The Chi-Square Test of Distance Correlation

Quantitative image quality metrics enable resource-efficient quality control of clinically applied AI-based reconstructions in MRI

OBJECTIVE

AI-based MRI reconstruction techniques improve efficiency by reducing acquisition times whilst maintaining or improving image quality. Recent recommendations from professional bodies suggest centres should perform quality assessments on AI tools. However, monitoring long-term performance presents challenges, due to model drift or system updates. Radiologist-based assessments are resource-intensive and may be subjective, highlighting the need for efficient quality control (QC) measures. This study explores using image quality metrics (IQMs) to assess AI-based reconstructions.

MATERIALS AND METHODS

58 patients undergoing standard-of-care rectal MRI were imaged using AI-based and conventional T2-weighted sequences. Paired and unpaired IQMs were calculated. Sensitivity of IQMs to detect retrospective perturbations in AI-based reconstructions was assessed using control charts, and statistical comparisons between the four MR systems in the evaluation were performed. Two radiologists evaluated the image quality of the perturbed images, giving an indication of their clinical relevance.

RESULTS

Paired IQMs demonstrated sensitivity to changes in AI-reconstruction settings, identifying deviations outside ± 2 standard deviations of the reference dataset. Unpaired metrics showed less sensitivity. Paired IQMs showed no difference in performance between 1.5 T and 3 T systems (p > 0.99), whilst minor but significant (p < 0.0379) differences were noted for unpaired IQMs.

DISCUSSION

IQMs are effective for QC of AI-based MR reconstructions, offering resource-efficient alternatives to repeated radiologist evaluations. Future work should expand this to other imaging applications and assess additional measures.

Distinguishing cause from effect in psychological research: An independence-based approach under linear non-Gaussian models

Distinguishing cause from effect - that is, determining whether x causes y (x → y) or, alternatively, whether y causes x (y → x) - is a primary research goal in many psychological research areas. Despite its importance, determining causal direction with observational data remains a difficult task. In this study, we introduce an independence-based approach for causal discovery between two variables of interest under a linear non-Gaussian model framework. We propose a two-step algorithm based on distance correlations that provides empirical conclusions on the causal directionality of effects under realistic conditions typically seen in psychological studies, that is, in the presence of hidden confounders. The performance of the proposed algorithm is evaluated using Monte-Carlo simulations. Findings suggest that the algorithm can effectively detect the causal direction between two variables of interest, even in the presence of weak hidden confounders. Moreover, distance correlations provide useful insights into the magnitude of hidden confounding. We provide an empirical example to demonstrate the application of our proposed approach and discuss practical implications and future directions.

Network Biomarkers of Alzheimer's Disease Risk Derived from Joint Volume and Texture Covariance Patterns in Mouse Models

Alzheimer's disease (AD) lacks effective cures and is typically detected after substantial pathological changes have occurred, making intervention challenging. Early detection and understanding of risk factors and their downstream effects are therefore crucial. Animal models provide valuable tools to study these prodromal stages. We investigated various levels of genetic risk for AD using mice expressing the three major human APOE alleles in place of mouse APOE. We leverage these mouse models utilizing high-resolution magnetic resonance diffusion imaging, due to its ability to provide multiple parameters that can be analysed jointly. We examine how APOE genotype interacts with age, sex, diet, and immunity to yield jointly discernable changes in regional brain volume and fractional anisotropy, a sensitive metric for brain water diffusion. Our results demonstrate that genotype strongly influences the caudate putamen, pons, cingulate cortex, and cerebellum, while sex affects the amygdala and piriform cortex bilaterally. Immune status impacts numerous regions, including the parietal association cortices, thalamus, auditory cortex, V1, and bilateral dentate cerebellar nuclei. Risk factor interactions particularly affect the amygdala, thalamus, and pons. APOE2 mice on a regular diet exhibited the fewest temporal changes, suggesting resilience, while APOE3 mice showed minimal effects from a high-fat diet (HFD). HFD amplified aging effects across multiple brain regions. The interaction of AD risk factors, including diet, revealed significant changes in the periaqueductal gray, pons, amygdala, inferior colliculus, M1, and ventral orbital cortex. Future studies should investigate the mechanisms underlying these coordinated changes in volume and texture, potentially by examining network similarities in gene expression and metabolism, and their relationship to structural pathways involved in neurodegenerative disease progression.

GradeDiff-IM: an ensembles model-based grade classification of breast cancer

Cancer grade classification is a challenging task identified from the cell structure of healthy and abnormal tissues. The practitioners learns about the malignant cell through the grading and plans the treatment strategy accordingly. A major portion of researchers used DL models for grade classification. However, the behavior of DL models is hidden type, it is unknown which features contribute to the accuracy and how the features are chosen for grading. To address the issue the study proposes a Grade Differentiation Integrated Model (GradeDiff-IM) to classify the grades G1, G2, and G3. In GradeDiff-IM, different ML models, are used for grade classification from clinical and pathological reports. The biological-significant features with ranking technique prioritize influential features are used to identify grades G. Subsequently, histopathological images are used by DL models for grade classification and compared with ML models. Instead of employing a single ML model, the GradeDiff-IM model uses the stack-ensembled approach to improve the grade G classification performance. The maximum accuracy is attained by stacking G1-98.2, G2-97.6, and G3-97.5. The proposed study shows that the ML ensemble model is more accurate than the DL models. As a result, the proposed model achieved higher accuracy for G by implementing the stacking technique than the other state-of-the-art models.

An autonomous vehicles' test case extraction method: Example of vehicle-to-pedestrian scenarios

Testing autonomous vehicles (AVs) in hazardous scenarios is a crucial technical approach to ensure their safety. A key aspect of this process is the generation of hazard scenarios. In general, such scenarios are generated through cluster analysis of traffic accident data. However, this approach may not fully capture the criticality of the generated scenarios, as it tends to emphasize the statistical characteristics of the data rather than its real-world applicability. This paper proposes a novel method to enhance scenario adaptation by integrating quantization weights with a new clustering algorithm. These weights, representing the correlation between scenario elements and the AV system, are calculated using fuzzy comprehensive evaluation (FCE). The proposed method is applied to 1044 pedestrian accident cases in China, resulting in the identification of nine categories of typical scenarios and corresponding test schemes for both the perception and decision-making systems of AVs. The results show that the new method increases the proportion of critical scenarios by 17.4 % and 13.6 %, respectively, compared to traditional methods. Overall, the critical scenarios generated in this paper can significantly improve the testing efficiency and safety of AVs.

Universally Consistent K-Sample Tests via Dependence Measures

The K-sample testing problem involves determining whether K groups of data points are each drawn from the same distribution. Analysis of variance is arguably the most classical method to test mean differences, along with several recent methods to test distributional differences. In this paper, we demonstrate the existence of a transformation that allows K-sample testing to be carried out using any dependence measure. Consequently, universally consistent K-sample testing can be achieved using a universally consistent dependence measure, such as distance correlation and the Hilbert-Schmidt independence criterion. This enables a wide range of dependence measures to be easily applied to K-sample testing.

Assessing body mass index stages, individual diabetes and hypertension history effects on the risk of developing hypertension among Ghanaians: A cross-sectional study

Background and Aims

This study aimed to understand the relationship between body mass index (BMI), diabetes and hypertension history, and other risk of hypertension among Ghanaians.

Methods

The BMI data are categorized according to the World Health Organization (WHO) definition. The data were obtained from the WHO Study on global AGEing and adult health (WHO SAGE) Ghana Wave 2. Descriptive statistics were used to summarize the variables, and the association between these variables and hypertension was assessed using the χ². Multivariable logistic regression was used to examine the relationship between hypertension and different BMI levels and other variables.

Results

Obesity class II individuals have about a 4-fold higher risk of developing hypertension compared to underweight individuals. Obesity class III, class I, and preobesity individuals have approximately a 3-fold higher risk. Normal weight is associated with increased hypertension risk. Both males and females show a significant increase in hypertension risk across all BMI categories. History of hypertension is linked to a 2.2-fold increased risk. Diabetes history is associated with hypertension when considering other factors. Elevated hypertension risk is observed among married, divorced, and widowed males then never married males. Only widowed females showed an increased risk. Older age significantly increases hypertension risk, particularly in females. Vegetable servings reduce hypertension risk, while fruit servings are associated with an increased risk. Vigorous exercise increases hypertension risk, particularly in females.

Conclusion

Regular check-ups are recommended for married, divorced, and widowed males, focusing on blood pressure (BP) levels. Regular exercise from young age helps lower BP in later years. Individuals with a history of hypertension should follow BP control measures. Encouraging the consumption of the right combination of vegetables and fruits can help lower BP. Female tobacco smoking should be strongly discouraged due to a 54% increased risk of developing hypertension.

Mixed effects logistic regression analysis of blood pressure among Ghanaians and associated risk factors

Blood pressure (BP) control is a global health issue with an increase in BP beyond the normal BP leading to different stages of hypertension in humans and hence the need to identify risk factors of BP for efficient and effective control. Multiple BP measurement have proven to provide BP readings close to the true BP status of the individual. In this study, we used multiple BP measurement data on 3809 Ghanaians to determine risk factors associated with BP. The data were obtained from World Health Organization study on Global AGEing and Adult Health. We defined high blood pressure (HBP) as [Formula: see text] 130/80 mmHg or normal as [Formula: see text] 130/80 mmHg. We provide summary statistics and also used the Chi-Square test to assess significance of association between HBP versus risk factors of HBP. The aim of this study is to identify risk factors of BP using the mixed effects logistic regression model. Data were analyzed using R version 4.2.2. The results showed that the risk of high blood pressure (HBP) decreases across the three measurement periods. There is reduced risk (OR = 0.274, 95% CI = 0.2008, 0.405) of HBP among male participants relative to female participants. The risk (OR = 2.771, 95% CI = 1.8658, 4.1145) of HBP increased by 2.771-folds among those who are 60 years and above relative to those below the age of 60 years. Those whose work involves/requires vigorous exercise has 1.631-fold increase in the risk (OR = 1.631, 95% CI = 1.1151, 2.3854) of HBP relative to those whose work does not involve vigorous exercise. There is approximately 5-folds increased in the risk (OR = 4.896, 95% CI = 1.9535, 12.2268) of among those who have ever been diagnosed with diabetes. The results also revealed high risk (OR = 1.649, 95%CI = 1.1108, 2.4486) of HBP among those who have formal education. The risk (OR = 1.009, 95% CI = 1.0044, 1.0137) of HBP increases with increasing weight and a reduced risk (OR = 0.996, 95% CI = 0.9921, 0.9993) of HBP with increasing height. We found that sad experience, either mild, moderate or severe, is associated with a reduced risk of HBP. Those who have vegetable servings at least 2 cups per day have increased risk of HBP and those who have fruits servings at least 2 cups per day is associated with a reduced risk of HBP, however this is not statistically significant. To achieve success in BP control, programs should be designed with the aim of reducing weight, educate those with formal eduction on issues relating to HBP. Those whose work requires vigorous exercise are recommended to have regular check-ups to ensure that pressure build-up in the lungs is cleared. SBP is lower for women at young age but continue to increase after menopause as their BP increase becomes salt-sensitive. Hence there is need to give more attention to menopausal women so as to improve BP. Both young and old individuals are recommended to practice regular exercise since this has shown to reduce risk of being overweight or becoming diabetic and reduces the risk of HBP at yong age and old age. Also, to improve blood pressure control, programs for management of blood pressure or hypertension should focus more short stature individuals since such people are more likely to experience HBP.

Investigating the performance of machine learning algorithms in predicting the survival of COVID-19 patients: A cross section study of Iran

Background and Aims

Like early diagnosis, predicting the survival of patients with Coronavirus Disease 2019 (COVID-19) is of great importance. Survival prediction models help doctors be more cautious to treat the patients who are at high risk of dying because of medical conditions. This study aims to predict the survival of hospitalized patients with COVID-19 by comparing the accuracy of machine learning (ML) models.

Methods

It is a cross-sectional study which was performed in 2022 in Fasa city in Iran country. The research data set was extracted from the period February 18, 2020 to February 10, 2021, and contains 2442 hospitalized patients' records with 84 features. A comparison was made between the efficiency of five ML algorithms to predict survival, includes Naive Bayes (NB), K-nearest neighbors (KNN), random forest (RF), decision tree (DT), and multilayer perceptron (MLP). Modeling steps were done with Python language in the Anaconda Navigator 3 environment.

Results

Our findings show that NB algorithm had better performance than others with accuracy, precision, recall, F-score, and area under receiver operating characteristic curve of 97%, 96%, 96%, 96%, and 97%, respectively. Based on the analysis of factors affecting survival, heart disease, pulmonary diseases and blood related disease were the most important disease related to death.

Conclusion

The development of software systems based on NB will be effective to predict the survival of COVID-19 patients.

Designing an Embedded Feature Selection Algorithm for a Drowsiness Detector Model Based on Electroencephalogram Data

Driver fatigue reduces the safety of traditional driving and limits the widespread adoption of self-driving cars; hence, the monitoring and early detection of drivers' drowsiness plays a key role in driving automation. When representing the drowsiness indicators as large feature vectors, fitting a machine learning model to the problem becomes challenging, and the problem's perspicuity decreases, making dimensionality reduction crucial in practice. For this reason, we propose an embedded feature selection algorithm that can be later utilized as a building block in the system development of a neural network-based drowsiness detector. We have adopted a technique: a so-called Feature Prune Layer is placed in front of the first layer in the architecture; as a result, its weights change regarding the importance of the corresponding input features and are deleted iteratively until the desired number is reached. We test the algorithm on EEG data, as it is one of the best indicators of drowsiness based on the literature. The proposed FS algorithm is able to reduce the original feature set by 95% with only 1% degradation in precision, while the precision increases by 1.5% and 2.7% respectively when selecting the top 10% and top 20% of the initial features. Moreover, the proposed method outperforms the widely popular Principal Component Analysis and the Chi-squared test when reducing the original feature set by 95%: it achieves 24.3% and 3.2% higher precision respectively.

Significance of Nuclear Factor-Kappa B (NF-κB) and Survivin in Breast Cancer and Their Association with Radiosensitivity and Prognosis

Purpose

Analyze the expression of NF-κB and survivin genes and mRNAs in breast cancer, and evaluate their impact on prognosis. Investigate their association with radiosensitivity in breast cancer.

Methods

The expression levels of NF-κB and survivin genes in breast cancer were analyzed by bioinformatics, NF-κB and survivin mRNA was verified by RTRCR, and their association with prognosis were assessed. Knockdown of survivin by siRNA was used to analyze its association with radiosensitivity in breast cancer.

Results

The gene expression of NFKB1 and BIRC5 are differentially expressed in a variety of tumours and their corresponding normal tissue species. In breast cancer tissues, NFKB1 expression levels were reduced compared to normal tissue, while BIRC5 expression levels were increased (P<0.05). In different molecular subtypes of breast cancer, NFKB1 and BIRC5 were differentially expressed (P<0.05), NFKB1 was highly expressed in the luminal subtype and BIRC5 was highly expressed in the TNBC subtype. In TNBC subtype, NFKB1 expression is higher in IM subtype than other subtypes (P<0.05), and BIRC5 expression is higher in BL-2 than other subtypes (P<0.05). NFKB1 was not associated with tumour size, lymph node stage and distant metastasis (P≥0.05), while BRIC5 was associated with these clinical features (P<0.05). NF-κB and survivin genes were negatively correlated (R = - 0.193, P<0.05). The mRNA levels of NF-κB and survivin are expressed in the same trend in breast cancer patients. NF-κB and survivin were not significantly different in recurrent and non-recurrent patients (P≥0.05). The mRNA levels of the both were not correlated with breast cancer subtypes (P≥0.05). The mRNA expression of NF-κB and survivin correlated with distant metastasis. NF-κB and survivin mRNAs were positively correlated (R=0.903, P<0.05). Gene and mRNA expression of NF-κB and survivin were not associated with patients' survival overall survival (OS) (P≥0.05). Down-regulation of survivin has little effect on the proliferation rate of breast cancer cells (P≥0.05), but increase the apoptosis rate of breast cancer cells (P<0.05).The proliferation rate of cells decreased and the apoptosis rate increased significantly (P<0.05) after the implementation of radiotherapy, and this technique could improve the radiosensitivity of breast cancer cells.

Conclusion

NF-κB and survivin interact at the gene and mRNA levels. Regulation of mRNA expression of NF-κB or survivin may help to improve the radiosensitivity of breast cancer cells, more experiments are needed to verify this in the future.

The FEDHC Bayesian Network Learning Algorithm

The paper proposes a new hybrid Bayesian network learning algorithm, termed Forward Early Dropping Hill Climbing (FEDHC), devised to work with either continuous or categorical variables. Further, the paper manifests that the only implementation of MMHC in the statistical software R is prohibitively expensive, and a new implementation is offered. Further, specifically for the case of continuous data, a robust to outliers version of FEDHC, which can be adopted by other BN learning algorithms, is proposed. The FEDHC is tested via Monte Carlo simulations that distinctly show that it is computationally efficient, and that it produces Bayesian networks of similar to, or of higher accuracy than MMHC and PCHC. Finally, an application of FEDHC, PCHC and MMHC algorithms to real data, from the field of economics, is demonstrated using the statistical software R.

Telework Systematic Model Design for the Future of Work

The practice and popularity of telework has expanded significantly in the past few years, mostly due to the COVID-19 pandemic. As a critical factor contributing to business resilience, the new work model challenged companies to figure out innovative ways to address contemporary organizational and employee needs. To address this gap, this study approaches the telework concept from a broader perspective, integrating inputs, outputs and outcomes in an analytical framework. Drawing from data collected based on interviews and questionnaires addressed to professionals in the business service industry who experienced telework, frequency analysis, discourse analysis and chi-square test were used to synthesize the findings. Results show that resource availability and professional relationships represent the basic factors, while technology may be more than a facilitator. Moreover, knowledge exchange, work–life balance and professional isolation are critical factors emerging from the virtual environment that influence work goals achievement. This study contributes to research by proposing a Telework Systematic Model (TSM), which addresses the interaction of various organizational dynamics factors as a result of mixed working patterns. The discussions address the future of work by including the hybrid work model, platform innovation and new business opportunities to enhance organizational resilience for sustainable innovation and change through digital technology.

Understanding the Drivers of Production in South African Farming Systems: A Case Study of the Vhembe District, Limpopo South Africa

Farming systems in South Africa operate against the backdrop of constantly changing environmental, political, and socio-economic conditions. Farming systems are commonly defined by the Food and Agriculture Organization (FAO) as a population of individual farm systems that have broadly similar resource bases, enterprise patterns, household livelihoods and constraints, and for which similar development strategies and interventions would be appropriate. Historically farming systems in South Africa have been characterised by dualism in which large-scale commercial farmers co-exist with small-scale farmers. Although the two farming systems are impacted by the same drivers of production (land, labour, capital, and enterprise), however, they respond to these drivers differently and the nature of the responses reveal their connectivity and possible approaches to sustaining them. A systems thinking approach is best suited to draw possible scenarios of how farming systems in the Vhembe district located in the Limpopo Province of South Africa will respond to changes with respect to the four drivers. In this area, large-scale commercial farming forms a significant component of the production of a number of subtropical crops that contribute to the country's agricultural economy particularly through exports. Simultaneously 90% of rural communities in the district depend mainly on small-scale agriculture to sustain their livelihoods and generate income. The paper provides an overview of the drivers of production for the two farming systems in the Vhembe district and explores how the government can successfully promote development through agriculture by building capacity for the joint success of the two farming systems.

A Comparison of Functional Features of Chinese and US Mobile Apps for Pregnancy and Postnatal Care: A Systematic App Store Search and Content Analysis

BACKGROUND

Pregnancy to postpartum (PtP) applications (apps) are becoming more common tools to document everything from pregnancy and delivery to nutrient allocation, life taboos, and infant medical examinations. However, the dependability, quality, and efficacy of these apps remain unclear. This study examined the features and functions of mobile PtP care apps accessible in China and the United States and to identify the major gaps that need to be addressed.

METHODS

Apps were selected by searching the Apple App Store and Android Markets (in the US and China) for the terms "pregnancy" and "postpartum" in Chinese and English. The apps' security, quality, and effectiveness were investigated, and chi-square tests and analysis of variance were performed to examine the differences in characteristics between apps available in the US and China.

RESULTS

A total of 84 mobile PtP care apps (45 from the US and 39 from China) were included. A total of 89.7% (35/39) of Chinese mobile apps did not provide safety statements or supporting evidence. The objective app quality ratings for Chinese and US apps were 3.20 ± 0.48 (mean ± standard deviation) and 3.56 ± 0.45, respectively (p > 0.05). A greater number of Chinese apps provided app-based monitoring functions, namely recording fetal size (n = 18, 46.2% in China vs. n = 3, 6.7% in the US), contractions (n = 11, 28.2% in China vs. n = 0, 0% in the US), pregnancy weight (n = 11, 28.2% in China vs. 0, 0% in the US), and pregnancy check-up reminders (n = 10, 25.6% in China vs. n = 0, 0% in the US). Meanwhile, a greater number of US apps provided exercise modules, namely pregnancy yoga (n = 2, 5.1% in China vs. n = 21, 46.7% in the US), pregnancy workouts (n = 2, 5.1% in China vs. n = 13, 28.9% in the US), and pregnancy meditation (n = 0, 0% in China vs. 10, 22.2% in the US) (p < 0.01). A medium security risk was identified for 40% (18/45) of apps in the US and 82.1% (32/39) of apps in China (p < 0.01).

CONCLUSIONS

The functionality and characteristics of in-store mobile apps for PtP care varied between China and the US. Both countries' apps, particularly Chinese apps, encountered issues related to a lack of evidence-based information, acceptable content risk, and program evaluations. Both countries' apps lacked proper mental health care functions. The findings suggest that the design of app features should be enhanced in both countries, and increased interaction between app creators and users is recommended.