Smart Seizure Detection System: Machine Learning Based Model in Healthcare IoT.

AIMS: Epilepsy, the tendency to have recurrent seizures, can have various causes, including brain tumors, genetics, stroke, brain injury, infections, and developmental disorders. Epileptic seizures are usually transient events. They normally leave no trace after the postictal recovery period has passed. BACKGROUND: An electroencephalogram (EEG) can only detect brain activity during the recording. It will be detected if an epileptogenic focus or generalized abnormality is active during the recording. OBJECTIVE: This work demonstrated a smart seizure detection system for Healthcare IoT, which is a challenging problem of EEG data analysis. METHOD: The study suggested an integrated methodology in recognition of the drawbacks of manual identification and the significant negative effects of uncontrollable seizures on patients' lives. RESULT: The research shows remarkable accuracy, up to 100% in some experiments, by combining classifier ensembles like Decision Trees, Logistic Regression, and Support Vector Machine with different signal processing techniques like Discrete Wavelet Transform, Hjorth Parameters, and statistical features. The results were compared using the kNN classifier, compared with other datasets and other state-of-the-art techniques. CONCLUSION: Healthcare IoT is further utilized by the methodology, which takes a comprehensive approach using classifier ensembles and signal processing approaches resulting in real-time data to help them make better decisions. This demonstrates how well the suggested method works for smart seizure detection, which is a crucial development for better patient outcomes.

A Novel Detection of Cerebrovascular Disease using Multimodal Medical Image Fusion.

BACKGROUND: Diseases are medical situations that are allied with specific signs and symptoms. A disease may be instigated by internal dysfunction or external factors like pathogens. Cerebrovascular disease can progress from diverse causes, comprising thrombosis, atherosclerosis, cerebral venous thrombosis, or embolic arterial blood clot. OBJECTIVE: In this paper, authors have proposed a robust framework for the detection of cerebrovascular diseases employing two different proposals which were validated by use of other dataset. METHODS: In proposed model 1, the Discrete Fourier transform is used for the fusion of CT and MR images which was classified them using machine learning techniques and pre-trained models while in proposed model 2, the cascaded model was proposed. The performance evaluation parameters like accuracy and losses were evaluated. RESULTS: 92% accuracy was obtained using Support Vector Machine using Gray Level Difference Statistics and Shape features with Principal Component Analysis as a feature selection technique while Inception V3 resulted in 95.6% accuracy while the cascaded model resulted in 96.21% accuracy. CONCLUSION: The cascaded model is later validated on other datasets which results in 0.11% and 0.14% accuracy improvement for TCIA and BRaTS datasets respectively.

Influence of scanning protocol on the accuracy of complete-arch digital implant scans: An in vitro study.

OBJECTIVE: This in-vitro study assessed the influence of two intraoral scanning (IOS) protocols on the accuracy (trueness and precision) of digital scans performed in edentulous arches. METHODS: Twenty-two abutment-level master casts of edentulous arches with at least four implants were scanned repeatedly five times, each with two different scanning protocols. Protocol A (IOS-A) consisted of scanning the edentulous arch before inserting the implant scan bodies, followed by their insertion and its subsequent digital acquisition. Protocol B (IOS-B) consisted of scanning the edentulous arch with the scan bodies inserted from the outset. A reference scan from each edentulous cast was obtained using a laboratory scanner. Trueness and precision were calculated using the spatial fit analysis, cross-arch distance, and virtual Sheffield test. Statistical analysis was performed using generalized estimating equations (GEEs). Statistical significance was set at α = .05. RESULTS: In the spatial fit test, the precision of average 3D distances was 45 µm (±23 µm) with protocol IOS-A and 25 µm (±10 µm) for IOS-B (p < .001), and the trueness of average 3D distances was 44 µm (±24 µm) with protocol IOS-A and 24 µm (±7 µm) for IOS-B (p < .001). Cross-arch distance precision was 59 µm (±53 µm) for IOS-A and 41 µm (±43 µm) for IOS-B (p = .0035), and trueness was 64 µm (±47 µm) for IOS-A and 50 µm (±40 µm) for IOS-B (p = .0021). Virtual Sheffield precision was 286 µm (±198 µm) for IOS-A and 146 µm (±92 µm) for IOS-B (p < .001), and trueness was 228 µm (±171 µm) for IOS-A and 139 µm (±92 µm) for IOS-B (p < .001). CONCLUSIONS: The IOS-B protocol demonstrated significantly superior accuracy. Placement of scan bodies before scanning the edentulous arch is recommended to improve the accuracy of complete-arch intraoral scanning.

Post-partum cerebral venous sinus thrombosis: A case report.

INTRODUCTION: Cerebral venous sinus thrombosis (CVST) is a rare and life-threatening condition that may be encountered during pregnancy and puerperium. The diagnosis of CVST is a challenge because of its varied presentation. CASE REPORT: A 28-year-old woman presented with headache, projectile vomiting, and generalized tonic-clonic seizures 10 days after delivery by cesarean section. She had an uneventful antenatal period of 38 weeks of gestation. High clinical suspicion and the availability of magnetic resonance venography helped in making a diagnosis of CVST. She was successfully managed with a low-molecular-weight heparin (LMWH) and anti-epileptic therapy with no residual complications. DISCUSSION: Pregnancy induces several prothrombotic changes in the coagulation system that predispose to CVST. These changes persist for six to eight weeks after birth. Infection and cesarean section are the additional risk factors for CVST during puerperium. The symptoms of CVST depend on the sinuses and veins involved, raised intracranial pressure, and the extent of brain parenchymal injury. CONCLUSION: Greater awareness of the disease and the availability of imaging modalities have contributed to the early diagnosis and favorable outcomes in these cases. LMWH is the main stay of treatment in this disease.

A Study of the Long-Term Electrochemical Stability of Thin-Film Titanium-Platinum Microelectrodes and Their Comparison to Classic, Wire-Based Platinum Microelectrodes in Selected Inorganic Electrolytes.

This paper discusses the electrochemical properties of thin-film, planar, titanium-platinum (Ti-Pt) microelectrodes fabricated using glass or silicon substrates and compares their performance to the classic platinum (Pt) microelectrodes embedded in glass. To analyze the possible differences coming both from the size of the tested electrodes as well as from the substrate, short- and long-term electrochemical tests were performed on selected water electrolytes (KCl, HCl, KOH). To study the electrochemical response of the electrodes, the cyclic voltammetry (CV) measurements were carried out at different scanning rates (from 5 to 200 mV/s). Long-term tests were also conducted, including one thousand cycles with a 100 mV/s scan rate to investigate the stability of the tested electrodes. Before and after electrochemical measurements, the film morphology was analyzed using a scanning electron microscope (SEM). The good quality of the thin-film Pt electrodes and the high repeatability in electrochemical response have been shown. There are minor differences in standard deviation values taken from electrochemical measurements, comparing thin-film and wire-based electrodes. Damages or any changes on the electrodes' surfaces were revealed by SEM observations after long-term electrochemical tests.

Attention-deficit hyperactivity disorder and psychostimulant use in patients seeking dental care-Associations with common orofacial pain complaints.

BACKGROUND: Dental medicine should expand its scope to properly assess medical and psychosocial factors that might have an impact on patients' oral health. Based on previous literature and clinical experience, attention-deficit/hyperactivity disorder and psychostimulant medications might represent factors associated with orofacial pain symptoms. OBJECTIVE: The aim of the study was to assess whether common orofacial pain complaints such as jaw pain, jaw clicking, teeth clenching and headaches are more prevalent in dental patients who have an ADHD diagnosis and/or use psychostimulant medications. METHODS: Orofacial pain symptoms prevalence was compared among four groups from a sample of new patients seeking dental care at Tufts University School of Dental Medicine (n = 11 699) based on ADHD diagnosis and psychostimulants intake: G1: no ADHD, no stimulants; G2: yes ADHD, yes stimulants; G3: yes ADHD, no stimulants; G4: no ADHD, yes stimulants. RESULTS: In multivariable logistic regression models adjusting for age, gender, tobacco use, and alcohol consumption, significant differences were found for clenching (p < .0001), jaw pain (p < .0001), and headache (p < .0001). Compared to G1, two groups (G2 and G4) exhibited significantly higher odds of clenching and headaches, whereas only G2 exhibited significantly higher odds of jaw pain. CONCLUSIONS: In comparison with patients without ADHD and not taking psychostimulants medications, dental patients using psychostimulants with and without ADHD diagnosis report headaches and teeth clenching more frequently, while jaw pain is reported more frequently only by those taking psychostimulants with an ADHD diagnosis. Further research is necessary to assess the nature of these associations and their clinical relevance.

Evaluation of marginal and internal adaptation of veneers generated by the guided prosthetic tooth preparation system.

OBJECTIVE: This in vitro study aimed to assess and contrast the marginal and internal adaptation of all-ceramic prefabricated veneers manufactured via the FirstFit guided tooth preparation system against all-ceramic veneers produced using the chairside Computer-Aided Design/Computer Aided Manufacture (CAD/CAM) system following identical guided preparation protocols. MATERIALS AND METHODS: Two main groups were included, with 16 lithium disilicate veneers per group. Four typodonts were used for the test (FirstFit) and control CAD/CAM groups. Intraoral scans created master casts and preparation guides. Guides performed preparations on typodont teeth (two central incisors and two lateral incisors). Prepared teeth were scanned (CEREC Omnicam) to design and mill CAD/CAM veneers. Marginal gap thickness and cement space thickness were measured using light microscopy at four locations: marginal, cervical internal, middle internal, and incisal internal. RESULTS: No significant difference existed between groups for marginal adaptation (p = 0.058) or incisal internal adaptation (p = 0.076). The control group had significantly lower values for middle internal adaptation (p = 0.023) and cervical internal adaptation (p = 0.019). CONCLUSIONS: Guided preparation evaluation showed no significant differences in marginal or incisal internal adaptation. The CAD/CAM group had significantly lower middle and cervical internal adaptation values.

Computational Model for the Detection of Diabetic Retinopathy in 2-D Color Fundus Retina Scan.

BACKGROUND: Diabetic Retinopathy (DR) is a growing problem in Asian countries. DR accounts for 5% to 7% of all blindness in the entire area. In India, the record of DR-affected patients will reach around 79.4 million by 2030. AIMS: The main objective of the investigation is to utilize 2-D colored fundus retina scans to determine if an individual possesses DR or not. In this regard, Engineering-based techniques such as deep learning and neural networks play a methodical role in fighting against this fatal disease. METHODS: In this research work, a Computational Model for detecting DR using Convolutional Neural Network (DRCNN) is proposed. This method contrasts the fundus retina scans of the DR-afflicted eye with the usual human eyes. Using CNN and layers like Conv2D, Pooling, Dense, Flatten, and Dropout, the model aids in comprehending the scan's curve and color-based features. For training and error reduction, the Visual Geometry Group (VGG-16) model and Adaptive Moment Estimation Optimizer are utilized. RESULTS: The variations in a dataset like 50%, 60%, 70%, 80%, and 90% images are reserved for the training phase, and the rest images are reserved for the testing phase. In the proposed model, the VGG-16 model comprises 138M parameters. The accuracy is achieved maximum rate of 90% when the training dataset is reserved at 80%. The model was validated using other datasets. CONCLUSION: The suggested contribution to research determines conclusively whether the provided OCT scan utilizes an effective method for detecting DRaffected individuals within just a few moments.

Empirical Curvelet-ridgelet Wavelet Transform for Multimodal Fusion of Brain Images.

BACKGROUND: Empirical curvelet and ridgelet image fusion is an emerging technique in the field of image processing that aims to combine the benefits of both transforms. OBJECTIVE: The proposed method begins by decomposing the input images into curvelet and ridgelet coefficients using respective transform algorithms for Computerized Tomography (CT) and magnetic Resonance Imaging (MR) brain images. METHODS: An empirical coefficient selection strategy is then employed to identify the most significant coefficients from both domains based on their magnitude and directionality. These selected coefficients are coalesced using a fusion rule to generate a fused coefficient map. To reconstruct the image, an inverse curvelet and ridgelet transform was applied to the fused coefficient map, resulting in a high-resolution fused image that incorporates the salient features from both input images. RESULTS: The experimental outcomes on real-world datasets show how the suggested strategy preserves crucial information, improves image quality, and outperforms more conventional fusion techniques. For CT Ridgelet-MR Curvelet and CT Curvelet-MR Ridgelet, the authors' maximum PSNRs were 58.97 dB and 55.03 dB, respectively. Other datasets are compared with the suggested methodology. CONCLUSION: The proposed method's ability to capture fine details, handle complex geometries, and provide an improved trade-off between spatial and spectral information makes it a valuable tool for image fusion tasks.

Effect of varying hydrologic regime on seasonal total maximum daily loads (TDML) in an agricultural watershed.

Rising hypoxia due to the eutrophication of riverine ecosystems is primarily caused by the transport of nutrients. The majority of existing TMDL models cannot be efficienty applied to represent nutrient concentrations in riverine ecosystems having varying flow regimes due to seasonal differences. Accurate TMDL assessment requires nutrient loads and suspended matter estimation under varying flow regimes with minimal uncertainty. Though a large database can enhance accuracy, it can be resource intensive. This study presents the design of an innovative modeling strategy to optimize the use of existing datasets to effectively represent streamflow-load dynamics while minimizing uncertainty. The study developed an approach to assess TMDLs using six different flux models and kriging techniques (i) to enhance the accuracy of nutrient load estimation under different hydrologic regimes (flow stratifications) and (ii) to derive an optimal modeling strategy and sampling scheme for minimizing uncertainty. The flux models account for uncertainty in load prediction across varying flow strata, and the deployment of multiple load calculation procedures. Further, the proposed flux approach allows the determination of load exceedance under different TMDL scenarios aimed at minimizing uncertainty to achieve reliable load predictions. The study employed a 10-year dataset (2009-2018) consisting of daily flow data (m3/sec) and weekly data (mg/L) for nitrogen (N), phosphorus (P) and total suspended solids (TSS) concentrations in three distinct agricultural sites in+ the Minnesota River Watershed. The outcomes were analyzed geospatially in a Geographic Information System (GIS) environment using the kriging interpolation technique. The study recommends (i) triple stratification of flows to obtain accurate load estimates, and (ii) an optimal sampling scheme for nitrogen and phosphorous with 30.6 % and 49.8 % datapoints from high flow strata. The study outcomes are expected to contribute to the planning of economically and technically sound combinations of best management practices (BMPs) required for achieving total maximum daily loads (TMDL) in a watershed.

Bandgaps of long-period polytypes of IV, IV-IV, and III-V semiconductors estimated with an Ising-type additivity model.

We apply an Ising-type model to estimate the bandgaps of the polytypes of group IV elements (C, Si, and Ge) and binary compounds of groups: IV-IV (SiC, GeC, and GeSi), and III-V (nitride, phosphide, and arsenide of B, Al, and Ga). The models use reference bandgaps of the simplest polytypes comprising 2-6 bilayers calculated with the hybrid density functional approximation, HSE06. We report four models capable of estimating bandgaps of nine polytypes containing 7 and 8 bilayers with an average error of ≲0.05 eV. We apply the best model with an error of <0.04 eV to predict the bandgaps of 497 polytypes with up to 15 bilayers in the unit cell, providing a comprehensive view of the variation in the electronic structure with the degree of hexagonality of the crystal structure. Within our enumeration, we identify four rhombohedral polytypes of SiC-9R, 12R, 15R(1), and 15R(2)-and perform detailed stability and band structure analysis. Of these, 15R(1) that has not been experimentally characterized has the widest bandgap (>3.4 eV); phonon analysis and cohesive energy reveal 15R(1)-SiC to be metastable. Additionally, we model the energies of valence and conduction bands of the rhombohedral SiC phases at the high-symmetry points of the Brillouin zone and predict band structure characteristics around the Fermi level. The models presented in this study may aid in identifying polytypic phases suitable for various applications, such as the design of wide-gap materials, that are relevant to high-voltage applications. In particular, the method holds promise for forecasting electronic properties of long-period and ultra-long-period polytypes for which accurate first-principles modeling is computationally challenging.

CAD System Design for Pituitary Tumor Classification based on Transfer Learning Technique.

BACKGROUND: A brain tumor is an asymmetrical expansion by cells inevitably emulating amid them. Image processing is a vibrant research area where the handing out of the image in the medical field is an exceedingly tricky field. In this paper, an expert algorithm is suggested for the detection of pituitary brain tumors from MR images. METHODS: The preprocessing techniques (smoothing, edge detection, filtering) and segmentation techniques (watershed) are applied to the online data set. The transfer learning technique is used as a classifier whose performance is measured in terms of classification accuracy. Resnet 50, Inception V3VGG16, and VGG19 models are used as classification algorithms. The proposed model is validated using different machine learning techniques considering hybrid features. RESULTS: 96% accuracy was obtained employing the Inception V3 model & 95% accuracy was attained using hybrid GLDS and GLCM features employing Support Vector Machine algorithm while 93% was attained using Probabilistic Neural Network and k Nearest Neighbor techniques. CONCLUSION: Computer-aided systems gave much faster and more accurate results than image processing techniques.1.0% accuracy improvement was observed while using Inception V3 over GLDS + GLCM + SVM and 2.1% accuracy improvement using GLDS + GLCM + SVM over GLDS + GLCM + kNN.

Detection of Brain Tumor Employing Residual Network-based Optimized Deep Learning.

BACKGROUND: Diagnosis and treatment planning play a very vital role in improving the survival of oncological patients. However, there is high variability in the shape, size, and structure of the tumor, making automatic segmentation difficult. The automatic and accurate detection and segmentation methods for Brain tumors are proposed in this paper. METHODS: A modified ResNet50 model was used for tumor detection, and a ResUNetmodel-based convolutional neural network for segmentation is proposed in this paper. The detection and segmentation were performed on the same dataset consisting of pre-contrast, FLAIR, and postcontrast MRI images of 110 patients collected from the Cancer Imaging Archive. Due to the use of Residual Networks, the authors observed improvement in evaluation parameters, such as accuracy for tumor detection and dice similarity coefficient for tumor segmentation. RESULTS: The accuracy of tumor detection and Dice Similarity Coefficient achieved by the segmentation model were 96.77% and 0.893, respectively, for the TCIA dataset. The results were compared based on manual segmentation and existing segmentation techniques. The tumor mask was also individually compared to the ground truth using the SSIM value. The proposed detection and segmentation models were validated on BraTS2015 and BraTS2017 datasets, and the results were consensus. CONCLUSION: The use of residual networks in both the detection and the segmentation model resulted in improved accuracy and DSC score. DSC score was increased by 5.9% compared to the UNet model, and the accuracy of the model was increased from 92% to 96.77% for the test set.

Flow cytometry profiling of cellular immune response in COVID-19 infected, recovered and vaccinated individuals.

INTRODUCTION: SARS-CoV-2 has infected over 753 million individuals and caused more than 6.8 million deaths globally to date. COVID-19 disease severity has been associated with SARS-CoV-2 induced hyper inflammation and the immune correlation with its pathogenesis remains unclear. Acute viral infection is characterised by vigorous coordinated innate and adaptive activation, including an early cellular response that correlates well with the amplitude of virus specific humoral response. OBJECTIVE: The present study covers a wide spectrum of cellular immune response against COVID-19, irrespective of infection and vaccination. METHODS: We analysed immune status of (a) COVID-19 hospitalised patients including deceased and recovered patients, and compared with home isolated and non-infected healthy individuals, and (b) infected home isolated individuals with vaccinated individuals, using flow cytometry. We performed flow cytometry analysis of PBMCs to determine non-specific cell-mediated immune response. RESULTS: The immune response revealed extensive induction and activation of multiple immune lineages, including T and B cells, Th17 regulatory subsets and M1, M2 macrophages in deceased and hospitalised recovered patients, vaccinated and healthy individuals. Compromised immune cell expression was observed in deceased patients even in later stages, while expression was restored in hospitalised recovered patients and home isolated individuals. CONCLUSION: The findings associated with recovery and convalescence define a new signature of cellular immune response that persists in individuals with SARS-CoV-2 infection and vaccination. The findings will help in providing a better understanding of COVID-19 disease and will aid in developing better therapeutic strategies for treatment.

Aberrant glycosylation of α3 integrins as diagnostic markers in epithelial ovarian cancer.

BACKGROUND: Glycans are strongly involved in stability and function of integrins (ITG) and tetraspanin protein CD63 and their respective interaction partners as they are dysregulated in the tumorigenic processes. Glycosylation changes is a universal phenomenon of cancer cells. In this study, glycosylation changes in epithelial ovarian cancer (EOC) are explored using tetraspanin and integrin molecules. METHODS: ITG and CD63 were immobilized from 10 EOC and 5 benign ovarian cyst fluid on microtiter wells and traced with 3 glycan binding proteins (STn, WGA, UEA) conjugated on europium nanoparticles. Total protein measurements (ITG & CD63 immunoassays) were also performed. The most promising glycovariant candidates identified were then clinically evaluated on the whole cohort of 77 ovarian cyst fluids. Additional testing was performed in ascites fluid samples of liver cirrhosis (n = 2) and EOC (n = 4). RESULTS: Sialylated Tn antibody based glycovariants of ITGα3 (ITGα3STn) and CD63 (CD63STn) performed better than corresponding protein epitope-based immunoassays, ITGα3IA and CD63IA respectively. Combined ITGα3 based assays (ITGα3IA + ITGα3STn) detected 49 out of 55 malignant & borderline cases without detecting any of the 22 benign and healthy cysts. CONCLUSION: Our findings indicate the potential diagnostic application of ITGα3STn along with total ITGα3IA, which could help reduce the unnecessary surgeries. The results encourage studying further the potential use of these novel assays to detect EOC at earlier clinical stages.

Novel predictive model of cell survival/death related effects of Extracellular Signal-Regulated kinase protein.

Computational modelling is a technique for modelling and solving real-world problems by utilising computing to provide solutions. This paper presents a novel predictive model of cell survival/death-related effects of Extracellular Signal-Regulated Kinase Protein. The computational model was designed using Neural Networks and fuzzy system. Three hundred ERK samples were examined using ten different concentrations of three input proteins: EGF, TNF, and insulin. Based on the different concentrations of input proteins and different samples of ERK protein, adjustment Anderson darling (AD) statistics for multiple distribution functions were computed considering different test such as visual test, Pearson correlation coefficient, and uniformity tests. The results reveal that utilising different concentrations and samples, values such as 7.55 AD and 18.4 AD were obtained using the Weibull distribution function for 0 ng/ml of TNF, 100 ng/ml of EGF, and 0 ng/mL of insulin concentrations. The model was validated by predicting the various ERK protein values that fall within the observed range. The proposed model agrees with the deterministic model, which was developed using difference equations.

The Pulpal Response to Crown Preparation and Cementation.

INTRODUCTION: This study aimed to evaluate the risk factors and occurrence of pulpal disease in patients who received either full-coverage (crowns) or large noncrown restorations (fillings, inlays, or onlays involving ≥3 surfaces). METHODS: A retrospective chart review identified 2177 cases of large restorations placed on vital teeth. Based on the restoration type, patients were stratified into various groups for statistical analysis. After restoration placement, those who required endodontic intervention or extraction were classified as having pulpal disease. RESULTS: Over the course of the study, 8.77% (n = 191) of patients developed pulpal disease. Pulpal disease was slightly more common in the large noncrown group than the full-coverage group (9.05% vs 7.54%, respectively). For patients who received large fillings, there was not a statistically significant difference based on operative material (amalgam vs composite: odds ratio = 1.32 [95% confidence interval, 0.94-1.85], P > .05) or the number of surfaces involved (3 vs 4: odds ratio = 0.78 [95% confidence interval, 0.54-1.12], P > .05). The association between the restoration type and the pulpal disease treatment performed was statistically significant (P < .001). The full-coverage group more frequently underwent endodontic treatment than extraction (5.78% vs 3.37%, respectively). Only 1.76% (n = 7) of teeth in the full-coverage group were extracted compared with 5.68% (n = 101) in the large noncrown group. CONCLUSIONS: It appears that ∼9% of patients who receive large restorations will go on to develop pulpal disease. The risk of pulpal disease tended to be highest in older patients who receive large (4 surface) amalgam restorations. However, teeth with full-coverage restorations were less likely to be extracted.

Evaluation of altered miRNA expression pattern to predict COVID-19 severity.

Outbreak of COVID-19 pandemic in December 2019 affected millions of people globally. After substantial research, several biomarkers for COVID-19 have been validated however no specific and reliable biomarker for the prognosis of patients with COVID-19 infection exists. Present study was designed to identify specific biomarkers to predict COVID-19 severity and tool for formulating treatment. A small cohort of subjects (n = 43) were enrolled and categorized in four study groups; Dead (n = 16), Severe (n = 10) and Moderate (n = 7) patients and healthy controls (n = 10). Small RNA sequencing was done on Illumina platform after isolation of microRNA from peripheral blood. Differential expression (DE) of miRNA (patients groups compared to control) revealed 118 down-regulated and 103 up-regulated known miRNAs with fold change (FC) expression ≥2 folds and p ≤ 0.05. DE miRNAs were then subjected to functional enrichment and network analysis. Bioinformatic analysis resulted in 31 miRNAs (24 Down-regulated; 7 up-regulated) significantly associated with COVID-19 having AUC>0.8 obtained from ROC curve. Seventeen out of 31 DE miRNAs have been linked to COVID-19 in previous studies. Three miRNAs, hsa-miR-147b-5p and hsa-miR-107 (down-regulated) and hsa-miR-1299 (up-regulated) showed significant unique DE in Dead patients. Another set of 4 miRNAs, hsa-miR-224-5p (down-regulated) and hsa-miR-4659b-3p, hsa-miR-495-3p and hsa-miR-335-3p were differentially up-regulated uniquely in Severe patients. Members of three miRNA families, hsa-miR-20, hsa-miR-32 and hsa-miR-548 were significantly down-regulated in all patients group in comparison to healthy controls. Thus a distinct miRNA expression profile was observed in Dead, Severe and Moderate COVID-19 patients. Present study suggests a panel of miRNAs which identified in COVID-19 patients and could be utilized as potential diagnostic biomarkers for predicting COVID-19 severity.

Detection of Cerebrovascular Diseases using Novel Discrete Component Wavelet Cosine Transform.

AIMS: Detecting and classifying a brain tumor amid a sole image can be problematic for doctors, although improvements can be made with medical image fusions. BACKGROUND: A brain tumor develops in the tissues surrounding the brain or the skull and has a major impact on human life. Primary tumors begin within the brain, whereas secondary tumors, identified as brain metastasis tumors, are generated outside the brain. OBJECTIVE: This paper proposes hybrid fusion techniques to fuse multi-modal images. The evaluations are based on performance metrics, and the results are compared with conventional ones. METHODS: In this paper, pre-processing is done considering enhancement methods like Binarization, Contrast Stretching, Median Filter, & Contrast Limited Adaptive Histogram Equalization (CLAHE). Authors have proposed three techniques, PCA-DWT, DCT-PCA, and Discrete ComponentWaveletCosine Transform (DCWCT), which were used to fuse CT-MR images of brain tumors. RESULTS: The different features were evaluated from the fused images, which were classified using various machine learning approaches. Maximum accuracy of 97.9% and 93.5% is obtained using DCWCT for Support Vector Machine (SVM) and k Nearest Neighbor (kNN), respectively, considering the combination of both feature's shape & Grey Level Difference Statistics. The model is validated using another online dataset. CONCLUSION: It has been observed that the classification accuracy for detecting cerebrovascular disease is better after employing the proposed image fusion technique.