ABSTRACT
Although presepsin, a crucial biomarker for the diagnosis and management of sepsis, has gained prominence in contemporary medical research, its relationship with routine laboratory parameters, including demographic data and hospital blood test data, remains underexplored. This study integrates machine learning with explainable artificial intelligence (XAI) to provide insights into the relationship between presepsin and these parameters. Advanced machine learning classifiers provide a multilateral view of data and play an important role in highlighting the interrelationships between presepsin and other parameters. XAI enhances analysis by ensuring transparency in the model's decisions, especially in selecting key parameters that significantly enhance classification accuracy. Utilizing XAI, this study successfully identified critical parameters that increased the predictive accuracy for sepsis patients, achieving a remarkable ROC AUC of 0.97 and an accuracy of 0.94. This breakthrough is possibly attributed to the comprehensive utilization of XAI in refining parameter selection, thus leading to these significant predictive metrics. The presence of missing data in datasets is another concern; this study addresses it by employing Extreme Gradient Boosting (XGBoost) to manage missing data, effectively mitigating potential biases while preserving both the accuracy and relevance of the results. The perspective of examining data from higher dimensions using machine learning transcends traditional observation and analysis. The findings of this study hold the potential to enhance patient diagnoses and treatment, underscoring the value of merging traditional research methods with advanced analytical tools.
ABSTRACT
Recent advancements in deep learning have achieved significant progress by increasing the number of parameters in a given model. However, this comes at the cost of computing resources, prompting researchers to explore model compression techniques that reduce the number of parameters while maintaining or even improving performance. Convolutional neural networks (CNN) have been recognized as more efficient and effective than fully connected (FC) networks. We propose a column row convolutional neural network (CRCNN) in this letter that applies 1D convolution to image data, significantly reducing the number of learning parameters and operational steps. The CRCNN uses column and row local receptive fields to perform data abstraction, concatenating each direction's feature before connecting it to an FC layer. Experimental results demonstrate that the CRCNN maintains comparable accuracy while reducing the number of parameters and compared to prior work. Moreover, the CRCNN is employed for one-class anomaly detection, demonstrating its feasibility for various applications.
ABSTRACT
Human fingerprints are randomly created during fetal activity in the womb, resulting in unique and physically irreproducible fingerprint patterns that are applicable as a biological cryptographic primitive. Similarly, stochastically knitted single-walled carbon nanotube (SWNT) network surfaces exhibit inherently random and unique electrical characteristics that can be exploited as a physical unclonable function (PUF) in the authentication. In this study, filamentous M13 bacteriophages are used as a biological gluing template to create a random SWNT network surface with mechanical flexibility, with electrical properties determined by random variation during fabrication. The resistance profile between two adjacent electrodes was mapped for these M13-mediated SWNT network surfaces, with the results demonstrating a unique resistance profile for each M13-SWNT device, similar to that of human fingerprints. Randomness and uniqueness measures were evaluated as respectively 50.5% and 50% using generated challenge-response pairs. Min-entropy for unpredictability evaluation of the M13-SWNT based PUFs resulted in 0.98. Our results showed that M13-SWNT random network exhibits cryptographic characteristics when used in a bio-inspired PUF device.
Subject(s)
Bacteriophages , Biomimetic Materials , Nanotubes, Carbon , Humans , Electronics , ElectrodesABSTRACT
Dendritic network implementable organic neurofiber transistors with enhanced memory cyclic endurance for spatiotemporal iterative learning are proposed. The architecture of the fibrous organic electrochemical transistors consisting of a double-stranded assembly of electrode microfibers and an iongel gate insulator enables the highly sensitive multiple implementation of synaptic junctions via simple physical contact of gate-electrode microfibers, similar to the dendritic connections of a biological neuron fiber. In particular, carboxylic-acid-functionalized polythiophene as a semiconductor channel material provides stable gate-field-dependent multilevel memory characteristics with long-term stability and cyclic endurance, unlike the conventional poly(alkylthiophene)-based neuromorphic electrochemical transistors, which exhibit short retention and unstable endurance. The dissociation of the carboxylic acid of the polythiophene enables reversible doping and dedoping of the polythiophene channel by effectively stabilizing the ions that penetrate the channel during potentiation and depression cycles, leading to the reliable cyclic endurance of the device. The synaptic weight of the neurofiber transistors with a dendritic network maintains the state levels stably and is independently updated with each synapse connected with the presynaptic neuron to a specific state level. Finally, the neurofiber transistor demonstrates successful speech recognition based on iterative spiking neural network learning in the time domain, showing a substantial recognition accuracy of 88.9%.
ABSTRACT
Ball-milling process was applied to increase sulfur content in sulfur/polyacrylonitrile (SPAN) composites and improve electrochemical properties of Li/S batteries. In contrast to as-received PAN, pre-heated PAN was pulverized by the ball-milling, resulting in the continuous size reduction with increasing the milling time. Sulfur contents in SPAN composites synthesized with a pre-heated and milled PAN were increased with prolonging the milling time and the maximum content reached 44.5% for the milling time of 10 h. Li/S cells with SPAN electrodes delivered the first discharge capacities of 1356, 1409, 1512, and 1610 mAh/g-sulfur for milling times of 0, 1, 5, and 10 h. The 10 h-milled SPAN electrode with the highest sulfur content exhibited poor initial efficiency and low capacity retention at 100 cycles, whereas from a comprehensive viewpoint of the specific capacity and capacity retention, the 6 h-milled SPAN electrode exhibited the best electrochemical performance due to the suitable size and sulfur content.
ABSTRACT
PURPOSE: This study aims to evaluate early changes in retinal structure and BMP2 expression in the retina and crystalline lens by comparing streptozotocin-induced diabetic pigs and normal control group pigs. METHODS: Five eye samples from five diabetic Micro-pigs (Medikinetics, Pyeongtaek, Korea) and five eye samples from five control pigs bred in a specific pathogen-free area were used. Diabetes was developed through intravenous injection of nicotinamide and streptozotocin, and the average fasting glucose level was maintained at 250 mg/dL or higher for 16 weeks. To evaluate BMP2 expression in the retina and crystalline lens, Western blotting was performed. RESULTS: In Hematoxylin and Eosin staining, most diabetic pigs showed structural abnormalities in the inner plexiform layer. The number of nuclei in the ganglion cell layer within the range of 104 µm2 was 3.78±0.60 for diabetic pigs and 5.57±1.07 for control group pigs, showing a statistically significant difference. In immunohistochemical staining, diabetic retinas showed an overall increase in BMP2 expression. In Western blotting, the average BMP2/actin level of diabetic retinas was 1.19±0.05, showing a significant increase compared to the 1.06±0.03 of the control group retinas (P=0.016). The BMP2/actin level of diabetic crystalline lenses was similar to the control group crystalline lenses (P=0.730). CONCLUSIONS: Compared to control group pigs, the number of nuclei in the inner nuclear layer of retinas from streptozotocin-induced diabetic pigs decreased, while an increase in BMP2 expression was observed in the retina of diabetic pigs.
ABSTRACT
There are four prerequisites when applying all types of thin-film solar cells to power-generating window photovoltaics (PVs): high power-generation efficiency, longevity and high durability, semitransparency or partial-light transmittance, and colorful and aesthetic value. Solid-type thin-film Cu(In,Ga)S2 (CIGS) or Cu(In,Ga)(S,Se)2 (CIGSSe) PVs nearly meet the first two criteria, making them promising candidates for power-generating window applications if they can transmit light to some degree and generate color with good aesthetic value. In this study, the mechanical scribing process removes 10% of the window CIGSSe thin-film solar cell with vacant line patterns to provide a partial-light-transmitting CIGSSe PV module to meet the third requirement. The last concept of creating distinct colors could be met by the addition of reflectance colors of one-dimensional (1D) photonic crystal (PC) dichroic film on the black part of a partial-light-transmitting CIGSSe PV module. Beautiful violets and blues were created on the cover glass of a black CIGSSe PV module via the addition of 1D PC blue-mirror-yellow-pass dichroic film to improve the aesthetic value of the outside appearance. As a general result from the low external quantum efficiency (EQE) and absorption of CIGSSe PVs below a wavelength of 400 nm, the harvesting efficiency and short-circuit photocurrent of CIGSSe PVs were reduced by only â¼10% without reducing the open-circuit voltage (VOC) because of the reduced overlap between the absorption spectrum of CIGSSe PV and the reflectance spectrum of the 1D PC blue-mirror-yellow-pass dichroic film. The combined technology of partial-vacancy-scribed CIGSSe PV modules and blue 1D PC dichroic film can provide a simple strategy to be applied to violet/blue power-generating window applications, as such a strategy can improve the transparency and aesthetic value without significantly sacrificing the harvesting efficiency of the CIGSSe PV modules.
ABSTRACT
PURPOSE: To elucidate patterns of macular ganglion cell-inner plexiform layer (GCIPL) defects by Cirrus optical coherence tomography (OCT) and examine the spatial relationship between GCIPL defect and visual field (VF) defect patterns. METHODS: A total of 116 eyes of 116 normal subjects and 111 eyes of 111 glaucoma patients were included. The 227 study subjects underwent Cirrus OCT imaging in macular cube mode and reliable standard VF testing. Two ophthalmologists blindly classified GCIPL defect patterns and VF defects. The frequency distribution of GCIPL defect patterns and spatial relationships between GCIPL defects and VF defects were investigated. RESULTS: GCIPL defect patterns were classified as minimal, inner, outer, diffuse mild, diffuse severe, inferior confined, inferior dominant, superior confined, and superior dominant defects in normal controls (71.6%, 7.8%, 4.3%, 1.7%, 0%, 10.3%, 1.7%, 1.7%, and 0.9%, respectively) and in glaucoma patients (11.7%, 3.6%, 4.5%, 7.2%, 21.6%, 22.5%, 18.0%, 4.5%, and 6.3%, respectively). In mild and moderate glaucoma patients, the inferior confined type was most frequent (21.9% and 50.0%, respectively). However, the diffuse severe type was most frequent (59.1%) in advanced glaucoma patients. The locations of the VF defects corresponded to the locations of the GCIPL defects in glaucoma patients (P=0.012). CONCLUSIONS: Glaucomatous damage of the macula was common and more frequent in the inferior retina. GCIPL defect patterns as determined by SD-OCT imaging corresponded well with central VF defects. It seems macular GCIPL analysis may be useful for evaluating glaucomatous optic neuropathy.
Subject(s)
Glaucoma/diagnosis , Nerve Fibers/pathology , Optic Nerve Diseases/diagnosis , Retinal Ganglion Cells/pathology , Tomography, Optical Coherence/methods , Aged , Corneal Pachymetry , Female , Glaucoma/physiopathology , Healthy Volunteers , Humans , Intraocular Pressure/physiology , Male , Middle Aged , Refraction, Ocular/physiology , Tonometry, Ocular , Vision Disorders/physiopathology , Visual Field Tests , Visual Fields/physiologyABSTRACT
The goal of this study was to prove the possibility of using silicone stents broadly used for human medicine in canine obstructive tracheal disease. A silicone stent anatomically designed for canine trachea was tested on 5 beagle dogs for 8 weeks. The stent was carefully inserted using a newly developed delivery device under fluoroscopic guidance. There were no technical difficulties in placing the stent during the procedure. Previously reported complications of airway stenting such as stent migration or granulation tissue formation did not occur in any of the cases. In addition, removal of the stent was as simple as inserting it, and complications were absent. The stent introduced in this study could possibly be applied to various canine obstructive tracheal diseases.
Subject(s)
Dog Diseases/surgery , Silicones/therapeutic use , Stents/veterinary , Tracheal Diseases/veterinary , Animals , Bronchoscopy/instrumentation , Bronchoscopy/veterinary , Dog Diseases/diagnostic imaging , Dogs , Male , Pilot Projects , Radiography , Tracheal Diseases/diagnostic imaging , Tracheal Diseases/surgeryABSTRACT
Chronic alcohol consumption contributes to the development of type 2 diabetes mellitus (T2DM) while decreasing the level of brain-derived neurotrophic factor (BDNF). BDNF may be an important regulator of glucose metabolism, so it may be associated with an increased risk for T2DM in alcoholism. We evaluated the association of chronic heavy alcohol exposure, T2DM and BDNF level. Ten week-old type 2 diabetic OLETF rats and non-diabetic LETO rats of similar weight were used. The rats were randomized by weight into four treatment groups: (1) OLETF-Ethanol (O-E, n=13), (2) OLETF-Control (O-C, n=15), (3) LETO-Ethanol (L-E, n=11), and (4) LETO-Control (L-C, n=14). The ethanol groups were fed an isocaloric liquid diet containing ethanol while the control groups were fed with the same diet containing maltose-dextran over a 6-week period using a pair-feeding control model in order to regulate different caloric ingestion. After 6 weeks of feeding, an Intraperitoneal Glucose Tolerance Test (IP-GTT) was performed and BDNF levels were analyzed. Prior to IP-GTT, the mean glucose levels in the O-E, O-C, L-E, and L-C groups were 90.38±12.84, 102.13±5.04, 95.18±6.43, and 102.36±4.43mg/dL, respectively. Thirty minutes after intraperitoneal injection, the mean glucose levels were 262.62±63.77, 229.07±51.30, 163.45±26.63, and 156.64±34.42mg/dL, respectively; the increased amount of the mean glucose level in the O-E group was significantly higher than that in the O-C group (p<0.05). One hundred twenty minutes after intraperitoneal injection, the mean glucose levels were 167.38±45.37, 121.20±18.54, 106.73±6.94, and 104.57±9.49mg/dL, respectively; the increased amount of the mean glucose level in the O-E group was significantly higher than that in the O-C group (p<0.01). The difference in mean glucose levels between the O-E group and O-C group was still significant even after adjusting for time (p<0.05). Mean BDNF levels were 405.95±326.16, 618.23±462.15, 749.18±599.93, and 1172.00±839.17pg/mL, respectively; mean BDNF level in the O-E group was significantly lower than the L-C group (p<0.05). In conclusion, the results of the present study suggest that chronic heavy alcohol ingestion may aggravate T2DM and may possibly lower BDNF level.
Subject(s)
Alcoholism/blood , Brain-Derived Neurotrophic Factor/blood , Diabetes Mellitus, Type 2/blood , Ethanol/administration & dosage , Alcoholism/complications , Animals , Biomarkers/blood , Blood Glucose/metabolism , Chronic Disease , Diabetes Mellitus, Type 2/etiology , Male , Random Allocation , Rats , Rats, Inbred OLETF , Rats, Long-EvansABSTRACT
AIM: To investigate the relationships of serum testosterone, insulin-like growth factor (IGF)-1 and IGF-binding protein (IGFBP)-3 levels with prostate cancer risk and also with known prognostic parameters of prostate cancer in Korean men who received radical retropubic prostatectomy (RRP) for clinically-localized prostate cancer. METHODS: Serum levels of total testosterone, free testosterone, IGF-1 and IGFBP-3 were determined in 592 patients who subsequently received prostate biopsy. Results were compared between patients who eventually received RRP for prostate cancer (n=159) and those who were not diagnosed with prostate cancer from biopsy (control group, n=433). Among the prostate cancer only patients, serum hormonal levels obtained were analyzed in relation to serum prostate specific antigen (PSA), pathological T stage and pathological Gleason score. RESULTS: Prostate cancer patients and the control group demonstrated no significant differences regarding serum levels of total testosterone, free testosterone, IGF-1 and IGFBP-3 across the different age groups. Among the cancer only patients, no significant associations were observed for serum levels of total testosterone, free testosterone, IGF-1 and IGFBP-3 levels with pathological T stage, pathological Gleason score and preoperative PSA. CONCLUSION: Our data indicate that simple quantifications of serum testosterone and IGF-1 along with IGFBP-3 levels might not provide useful clinical information in the diagnosis of clinically localized prostate cancer in Korean men. Also, our results suggest that serum levels of testosterone, IGF-1 and IGFBP-3 might not be significantly associated with known prognostic factors of clinically localized prostate cancer in Korean men.