Balanced Metrics Research Articles

Impact of nutrition-related laboratory tests on mortality of patients who are critically ill using artificial intelligence: A focus on trace elements, vitamins, and cholesterol.

This study aimed to understand the collective impact of trace elements, vitamins, cholesterol, and prealbumin on patient outcomes in the intensive care unit (ICU) using an advanced artificial intelligence (AI) model for mortality prediction. Data from ICU patients (December 2016 to December 2021), including serum levels of trace elements, vitamins, cholesterol, and prealbumin, were retrospectively analyzed using AI models. Models employed included category boosting (CatBoost), extreme gradient boosting (XGBoost), light gradient boosting machine (LGBM), and multilayer perceptron (MLP). Performance was evaluated using area under the receiver operating characteristic curve (AUROC), accuracy, precision, recall, and F1-score. The performance was evaluated using 10-fold crossvalidation. The SHapley Additive exPlanations (SHAP) method provided interpretability. CatBoost emerged as the top-performing individual AI model with an AUROC of 0.756, closely followed by LGBM, MLP, and XGBoost. Furthermore, the ensemble model combining these four models achieved the highest AUROC of 0.776 and more balanced metrics, outperforming all models. SHAP analysis indicated significant influences of prealbumin, Acute Physiology and Chronic Health Evaluation II score, and age on predictions. Notably, the ratios of selenium to age and low-density lipoprotein to total cholesterol also had a notable impact on the models' output. The study underscores the critical role of nutrition-related parameters in ICU patient outcomes. Advanced AI models, particularly in an ensemble approach, demonstrated improved predictive accuracy. SHAP analysis offered insights into specific factors influencing patient survival, highlighting the need for broader consideration of these biomarkers in critical care management.

G4 & the balanced metric family – a novel approach to solving binary classification problems in medical device validation & verification studies

BackgroundIn medical device validation and verification studies, the area under the receiver operating characteristic curve (AUROC) is often used as a primary endpoint despite multiple reports showing its limitations. Hence, researchers are encouraged to consider alternative metrics as primary endpoints. A new metric called G4 is presented, which is the geometric mean of sensitivity, specificity, the positive predictive value, and the negative predictive value. G4 is part of a balanced metric family which includes the Unified Performance Measure (also known as P4) and the Matthews’ Correlation Coefficient (MCC). The purpose of this manuscript is to unveil the benefits of using G4 together with the balanced metric family when analyzing the overall performance of binary classifiers.ResultsSimulated datasets encompassing different prevalence rates of the minority class were analyzed under a multi-reader-multi-case study design. In addition, data from an independently published study that tested the performance of a unique ultrasound artificial intelligence algorithm in the context of breast cancer detection was also considered. Within each dataset, AUROC was reported alongside the balanced metric family for comparison. When the dataset prevalence and bias of the minority class approached 50%, all three balanced metrics provided equivalent interpretations of an AI’s performance. As the prevalence rate increased / decreased and the data became more imbalanced, AUROC tended to overvalue / undervalue the true classifier performance, while the balanced metric family was resistant to such imbalance. Under certain circumstances where data imbalance was strong (minority-class prevalence < 10%), MCC was preferred for standalone assessments while P4 provided a stronger effect size when evaluating between-groups analyses. G4 acted as a middle ground for maximizing both standalone assessments and between-groups analyses.ConclusionsUse of AUROC as the primary endpoint in binary classification problems provides misleading results as the dataset becomes more imbalanced. This is explicitly noticed when incorporating AUROC in medical device validation and verification studies. G4, P4, and MCC do not share this limitation and paint a more complete picture of a medical device’s performance in a clinical setting. Therefore, researchers are encouraged to explore the balanced metric family when evaluating binary classification problems.

A Kummer construction for Chern–Ricci flat balanced manifolds

Given a non-Kähler Calabi–Yau compact orbifold with isolated singularities endowed with a Chern–Ricci flat balanced metric, we study, via a gluing construction, the existence of Chern–Ricci flat balanced metrics on its crepant resolutions, and discuss applications to the search of solutions for the Hull–Strominger system. We also describe the scenario of singular threefolds with ordinary double points, and see that similarly is possible to obtain balanced approximately Chern–Ricci flat metrics.

Watershed-scale spatial prediction of agricultural land phosphorus mass balance and soil phosphorus metrics: A bottom-up approach.

Analysis of nutrient balance at the watershed scale, including for phosphorus (P), is typically accomplished using aggregate input datasets, resulting in an inability to capture the variability of P status across the study region. This study presents a set of methods to predict and visualize partial P mass balance, soil P saturation ratio (PSR), and soil test P for agricultural parcels across a watershed in the Lake Champlain Basin (Vermont, USA) using granular, field-level data. K-means cluster analyses were used to group agricultural parcels by soil texture, average slope, and crop type. Using a set of parcels accounting for ∼21% of the watershed's agricultural land and having known soil test and nutrient management parameters, predictions of partial P mass balance, PSR, and soil test P for agricultural land across the watershed were made by cluster, incorporating uncertainty. This resulted in an average partial P balance of 5.5±0.2kg P ha-1year-1 and an average PSR of 0.0399±0.0002. Furthermore, approximately 30% of agricultural land had predicted soil test P values above optimum levels. Results were used to visualize areas with high P loss potential. Such data and visualizations can inform watershed P modeling and assist practitioners in nutrient management decision making. These techniques can also serve as a framework for bottom-up modeling of nutrient mass balance and soil metrics in other regions.

Prognostic score-based model averaging approach for propensity score estimation

BackgroundPropensity scores (PS) are typically evaluated using balance metrics that focus on covariate balance, often without considering their predictive power for the outcome. This approach may not always result in optimal bias reduction in the treatment effect estimate. To address this issue, evaluating covariate balance through prognostic scores, which account for the relationship between covariates and the outcome, has been proposed. Similarly, using a typical model averaging approach for PS estimation that minimizes prediction error for treatment status and covariate imbalance does not necessarily optimize PS-based confounding adjustment. As an alternative approach, using the averaged PS model that minimizes inter-group differences in the prognostic score may further reduce bias in the treatment effect estimate. Moreover, since the prognostic score is also an estimated quantity, model averaging in the prognostic scores can help identify a better prognostic score model. Utilizing the model-averaged prognostic scores as the balance metric for constructing the averaged PS model can contribute to further decreasing bias in treatment effect estimates. This paper demonstrates the effectiveness of the PS model averaging approach based on prognostic score balance and proposes a method that uses the model-averaged prognostic score as a balance metric, evaluating its performance through simulations and empirical analysis.MethodsWe conduct a series of simulations alongside an analysis of empirical observational data to compare the performances of weighted treatment effect estimates using the proposed and existing approaches. In our examination, we separately provid four candidate estimates for the PS and prognostic score models using traditional regression and machine learning methods. The model averaging of PS based on these candidate estimators is performed to either maximize the prediction accuracy of the treatment or to minimize intergroup differences in covariate distributions or prognostic scores. We also utilize not only the prognostic scores from each candidate model but also an averaged score that best predicted the outcome, for the balance assessment.ResultsThe simulation and empirical data analysis reveal that our proposed model-averaging approaches for PS estimation consistently yield lower bias and less variability in treatment effect estimates across various scenarios compared to existing methods. Specifically, using the optimally averaged prognostic scores as a balance metric significantly improves the robustness of the weighted treatment effect estimates.DiscussionThe prognostic score-based model averaging approach for estimating PS can outperform existing model averaging methods. In particular, the estimator using the model averaging prognostic score as a balance metric can produce more robust estimates. Since our results are obtained under relatively simple conditions, applying them to real data analysis requires adjustments to obtain accurate estimates according to the complexity and dimensionality of the data.ConclusionsUsing the prognostic score as the balance metric for the PS model averaging enhances the performance of the treatment effect estimator, which can be recommended for a wide variety of situations. When applying the proposed method to real-world data, it is important to use it in conjunction with techniques that mitigate issues arising from the complexity and high dimensionality of the data.

THE TRANSFORMATIVE IMPACT OF JUDO ON ATHLETES WITH DISABILITIES: INCLUSION, ADAPTATION, AND PERFORMANCE INSIGHTS

Judo is celebrated for its inclusivity and adaptability, offering significant benefits for athletes with disabilities (PwD). It fosters physical, emotional, social, and cognitive growth, enhancing motor skills, coordination, and balance. For athletes with physical disabilities, mastering techniques such as falling safely promotes strength and agility. Moreover, engaging in judo boosts self-esteem and confidence as athletes achieve milestones like advancing through belt ranks and earning medals. The supportive environment encourages personal accomplishment and helps develop social skills through interaction with peers. Research highlights the diverse range of disabilities in judo, including visual, auditory, and intellectual impairments. Studies by Oblak et al. (2020) reveal the therapeutic benefits of judo for individuals with intellectual disabilities, emphasizing the need for more research in this area. Bocioacă and Marin (2023) explore the role of adapted judo for individuals with Autism Spectrum Disorder (ASD) and Down syndrome, demonstrating improvements in integration skills through tailored practices. Additional research by Kons et al. (2021) and Krabben et al. (2017) addresses the dynamics of Paralympic judo, revealing challenges related to classification systems for visually impaired athletes. Kons et al. (2022) identify a concerning rise in injury rates among judo athletes with disabilities, indicating the need for improved safety measures. Lastly, the comparative study of neuromuscular control by Kons et al. (2019) shows that visually impaired judokas can excel in balance and certain strength metrics. Overall, these studies underscore judo’s role in promoting physical development and social integration for PwD, while also highlighting the need for further research to ensure equitable and inclusive practices in the sport.

Energy‐Water Asynchrony Principally Determines Water Available for Runoff From Snowmelt in Continental Montane Forests

ABSTRACTChanges in the volume, rate, and timing of the snowmelt water pulse have profound implications for seasonal soil moisture, evapotranspiration (ET), groundwater recharge, and downstream water availability, especially in the context of climate change. Here, we present an empirical analysis of water available for runoff using five eddy covariance towers located in continental montane forests across a regional gradient of snow depth, precipitation seasonality, and aridity. We specifically investigated how energy‐water asynchrony (i.e., snowmelt timing relative to atmospheric demand), surface water input intensity (rain and snowmelt), and observed winter ET (winter AET) impact multiple water balance metrics that determine water available for runoff (WAfR). Overall, we found that WAfR had the strongest relationship with energy‐water asynchrony (adjusted r2 = 0.52) and that winter AET was correlated to total water year evapotranspiration but not to other water balance metrics. Stepwise regression analysis demonstrated that none of the tested mechanisms were strongly related to the Budyko‐type runoff anomaly (highest adjusted r2 = 0.21). We, therefore, conclude that WAfR from continental montane forests is most sensitive to the degree of energy‐water asynchrony that occurs. The results of this empirical study identify the physical mechanisms driving variability of WAfR in continental montane forests and are thus broadly relevant to the hydrologic management and modelling communities.

Artificial intelligent-driven decision-making for automating root fracture detection in periapical radiographs

BackgroundThe detection and early diagnosis of root fractures can be challenging; this difficulty applies particularly to newly qualified dentists. Aside from clinical examination, diagnosis often requires radiographic assessment. Nonetheless, human fallibility can introduce errors due to a lack of experience.AimThe proposed system aimed to assist in detecting root fractures through the integration of artificial intelligence techniques into the diagnosis process as a step for automating dental diagnosis and decision-making processes.Materials and methodA total of 400 radiographic images of fractured and unfractured teeth were obtained for the present research. Data handling techniques were implemented to balance the distribution of the samples. The AI-based system used the voting technique for five different pretrained models namely, VGG16, VGG19, ResNet50. DenseNet121, and DenseNet169 to perform the analysis. The parameters used for the analysis of the models are loss and accuracy curves.ResultsVGG16 exhibited notable success with low training and validation losses (0.09% and 0.18%, respectively), high specificity, sensitivity, and positive predictive value (PPV). VGG19 showed potential overfitting concerns, while ResNet50 displayed progress in minimizing loss but exhibited bias toward unfractured cases. DenseNet121 effectively addressed overfitting and noise issues, achieving balanced metrics and impressive PPVs for both fractured and unfractured cases (0.933 and 0.898 respectively). With increased depth, DenseNet169 demonstrated enhanced generalization capability.ConclusionThe proposed AI- based system demonstrated high precision and sensitivity for detecting root fractures in endodontically treated teeth by utilizing the voting method.

Frölicher spectral sequence of compact complex manifolds with special Hermitian metrics

In this paper we focus on the interplay between the behaviour of the Frölicher spectral sequence and the existence of special Hermitian metrics on the manifold, such as balanced, SKT or generalized Gauduchon. The study of balanced metrics on nilmanifolds endowed with strongly non-nilpotent complex structures allows us to provide infinite families of compact balanced manifolds with Frölicher spectral sequence not degenerating at the second page. Moreover, this result is extended to non-degeneration at any arbitrary page. Similar results are obtained for the Frölicher spectral sequence of compact generalized Gauduchon manifolds. We also find a compact SKT manifold whose Frölicher spectral sequence does not degenerate at the second page, thus providing a counterexample to a conjecture by Popovici.

Utilizing Cobit 4.1 and Balance Scorecard to Manage Information Technology Business Process in Higher Education Institution

Pamulang University's Information Technology Strategy Plan aims to develop a comprehensive strategy to maximize IT efficiency and effectiveness in the institution. Using the COBIT 4.1 approach, SWOT analysis, and Balanced Scorecard, this plan identifies strengths, weaknesses, opportunities, and threats in the current IT infrastructure and develops an integrated strategy for IT development. COBIT 4.1 provides a framework for managing and controlling IT by ensuring that IT processes support business objectives and meet stakeholder needs. The SWOT analysis explores the internal and external factors that influence IT at Pamulang University, while the Balanced Scorecard helps in establishing balanced strategic metrics and objectives across multiple perspectives, including financial, customer, internal processes, and learning and growth. This strategic plan is expected to improve IT management, support the achievement of Pamulang University's goals and facilitate data-based decisions for innovation and continuous improvement.

The Impact of Global Oil Market Shocks on India’s Merchandise Trade Balances

This paper examines the influence of global oil market shocks on India’s aggregate, non-oil, and oil trade balances using the structural vector autoregression model. The impulse response analysis suggests that aggregate and non-oil trade balances deteriorate in response to adverse oil-specific demand shocks, while the oil trade balance improves. Among oil market shocks, only oil-specific demand shocks have a significant impact on all trade balance metrics. This research has crucial policy implications for the Indian economy.

Evaluation and comparison of covariate balance metrics in studies with time-dependent confounding.

Marginal structural models have been increasingly used by analysts in recent years to account for confounding bias in studies with time-varying treatments. The parameters of these models are often estimated using inverse probability of treatment weighting. To ensure that the estimated weights adequately control confounding, it is possible to check for residual imbalance between treatment groups in the weighted data. Several balance metrics have been developed and compared in the cross-sectional case but have not yet been evaluated and compared in longitudinal studies with time-varying treatment. We have first extended the definition of several balance metrics to the case of a time-varying treatment, with or without censoring. We then compared the performance of these balance metrics in a simulation study by assessing the strength of the association between their estimated level of imbalance and bias. We found that the Mahalanobis balance performed best. Finally, the method was illustrated for estimating the cumulative effect of statins exposure over one year on the risk of cardiovascular disease or death in people aged 65 and over in population-wide administrative data. This illustration confirms the feasibility of employing our proposed metrics in large databases with multiple time-points.

Addressing the sustainable urbanism paradox: tipping points for the operational reconciliation of dense and green morphologies

To improve biodiversity and human living conditions in the Anthropocene, urban forms must reconcile density with vegetation to meet the dual sustainability-liveability challenge. This paradox poses a dilemma for urban planners and is a critical research issue requiring comprehensive analyses. Multi-family residential housing holds the potential to achieve balanced density-greening, proximity ecosystem services and human-nature connectedness, but meeting such objectives relies on finding balanced morphologies and metrics at an operational scale. Analysing 11,593 plots in the Lyon metropolitan area (France) using a systemic approach, we identified critical tipping points in morphology and greening. Density explained only 6% of Plot Greening, while morphology and landscaping accounted for 94%. We identified an open-space ratio (unbuilt area/floor area) >0.3 as a morphological threshold to achieve sustainable green supply. Operational morphologies balancing density and greening were modelled and illustrated across building heights, providing guidelines for emerging regulatory tools in sustainable urban planning.

Effects of retained node numbers on berry maturity and yield components of cane-pruned Sauvignon blanc

Cane pruning is used in most New Zealand Sauvignon blanc vineyards to manage yield, vine balance (relationship between vegetative growth and fruit growth) and fruit primary and secondary metabolites. The source–sink ratio (TLA/FM—total leaf area to fruit mass or ELA/FM—exposed leaf area to fruit mass), the fruit mass to pruning mass (FM/PM), the fruit mass to cane mass (FM/CM) and fruit composition provide an assessment of the vine performance and balance. The interpretation of these metrics (i.e., TLA/FM, ELA/FM, FM/PM, FM/CM) requires their comparison with known optimal ranges specific to cultivars, locations and growing conditions. More often, such context- and cultivar-specific optimal ranges do not exist, thus warranting research to investigate them. To understand the influence of retained node numbers on the vegetative and fruit development of Sauvignon blanc, grapevines were pruned across three vineyard sites (two in Marlborough—Site 1 and 2, and one in Waipara—Site 3) over two growing seasons, retaining 10, 20, 30, 40 and 50 nodes on one to four canes (each cane carrying ten nodes, with 50-node vines carrying on average 12.5 nodes on each of the four canes). The accumulation of soluble solids (TSS) generally increased at lower node numbers and vine yields, reflecting an increase in ELA/FM, with 50-node vines having the least TSS concentration at harvest. The average berry mass, titrable acidity (TA) and pH were unaffected by node numbers over the two seasons. A low source–sink ratio induced by high node numbers not only reduced the vine capacity to ripen the current crop but also reduced the following season’s bunch number per shoot (from 1.8 to 1.6 bunches per shoot; p &lt; 0.05), average bunch mass (from 82.0 ± 6 g to 67.7 ± 3 g; p &lt; 0.01) and bunch mass per shoot (from 153.5 ± 15 g to 106.7 ± 9 g; p &lt; 0.05). When compared to 50-node vines, 10-node vines had a two-fold increase in the average cane mass (from 30.1 ± 3.9 g to 69.2 ± 8.7 g; p &lt; 0.001) and average old cane mass (from 82.4 ± 6.9 g to 163.8 ± 21 g; p &lt; 0.001). The ELA/FM and TLA/FM required for optimal TSS accumulation were 0.75 m2 kg-1 and 2.0 m2 kg-1, respectively, across all sites. A source–sink ratio above these values resulted in high average cane mass and average old cane mass (an indication of excess vigour), while lower values indicated reduced vigour and slowed TSS accumulation. This research provides useful optimal ranges to compare and interpret vine balance metrics measured at those sites.

Wearable sensor and smartphone assisted vestibular physical therapy for multiple sclerosis: usability and outcomes.

Vertigo, dizziness, gaze instability and disequilibrium are highly prevalent in people with MS (PwMS) and head movement induced dizziness is commonly reported. Vestibular physical therapy (VPT) is a specialised, non-invasive and effective therapy for these problems but usually involves travel for the person to a specialist center with both personal and carbon costs. The use of wearable sensors to track head movement and smartphone applications to deliver and track programs has potential to improve VPT in MS. This study investigated the usability and effects of a commercially available digital VPT system (wearable head sensor, smartphone app and clinician software) to deliver VPT to PwMS. A pre/post treatment design was employed and the primary outcome was the System Usability Scale (SUS). Other patient reported outcomes were the Service User Acceptability Questionnaire (SUTAQ), the Patient Enablement Instrument (PEI) and the Dizziness Handicap Inventory (DHI). Physical outcomes measurements included Mini-BESTest (MB), Modified Dynamic Gait Index (mDGI), Gait Speed (GS), Dynamic Visual Acuity (DVA) and head kinematics and symptoms during exercise. Sixteen PwMS (14 female), mean age 44(±14) years were recruited to the study and twelve completed VPT. Mean adherence to exercise, measured digitally was 60% (±18.4). SUS scores were high at 81 (±14) and SUTAQ scores also demonstrated high levels of satisfaction and acceptability of the system. Statistically significant improvements in MB (mean change 2.25; p = 0.004), mDGI (median change 1.00; p = 0.008), DVA (median change -1.00; p = 0.004) were found. Head frequencies significantly improved with concurrent decreased intensity of dizziness during head movements (mean change across 4 gaze stabilization exercises was 23 beats per minute; p < 0.05). Non-significant improvements were seen in DHI (p = 0.07) and GS (p = 0.15). 64.5% of follow up visits were conducted remotely (video or phone), facilitated by the system. This study had two main outcomes and benefits for PwMS. Firstly, we showed that the system used was both acceptable and could be used by PwMS. Secondly, we demonstrated an improvement in a range of dizziness, balance and gait metrics with remotely delivered care. This system has the potential to positively impact on MS physiotherapy service provision with the potential to deliver effective remote care.

Outpatient Factors Implicated in Timing of Delayed Elective Inguinal Hernia Repair in Premature Neonates: Single Center Analysis Brings Up a Touchy Issue and Balancing Metric

IntroductionAmong premature infants, the incidence of inguinal hernias (IH) has been reported to be as high as 10–30%. We performed this study to characterize the association between individual and systemic variables that may affect diagnosis to definitive operative repair of the premature neonatal IH in the outpatient setting. MethodsA single center cohort retrospective review analyzing IH repair in the premature neonatal (<37 GA) population was performed. Data was collected between 2013 and 2022. The cohort was defined as patients who underwent repair before the age of 1 and excluded patients with major medical comorbidities or underwent simultaneous major abdominal surgeries. ResultsOf the 836 premature neonates who underwent IH repair, the majority (73%) were repaired electively. Patients were characterized into risk cohorts a-priori. High-risk patients (HR, n = 43) were more likely to have Government insurance (67%). There was a significant difference in HR patient time to surgery between Government versus Commercial insurance, 10.6 versus 4.7 days, respectively (95% CI -11.09 to −0.4396, p = 0.0345). HR patients were also seen more frequently (clinic or emergency department) prior to operative repair (2.51 vs 1.72 95% CI –1.296% to −0.289%, p = 0.0021). A multivariate linear regression model demonstrated that risk class (p = 0.0244), touches (p < 0.0001), GA (p < 0.001), and prior authorization (p < 0.0001) were significantly associated with time to hernia repair. ConclusionsSystemic variables such as insurance type may increase average wait time for elective outpatient IH repair. This increase in wait time is associated with an increased number of healthcare visits. Therefore, timely access to surgical care prevents potential harm and may decrease health system burden. Type of StudyRetrospective comparative study/cohort study. Level of EvidenceIII.

Empowering healthcare with NLP-driven deep learning unveiling biomedical materials through text mining

This study presents a comprehensive approach to automated biomedical materials discovery in healthcare applications by integrating text mining and deep learning techniques. The research methodology encompasses two main components: exploration of key research questions through graphical representation and evaluation of model performance using precision, recall, and F1-score metrics. The identification of pertinent research questions is visualized using bar charts, offering insights into the distribution of studies across domains such as data harmonization, heterogeneity, industrial textual data, and sequential data performance. The precision comparison chart highlights the strengths and weaknesses of different models, with model 1 demonstrating notable precision. The recall comparison chart emphasizes model 2’s outstanding performance in capturing relevant information, while the f1-score comparison chart showcases the balanced metrics of model 2 and 4. These visual analyses contribute to a nuanced understanding of the research landscape and guide the development of the proposed NLP-ML pipeline. The study’s findings underscore the significance of addressing data harmonization challenges and extracting insights from industrial textual data in advancing biomedical materials discovery. Overall, this research amalgamates exploratory data analysis and quantitative model evaluation to contribute to the evolving field of text mining and deep learning applications in biomedical material discovery for healthcare applications.

Leveraging machine learning for predicting acute graft-versus-host disease grades in allogeneic hematopoietic cell transplantation for T-cell prolymphocytic leukaemia

Orphan diseases, exemplified by T-cell prolymphocytic leukemia, present inherent challenges due to limited data availability and complexities in effective care. This study delves into harnessing the potential of machine learning to enhance care strategies for orphan diseases, specifically focusing on allogeneic hematopoietic cell transplantation (allo-HCT) in T-cell prolymphocytic leukemia. The investigation evaluates how varying numbers of variables impact model performance, considering the rarity of the disease. Utilizing data from the Center for International Blood and Marrow Transplant Research, the study scrutinizes outcomes following allo-HCT for T-cell prolymphocytic leukemia. Diverse machine learning models were developed to forecast acute graft-versus-host disease (aGvHD) occurrence and its distinct grades post-allo-HCT. Assessment of model performance relied on balanced accuracy, F1 score, and ROC AUC metrics. The findings highlight the Linear Discriminant Analysis (LDA) classifier achieving the highest testing balanced accuracy of 0.58 in predicting aGvHD. However, challenges arose in its performance during multi-class classification tasks. While affirming the potential of machine learning in enhancing care for orphan diseases, the study underscores the impact of limited data and disease rarity on model performance.

Using decision tree models and comprehensive statewide data to predict opioid overdoses following prison release

PurposeIdentifying predictors of opioid overdose following release from prison is critical for opioid overdose prevention. MethodsWe leveraged an individually linked, state-wide database from 2015–2020 to predict the risk of opioid overdose within 90 days of release from Massachusetts state prisons. We developed two decision tree modeling schemes: a model fit on all individuals with a single weight for those that experienced an opioid overdose and models stratified by race/ethnicity. We compared the performance of each model using several performance measures and identified factors that were most predictive of opioid overdose within racial/ethnic groups and across models. ResultsWe found that out of 44,246 prison releases in Massachusetts between 2015–2020, 2237 (5.1%) resulted in opioid overdose in the 90 days following release. The performance of the two predictive models varied. The single weight model had high sensitivity (79%) and low specificity (56%) for predicting opioid overdose and was more sensitive for White non-Hispanic individuals (sensitivity = 84%) than for racial/ethnic minority individuals. ConclusionsStratified models had better balanced performance metrics for both White non-Hispanic and racial/ethnic minority groups and identified different predictors of overdose between racial/ethnic groups. Across racial/ethnic groups and models, involuntary commitment (involuntary treatment for alcohol/substance use disorder) was an important predictor of opioid overdose.

Abnormalities in both stimulus-induced and baseline MEG alpha oscillations in the auditory cortex of children with Autism Spectrum Disorder.

The neurobiology of Autism Spectrum Disorder (ASD) is hypothetically related to the imbalance between neural excitation (E) and inhibition (I). Different studies have revealed that alpha-band (8-12Hz) activity in magneto- and electroencephalography (MEG and EEG) may reflect E and I processes and, thus, can be of particular interest in ASD research. Previous findings indicated alterations in event-related and baseline alpha activity in different cortical systems in individuals with ASD, and these abnormalities were associated with core and co-occurring conditions of ASD. However, the knowledge on auditory alpha oscillations in this population is limited. This MEG study investigated stimulus-induced (Event-Related Desynchronization, ERD) and baseline alpha-band activity (both periodic and aperiodic) in the auditory cortex and also the relationships between these neural activities and behavioral measures of children with ASD. Ninety amplitude-modulated tones were presented to two groups of children: 20 children with ASD (5 girls, Mage = 10.03, SD = 1.7) and 20 typically developing controls (9 girls, Mage = 9.11, SD = 1.3). Children with ASD had a bilateral reduction of alpha-band ERD, reduced baseline aperiodic-adjusted alpha power, and flattened aperiodic exponent in comparison to TD children. Moreover, lower raw baseline alpha power and aperiodic offset in the language-dominant left auditory cortex were associated with better language skills of children with ASD measured in formal assessment. The findings highlighted the alterations of E / I balance metrics in response to basic auditory stimuli in children with ASD and also provided evidence for the contribution of low-level processing to language difficulties in ASD.