Stabilization of High-Pressure Phase of Face-Centered Cubic Lutetium Trihydride at Ambient Conditions.

Superconductivity at room temperature and near-ambient pressures is a highly sought-after phenomenon in physics and materials science. A recent study reported the presence of this phenomenon in N-doped lutetium hydride [Nature 615, 244 (2023)], however, subsequent experimental and theoretical investigations have yielded inconsistent results. This study undertakes a systematic examination of synthesis methods involving high temperatures and pressures, leading to insights into the impact of the reaction path on the products and the construction of a phase diagram for lutetium hydrides. Notably, the high-pressure phase of face-centered cubic LuH3 (fcc-LuH3) is maintained to ambient conditions through a high-temperature and high-pressure method. Based on temperature and anharmonic effects corrections, the lattice dynamic calculations demonstrate the stability of fcc-LuH3 at ambient conditions. However, no superconductivity is observed above 2 K in resistance and magnetization measurements in fcc-LuH3 at ambient pressure. This work establishes a comprehensive synthesis approach for lutetium hydrides, thereby enhancing the understanding of the high-temperature and high-pressure method employed in hydrides with superconductivity deeply.

One-Dimensional Non-coplanar Nitrogen Chains in Manganese Tetranitride under High Pressure.

Transition metal nitrides have great potential applications as incompressible and high energy density materials. Various polymeric nitrogen structures significantly affect their properties, contributing to their complex bonding modes and coordination conditions. Herein, we first report a new manganese polynitride MnN4 with bifacial trans-cis [N4]n chains by treating with high-pressure and high-temperature conditions in a diamond anvil cell. Our experiments reveal that MnN4 has a P-1 symmetry and could stabilize in the pressure range of 56-127 GPa. Detailed pressure-volume data and calculations of this phase indicate that MnN4 is a potential hard (255 GPa) and high energy density (2.97 kJ/g) material. The asymmetric interactions impel N1 and N4 atoms to hybridize to sp2-3, which causes distortions of [N4]n chains. This work discovers a new polynitride material, fills the gap for the study of manganese polynitride under high pressure, and offers some new insights into the formation of polymeric nitrogen structures.

Restoration of saturated outputs from microchannel plate photomultiplier tubes in sub-microsecond single-pulse-current mode.

Microchannel plate (MCP) photomultiplier tubes (PMTs) are frequently used in experimental diagnostics, where they are operated in single-pulse current measurement mode. However, considering the significant amplitude fluctuations in the measured signal, the resulting output signal from the MCP-PMT is inevitably distorted by gain saturation. Therefore, understanding the correlation between the MCP-PMT output signal and gain saturation is critical in assessing the extent of output signal distortion and determining the MCP-PMT saturation level. This knowledge allows for a more precise assessment of the input signal's features. In this paper, we present an experimental method for restoring the initial waveform from the saturated MCP-PMT signal. To correct the amplitude-drop caused by gain saturation, our technique involves calibrating the MCP-PMT's relative gain as a function of the accumulated output charge using a square-wave light source. We then applied this approach to restore a ∼500 ns saturated pulse from a double-layer 10 mm diameter MCP-PMT. The restored signal showed a deviation of less than 6% from the reference waveform, which validates the effectiveness of the technique.

RIT1 regulates mitosis and promotes proliferation by interacting with SMC3 and PDS5 in hepatocellular carcinoma.

BACKGROUND: As a small G protein of Ras family, Ras-like-without-CAAX-1 (RIT1) plays a critical role in various tumors. Our previous study has demonstrated the involvement of RIT1 in promoting malignant progression of hepatocellular carcinoma (HCC). However, its underlying mechanism remains unclear. METHODS: Gene set enrichment analysis (GSEA) was conducted in the TCGA LIHC cohort to investigate the underlying biological mechanism of RIT1. Live cell imaging, immunofluorescence (IF) and flow cytometry assays were used to verify biological function of RIT1 in HCC mitosis. Subcutaneous xenografting of human HCC cells in BALB/c nude mice was utilized to assess tumor proliferation in vivo. RNA-seq, co-immunoprecipitation (Co-IP), mass spectrometry analyses, western blot and IF assays were employed to elucidate the mechanisms by which RIT1 regulates mitosis and promotes proliferation in HCC. RESULTS: Our findings demonstrate that RIT1 plays a crucial role in regulating mitosis in HCC. Knockdown of RIT1 disrupts cell division, leading to G2/M phase arrest, mitotic catastrophe, and apoptosis in HCC cells. SMC3 is found to interact with RIT1 and knockdown of SMC3 attenuates the proliferative effects mediated by RIT1 both in vitro and in vivo. Mechanistically, RIT1 protects and maintains SMC3 acetylation by binding to SMC3 and PDS5 during mitosis, thereby promoting rapid cell division and proliferation in HCC. Notably, we have observed an upregulation of SMC3 expression in HCC tissues, which is associated with poor patient survival and promotion of HCC cell proliferation. Furthermore, there is a significant positive correlation between the expression levels of RIT1, SMC3, and PDS5. Importantly, HCC patients with high expression of both RIT1 and SMC3 exhibit worse prognosis compared to those with high RIT1 but low SMC3 expression. CONCLUSIONS: Our findings underscore the crucial role of RIT1 in regulating mitosis in HCC and further demonstrate its potential as a promising therapeutic target for HCC treatment.

Hybrid photonic deep convolutional residual spiking neural networks for text classification.

Spiking neural networks (SNNs) offer powerful computation capability due to its event-driven nature and temporal processing. However, it is still limited to shallow structure and simple tasks due to the training difficulty. In this work, we propose a deep convolutional residual spiking neural network (DCRSNN) for text classification tasks. In the DCRSNN, the feature extraction is achieved via a convolution SNN with residual connection, using the surrogate gradient direct training technique. Classification is performed by a fully-connected network. We also suggest a hybrid photonic DCRSNN, in which photonic SNNs are used for classification with a converted training method. The accuracy of hard and soft reset methods, as well as three different surrogate functions, were evaluated and compared across four different datasets. Results indicated a maximum accuracy of 76.36% for MR, 91.03% for AG News, 88.06% for IMDB and 93.99% for Yelp review polarity. Soft reset methods used in the deep convolutional SNN yielded slightly better accuracy than their hard reset counterparts. We also considered the effects of different pooling methods and observation time windows and found that the convergence accuracy achieved by convolutional SNNs was comparable to that of convolutional neural networks under the same conditions. Moreover, the hybrid photonic DCRSNN also shows comparable testing accuracy. This work provides new insights into extending the SNN applications in the field of text classification and natural language processing, which is interesting for the resources-restrained scenarios.

Drug Repurposing Flubendazole to Suppress Tumorigenicity via PCSK9-dependent Inhibition and Potentiate Lenvatinib Therapy for Hepatocellular Carcinoma.

Hepatocellular carcinoma (HCC) is one of the most lethal malignant cancers across the world. It has a poor prognosis and lacks effective therapies, especially for patients with advanced-stage cancer, indicating an urgent need for new therapies and novel therapeutic targets. Here, by screening the U.S. Food and Drug Administration drug library against HCC cell lines, we identified that flubendazole, a traditional anthelmintic drug, could prominently suppress HCC cells in vivo and in vitro. RNA sequence analysis and cellular thermal shift assays showed that flubendazole reduced the expression of PCSK9 protein by direct targeting. The increased expression of PCSK9 in HCC tissues was demonstrated to be correlated with poor prognosis, and the inhibitory ability of flubendazole was selectively dependent on PCSK9 expression. PCSK9 knockdown abolished the antitumor effects of flubendazole in HCC. Mechanistically, flubendazole inhibited the Hedgehog signaling pathway induced by PCSK9, resulting in the downregulation of smoothened (SMO) and GLI Family Zinc Finger 1 (Gli1). Moreover, combining flubendazole with lenvatinib was found more effective than administering lenvatinib only for HCC treatment in vivo and in vitro. These findings reveal the therapeutic potential of flubendazole against HCC and provide clues on new repurposed drugs and targets for cancer treatment.

Frontier in the diamond anvil cell techniques for ultrahigh pressure generation.

The diamond anvil cell (DAC) has become a principal tool for high-pressure research in muti-fields such as physics, earth, and planetary sciences, because of its ability to the realization of megabar pressures and thousands of degrees. Nevertheless, the strain on the culet of single crystal diamond at high loads leads to the conventional DAC having a 400 GPa limit. To date, based on the conventional DAC, several new designs were innovatively proposed, such as the double stage DAC (ds-DAC) and toroidal DAC. They are both capable to reach pressures above 600 GPa, and even static pressures of more than 1.0 TPa are achieved using ds-DAC. All these progresses promote the exploration of unique properties and new matters in a remarkable extended pressure range. Here, the typical characteristics and experimental methods of these interesting and important ultrahigh-pressure technologies are reviewed, the strengths and limitations are summarized, and an outlook on the development of ultrahigh-pressure technology is also provided. These exciting results will further stimulate breakthrough discoveries for ultrahigh-pressure studies.

[Predictive value of early lactate/albumin ratio in the prognosis of sepsis].

OBJECTIVE: To investigate the prognostic value of early serum lactate, albumin, and lactate/albumin ratio (L/A) on the 28-day prognosis of adult patients with sepsis. METHODS: A retrospective cohort study was conducted among adult patients with sepsis admitted to the First Affiliated Hospital of Xinjiang Medical University from January to December in 2020. Gender, age, comorbidities, lactate within 24 hours of admission, albumin, L/A, interleukin-6 (IL-6), procalcitonin (PCT), C-reactive protein (CRP) and 28-day prognosis were recorded. The receiver operator characteristic curve (ROC curve) was drawn to analyze the predictive value of lactate, albumin and L/A for 28-day mortality in patients with sepsis. Subgroup analysis of patients was performed according to the best cut-off value, Kaplan-Meier survival curves were drawn, and the 28-day cumulative survival of patients with sepsis was analyzed. RESULTS: A total of 274 patients with sepsis were included, and 122 patients died at 28 days, with a 28-day mortality of 44.53%. Compared with the survival group, the age, the proportion of pulmonary infection, the proportion of shock, lactate, L/A and IL-6 in the death group were significantly increased, and albumin was significantly decreased [age (years): 65 (51, 79) vs. 57 (48, 73), pulmonary infection: 75.4% vs. 53.3%, shock: 37.7% vs. 15.1%, lactate (mmol/L): 4.76 (2.95, 9.23) vs. 2.21 (1.44, 3.19), L/A: 0.18 (0.10, 0.35) vs. 0.08 (0.05, 0.11), IL-6 (ng/L): 337.00 (97.73, 2 318.50) vs. 55.88 (25.26, 150.65), albumin (g/L): 27.68 (21.02, 33.03) vs. 29.62 (25.25, 34.23), all P < 0.05]. The area under the ROC curve (AUC) and 95% confidence interval (95%CI) of lactate, albumin, and L/A were 0.794 (95%CI was 0.741-0.840), 0.589 (95%CI was 0.528-0.647), 0.807 (95%CI was 0.755-0.852) for predicting 28-day mortality in sepsis patients. The optimal diagnostic cut-off value of lactate was 4.07 mmol/L, the sensitivity was 57.38%, the specificity was 92.76%. The optimal diagnostic cut-off value of albumin was 22.28 g/L, the sensitivity was 31.15%, the specificity was 92.76%. The optimal diagnostic cut-off of L/A was 0.16, the sensitivity was 54.92%, and the specificity was 95.39%. Subgroup analysis showed that the 28-day mortality of sepsis patients in the L/A > 0.16 group was significantly higher than that in the L/A ≤ 0.16 group [90.5% (67/74) vs. 27.5% (55/200), P < 0.001]. The 28-day mortality of sepsis patients in the albumin ≤ 22.28 g/L group was significantly higher than that in the albumin > 22.28 g/L group [77.6% (38/49) vs. 37.3% (84/225), P < 0.001]. The 28-day mortality in the group with lactate > 4.07 mmol/L was significantly higher than that in the group with lactate ≤ 4.07 mmol/L [86.4% (70/81) vs. 26.9% (52/193), P < 0.001]. The three were consistent with the analysis results of Kaplan-Meier survival curve. CONCLUSIONS: The early serum lactate, albumin, and L/A were all valuable in predicting the 28-day prognosis of patients with sepsis, and L/A was better than lactate and albumin.

Regulation mechanism of ferroptosis and its research progress in tumor immunotherapy.

Ferroptosis is a novel regulatory cell death, which is characterized by iron dependency and mainly caused by accumulation of intracellular lipid peroxides and reactive oxygen species. Ferroptosis plays an important role in the occurrence and development of a variety of malignant tumors, especially in anti-tumor treatment. As an emerging treatment method, the immunotherapy has been widely applied in the clinical practice, and the role of ferroptosis in tumor immunotherapy has been gradually explored. This study aims to illustrate the features of ferroptosis, and its role in anti-tumor immunotherapy and potential clinical application.

Stabilization of hexazine rings in potassium polynitride at high pressure.

Polynitrogen molecules are attractive for high-energy-density materials due to energy stored in nitrogen-nitrogen bonds; however, it remains challenging to find energy-efficient synthetic routes and stabilization mechanisms for these compounds. Direct synthesis from molecular dinitrogen requires overcoming large activation barriers and the reaction products are prone to inherent inhomogeneity. Here we report the synthesis of planar N62- hexazine dianions, stabilized in K2N6, from potassium azide (KN3) on laser heating in a diamond anvil cell at pressures above 45 GPa. The resulting K2N6, which exhibits a metallic lustre, remains metastable down to 20 GPa. Synchrotron X-ray diffraction and Raman spectroscopy were used to identify this material, through good agreement with the theoretically predicted structural, vibrational and electronic properties for K2N6. The N62- rings characterized here are likely to be present in other high-energy-density materials stabilized by pressure. Under 30 GPa, an unusual N20.75--containing compound with the formula K3(N2)4 was formed instead.

Genome optimization via virtual simulation to accelerate maize hybrid breeding.

The employment of doubled-haploid (DH) technology in maize has vastly accelerated the efficiency of developing inbred lines. The selection of superior lines has to rely on genotypes with genomic selection (GS) model, rather than phenotypes due to the high expense of field phenotyping. In this work, we implemented 'genome optimization via virtual simulation (GOVS)' using the genotype and phenotype data of 1404 maize lines and their F1 progeny. GOVS simulates a virtual genome encompassing the most abundant 'optimal genotypes' or 'advantageous alleles' in a genetic pool. Such a virtually optimized genome, although can never be developed in reality, may help plot the optimal route to direct breeding decisions. GOVS assists in the selection of superior lines based on the genomic fragments that a line contributes to the simulated genome. The assumption is that the more fragments of optimal genotypes a line contributes to the assembly, the higher the likelihood of the line favored in the F1 phenotype, e.g. grain yield. Compared to traditional GS method, GOVS-assisted selection may avoid using an arbitrary threshold for the predicted F1 yield to assist selection. Additionally, the selected lines contributed complementary sets of advantageous alleles to the virtual genome. This feature facilitates plotting the optimal route for DH production, whereby the fewest lines and F1 combinations are needed to pyramid a maximum number of advantageous alleles in the new DH lines. In summary, incorporation of DH production, GS and genome optimization will ultimately improve genomically designed breeding in maize. Short abstract: Doubled-haploid (DH) technology has been widely applied in maize breeding industry, as it greatly shortens the period of developing homozygous inbred lines via bypassing several rounds of self-crossing. The current challenge is how to efficiently screen the large volume of inbred lines based on genotypes. We present the toolbox of genome optimization via virtual simulation (GOVS), which complements the traditional genomic selection model. GOVS simulates a virtual genome encompassing the most abundant 'optimal genotypes' in a breeding population, and then assists in selection of superior lines based on the genomic fragments that a line contributes to the simulated genome. Availability of GOVS (https://govs-pack.github.io/) to the public may ultimately facilitate genomically designed breeding in maize.

Flavonoids derived from Anemarrhenae Rhizoma ameliorate inflammation of benign prostatic hyperplasia via modulating COX/LOX pathways.

ETHNOPHARMACOLOGICAL RELEVANCE: Flavonoids are the main components of the traditional Chinese medicine Anemarrhenae Rhizoma (dried rhizome of Anemarrhena asphodeloides Bge.), which has been reported to possess activity against inflammation and tumor. AIM OF STUDY: Regulation of the arachidonic acid (AA) cascade through cyclooxygenase (COX) and lipoxygenase (LOX) represent the two major pathways to treat inflammatory of benign prostatic hyperplasia (BPH). In this study, Anemarrhenae Rhizoma flavonoids and its main compounds (mangiferin, neomangiferin and isomangiferin) were investigated for effects on AA metabolism. METHODS: Ultra-performance liquid chromatography tandem mass spectrometry (UPLC-MS/MS) was used to monitor AA metabolites in BPH rats and in PC-3 cells. COX-2 and 5-LOX protein and mRNA levels were measured by Western blot and qPCR, respectively, along with histopathological assessment of prostate tissues. RESULTS: Treatment with flavonoids significantly ameliorated BPH-associated prostate inflammation and inhibited the expression of COX-2 and 5-LOX at the protein and mRNA levels. Quantitative metabolomic analysis of blood plasma showed flavonoids treatment decreased AA levels and its metabolites associated with the COX and LOX pathways. Further exploration of the flavonoid compounds mangiferin, neomangiferin and isomangiferin showed they inhibited AA metabolism to varying degrees in PC-3 cell cultures. CONCLUSION: Anemarrhenae Rhizoma flavonoids act to inhibit BPH-related inflammation in vivo and in vitro by targeting AA metabolism and interfering with COX and LOX pathways. The identification of mangiferin, neomangiferin and isomangiferin as anti-inflammatory components suggests flavonoids interventions represent a promising therapeutic approach for BPH.

Compression-Driven Internanocluster Reaction for Synthesis of Unconventional Gold Nanoclusters.

Can the active kernels in ultrasmall metal nanoparticles (nanoclusters, NCs) react with one another, or can the internanocluster reaction occur when they are in close enough proximity? To resolve this fundamental issue, we investigated the solid-state internanocluster reaction of the most studied gold NC Au25 (SR)18 (SR: thiolate). A novel NC was produced by increasing the pressure to 5 GPa, whose composition was determined to be Au32 (SC2 H4 Ph)24 by mass spectrometry and thermogravimetric analysis. As revealed by single-crystal X-ray crystallography, the structure, a bicuboid Au14 kernel and three pairs of interlocked trimetric staples, has not been reported and endows the NC with obvious photoluminescence. DFT calculations indicate that the staples contribute substantially to the absorption properties. Further experiments reveal the pressure (internanocluster distance) can tune the internanocluster reaction, and the resulting product is not necessarily the thermodynamic product.

Hydrogen Pentagraphenelike Structure Stabilized by Hafnium: A High-Temperature Conventional Superconductor.

The recent discovery of H_{3}S and LaH_{10} superconductors with record high superconducting transition temperatures T_{c} at high pressure has fueled the search for room-temperature superconductivity in the compressed superhydrides. Here we introduce a new class of high T_{c} hydrides with a novel structure and unusual properties. We predict the existence of an unprecedented hexagonal HfH_{10}, with remarkably high value of T_{c} (around 213-234 K) at 250 GPa. As concerns the novel structure, the H ions in HfH_{10} are arranged in clusters to form a planar "pentagraphenelike" sublattice. The layered arrangement of these planar units is entirely different from the covalent sixfold cubic structure in H_{3}S and clathratelike structure in LaH_{10}. The Hf atom acts as a precompressor and electron donor to the hydrogen sublattice. This pentagraphenelike H_{10} structure is also found in ZrH_{10}, ScH_{10}, and LuH_{10} at high pressure, each material showing a high T_{c} ranging from 134 to 220 K. Our study of dense superhydrides with pentagraphenelike layered structures opens the door to the exploration of a new class of high T_{c} superconductors.

Distance makes a difference in crystalline photoluminescence.

Crystallization-induced photoluminescence weakening was recently revealed in ultrasmall metal nanoparticles. However, the fundamentals of the phenomenon are not understood yet. By obtaining conformational isomer crystals of gold nanoclusters, we investigate crystallization-induced photoluminescence weakening and reveal that the shortening of interparticle distance decreases photoluminescence, which is further supported by high-pressure photoluminescence experiments. To interpret this, we propose a distance-dependent non-radiative transfer model of excitation electrons and support it with additional theoretical and experimental results. This model can also explain both aggregation-induced quenching and aggregation-induced emission phenomena. This work improves our understanding of aggregated-state photoluminescence, contributes to the concept of conformational isomerism in nanoclusters, and demonstrates the utility of high pressure studies in nanochemistry.

300 kV/6 mA integrated Cockcroft-Walton high voltage power supply for a compact neutron generator.

A high voltage power supply system for a compact neutron generator is developed. A four-stage symmetrical Cockcroft-Walton voltage multiplier circuit is adopted to produce 300 kV direct current high voltage. A two-stage 360 kV isolation transformer system is used to drive the ion source power supply. The high voltage power supply and the isolation transformer system are integrated in an epoxy bucket with a size of ϕ360 × 700 containing No. 25 transformer oil. The maximum output voltage of the power supply can reach 300 kV. The variation in the high voltage power supply is less than 0.5% when the power supply works at 300 kV/6 mA with an input voltage variation of ±8%. Meanwhile, the isolation transformer system can withstand more than 360 kV, with its output power being about 2.5 kW. No overvoltage protection devices are used in the power supply, and the protection resistors are connected in series to two pairs of rectifiers at the highest and lowest potential terminals and to the output terminal of the voltage multiplier to prevent overcurrent.

Polymorphism of polymeric nitrogen at high pressures.

Polymeric nitrogen at 120 GPa-180 GPa is known in two monatomic crystalline cubic gauche (cg-N) and layered polymeric (LP-N) phases and one amorphous modification (η-N), and all these high-pressure phases attract considerable attention for their potential application as a high energy density material. Here, we investigated the stability of these modifications at high pressures in the laser heated diamond anvil cell upon decompression from 161 GPa. Pure LP-N was synthesized above 152 GPa upon laser heating of η-N to 2500 K, while cg-N forms below 150 GPa. Upon laser heating at 129 GPa and 123 GPa, the LP-N clearly diminished, indicating that the synthesis of cg-N becomes more favorable in a mixed phase region below 129 GPa. Upon unloading, cg-N and LP-N were metastable to at least 71 GPa at up to 2500 K and at room temperature, respectively. These observations clarified a complicated polymorphism of monatomic nitrogen at high pressures and large hysteretic phenomena related to a transition to nonmolecular nitrogen.

Subchronic feeding toxicity studies of drought-tolerant transgenic wheat MGX11-10 in Wistar Han RCC rats.

A subchronic toxicity study were conducted in Wistar Han RCC rats to evaluate the potential health effects of genetically modified (GM), drought-tolerant wheat MGX11-10. Rats were fed a rodent diet formulated with MGX11-10 and were compared with rats fed a diet formulated with its corresponding non-transgenic control Jimai22 and rats fed a basal diet. MGX11-10 and Jimai22 were ground into flour and formulated into diets at concentrations of 16.25, 32.5, or 65%, w/w% and fed to rats (10/sex/group) for 13 weeks. Compared with rats fed Jimai22 and the basal-diet group, no biologically relevant differences were observed in rats fed the GM diet with respect to body weight/gain, food consumption/efficiency, clinical signs, mortality, ophthalmology, clinical pathology (hematology, prothrombin time, urinalysis, clinical chemistry), organ weights, and gross and microscopic pathology. Under the conditions of this study, the MGX11-10 diets did not cause any treatment-related effects in rats following at least 90 days of dietary administration as compared with rats fed diets with the corresponding non-transgenic control diet and the basal-diet group. The MGX11-10 diets are considered equivalent to the diets prepared from conventional comparators. The results demonstrated that MGX11-10 wheat is as safe and wholesome as the corresponding non-transgenic control wheat.

A Spectroscopic Study of the Insulator-Metal Transition in Liquid Hydrogen and Deuterium.

The insulator-to-metal transition in dense fluid hydrogen is an essential phenomenon in the study of gas giant planetary interiors and the physical and chemical behavior of highly compressed condensed matter. Using direct fast laser spectroscopy techniques to probe hydrogen and deuterium precompressed in a diamond anvil cell and laser heated on microsecond timescales, an onset of metal-like reflectance is observed in the visible spectral range at P >150 GPa and T ≥ 3000 K. The reflectance increases rapidly with decreasing photon energy indicating free-electron metallic behavior with a plasma edge in the visible spectral range at high temperatures. The reflectance spectra also suggest much longer electronic collision time (≥1 fs) than previously inferred, implying that metallic hydrogen at the conditions studied is not in the regime of saturated conductivity (Mott-Ioffe-Regel limit). The results confirm the existence of a semiconducting intermediate fluid hydrogen state en route to metallization.

Use of a Sparse-Response Deep Belief Network and Extreme Learning Machine to Discriminate Alzheimer's Disease, Mild Cognitive Impairment, and Normal Controls Based on Amyloid PET/MRI Images.

In recent years, interest has grown in using computer-aided diagnosis (CAD) for Alzheimer's disease (AD) and its prodromal stage, mild cognitive impairment (MCI). However, existing CAD technologies often overfit data and have poor generalizability. In this study, we proposed a sparse-response deep belief network (SR-DBN) model based on rate distortion (RD) theory and an extreme learning machine (ELM) model to distinguish AD, MCI, and normal controls (NC). We used [18F]-AV45 positron emission computed tomography (PET) and magnetic resonance imaging (MRI) images from 340 subjects enrolled in the ADNI database, including 116 AD, 82 MCI, and 142 NC subjects. The model was evaluated using five-fold cross-validation. In the whole model, fast principal component analysis (PCA) served as a dimension reduction algorithm. An SR-DBN extracted features from the images, and an ELM obtained the classification. Furthermore, to evaluate the effectiveness of our method, we performed comparative trials. In contrast experiment 1, the ELM was replaced by a support vector machine (SVM). Contrast experiment 2 adopted DBN without sparsity. Contrast experiment 3 consisted of fast PCA and an ELM. Contrast experiment 4 used a classic convolutional neural network (CNN) to classify AD. Accuracy, sensitivity, specificity, and area under the curve (AUC) were examined to validate the results. Our model achieved 91.68% accuracy, 95.47% sensitivity, 86.68% specificity, and an AUC of 0.87 separating between AD and NC groups; 87.25% accuracy, 79.74% sensitivity, 91.58% specificity, and an AUC of 0.79 separating MCI and NC groups; and 80.35% accuracy, 85.65% sensitivity, 72.98% specificity, and an AUC of 0.71 separating AD and MCI groups, which gave better classification than other models assessed.