Pilot Dataset Research Articles

Abstract Su801: Use of Large Language Models to Optimize Clinical Text Analysis for In-Hospital Cardiac Arrest Identification

Introduction/Background: In-hospital cardiac arrest (IHCA) is experienced by approximately 300,000 patients annually in the US. While individual care teams can readily identify IHCA at the bedside, the reporting of these events after the event often lacks consistency and poses a major challenge for hospitals across the US. Accurate and timely reporting of IHCA events is crucial for facilitating QI initiatives, such as CPR quality review, optimizing team performance, and benchmarking IHCA outcomes. Hypothesis: Large language models (LLMs) can be used to identify IHCA events at the patient encounter-level by analyzing targeted note types from the inpatient record with high performance. Methods: Adult (≥18 years old) inpatient encounters at the Hospital of the University of Pennsylvania from 06/2018 to 03/2022 with a reported clinical emergency were identified from a QI database. Discharge summaries and notes associated with the encounters of interest were extracted from the EHR and deidentified using a natural language processing tool. Notes were reviewed by research staff to ascertain true IHCA events. Positive IHCA was defined as the loss of pulses followed by the delivery of CPR. An LLM, GPT4 v. 32K-Chat, was employed on the Penn Medicine Microsoft Azure Databricks environment to compare zero- and few-shot prompting methods to identify IHCA. Model performance was measured using accuracy, precision, recall and F1 scores. Results: Manual chart review determined 96.8% and 3.2% of cases to be IHCA+ and IHCA-, respectively. A balanced pilot dataset of positive and negative cases was created and the accuracy values across strategically sampled note types was checked. Using zero-shot learning with Chain-of-Thought (COT) prompting, the LLM performed with an accuracy of 86%, precision of 77%, recall of 100%, and F1-score of 87%. Using three-shot learning with COT, the LLM performed with an accuracy of 90%, precision of 83%, recall of 100%, and F1-score of 91%. Conclusions: This study demonstrates the potential efficacy of leveraging LLMs to automatically classify IHCA from strategically sampled note types using both zero and few-shot learning approaches. In the future, we will implement temporal sampling of note types and refine model parameters such as output token size and temperature. These findings suggest that such technologies can be effective in real-world clinical settings, providing a scalable solution to improve patient outcomes and quality improvement efforts.

Navigating the Social Shift: A Preliminary Study of Consumer Behavior on Malaysia's Social Commerce Platforms

The proliferation of online businesses coupled with the advancement of information and communication technologies (ICT) has led to the emergence of social commerce. The integration of social media platforms into e-commerce offers businesses a convenient avenue for engaging with consumers and a novel method for capturing additional sales volume. This seamless integration of social media platforms into current e-commerce practices has significantly increased its success. Research on the factors that affect purchase intentions in social commerce is still sparse, especially in the context of Malaysian consumers. In general, the causal model of purchase intentions on social commerce in this study aims to investigate the impact of the constructs of visibility and metavoicing on purchase intentions among Malaysian millennial consumers. Henceforth, this initial study specifically aims to evaluate the reliability and validity of the IT affordance constructs of visibility and metavoicing through rigorous assessments. At the items-construct level, the internal consistency assessment was conducted using McDonald’s Omega, which is known for its robustness in measuring the reliability of multi-item constructs. At the construct level, discriminant validity and the strength of linear relationships among variables were assessed using Pearson correlation analysis. The pilot dataset gathered was analyzed with SPSS 29, and the results have validated that the predictive model achieved high levels of reliability and validity at both the item and construct levels. Furthermore, the analysis revealed that the independent variables, metavoicing and visibility, have a significant association with purchase intentions, underscoring their importance in understanding consumer behavior.

Developing a codebook to characterize barriers, enablers, and strategies for implementing learning health systems from a multilevel perspective

AbstractIntroductionLearning health systems (LHSs) play a crucial role in improving healthcare delivery and outcomes through continuous learning and data‐driven decision‐making. Implementation of LHSs spans individual, organization, and systemic levels of healthcare. This paper outlines a systematic approach for developing a comprehensive codebook to identify barriers, enablers, and strategies associated with the establishment and operation of LHS from a multilevel perspective.MethodsThe codebook development process was divided into two phases and employed a coding team. Phase 1 involved the synthesis of previous literature, which drove the development of initial codes. Phase 2 included the testing of the codebook with a pilot dataset to derive new codes or iterative refinement, ensuring robustness, and validity.ResultsThe literature search revealed 12 papers that detailed the barriers, enablers, and strategies for LHS implementation. Micro‐level codes were derived from a mixture of existing literature and our pilot dataset. Most meso‐level codes barriers and enablers were derived from the literature, with some subcodes derived from participant interviews. All strategies for implementation at the meso‐level were identified in the literature. At the macro‐level, all codes and subcodes were from the literature.ConclusionsThe codebook contributes to the advancement of implementation science in LHS. The codebook facilitates effective analysis and understanding of the key factors influencing the success of LHS implementation, offering practical insights for policymakers, healthcare practitioners and researchers engaged in the ongoing evolution of LHS.

Leveraging multi-site electronic health data for characterization of subtypes: a pilot study of dementia in the N3C Clinical Tenant.

To provide a foundational methodology for differentiating comorbidity patterns in subphenotypes through investigation of a multi-site dementia patient dataset. Employing the National Clinical Cohort Collaborative Tenant Pilot (N3C Clinical) dataset, our approach integrates machine learning algorithms-logistic regression and eXtreme Gradient Boosting (XGBoost)-with a diagnostic hierarchical model for nuanced classification of dementia subtypes based on comorbidities and gender. The methodology is enhanced by multi-site EHR data, implementing a hybrid sampling strategy combining 65% Synthetic Minority Over-sampling Technique (SMOTE), 35% Random Under-Sampling (RUS), and Tomek Links for class imbalance. The hierarchical model further refines the analysis, allowing for layered understanding of disease patterns. The study identified significant comorbidity patterns associated with diagnosis of Alzheimer's, Vascular, and Lewy Body dementia subtypes. The classification models achieved accuracies up to 69% for Alzheimer's/Vascular dementia and highlighted challenges in distinguishing Dementia with Lewy Bodies. The hierarchical model elucidates the complexity of diagnosing Dementia with Lewy Bodies and reveals the potential impact of regional clinical practices on dementia classification. Our methodology underscores the importance of leveraging multi-site datasets and tailored sampling techniques for dementia research. This framework holds promise for extending to other disease subtypes, offering a pathway to more nuanced and generalizable insights into dementia and its complex interplay with comorbid conditions. This study underscores the critical role of multi-site data analyzes in understanding the relationship between comorbidities and disease subtypes. By utilizing diverse healthcare data, we emphasize the need to consider site-specific differences in clinical practices and patient demographics. Despite challenges like class imbalance and variability in EHR data, our findings highlight the essential contribution of multi-site data to developing accurate and generalizable models for disease classification.

EMOKINE: A software package and computational framework for scaling up the creation of highly controlled emotional full-body movement datasets

EMOKINE is a software package and dataset creation suite for emotional full-body movement research in experimental psychology, affective neuroscience, and computer vision. A computational framework, comprehensive instructions, a pilot dataset, observer ratings, and kinematic feature extraction code are provided to facilitate future dataset creations at scale. In addition, the EMOKINE framework outlines how complex sequences of movements may advance emotion research. Traditionally, often emotional-‘action’-based stimuli are used in such research, like hand-waving or walking motions. Here instead, a pilot dataset is provided with short dance choreographies, repeated several times by a dancer who expressed different emotional intentions at each repetition: anger, contentment, fear, joy, neutrality, and sadness. The dataset was simultaneously filmed professionally, and recorded using XSENS® motion capture technology (17 sensors, 240 frames/second). Thirty-two statistics from 12 kinematic features were extracted offline, for the first time in one single dataset: speed, acceleration, angular speed, angular acceleration, limb contraction, distance to center of mass, quantity of motion, dimensionless jerk (integral), head angle (with regards to vertical axis and to back), and space (convex hull 2D and 3D). Average, median absolute deviation (MAD), and maximum value were computed as applicable. The EMOKINE software is appliable to other motion-capture systems and is openly available on the Zenodo Repository. Releases on GitHub include: (i) the code to extract the 32 statistics, (ii) a rigging plugin for Python for MVNX file-conversion to Blender format (MVNX=output file XSENS® system), and (iii) a Python-script-powered custom software to assist with blurring faces; latter two under GPLv3 licenses.

An exploratory characterization of speech- and fine-motor coordination in verbal children with Autism spectrum disorder

Autism spectrum disorder (ASD) is a neurodevelopmental disorder often associated with difficulties in speech production and fine-motor tasks. Thus, there is a need to develop objective measures to assess and understand speech production and other fine-motor challenges in individuals with ASD. In addition, recent research suggests that difficulties with speech production and fine-motor tasks may contribute to language difficulties in ASD. In this paper, we explore the utility of an off-body recording platform, from which we administer a speech- and fine-motor protocol to verbal children with ASD and neurotypical controls. We utilize a correlation-based analysis technique to develop proxy measures of motor coordination from signals derived from recordings of speech- and fine-motor behaviors. Eigenvalues of the resulting correlation matrix are inputs to Gaussian Mixture Models to discriminate between highly-verbal children with ASD and neurotypical controls. These eigenvalues also characterize the complexity (underlying dimensionality) of representative signals of speech- and fine-motor movement dynamics, and form the feature basis to estimate scores on an expressive vocabulary measure. Based on a pilot dataset (15 ASD and 15 controls), features derived from an oral story reading task are used in discriminating between the two groups with AUCs > 0.80, and highlight lower complexity of coordination in children with ASD. Features derived from handwriting and maze tracing tasks led to AUCs of 0.86 and 0.91, however features derived from ocular tasks did not aid in discrimination between the ASD and neurotypical groups. In addition, features derived from free speech and sustained vowel tasks are strongly correlated with expressive vocabulary scores. These results indicate the promise of a correlation-based analysis in elucidating motor differences between individuals with ASD and neurotypical controls.

Prevention of post-operative adhesions: Model development and pilot outcomes of human placental stem cell-based interventions.

Abdominal adhesions are the most common surgical complication and without reliable prophylactics. This study presents a novel rat model for abdominal adhesions and reports pilot results of human placental stem cell (hPSC)-based therapies. Forty-four (n = 44) male Sprague-Dawley rats (250-350 g) were used in the experiment. Of these, thirty-eight (n = 38) were included in a preliminary data set to determine a minimum treatment effect. Adhesions were created in a reproducible model to the abdominal wall and between organs. Experimental groups included the control group (Model No Treatment, MNT), Plasmalyte A (Media Alone, MA, 10 mL), hPSC (5 × 106 cells/10 mL Plasmalyte A), hPSC-CM (hPSC secretome, conditioned media) in 10 mL Plasmalyte A, Seprafilm™ (Baxter, Deerfield, IL), and sham animals (laparotomy only). Treatments were inserted intraperitoneally (IP) and the study period was 14 days post-operation. Results are reported as the difference between means of an index statistic (AIS, Animal Index Score) and compared by ANOVA with pairwise comparison. The overall mean AIS was 23 (SD 6.16) for the MNT group with an average of 75% of ischemic buttons involved in abdominal adhesions. Treatment groups MA (mean overall AIS 17.33 SD 6.4), hPSC (mean overall AIS 13.86 SD 5.01), hPSC-CM (mean overall AIS 13.13 SD 6.15), and Seprafilm (mean overall AIS 13.43 SD 9.11) generated effect sizes of 5.67, 9.14, 9.87, and 9.57 decrease in mean overall AIS, respectively, versus the MNT. The presented rat model and scoring system represent the clinical adhesion disease process. hPSC-based interventions significantly reduce abdominal adhesions in this pilot dataset.

Neuroimaging data repositories and AI-driven healthcare-Global aspirations vs. ethical considerations in machine learning models of neurological disease.

Neuroimaging data repositories are data-rich resources comprising brain imaging with clinical and biomarker data. The potential for such repositories to transform healthcare is tremendous, especially in their capacity to support machine learning (ML) and artificial intelligence (AI) tools. Current discussions about the generalizability of such tools in healthcare provoke concerns of risk of bias-ML models underperform in women and ethnic and racial minorities. The use of ML may exacerbate existing healthcare disparities or cause post-deployment harms. Do neuroimaging data repositories and their capacity to support ML/AI-driven clinical discoveries, have both the potential to accelerate innovative medicine and harden the gaps of social inequities in neuroscience-related healthcare? In this paper, we examined the ethical concerns of ML-driven modeling of global community neuroscience needs arising from the use of data amassed within neuroimaging data repositories. We explored this in two parts; firstly, in a theoretical experiment, we argued for a South East Asian-based repository to redress global imbalances. Within this context, we then considered the ethical framework toward the inclusion vs. exclusion of the migrant worker population, a group subject to healthcare inequities. Secondly, we created a model simulating the impact of global variations in the presentation of anosmia risks in COVID-19 toward altering brain structural findings; we then performed a mini AI ethics experiment. In this experiment, we interrogated an actual pilot dataset (n = 17; 8 non-anosmic (47%) vs. 9 anosmic (53%) using an ML clustering model. To create the COVID-19 simulation model, we bootstrapped to resample and amplify the dataset. This resulted in three hypothetical datasets: (i) matched (n = 68; 47% anosmic), (ii) predominant non-anosmic (n = 66; 73% disproportionate), and (iii) predominant anosmic (n = 66; 76% disproportionate). We found that the differing proportions of the same cohorts represented in each hypothetical dataset altered not only the relative importance of key features distinguishing between them but even the presence or absence of such features. The main objective of our mini experiment was to understand if ML/AI methodologies could be utilized toward modelling disproportionate datasets, in a manner we term "AI ethics." Further work is required to expand the approach proposed here into a reproducible strategy.

Prediction of the effluent chemical oxygen demand and volatile fatty acids for anaerobic treatment based on different feature selections machine-learning methods from lab-scale to pilot-scale

The effluent chemical oxygen demand (COD) and volatile fatty acids (VFA) are important indicators for measuring wastewater anaerobic treatment systems. Through lab and pilot experiments on the anaerobic treatment process of high-salt wastewater, combined with different prediction requirements, the feature variables are classified. The effluent-COD and VFA prediction models of five different Machine-learning methods (i.e., back propagation neural network (BPNN), Genetic algorithm optimizes BPNN(GA-BP), support vector machine (SVM), Least squares support vector machine (LSSVM), and random forest (RF)) were established. The modeling results on different dataset (i.e., lab dataset, pilot dataset and mixed dataset) showed that the SVM and LSSVM methods based on statistical learning had the best model prediction performance in lab and pilot datasets, and the RF method performed well in mixed dataset. Furthermore, the genetic algorithm was adopted to optimize the RF model. After optimization of mixed dataset, the effluent-COD prediction (i.e., R2 = 0.966, RMSE/MAE are 61.227/48.584 mg/L) and VFA prediction (i.e., R2 = 0.639, RMSE/MAE are 26.838/21.831 mg/L) have been improved compared with the original RF model (i.e., R2 = 0.918, RMSE/MAE are 86.232/65.03 mg/L of effluent-COD, and R2 = 0.539, RMSE/MAE are 35.489/26.726 mg/L of VFA). This research can provide a reference for the prediction of the anaerobic treatment of actual large-scale industrial wastewater.

Molecular Identification and Phylogenetic Study Based on the Fusion Gene of Newcastle Disease Virus Isolated from Broiler Poultry Farms in Markazi Province, Iran.

Newcastle disease (ND) is an economically significant and extremely spreadable viral illness affecting a wide variety of avian species. ND can rapidly spread within poultry farms and result in considerable economic losses for the global poultry industry. This disease is endemic in Iran, and despite intensive vaccination efforts in the poultry industry, outbreaks of ND occur unexpectedly. This study aimed to isolate the Newcastle disease virus (NDV) from poultry farms with breathing problems in Markazi province, Iran, and investigate the evolutionary relationship and molecular characteristics of the isolates during 2017-2019. To this end, tissue samples (lung, brain, and trachea) were taken from 42 broiler farms exhibiting respiratory symptoms. The samples were inoculated into 9-11-day-old embryonated eggs, and the virus was isolated from 20 (47.6%) of the 42 farms. Subsequently, RT-PCR was used to amplify partial fusion gene sequences from the new isolates. The amplified products were sequenced and compared phylogenetically to the standard pilot dataset (125 selected sequences) generated by the NDV consortium. As determined by phylogenetic analysis, all nine isolates belonged to subgenotype VII.1.1 of genotype VII and were highly similar to isolates from other parts of Iran and China. Moreover, all isolates possessed a polybasic cleavage site motif (112RRQKRF117), characteristic of virulent strains. Furthermore, the present isolates shared a high nucleotide identity (96%) with viruses previously isolated from other provinces of Iran, as determined by BLAST searches and multiple alignments. In addition, they shared a high degree of sequence similarity but were distinct from the existing NDV vaccines. Therefore, the genetic dissimilarity between current vaccine strains and circulating NDVs must be considered in vaccination programs.

Impact of Anaemia on HbA1c Interpretation Among Normogylcaemic Individuals With Diabetes Mellitus: A Clinical Laboratory Observational Study.

Over the past decades, glycosylated haemoglobin (HbA1c) has been a gold standard for monitoring diabetes control over a long period, relative to blood glucose level (BGL) which measures short-term results. It is speculated that anaemia and factors that induce haemolysis may cause falsely elevated HbA1c leading to 'false positive' interpretations. This study aimed to investigate how anaemia impacts HbA1c. This was a pathology-based observational pilot study using archived data of diabetic subjects monitored with both BGL and HbA1c in regional New South Wales (NSW), Australia. A total of 28,487 cases of blood glucose results were pooled and those with HbA1c and anaemia results were evaluated for correlation with the BGL results. Data collection was limited to de-identified information from the laboratory information system, hence details on the ethnicity and medical history were unavailable. Descriptive frequencies and Pearson correlations were performed. In the pooled data, 53.36% of individuals were females, and 50.54% had BGL ≥5.6 mmol/L. In the pilot dataset, the majority (64.86%) were males, 18.92% with BGL ≤5.6 mmol/L and 67.57% had HbA1c (≥6.5%). In the entire dataset, BGL was moderately and positively correlated with HbA1c (r = 0.6), whereas in the subset of individuals with normo-BGL and anaemia, the correlation was negative (r = -0.2). This pilot investigation observed a pertinent issue, which is a negative correlation between glycaemia and HbA1c in patients with anaemia. HbA1c was falsely increased despite normal blood glucose levels in individuals with anaemia. This advances the speculation that anaemia falsely increases HbA1c. Therefore, caution is necessary while interpreting HbA1c results for patients with anaemia, and new methods for interpretation are required.

Exploring the Gut-Brain Connection in Gastroparesis With Autonomic and Gastric Myoelectric Monitoring.

The goal of this study was to identify autonomic and gastric myoelectric biomarkers from throughout the day that differentiate patients with gastroparesis, diabetics without gastroparesis, and healthy controls, while providing insight into etiology. We collected 19 24-hour recordings of electrocardiogram (ECG) and electrogastrogram (EGG) data from healthy controls and patients with diabetic or idiopathic gastroparesis. We used physiologically and statistically rigorous models to extract autonomic and gastric myoelectric information from the ECG and EGG data, respectively. From these, we constructed quantitative indices which differentiated the distinct groups and demonstrated their application in automatic classification paradigms and as quantitative summary scores. We identified several differentiators that separate healthy controls from gastroparetic patient groups, specifically around sleep and meals. We also demonstrated the downstream utility of these differentiators in automatic classification and quantitative scoring paradigms. Even with this small pilot dataset, automated classifiers achieved an accuracy of 79% separating autonomic phenotypes and 65% separating gastrointestinal phenotypes. We also achieved 89% accuracy separating controls from gastroparetic patients in general and 90% accuracy separating diabetics with and without gastroparesis. These differentiators also suggested varying etiologies for different phenotypes. The differentiators we identified were able to successfully distinguish between several autonomic and gastrointestinal (GI) phenotypes using data collected while at-home with non-invasive sensors. Autonomic and gastric myoelectric differentiators, obtained using at-home recording of fully non-invasive signals, can be the first step towards dynamic quantitative markers to track severity, disease progression, and treatment response for combined autonomic and GI phenotypes.

0863 Evaluation of sleep apnea among consecutive patients with all patterns of atrial fibrillation

Abstract Introduction Atrial fibrillation (AF) is the most common arrhythmia encountered in clinical practice and is associated with significant morbidity, mortality, and financial burden. Obstructive sleep apnea is more common in individuals with AF and may impair AF treatment efficacy including with catheter ablation. However, the prevalence of undiagnosed OSA in all-comers with AF is unknown. Methods This pragmatic, phase IV prospective cohort study will test 250-300 consecutive ambulatory AF patients with all patterns of atrial fibrillation (paroxysmal, persistent, and long-term persistent) and no prior sleep testing for OSA using the WatchPAT system, a home sleep test. Results We report the design, methodology, and results from the initial pilot enrollment of approximately 25% (N=52) of the planned sample size. The primary outcome is the prevalence of undiagnosed OSA in all-comers with AF. Additional outcomes of interest include the sensitivity and specificity of clinical screening instruments for OSA in the AF population and AF-related quality of life measurements among those with and without OSA. Results of the pilot dataset demonstrate a 77.5% prevalence of at least mild (AHI≥5) OSA or greater in all-comers with AF. Conclusion Based on our preliminary dataset representing one quarter of the planned total enrollment size of 250-300, there is a high prevalence of OSA in all-comers with AF. Support (if any) Zoll Medical

Lemon: LEns MOdelling with Neural networks – I. Automated modelling of strong gravitational lenses with Bayesian Neural Networks

ABSTRACT The unprecedented number of gravitational lenses expected from new-generation facilities such as the ESA Euclid telescope and the Vera Rubin Observatory makes it crucial to rethink our classical approach to lens-modelling. In this paper, we present lemon (Lens Modelling with Neural networks): a new machine-learning algorithm able to analyse hundreds of thousands of gravitational lenses in a reasonable amount of time. The algorithm is based on a Bayesian Neural Network: a new generation of neural networks able to associate a reliable confidence interval to each predicted parameter. We train the algorithm to predict the three main parameters of the singular isothermal ellipsoid model (the Einstein radius and the two components of the ellipticity) by employing two simulated data sets built to resemble the imaging capabilities of the Hubble Space Telescope and the forthcoming Euclid satellite. In this work, we assess the accuracy of the algorithm and the reliability of the estimated uncertainties by applying the network to several simulated data sets of 104 images each. We obtain accuracies comparable to previous studies present in the current literature and an average modelling time of just ∼0.5 s per lens. Finally, we apply the lemon algorithm to a pilot data set of real lenses observed with HST during the SLACS program, obtaining unbiased estimates of their SIE parameters. The code is publicly available on GitHub (https://github.com/fab-gentile/LEMON).

177 Education and Its Effects on Barriers to Clinical Trial Participation in Alzheimer’s Disease Studies by Underrepresented Communities

OBJECTIVES/GOALS: Minoritized populations experience a large burden of Alzheimer’s Disease; interventions are often delayed and underrepresented communities’ participation in clinical trial research is low. Lack of information has been proposed as a barrier to clinical research enrollment of minoritized populations. METHODS/STUDY POPULATION: Brain Train, a cluster randomization trial, evaluated the role of education in increasing willingness to participate in research. A 3 segment program was developed whereas segment 3 had two versions: clinical trial education or healthy brain aging video. Brain Train was presented to multicultural communities with participants of at least 50 y/o throughout WNY. The primary outcome measure is evaluating the percent change in responding yes to would you be interested in participating in a clinical trial? before and after segment 3. The secondary outcome measure and exploration of barriers are measured by a Research Attitude Questionnaire. Demographic information such as age, race, gender, socioeconomic status, and educational attainment is collected. The statistical model is a generalized linear mixed model. RESULTS/ANTICIPATED RESULTS: Sixteen sessions with 281 participants were completed to generate the pilot dataset. We see 59.29 % of individuals answered yes before segment 3 and 46.02% answered yes after the educational intervention. Our data shows there is a 13.3% decrease in clinical trial participation interest after the educational intervention. Through our RAQ responses, our data shows significance when it comes to our participants’answers to questions regarding society needing to devote more resources to medical research (p=0.04). Trust emerged as the most significant barrier when it comes to one’s willingness to participate in medical research and clinical trials (p=0.03). DISCUSSION/SIGNIFICANCE: Our preliminary results from the first sixteen events suggest that the power of education is not sufficient to overcome barriers to clinical trial participation for underrepresented communities. Instead, trust appears to be the most significant barrier. Trust building strategies should be explored to answer this research question.

National CO2budgets (2015–2020) inferred from atmospheric CO2observations in support of the global stocktake

Abstract. Accurate accounting of emissions and removals of CO2 is critical for the planning and verification of emission reduction targets in support of the Paris Agreement. Here, we present a pilot dataset of country-specific net carbon exchange (NCE; fossil plus terrestrial ecosystem fluxes) and terrestrial carbon stock changes aimed at informing countries' carbon budgets. These estimates are based on “top-down” NCE outputs from the v10 Orbiting Carbon Observatory (OCO-2) modeling intercomparison project (MIP), wherein an ensemble of inverse modeling groups conducted standardized experiments assimilating OCO-2 column-averaged dry-air mole fraction (XCO2) retrievals (ACOS v10), in situ CO2 measurements or combinations of these data. The v10 OCO-2 MIP NCE estimates are combined with “bottom-up” estimates of fossil fuel emissions and lateral carbon fluxes to estimate changes in terrestrial carbon stocks, which are impacted by anthropogenic and natural drivers. These flux and stock change estimates are reported annually (2015–2020) as both a global 1∘ × 1∘ gridded dataset and a country-level dataset and are available for download from the Committee on Earth Observation Satellites' (CEOS) website: https://doi.org/10.48588/npf6-sw92 (Byrne et al., 2022). Across the v10 OCO-2 MIP experiments, we obtain increases in the ensemble median terrestrial carbon stocks of 3.29–4.58 Pg CO2 yr−1 (0.90–1.25 Pg C yr−1). This is a result of broad increases in terrestrial carbon stocks across the northern extratropics, while the tropics generally have stock losses but with considerable regional variability and differences between v10 OCO-2 MIP experiments. We discuss the state of the science for tracking emissions and removals using top-down methods, including current limitations and future developments towards top-down monitoring and verification systems.

Combining isotope ratios for provenancing Viking Age iron artefacts in the British Isles: a pilot study.

Stable and radiogenic isotope analysis - particularly using lead isotope analysis (LIA) - has previously been shown to be a useful tool for the provenancing of ancient metal artefacts of silver and copper and its alloys, but less progress has been made in the provenancing of iron artefacts, despite their importance and frequency in the archaeological record. In this pilot study we investigate for the first time the possibilities of iron isotope analysis in combination with trace strontium isotope analysis and LIA for the provenancing of iron objects believed to be from the Viking Age in the British Isles. Previous studies have shown that analysis of each of these isotopes can contribute to provenancing iron artefacts, but they are not individually resolutory. In this proof-of-concept study, we examine the Fe, Sr and Pb isotopes of 7 artefacts believed to derive from the Viking Age: 3 from Meols - a former Viking seaport on Wirral and 4 samples from the probable location of the AD 1066 Battle of Fulford in North Yorkshire. We also examine an additional artefact of unknown antiquity from Bebington Heath - a possible location of the AD 937 Battle of Brunanburh. Although the pilot data set is too small to make definitive conclusions, it has paved the way for a fuller study involving 100 samples (including 30 from the former Viking camp of Torksey, Lincolnshire) funded by the NEIF fund of the UK National Environmental Research Council. The high range of 87Sr/86Sr values in the present data set of 8 is beyond what would be expected for bog iron (with a cut-off ∼ 0.709) and suggests that mined ore was being used, a preliminary conclusion supported by the narrow range of Fe isotope data.

Exploring Radiomics for Classification of Supraglottic Tumors: A Pilot Study in a Tertiary Care Center

Accurate classification of laryngeal cancer is a critical step for diagnosis and appropriate treatment. Radiomics is a rapidly advancing field in medical image processing that uses various algorithms to extract many quantitative features from radiological images. The high dimensional features extracted tend to cause overfitting and increase the complexity of the classification model. Thereby, feature selection plays an integral part in selecting relevant features for the classification problem. In this study, we explore the predictive capabilities of radiomics on Computed Tomography (CT) images with the incidence of laryngeal cancer to predict the histopathological grade and T stage of the tumour. Working with a pilot dataset of 20 images, an experienced radiologist carefully annotated the supraglottic lesions in the three-dimensional plane. Over 280 radiomic features that quantify the shape, intensity and texture were extracted from each image. Machine learning classifiers were built and tested to predict the stage and grade of the malignant tumour based on the calculated radiomic features. To investigate if radiomic features extracted from CT images can be used for the classification of laryngeal tumours. Out of 280 features extracted from every image in the dataset, it was found that 24 features are potential classifiers of laryngeal tumour stage and 12 radiomic features are good classifiers of histopathological grade of the laryngeal tumor. The novelty of this paper lies in the ability to create these classifiers before the surgical biopsy procedure, giving the clinician valuable, timely information.

SpeakerPool: A remote speech data collection platform

The collection of speech production data for academic use has traditionally been carried out by recording participants in-person. While certain types of research require traditional data collection methods, many factors, such as distance and world events, make remote collection a valuable alternative for other types of research. Despite the existence of capable technologies and the advantages of remote collection methods, to the authors' knowledge there is no freely accessible, reusable, and research-oriented platform for the remote collection of speech production data. We aim to address this with SpeakerPool: a web application that, in addition to providing an easy-to-use recording interface for participants, has built-in functionalities for the automation of many tasks. Users are able to access SpeakerPool regardless of platform using modern web browsers on both mobile and non-mobile devices. This greatly reduces compatibility limitations and bypasses the need to download software. In the present study, we discuss the general advantages and disadvantages of remote speech data collection and investigate the usefulness of SpeakerPool by collecting and analyzing a pilot data set of Malagasy speech.

Remote photoplethysmography with consumer smartphone reveals temporal differences between glabrous and nonglabrous skin: Pilot in vivo study.

Photoplethysmography (PPG) is a noninvasive optical technology, with applications including vital sign extraction and patient monitoring. The PPG acquisition skin type may be of importance. Skin is either nonglabrous (~90%) or glabrous (~10%). Clinical PPG collection is typically from glabrous (fingerpad), while proliferating wearables collecting PPG, which may perform critical functions like arrythmia detection, often acquire from atypical sites. Glabrous skin has significant differences from nonglabrous, including microcirculation, yet comparisons between their PPG signals have not been well reported. Using a smartphone-based remote/contactless PPG, a pilot dataset was collected from the hands (palmar/dorsal) of five healthy volunteers. The data shows statistically significant lead time (52 ± 36 ms) of glabrous over nonglabrous. Further, a trend of glabrous amplitude increase over nonglabrous (31%) was found. Although our study has a small number of participants, these results further the characterization of PPG skin differences, and can be used to inform development of PPG-based devices.