monitoring system based on readily available components suitable for local manufacture. Open-source design and manufacturing documents are provided. The monitoring system comprises flow, pressure, and CO2 sensors on the expiratory path of the helmet circuit and a central remote station to monitor up to 20 patients. The system is validated in bench tests, in human-subject tests on healthy volunteers, and in experiments that compare respiratory features obtained at the expiratory path to simultaneous ground-truth measurements from proximal sensors. Measurements of flow and pressure at the expiratory path are shown to deviate at high flow rates, and the tidal volumes reported via the expiratory path are systematically underestimated. Helmet monitoring systems exhibit high-flow rate, nonlinear effects from flow and helmet dynamics. These deviations are found to be within a reasonable margin and should, in principle, allow for calibration, correction, and deployment of clinically accurate derived quantities