Large Variation Research Articles

Impacts of genetic polymorphisms and cancer cachexia on naldemedine pharmacokinetics and bowel movements in patients receiving opioid analgesics.

Clinical responses to naldemedine vary between individuals with advanced cancer. This is a prospective, single-center, observational study aimed to evaluate the influence of genetic polymorphisms and cachexia status on plasma naldemedine and clinical responses. Forty-eight patients being treated with naldemedine for opioid-induced constipation under treatment of cancer pain were enrolled. Plasma naldemedine concentrations were determined on the fourth day or later after administration of naldemedine, and the associations with genotypes, cachexia status, and clinical responses were assessed. Cancer patients exhibited a large variation in the plasma naldemedine concentrations, and it was correlated with serum total protein level. Patients who were homozygous CYP3A5*3 had a higher plasma concentration of naldemedine than those with the *1 allele. ABCB1 genotypes tested in this study were not associated with plasma naldemedine. A negative correlation was observed between the plasma naldemedine concentration and 4β-hydroxycholesterol level. The plasma naldemedine concentration was lower in patients with refractory cachexia than in those with precachexia and cachexia. While serum levels of interleukin-6 (IL-6) and acute-phase proteins were higher in patients with refractory cachexia, they were not associated with plasma naldemedine. A higher plasma concentration of naldemedine, CYP3A5*3/*3, and an earlier naldemedine administration after starting opioid analgesics were related to improvement of bowel movements. Plasma naldemedine increased under deficient activity of CYP3A5 in cancer patients. Cachectic patients with a higher serum IL-6 had a lower plasma naldemedine. Plasma naldemedine, related to CYP3A5 genotype, and the initiation timing of naldemedine were associated with improved bowel movements.

CRSANet: Class Representations Self-Attention network for the segmentation of thyroid nodules

Segmenting accurate thyroid nodules in medical ultrasound images is always non-trivial due to the large variation in size, shape and texture of nodules, and also the presence of surrounding tissues and organs with similar intensity values. Existing algorithms solve this by designing multi-scale spatial context modeling modules and introducing the emerging self-attention based paradigm, while failing to learn the class-wise relational context around nodules. To this end, we propose a novel coarse-to-fine framework, namely, CRSANet, which leverages both class-wise relational context and spatial multi-scale context for comprehensive context perception. Specifically, we first introduce the concept of Class Representations, which describe the overall representation of thyroid nodules in ultrasound images from a categorical perspective. Secondly, we propose a backbone-free dual-branch Class Representation Self-Attention Module (CRSA) to refine the coarse segmentation results from segmentation backbones. Concretely, Fine Class Region (FCR) branch is proposed to impose consistency between pixels and their corresponding Class Representations to correct the wrongly classified pixels. Meanwhile, Class Contextual Representation (CCR) branch is introduced to enhance the ability of pixel representations with the corresponding Class Representations. We evaluate the effectiveness and generality of CRSANet using three types of backbone networks. In the experiments, our proposed U-CRSANet achieves a mIoU of 78.86% and 81.35%, an accuracy of 97.11% and 98.63%, a Dice value of 74.26% and 87.57%, respectively, which outperforms most state-of-the-art methods on the public DDTI and TN-SCUI datasets.

Feasibility of OpenPose markerless motion analysis in a real athletics competition.

This study tested the performance of OpenPose on footage collected by two cameras at 200 Hz from a real-life competitive setting by comparing it with manually analyzed data in SIMI motion. The same take-off recording from the men's Long Jump finals at the 2017 World Athletics Championships was used for both approaches (markerless and manual) to reconstruct the 3D coordinates from each of the camera's 2D coordinates. Joint angle and Centre of Mass (COM) variables during the final step and take-off phase of the jump were determined. Coefficients of Multiple Determinations (CMD) for joint angle waveforms showed large variation between athletes with the knee angle values typically being higher (take-off leg: 0.727 ± 0.242; swing leg: 0.729 ± 0.190) than those for hip (take-off leg: 0.388 ± 0.193; swing leg: 0.370 ± 0.227) and ankle angle (take-off leg: 0.247 ± 0.172; swing leg: 0.155 ± 0.228). COM data also showed considerable variation between athletes and parameters, with position (0.600 ± 0.322) and projection angle (0.658 ± 0.273) waveforms generally showing better agreement than COM velocity (0.217 ± 0.241). Agreement for discrete data was generally poor with high random error for joint kinematics and COM parameters at take-off and an average ICC across variables of 0.17. The poor agreement statistics and a range of unrealistic values returned by the pose estimation underline that OpenPose is not suitable for in-competition performance analysis in events such as the long jump, something that manual analysis still achieves with high levels of accuracy and reliability.

Longevity, demographic characteristics, and socio-economic status are linked to triiodothyronine levels in the general population.

The hypothalamic-pituitary-thyroid (HPT) axis is fundamental to human biology, exerting central control over energy expenditure and body temperature. However, the consequences of normal physiologic HPT-axis variation in populations without diagnosed thyroid disease are poorly understood. Using nationally representative data from the 2007 to 2012 National Health and Nutrition Examination Survey, we explore relationships with demographic characteristics, longevity, and socio-economic factors. We find much larger variation across age in free T3 than other HPT-axis hormones. T3 and T4 have opposite relationships to mortality: free T3 is inversely related and free T4 is positively related to the likelihood of death. Free T3 and household income are negatively related, particularly at lower incomes. Finally, free T3 among older adults is associated with labor both in terms of unemployment and hours worked. Physiologic TSH/T4 explain only 1.7% of T3 variation, and neither are appreciably correlated to socio-economic outcomes. Taken together, our data suggest an unappreciated complexity of the HPT-axis signaling cascade broadly such that TSH and T4 may not be accurate surrogates of free T3. Furthermore, we find that subclinical variation in the HPT-axis effector hormone T3 is an important and overlooked factor linking socio-economic forces, human biology, and aging.

Modelling the physiological, muscular, and sensory characteristics in relation to beef quality from 15 cattle breeds

Sensory beef quality is an important parameter for consumers, but it is difficult to understand what determines quality because of the large variation due to breed, type of animal, sex, and farm management. To address these challenges, meat samples from Longissimus thoracis (LT) were collected 24 h after slaughter from a total of 436 young bulls representing 15 cattle breeds. The samples were analysed for physiological, muscular, and sensory characteristics to evaluate beef quality. Ten groups of variables were identified by hierarchical cluster analysis of individual variables and Principal Component Analysis (PCA). The variables described distinct characteristics, including Physiologicaltraits (animal maturity, growth rate, muscle mass), Sensorytraits (tenderness, juiciness, flavor, meat colour), and some Muscularcharacteristics such as lipid content, maturation enzymes and oxidative metabolism, which are likely to discriminate among breeds. The analyses showed that muscular characteristics are positively associated with beef sensory characteristics whereas physiological characteristics are negatively associated with sensory quality. Dairy breeds produced beef rich in lipids with a strong flavour, French breeds showed fast growth rate and highly flavoured meat, and Italian breeds were characterised by good muscular development. In contrast, British breeds have oxidative muscles and produce beef with a strong flavour. Danish breeds have an intermediate score for the sensorial and muscular characteristics.

Large Seasonal Variation in Nitrogen Isotopic Abundances of Ammonia Volatilized from a Cropland Ecosystem and Implications for Regional NH3 Source Partitioning.

Ammonia (NH3) volatilization from agricultural lands is a main source of atmospheric reduced nitrogen species (NHx). Accurately quantifying its contribution to regional atmospheric NHx deposition is critical for controlling regional air nitrogen pollution. The stable nitrogen isotope composition (expressed by δ15N) is a promising indicator to trace atmospheric NHx sources, presupposing a reliable nitrogen isotopic signature of NH3 emission sources. To obtain more specific seasonal δ15N values of soil NH3 volatilization for reliable regional seasonal NH3 source partitioning, we utilized an active dynamic sampling technique to measure the δ15N-NH3 values volatilized from maize cropping land in northeast China. These values varied from -38.0 to -0.2‰, with a significantly lower rate-weighted value observed in the early period (May-June, -30.5 ± 6.7‰) as compared with the late period (July-October, -8.5 ± 4.3‰). Seasonal δ15N-NH3 variations were related to the main NH3 production pathway, degree of soil ammonium consumption, and soil environment. Bayesian isotope mixing model analysis revealed that without considering the seasonal δ15N variation in soil-volatilized NH3 could result in an overestimate by up to absolute 38% for agricultural volatile NH3 to regional atmospheric bulk ammonium deposition during July-October, further demonstrating that it is essential to distinguish seasonal δ15N profile of agricultural volatile NH3 in regional source apportionment.

Laboratory practice is central to earlier myeloma diagnosis: Utilizing a primary care diagnostic tool and laboratory guidelines integrated into haematology services.

Treatment advances have greatly improved survival, but myeloma is among the worst of all cancers for delayed diagnosis, causing serious morbidities and early deaths. This delay is largely because the symptom profile of myeloma has very low specificity, and in primary care, myeloma is rare. However, initiating the journey to diagnosis simply requires considering myeloma and sending blood to test for monoclonal immunoglobulin. Laboratory tests reliably detect monoclonal immunoglobulin, which is present in 99% of myeloma cases, so why do health care systems have such a problem with delayed diagnosis? The Myeloma UK early diagnosis programme has brought together diverse expertise to investigate this problem, and this article was prepared by the programme's working group for laboratory best practice. It reviews evidence for test requesting, analysis and reporting, for which there is large variation in practice across the United Kingdom. It presents a 'GP Myeloma diagnostic tool' and how it can be integrated into laboratory practice alongside a laboratory best practice tool. It proposes improved requesting and integration with haematology services for reporting and interpretation. Here the laboratory has a central role in creating efficient and cost-effective pathways for appropriate and timely bone marrow examination for myeloma diagnosis.

Large regional variation in cardiac closure procedures to prevent ischemic stroke in Switzerland a population-based small area analysis.

Percutaneous closure of a patent foramen ovale (PFO) or the left atrial appendage (LAA) are controversial procedures to prevent stroke but often used in clinical practice. We assessed the regional variation of these interventions and explored potential determinants of such a variation. We conducted a population-based analysis using patient discharge data from all Swiss hospitals from 2013-2018. We derived hospital service areas (HSAs) using patient flows for PFO and LAA closure. We calculated age-standardized mean procedure rates and variation indices (extremal quotient [EQ] and systematic component of variation [SCV]). SCV values >5.4 indicate a high and >10 a very high variation. Because the evidence on the efficacy of PFO closure may differ in patients aged <60 years and ≥60 years, age-stratified analyses were performed. We assessed the influence of potential determinants of variation using multilevel regression models with incremental adjustment for demographics, cultural/socioeconomic, health, and supply factors. Overall, 2574 PFO and 2081 LAA closures from 10 HSAs were analyzed. The fully adjusted PFO and LAA closure rates varied from 3 to 8 and from 1 to 9 procedures per 100,000 persons per year across HSAs, respectively. The regional variation was high with respect to overall PFO closures (EQ 3.0, SCV 8.3) and very high in patients aged ≥60 years (EQ 4.0, SCV 12.3). The variation in LAA closures was very high (EQ 16.2, SCV 32.1). In multivariate analysis, women had a 28% lower PFO and a 59% lower LAA closure rate than men. French/Italian language areas had a 63% lower LAA closure rate than Swiss German speaking regions and areas with a higher proportion of privately insured patients had a 86% higher LAA closure rate. After full adjustment, 44.2% of the variance in PFO closure and 30.3% in LAA closure remained unexplained. We found a high to very high regional variation in PFO closure and LAA closure rates within Switzerland. Several factors, including sex, language area, and insurance status, were associated with procedure rates. Overall, 30-45% of the regional procedure variation remained unexplained and most probably represents differing physician practices.

Effectiveness of Spatiotemporal Data Fusion in Fine-Scale Land Surface Phenology Monitoring: A Simulation Study

Spatiotemporal data fusion technologies have been widely used for land surface phenology (LSP) monitoring since it is a low-cost solution to obtain fine-resolution satellite time series. However, the reliability of fused images is largely affected by land surface heterogeneity and input data. It is unclear whether data fusion can really benefit LSP studies at fine scales. To explore this research question, this study designed a sophisticated simulation experiment to quantify effectiveness of 2 representative data fusion algorithms, namely, pair-based Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM) and time series-based Spatiotemporal fusion method to Simultaneously generate Full-length normalized difference vegetation Index Time series (SSFIT) by fusing Landsat and Moderate Resolution Imaging Spectroradiometer (MODIS) data in extracting pixel-wise spring phenology (i.e., the start of the growing season, SOS) and its spatial gradient and temporal variation. Our results reveal that: (a) STARFM can improve the accuracy of pixel-wise SOS by up to 74.47% and temporal variation by up to 59.13%, respectively, compared with only using Landsat images, but it can hardly improve the retrieval of spatial gradient. For SSFIT, the accuracy of pixel-wise SOS, spatial gradient, and temporal variation can be improved by up to 139.20%, 26.36%, and 162.30%, respectively; (b) the accuracy improvement introduced by fusion algorithms decreases with the number of available Landsat images per year, and it has a large variation with the same number of available Landsat images, and (c) this large variation is highly related to the temporal distributions of available Landsat images, suggesting that fusion algorithms can improve SOS accuracy only when cloud-free Landsat images cannot capture key vegetation growth period. This study calls for caution with the use of data fusion in LSP studies at fine scales.

Analysis of corneal scrape outcomes in NHS Ayrshire and Arran

Aims/Purpose: Corneal scrape is a key diagnostic procedure in the management of microbial keratitis. Safe technique for corneal scrape is a requirement of the Royal College of Ophthalmology (RCOphth) curriculum; national and international best practice guidelines exist (1, 2). The purpose of our project is to evaluate the local method of scraping in Ayrshire and Aran, Scotland. The primary aims of this project are to determine: the current method of corneal scraping, the yield of positive samples and impact on safe patient care.Methods: The corneal scrape samples submitted to microbiology from Ayr Casualty Clinic were analysed between 1st November 2021–19th December 2022. An audit was taken of the equipment available in clinical areas of the Ophthalmology department and the technique used by on‐call clinicians. Through liaising with microbiology, a local poster was created with guidelines for performing corneal scrape and placed in the eye casualty department with appropriate equipment set up.Results: There was large variation in the clinical technique used to perform the corneal scrapes. None of the on‐call clinicians were able to perform the technique to the published standard. There were no slides, agar plates or needles available for use. One third of clinicians described the correct technique but had no access to the correct equipment. Two‐thirds did not describe the correct technique. Twenty‐two samples were submitted between 01/11/2021–19/12/2022. All samples were cultured, however there were only two positive samples.Conclusions: Uncertainty and clinical variation exist in the clinical technique used to perform this key diagnostic procedure. There are no guidelines available in the casualty clinic or on‐call room and the necessary equipment is unavailable. The next steps include the creation of a local protocol with correct equipment available to improve the diagnostic yield.

Genetic variations at BBX24 and MYB110a loci regulated anthocyanin accumulation in pear bud sports

Genetic variation contributes to the phenotypic diversity of plants. Most red pear cultivars that accumulate anthocyanins originated from somatic variation and are termed bud sports. ‘Hongzaosu’ is a bud sport derived from the green pear ‘Zaosu’. A new genetic map was constructed from a population derived from the cross of ‘Yuluxiang’ and ‘Hongzaosu’, from which PpBBX24 was identified as a crucial gene controlling anthocyanin accumulation. Genetic and phylogenetic evidence revealed that PpBBX24 is a repressor of anthocyanin biosynthesis in pear. A 14 bp deletion was detected in the third exon of PpBBX24 in ‘Hongzaosu’, resulting in premature termination of protein translation. The truncated PpBBX24 protein was a positive regulator of anthocyanin biosynthesis by activating the transcription of PpCHS and PpUFGT. Three independent variations in the BBX24 coding region that led to premature termination of translation were detected in other pear bud sports and the progeny of a bud sport that all accumulate anthocyanins. The genome of ‘Hongzaosu’ was assembled using both long-read and short-read sequences. By combining genomic and transcriptomic data, allele-specific expression of PpMYB110a was observed in the fruit skin of ‘Hongzaosu’, which was likely caused by large variation in the promoter. This work enhances understanding of coloration in red pears and provides a novel genomic resource for further studies on ‘Hongzaosu’ pear.

Spatiotemporal Changes of Evapotranspiration in Cities along the Yellow River Basin in Shandong Province from 2000 to 2018

Steapotranspiration is an important component of land surface water circulation and surface energy balance. The study of the spatial and temporal change of the cities along the Yellow River basin in Shandong Province is conducive to explore the regional water circulation and the response of the river basin to climate change. Based on the evapotranspiration data, the spatiotemporal distribution and change characteristics of evapotranspiration in cities along the Yellow River basin in Shandong Province from 2000 to 2018 were analyzed. The results show that (1), the mean value of evaporation in the Yellow River in Shandong Province from 2000 to 2018 was 469.43 mma-1, with large annual variation; (2) high in spring and summer, low in autumn and winter; (3) the highest spatial distribution in July; (5) the decreasing in Jining, Heze, Taian, Zibo, Jinan, Dongying, Liaocheng, Binzhou and Dezhou; and (6) increasing in eastern cities along the Yellow River; the decrease in the west.

Inheritance of feed intake-based resilience traits and their correlation with production traits in Finnish pig breeds.

High resilience against diseases, changing environmental conditions, and other stress factors and the ability to efficiently recover to normal status, is becoming increasingly important in pig production. Finding new phenotypes that relate to resilience is a crucial step for improving the resilience of pigs through selection. The objective of this study was to extract resilience-related phenotypes based on fluctuations in daily feed intake (DFI, g) and time spent in feeding per day (TPD, min) and to estimate the heritability of these traits and genetic correlations with production traits (PT). Resilience-related traits with high enough heritability and with either favorable or neutral genetic correlation with PT could be used in the selection program to improve the productivity and welfare of pigs. In this study, we used data from 7,347 Finnish Yorkshire, Landrace, and crossbred pigs raised at the test station. Six pig-specific resilience-related phenotypes were extracted from the individual DFI and TPD: root mean square error (RMSE), quantile regression (QR), and coefficient of variation (CV). RMSE was calculated from the differences between the actual DFI (or TPD) and the pig-specific predicted values. QR was based on the number of days that a pig belonged to the group with the lowest 5% of pigs based on DFI (or TPD), and CV was calculated over the daily observations of DFI (or TPD). PT included average daily gain (ADG, g), backfat thickness (BF, mm), and feed conversion rate (FCR, g/g). The heritability estimates for resilience-related traits varied between 0.07 ± 0.02 (QRDFI) and 0.20 ± 0.03 (RMSETPD). The genetic correlations between resilience-related traits and PT were mostly neutral, but for example, RMSEDFI had a favorable genetic correlation with FCR and BF but an unfavorable correlation with ADG. Lastly, we observed that pigs belonging to the lowest 10% group based on their breeding value (BV) for QRTPD had a lower proportion (10% incidence) of sick days compared to the highest 10% BV group (30% incidence). Therefore, pigs exhibiting small TPD variation (related to high resilience) tend to be less susceptible to sickness than pigs with large TPD variation (related to low resilience). Given its moderate heritability, neutral genetic correlation with PT, and positive effect on health, QRTPD can be considered the most promising resilience-related trait in the Finnish production system.

Understanding Factors Influencing Geographic Variation in Healthcare Expenditures: A Small Areas Analysis Study.

Dramatic geographic variations in healthcare expenditures were documented by developed countries, but little is known about such variations under China's context, and what causes such variations. This study aims to examine variations of healthcare expenditures among small areas and to determine the associations between demand-, supply-factors, and per capita inpatient expenditures. This cross-sectional study utilized hospital discharge data aggregated within delineated hospital service areas (HSAs) using the small-area analysis approach. Linear multivariate regression modeling with robust standard errors was used to estimate the sources of variation of per capita inpatient expenditures across HSAs covering the years 2017 to 2019; the Shapley value decomposition method was used to measure the respective contributions of demand-, supply-side to such variations. Among 149 HSAs, demand factors explained most of the (87.4%) overall geographic variation among HSAs. With each 1% increase in GDP per capita and urbanization rate was associated with 0.099% and 0.9% increase in inpatient expenditure per capita, respectively, while each 1% increase in the share of females and the unemployment rate was associated with a 0.7% and 0.4% reduction in the per capita inpatient expenditures, respectively. In supply-side, for every 1 increase in hospital beds per 1000 population, the per capita inpatient expenditures rose by 2.9%, while with every 1% increase in the share of private hospitals, the per capita inpatient expenditures would decrease by 0.4%. With Herfindahl-Hirschman Index decrease 10%, the per capita inpatient expenditures would increase 1.06%. This study suggests demand-side factors are associated with large geographic variation in per capita inpatient expenditures among HSAs, while supply-side factors played an important role. The evaluation of geographic variations in per capita inpatient expenditures as well as its associated factors have great potential to provide an indirect approach to identify possibly existing underutilized or overutilized healthcare procedures.

Multiscale Attention-Based Instance Segmentation for Measuring Crystals With Large Size Variation

Multiscale Attention-Based Instance Segmentation for Measuring Crystals With Large Size Variation

Multi-Surface Multi-Technique (MUST) Latent Fingerprint Database

Latent fingerprint recognition involves acquisition and comparison of latent fingerprints with an exemplar gallery of fingerprints. The diversity in the type of surface leads to different procedures to recover the latent fingerprint. The appearance of latent fingerprints vary significantly due to the development techniques, leading to large intra-class variation. Due to lack of large datasets acquired using multiple mechanisms and surfaces, existing algorithms for latent fingerprints enhancement and comparison may perform poorly. In this study, we propose a Multi-Surface Multi-Technique (MUST) Latent Fingerprint Database. The database consists of more than 16,000 latent fingerprint impressions from 120 unique classes (120 fingers from 12 participants). Including corresponding exemplar fingerprints (livescan and rolled) and extended gallery, the dataset has nearly 21,000 impressions. It has latent fingerprints acquired under 35 different scenarios and additional four subsets of exemplar prints captured using live scan sensor and inked-rolled prints. With 39 different subsets, the database illustrates intra-class variations in latent fingerprints. The database has a potential usage towards building robust algorithms for latent fingerprint enhancement, segmentation, comparison, and multi-task learning. We also provide annotations for manually marked minutiae, acquisition PPI, and semantic segmentation masks are also provided. We also present the experimental protocol and the baseline results for the proposed dataset. We assert that the availability of the proposed database can encourage research in handling intra-class variation in latent fingerprint recognition.

Dual-branch Branch Networks Based on Contrastive Learning for Long-Tailed Remote Sensing

Deep learning has been widely used in remote sensing image classification and achieves many excellent results. These methods are all based on relatively balanced data sets. However, in real-world scenarios, many data sets belong to the long-tailed distribution, resulting in poor performance. In view of the good performance of contrastive learning in long-tailed image classification, a new dual-branch fusion learning classification model is proposed to fuse the discriminative features of remote sensing images with spatial data, making full use of valuable image representation information in imbalance data. This paper also presents a hybrid loss, which solves the problem of poor discrimination of extracted features caused by large intra-class variation and inter-class ambiguity. Extended experiments on three long-tailed remote sensing image classification data sets demonstrate the advantages of the proposed dual-branch model based on contrastive learning in long-tailed image classification.

Iterative estimating equations for disease mapping with spatial zero‐inflated Poisson data

Spatial epidemiology often involves the analysis of spatial count data with an unusually high proportion of zero observations. While Bayesian hierarchical models perform very well for zero‐inflated data in many situations, a smooth response surface is usually required for the Bayesian methods to converge. However, for infectious disease data with excessive zeros, a Wombling issue with large spatial variation could make the Bayesian methods infeasible. To address this issue, we develop estimating equations associated with disease mapping by including over‐dispersion and spatial noises in a spatial zero‐inflated Poisson model. Asymptotic properties are derived for the parameter estimates. Simulations and data analysis are used to assess and illustrate the proposed method.

Robust multiferroicity and magnetic modulation of the ferroelectric imprint field in heterostructures comprising epitaxial Hf0.5Zr0.5O2 and Co.

Magnetoelectric multiferroics, either single-phase or composites comprising ferroelectric/ferromagnetic coupled films, are promising candidates for energy efficient memory computing. However, most of the multiferroic magnetoelectric systems studied so far are based on materials that are not compatible with industrial processes. Doped hafnia is emerging as one of the few CMOS-compatible ferroelectric materials. Thus, it is highly relevant to study the integration of ferroelectric hafnia into multiferroic systems. In particular, ferroelectricity in hafnia, and the eventual magnetoelectric coupling when ferromagnetic layers are grown atop of it, are very much dependent on quality of interfaces. Since magnetic metals frequently exhibit noticeable reactivity when grown onto oxides, it is expected that ferroelectricity and magnetoelectricity might be reduced in multiferroic hafnia-based structures. In this article, we present excellent ferroelectric endurance and retention in epitaxial Hf0.5Zr0.5O2 films grown on buffered silicon using Co as the top electrode. The crucial influence of a thin Pt capping layer grown on top of Co on the ferroelectric functional characteristics is revealed by contrasting the utilization of Pt-capped Co, non-capped Co and Pt. Magnetic control of the imprint electric field (up to 40% modulation) is achieved in Pt-capped Co/Hf0.5Zr0.5O2 structures, although this does not lead to appreciable tuning of the ferroelectric polarization, as a result of its high stability. Computation of piezoelectric and flexoelectric strain-mediated mechanisms of the observed magnetoelectric coupling reveal that flexoelectric contributions are likely to be at the origin of the large imprint electric field variation.

Quantized Magnetic Domain Wall Synapse for Efficient Deep Neural Networks.

The quantization of synaptic weights using emerging nonvolatile memory (NVM) devices has emerged as a promising solution to implement computationally efficient neural networks on resource constrained hardware. However, the practical implementation of such synaptic weights is hampered by the imperfect memory characteristics, specifically the availability of limited number of quantized states and the presence of large intrinsic device variation and stochasticity involved in writing the synaptic states. This article presents on-chip training and inference of a neural network using quantized magnetic domain wall (DW)-based synaptic array and CMOS peripheral circuits. A rigorous model of the magnetic DW device considering stochasticity and process variations has been utilized for the synapse. To achieve stable quantized weights, DW pinning has been achieved by means of physical constrictions. Finally, VGG8 architecture for CIFAR-10 image classification has been simulated by using the extracted synaptic device characteristics. The performance in terms of accuracy, energy, latency, and area consumption has been evaluated while considering the process variations and nonidealities in the DW device as well as the peripheral circuits. The proposed quantized neural network (QNN) architecture achieves efficient on-chip learning with 92.4% and 90.4% training and inference accuracy, respectively. In comparison to pure CMOS-based design, it demonstrates an overall improvement in area, energy, and latency by 13.8 × , 9.6 × , and 3.5 × , respectively.