Predicting the potential risk of Caragana shrub encroachment in the Eurasian steppe under anthropogenic climate change.

Climate change and human activities drive widespread shrub encroachment in global grassland ecosystems, particularly in the Eurasian steppe. Caragana shrubs, the primary contributors to shrub encroachment in this region, play a crucial role in shaping the ecosystem's structure and function. Future changes in the suitable distribution range of Caragana species will directly affect the ecological security and sustainable socio-economic development of the Eurasian steppe ecosystem. We used an ensemble modeling approach to predict Caragana shrub-dominated plant communities' current and future distribution in three major steppe subregions: the Black Sea-Kazakhstan steppe, the Tibetan Plateau steppe, and the Central Asian steppe. We assessed the potential risk of Caragana shrub encroachment by predicting changes in the suitable distribution area of 19 Caragana shrub species under future climate changes. Our research findings suggest that the expansion of Caragana species in different subregions of the Eurasian steppe is influenced by the effects of climate change in various ways. The distribution of Caragana species is primarily influenced by precipitation and temperature, and the global human modification (ghm) has a significant impact on the Central Asian and Tibetan Plateau subregions. Minimal changes are expected in the Black Sea-Kazakhstan subregion, a slight increase on the Tibetan Plateau, and a substantial rise in the Central Asian subregion, which suggests a higher potential risk of Caragana species shrub encroachment in that area. Our research provides valuable insights into the response of Caragana shrub encroachment to changing climates and human activities. It also has implications for the sustainable management of different areas of the vast Eurasian steppe ecosystem.

Distinguishing Laterality in Brain Injury in Rabbit Fetal Magnetic Resonance Imaging Using Novel Volume Rendering Techniques.

INTRODUCTION: Our laboratory has been exploring the MRI detection of fetal brain injury, which previously provided a prognostic biomarker for newborn hypertonia in an animal model of cerebral palsy (CP). The biomarker relies on distinct patterns of diffusion-weighted imaging-defined apparent diffusion coefficient (ADC) in fetal brains during uterine hypoxia-ischemia (H-I). Despite the challenges posed by small brains and tissue acquisition, our objective was to differentiate between left and right brain ADC changes. METHODS: A novel aspect involved utilizing three-dimensional rendering techniques to refine ADC measurements within spheroids encompassing fetal brain tissue. 25-day gestation age of rabbit fetuses underwent global hypoxia due to maternal uterine ischemia. RESULTS: Successful differentiation of left and right brain regions was achieved in 28% of the fetal brains. Ordinal analysis revealed predominantly higher ADC on the left side compared to the right at baseline and across the entire time series. During H-I and reperfusion-reoxygenation, the right side exhibited a favored percentage change. Among these fetal brains, 73% exhibited the ADC pattern predictive of hypertonia. No significant differences between left and right sides were observed in patterns predicting hypertonia, except for one timepoint during H-I. This study also highlights a balance between left-sided and right-sided alterations within the population. CONCLUSION: This study emphasizes the importance of investigating laterality and asymmetric hemispheric lesions for early diagnosis of brain injury, leading to CP. The technological limitations in obtaining a clear picture of the entire fetal brain for every fetus mirror the challenges encountered in human studies.

Effects of plant diversity and community structure on ecosystem multifunctionality under different grazing potentials in the eastern Eurasian steppe.

Grazing potential represents the potential carrying capacity of steppe livestock production. Understanding the impact of changes in plant diversity and community structure on ecosystem multifunctionality (EMF) at different grazing potentials is crucial for the sustainable management of steppe ecosystems. We examined the associations between plant diversity, community structure, above-ground ecosystem multifunctionality (AEMF), and below-ground ecosystem multifunctionality (BEMF) at various grazing potentials. Our assessment employed generalized linear mixed-effects models and structural equation models to determine the impact of these factors on ecosystem multifunctionality. Our study results indicated that ecosystem multifunctionality differed depending on the level of grazing potential and decreased as grazing potential declined. The impact of plant diversity and community structure on above- and below-ground ecosystem multifunctionality varied. Plant diversity and community structure correlated more with AEMF than BEMF. Plant diversity had the most significant effect on EMF under high grazing potential, while community structure had the greatest effect on EMF under moderate and low grazing potential. These improve our understanding of the correlation between steppe plant diversity, community structure, and above- and below-ground ecosystem multifunctionality. This understanding is necessary to develop strategies to increase plant diversity or regulate community structure and the sustainability of steppes.

Rodent disturbance reduces ecosystem stability through regulating plant and soil functions in Hulun Buir steppe.

Brandt's vole (Lasiopodomys brandtii), a typical rodent in the eastern Eurasian Steppe, has unclear impacts on ecosystem stability. In our field study in the Hulun Buir steppe, a multifunctional grazing ecosystem in this region, we used burrow entrance area and burrow density as alternative disturbance indices to derive a Disturbance Index (DI) for quantifying disturbance levels from rodents, and employed generalized linear mixed-effects model and the N-dimensional hypervolume framework to assess the influence of Brandt's vole disturbance on plant and soil functions, and then on the ecosystem functional stability. Our findings unequivocally illustrate that various plant functions including vegetation cover (Cover), aboveground biomass (ABG) and shoot carbon (ShootC) significantly declined with increasing disturbance, while shoot nitrogen (ShootN) and root nitrogen (RootN) show significantly positive responses. Soil functions such as soil nitrogen (SoilN), soil phosphorus (SoilP) and soil organic carbon (SoilC) showed significantly negative responses. Notably, the burrow entrance area exerts a more pronounced impact on both plant and soil functions in comparison to burrow density. Additionally, both disturbance indicators have a more significant influence on plant functions than on soil functions. Overall, the ecosystem functional stability progressively decreases with intensified disturbance, with varying response patterns for plant and soil functions, the former exhibited heightened stability as disturbance intensified, while the latter proved more stable at moderate disturbance levels. Our findings suggest that plant functions were more susceptible to disturbance by Brandt's vole compared to soils. Additionally, an ecosystem destabilization was synchronized with increasing Brandt's vole disturbance, although alterations in the functional stability of plants and soil show a different pattern.

Development of A New Scoring System in Higher Animals for Testing Cognitive Function in the Newborn Period: Effect of Prenatal Hypoxia-Ischemia.

Introduction Enhanced models for assessing cognitive function in the neonatal period are imperative in higher animals. Postnatal motor deficits, characteristic of cerebral palsy, emerge in newborn kits within our prenatal-rabbit model of hypoxia-ischemia (HI). In humans, prenatal HI leads to intellectual disability and cerebral palsy. In a study examining cognitive function in newborn rabbits, we explored several questions. Is there a distinction between conditioned and unconditioned kits? Can the kits discern the human face or the lab coat? Do motorically-normal kits, born after prenatal HI, exhibit cognitive deficits? Methods The conditioning protocol was randomly assigned to kits from each litter. For conditioning, the same human, wearing a lab coat, fed the rabbit kits for 9 days before the cognitive test. The 6-arm radial maze was chosen for its simplicity and ease of use. Normally appearing kits, born after uterine ischemia at 79% or 92% term in New Zealand White rabbits, were compared to Naïve kits. On postpartum day 22/23 or 29/30, the 6-arm maze helped determine if the kits recognized the original feeder from bystander (Test-1) or the lab coat on bystander (Test-2). The use of masks of feeder/bystander (Test-3) assessed confounding cues. A weighted score was devised to address variability in entry to maze arms, time, and repeated-trial learning. Results In conditioned kits, both Naïve and HI kits exhibited a significant preference for the face of the feeder, but not the lab coat. Cognitive deficits were minimal in normal-appearing HI kits. Conclusion The weighted score system was amenable to statistical manipulation.

Toward Lung Ventilation Imaging Using Hyperpolarized Diethyl Ether Gas Contrast Agent.

Hyperpolarized 129Xe gas was FDA-approved as an inhalable contrast agent for magnetic resonance imaging of a wide range of pulmonary diseases in December 2022. Despite the remarkable success in clinical research settings, the widespread clinical translation of HP 129Xe gas faces two critical challenges: the high cost of the relatively low-throughput hyperpolarization equipment and the lack of 129Xe imaging capability on clinical MRI scanners, which have narrow-bandwidth electronics designed only for proton (1H) imaging. To solve this translational grand challenge of gaseous hyperpolarized MRI contrast agents, here we demonstrate the utility of batch-mode production of proton-hyperpolarized diethyl ether gas via heterogeneous pairwise addition of parahydrogen to ethyl vinyl ether. An approximately 0.1-liter bolus of hyperpolarized diethyl ether gas was produced in 1 second and injected in excised rabbit lungs. Lung ventilation imaging was performed using sub-second 2D MRI with up to 2×2 mm2 in-plane resolution using a clinical 0.35 T MRI scanner without any modifications. This feasibility demonstration paves the way for the use of inhalable diethyl ether as a gaseous contrast agent for pulmonary MRI applications using any clinical MRI scanner.

Hematoma Enlargement After Intracerebral Hemorrhage: A Bibliometric Analysis.

OBJECTIVE: To conduct a quantitative analysis of published studies on hematoma enlargement after intracerebral hemorrhage. METHODS: Studies on hematoma enlargement after cerebral hemorrhage were retrieved from the Web of Science database on June 30, 2023. Microsoft Excel, VOSviewer, and CiteSpace software were used for bibliometric analysis and visualization, focusing on the quantitative characteristics of the literature. RESULTS: A total of 444 articles were published in 161 journals, with 2161 authors from 41 countries and 717 institutions. The most published authors, countries, and institutions were Goldstein, the USA, and Massachusetts General Hospital. Stroke published the most studies, but the average citation number per year of Lancet Neurology far exceeded that of other journals. The research field of hematoma enlargement is mainly divided into 3 focuses, including mechanisms, identification (computed tomography signs, predictive models), and treatment (hemostasis, antihypertensive therapy). Most bursts in publication number have been since 2010, where the highest burst was from research on spot signs, and the latest burst focused on tranexamic acid. Treatment using tranexamic acid based on different computed tomography signs is a focus of current research, but the effectiveness still requires further exploration. CONCLUSIONS: This bibliometric analysis analyzed the research framework and hotspots on hematoma enlargement after cerebral hemorrhage, which can help researchers better understand this field and provide potential suggestions for collaborations and research.

Development of a Skeletal Mechanism of a Four-Component Diesel Surrogate Fuel Using the Decoupling Method.

Alkylcyclohexanes with a long alkyl chain account for more than 30% of diesel fuel but seldom used in the oxidation mechanism of diesel surrogate fuel due to the lack of a reduced skeletal mechanism. Hence, a four-component diesel surrogate fuel was developed with n-butylcyclohexane (NBCH) as the representative of alkylcyclohexanes with a long alkyl chain in real diesel. The surrogate fuel can reproduce the physicochemical characteristics of real diesel, especially the distillation range. The reduced mechanism of NBCH was developed, and the skeletal mechanism of the surrogate fuel was formulated including 80 species and 251 reactions based on the decoupling method. The mechanism was validated under a wide range of conditions with the experimental results of ignition delay time (IDT), laminar flame speed, and species concentrations of both pure components and diesel. The accuracy of the mechanism on the spray and ignition performance was further validated against the experimental data obtained in a constant volume combustion chamber system. The calculated results showed a satisfactory agreement, in which the maximum error of flame lift-off length is 7.82 mm and that of IDTs is 0.16 ms. It was proven that the mechanism is suitable to reproduce the physicochemical properties of diesel and further predict the diesel spray and ignition performance.

Effect of Climate and Competition on Radial Growth of Pinus sylvestris var. mongolica Forest in Hulunbuir Sandy Land of Inner Mongolia, China.

(1) Background: The forest of Pinus sylvestris var. mongolica is an important semi-arid ecosystem in Hulunbuir sandy land that plays a key role in the carbon cycle and wind erosion control. It is crucial to explore the main factors affecting the radial growth of trees of P. sylvestris var. mongolica. (2) Methods: The study established the tree-ring chronology of P. sylvestris var. mongolica and analyzed the relationships among the radial growth, competition index, and climate variables using correlation analysis and a linear mixed effect model to explore the influence of competition and climate on radial growth of P. sylvestris var. mongolica. (3) Results: The results indicated that tree growth is mainly affected by the maximum average temperature (Tmax) and precipitation in June and July of the current year and that tree growth significantly decreased with increasing competition pressure. Analysis of the linear mixed effect model showed that tree age, competition intensity, self-calibrating Palmer drought severity index (scPDSI) from May to July, and vapor pressure deficit (VPD) have a significant impact on radial growth. (4) Conclusions: The competition plays a dominant role in radial growth of P. sylvestris var. mongolica compared to climate factors. This study helps to understand the growth mechanism of P. sylvestris var. mongolica forests under climate change and provides a scientific basis for effective management of semi-arid forests.

Shrub encroachment alters plant trait response to nitrogen addition in a semi-arid grassland.

Encroachment of shrubs over large regions of arid and semi-arid grassland can affect grassland traits and growth under a background of increasing nitrogen (N) deposition. However, the effects of N input rates on species traits and the growth of shrubs on grasslands remain unclear. We examined the effects of six different N addition rates on the traits of Leymus chinensis in an Inner Mongolia grassland encroached by the leguminous shrub, Caragana microphylla. We randomly selected 20 healthy L. chinensis tillers within shrubs and 20 tillers between shrubs in each plot, measuring the plant height, number of leaves, leaf area, leaf N concentration per unit mass (LNCmass), and aboveground biomass. Our results showed that N addition significantly enhanced the LNCmass of L. chinensis. The aboveground biomass, heights, LNCmass, leaf area, and leaf number of plants within the shrubs were higher than those between shrubs. For L. chinensis growing between shrubs, the LNCmass and leaf area increased with N addition rates, leaf number and plant height had binomial linear relationships to N addition rates. However, the number of leaves, leaf areas and heights of plants within shrubs did not vary under various N addition rates. Structural Equation Modelling revealed N addition had an indirect effect on the leaf dry mass through the accumulation of LNCmass. These results indicate that the response of dominant species to N addition could be regulated by shrub encroachment and provide new insights into management of shrub encroached grassland in the context of N deposition.

The CDK inhibitor AT7519 inhibits human glioblastoma cell growth by inducing apoptosis, pyroptosis and cell cycle arrest.

Glioblastoma multiforme (GBM) is the most lethal primary brain tumor with a poor median survival of less than 15 months. However, clinical strategies and effective therapies are limited. Here, we found that the second-generation small molecule multi-CDK inhibitor AT7519 is a potential drug for GBM treatment according to high-throughput screening via the Approved Drug Library and Clinical Compound Library (2718 compounds). We found that AT7519 significantly inhibited the cell viability and proliferation of U87MG, U251, and patient-derived primary GBM cells in a dose-dependent manner. Furthermore, AT7519 also inhibited the phosphorylation of CDK1/2 and arrested the cell cycle at the G1-S and G2-M phases. More importantly, AT7519 induced intrinsic apoptosis and pyroptosis via caspase-3-mediated cleavage of gasdermin E (GSDME). In the glioblastoma intracranial and subcutaneous xenograft assays, tumor volume was significantly reduced after treatment with AT7519. In summary, AT7519 induces cell death through multiple pathways and inhibits glioblastoma growth, indicating that AT7519 is a potential chemical available for GBM treatment.

Comparison of dural puncture and dural incision in deep brain stimulation surgery: A simple but worthwhile technique modification.

Background: The accuracy of the deep brain stimulation (DBS) electrode placement is influenced by a myriad of factors, among which pneumocephalus and loss of cerebrospinal fluid that occurs with dural opening during the surgery are considered most important. This study aimed to describe an effective method for decreasing pneumocephalus by comparing its clinical efficacy between the two different methods of opening the dura. Materials and methods: We retrospectively compared two different methods of opening the dura in 108 patients who underwent bilateral DBS surgery in our center. The dural incision group comprised 125 hemispheres (58 bilateral and 9 unilateral) and the dural puncture group comprised 91 (41 bilateral and 9 unilateral). The volume of intracranial air, dural opening time, intraoperative microelectrode recordings (MERs), postoperative electrode displacement, clinical efficacy, and complications were examined. Spearman correlation analysis was employed to identify factors associated with the volume of intracranial air and postoperative electrode displacement. Results: The volume of intracranial air was significantly lower (0.35 cm3 vs. 5.90 cm3) and dural opening time was significantly shorter (11s vs. 35s) in the dural puncture group. The volume of intracranial air positively correlated with dural opening time. During surgery, the sensorimotor area was longer (2.47 ± 1.36 mm vs. 1.92 ± 1.42 mm) and MERs were more stable (81.82% vs. 47.73%) in the dural puncture group. Length of the sensorimotor area correlated negatively with the volume of intracranial air. As intracranial air was absorbed after surgery, significant anterior, lateral, and ventral electrode displacement occurred; the differences between the two groups were significant (total electrode displacement, 1.0mm vs. 1.4mm). Electrode displacement correlated positively with the volume of intracranial air. Clinical efficacy was better in the dural puncture group than the dural incision group (52.37% ± 16.18% vs. 43.93% ± 24.50%), although the difference was not significant. Conclusion: Our data support the hypothesis that opening the dura via puncture rather than incision when performing DBS surgery reduces pneumocephalus, shortens dural opening time, enables longer sensorimotor area and more stable MERs, minimizes postoperative electrode displacement, and may permit a better clinical efficacy.

Tree characteristics and drought severity modulate the growth resilience of natural Mongolian pine to extreme drought episodes.

Global climate change and the increase in the frequency and intensity of drought have led to widespread forest decline and tree mortality. Studying the resilience components of tree growth to drought, including resistance (Rt), recovery (Rc), and resilience (Rs) and the influencing factors, helps assess forests' production and ecological stability under a changing climate. This study analyzed the responses of three resilience components of natural Mongolian pine (Pinus sylvestris var. mongolica) to drought events by examining individual-tree characteristics in two sites of Hulunbuir using the linear mixed effect model. The result showed that drought severity, diameter at breast height (dbh), pre-drought growth, and growth variability prior to drought had significant effects on the three resilience components of Mongolian pine growth. Specifically, as drought severity, dbh and growth variability increased, the Rt and Rs decreased, but Rc increased, showing a trade-off relationship with Rt. However, the Rt, Rc, and Rs decreased with pre-drought growth. Inter-tree competition and tree age also significantly impacted two resilience components. Besides, the interaction term between tree competition and tree age negatively affects Rt and Rs but positively affects Rc. Our findings highlight the influence of drought severity and individual-tree characteristics on drought resilience components, which can serve the adaptive management of natural Mongolian pine forests in the future.

Mimicking partial to total placental insufficiency in a rabbit model of cerebral palsy.

All placental abruptions begin as partial abruptions, which sometimes manifest as fetal bradycardia. The progression from partial to total abruption was mimicked by a new rabbit model of placental insufficiency, and we compared it, with sufficient statistical power, with the previous model mimicking total placental abruption. The previous model uses total uterine ischemia at E22 or E25 (70% or 79% term, respectively), in pregnant New Zealand white rabbits for 40 min (Full H-I). The new model, Partial+Full H-I, added a 30-min partial ischemia before the 40-min total ischemia. Fetuses were delivered either at E31.5 (full term) vaginally for neurobehavior testing, or by C-section at E25 for ex vivo brain cell viability evaluation. The onset of fetal bradycardia was within the first 2 min of either H-I protocol. There was no difference between Full H-I (n = 442 for E22, 312 for E25) and Partial+Full H-I (n = 154 and 80) groups in death or severely affected kits at E22 (76% vs. 79%) or at E25 (66% vs. 64%), or normal kits at E22 or E25, or any of the individual newborn neurobehavioral tests at any age. No sex differences were found. Partial+Full H-I (n = 6) showed less cell viability than Full H-I (n = 8) at 72-hr ex vivo in the brain regions studied. Partial+Full H-I insult produced similar cerebral palsy phenotype as our previous Full H-I model in a sufficiently powered study and may be more suitable for testing of potential neuroprotectants.

A systematic review of noninflammatory cerebrospinal fluid biomarkers for clinical outcome in neonates with perinatal hypoxic brain injury that could be biologically significant.

Neonatal encephalopathy (NE) that purportedly arises from hypoxia-ischemia is labeled hypoxic-ischemic encephalopathy (HIE). Perinatal asphyxia is a clinical syndrome involving acidosis, a low Apgar score and the need for resuscitation in the delivery room; asphyxia alerts one to the possibility of NE. In the present systematic review, we focused on the noninflammatory biomarkers in cerebrospinal fluid (CSF) that are involved in the development of possible brain injury in asphyxia or HIE. A literature search in PubMed and EMBASE for case-control studies was conducted and 17 studies were found suitable by a priori criteria. Statistical analysis used the Mantel-Haenszel model for dichotomous data. The pooled mean difference and 95% confidence intervals (CIs) were determined. We identified the best biomarkers, based on the estimation approach in evaluating the biological significance, out of hundreds in three categories: cell adhesion and proliferation, oxidants and antioxidants, and cell damage. The following subtotal-population comparisons were made: perinatal asphyxia versus no asphyxia, asphyxia with HIE versus asphyxia without HIE, asphyxia with HIE versus no asphyxia, and term versus preterm HIE newborn with asphyxia. Biological significance of the biomarkers was determined by using a modification of the estimation approach, by ranking the biomarkers according to the difference in the bounds of the CIs. The most promising CSF biomarkers for prognostication especially for the severest HIE include creatine kinase, xanthine oxidase, vascular endothelial growth factor, neuron-specific enolase, superoxide dismutase, and malondialdehyde. Future studies are recommended using such a combined test to prognosticate the most severely affected patients.

Tetrahydrobiopterin in Cell Function and Death Mechanisms.

Significance: Tetrahydrobiopterin (BH4) is most well known as a required cofactor for enzymes regulating cellular redox homeostasis, aromatic amino acid metabolism, and neurotransmitter synthesis. Less well known are the effects dependent on the cofactor's availability, factors governing its synthesis and recycling, redox implications of the cofactor itself, and protein-protein interactions that underlie cell death. This review provides an understanding of the recent advances implicating BH4 in the mechanisms of cell death and suggestions of possible therapeutic interventions. Recent Advances: The levels of BH4 often reflect the sum of synthetic and recycling enzyme activities. Enhanced expression of GTP cyclohydrolase, the rate-limiting enzyme in biosynthesis, increases BH4, leading to improved cell function and survival. Pharmacologically increasing BH4 levels has similar beneficial effects, leading to enhanced production of neurotransmitters and nitric oxide or reducing oxidant levels. The GTP cyclohydrolase-BH4 pairing has been implicated in a type of cell death, ferroptosis. At the cellular level, BH4 counteracts anticancer therapies directed to enhance ferroptosis via glutathione peroxidase 4 (GPX4) activity inhibition. Critical Issues: Because of the multitude of intertwined mechanisms, a clear relationship between BH4 and cell death is not well understood yet. The possibility that the cofactor directly influences cell viability has not been excluded in previous studies when modulating BH4-producing enzymes. Future Directions: The importance of cellular BH4 variations and BH4 biosynthetic enzymes to cell function and viability makes it essential to better characterize temporal changes, cofactor activity, and the influence on redox status, which in turn would help develop novel therapies. Antioxid. Redox Signal. 37, 171-183.

New Autophagy-Ferroptosis Gene Signature Predicts Survival in Glioma.

Background: Ferroptosis plays an important role in glioma and significantly affects the prognosis, but the specific mechanism has not yet been elucidated. Recent studies suggest that autophagy regulates the process of ferroptosis. This study aimed to find potential autophagy-ferroptosis genes and explore the prognostic significance in glioma. Methods: Ferroptosis and autophagy genes were obtained from two online databases (zhounan.org/ferrdb and autophagy.lu/). The RNAseq data and clinical information were obtained from the Chinese Glioma Genome Atlas (CGGA) database (http://www.cgga.org.cn/). Univariate, multivariate, lasso and Cox regression analysis screened out prognosis-related genes, and a risk model was constructed. Receiver operating characteristic (ROC) curve analysis evaluated the predictive efficiency of the model. Finally, a nomogram was constructed to more accurately predict the prognosis of glioma. Results: We developed a Venn diagram showing 23 autophagy-ferroptosis genes. A total of 660 cases (including RNA sequences and complete clinical information) from two different cohorts (training group n = 413, verification group n = 247) of the CGGA database was acquired. Cohorts were screened to include five prognosis-related genes (MTOR, BID, HSPA5, CDKN2A, GABARAPLA2). Kaplan-Meier curves showed that the risk model was a good prognostic indicator (p < 0.001). ROC analysis showed good efficacy of the risk model. Multivariate Cox analysis also revealed that the risk model was suitable for clinical factors related to prognosis, including type of disease (primary, recurrence), grade (III-IV), age, temozolomide treatment, and 1p19q state. Using the five prognosis-related genes and the risk score, we constructed a nomogram assessed by C-index (0.7205) and a calibration plot that could more accurately predict glioma prognosis. Conclusion: Using a current database of autophagy and ferroptosis genes, we confirmed the prognostic significance of autophagy-ferroptosis genes in glioma, and we constructed a prognostic model to help guide treatment for high grade glioma in the future.

Theory of deep convolutional neural networks III: Approximating radial functions.

We consider a family of deep neural networks consisting of two groups of convolutional layers, a downsampling operator, and a fully connected layer. The network structure depends on two structural parameters which determine the numbers of convolutional layers and the width of the fully connected layer. We establish an approximation theory with explicit approximation rates when the approximated function takes a composite form f∘Q with a feature polynomial Q and a univariate function f. In particular, we prove that such a network can outperform fully connected shallow networks in approximating radial functions with Q(x)=|x|2, when the dimension d of data from Rd is large. This gives the first rigorous proof for the superiority of deep convolutional neural networks in approximating functions with special structures. Then we carry out generalization analysis for empirical risk minimization with such a deep network in a regression framework with the regression function of the form f∘Q. Our network structure which does not use any composite information or the functions Q and f can automatically extract features and make use of the composite nature of the regression function via tuning the structural parameters. Our analysis provides an error bound which decreases with the network depth to a minimum and then increases, verifying theoretically a trade-off phenomenon observed for network depths in many practical applications.