RESUMEN

Recleaning phosphate tailings using the low-cost enhanced gravity separation method is beneficial for maximizing the recovery of phosphorus element. A machine learning framework was constructed to predict the target variables of the yield, grade, and recovery from the feature variables of slurry concentration, backwash water pressure, and rotational frequency of bowl, whose data came from the phosphate tailings separation experiments in the enhanced gravity field. The coefficient of determination R2 and mean squared error were used to evaluate the performance of seven machine learning models. After hyper-parameter optimization, GBR demonstrated the best performance in predicting yield, grade, and recovery, with prediction accuracy of 95.58 %, 90.72 %, and 94.25 %, respectively. SHapley Additive exPlanations interpretability analysis revealed that the rotational frequency of the bowl had the most significant impact on the grade and recovery of concentrates, while slurry concentration had the most significant effect on the yield. A lower rotational frequency of the bowl, a higher slurry concentration, and an increased backwash water pressure were positively correlated with both the yield and recovery. However, the grade was favorably correlated with a higher rotational frequency of bowl and a lower slurry concentration, whereas its correlation with the backwash water pressure could be positive or adverse, depending on its specific value. The limitations and implications of these findings were also demonstrated, and the constructed framework was anticipated to achieve higher prediction accuracy with reasonable interpretability in further studies.

RESUMEN

Scene Perception and Event Comprehension Theory (SPECT) posits that understanding picture stories depends upon a coordination of two processes: (1) integrating new information into the current event model that is coherent with it (i.e., mapping) and (2) segmenting experiences into distinct event models (i.e., shifting). In two experiments, we investigated competing hypotheses regarding how viewers coordinate the mapping process of bridging inference generation and the shifting process of event segmentation by manipulating the presence/absence of Bridging Action pictures (i.e., creating coherence gaps) in wordless picture stories. The Computational Effort Hypothesis says that experiencing a coherence gap prompts event segmentation and the additional computational effort to generate bridging inferences. Thus, it predicted a positive relationship between event segmentation and explanations when Bridging Actions were absent. Alternatively, the Coherence Gap Resolution Hypothesis says that experiencing a coherence gap prompt generating a bridging inference to close the gap, which obviates segmentation. Thus, it predicted a negative relationship between event segmentation and the production of explanations. Replicating prior work, viewers were more likely to segment and generate explanations when Bridging Action pictures were absent than when they were present. Crucially, the relationship between explanations and segmentation was negative when Bridging Action pictures were absent, consistent with the Coherence Gap Resolution Hypothesis. Unexpectedly, the relationship was positive when Bridging Actions were present. The results are consistent with SPECT's assumption that mapping and shifting processes are coordinated, but how they are coordinated depends upon the experience of a coherence gap.

RESUMEN

Accurately predicting the concentration of organochlorine pesticides (OCPs) presents a challenge due to their complex sources and environmental behaviors. In this study, we introduced a novel and advanced model that combined the power of three distinct techniques: Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN), Variational Mode Decomposition (VMD), and a deep learning network of Long Short-Term Memory (LSTM). The objective is to characterize the variation in OCPs concentrations with high precision. Results show that the hybrid two-stage decomposition coupled models achieved an average symmetric mean absolute percentage error (SMAPE) of 23.24 % in the empirical analysis of typical surface water. It exhibited higher predictive power than the given individual benchmark models, which yielded an average SMAPE of 40.88 %, and single decomposition coupled models with an average SMAPE of 29.80 %. The proposed CEEMDAN-VMD-LSTM model, with an average SMAPE of 13.55 %, consistently outperformed the other models, yielding an average SMAPE of 33.53 %. A comparative analysis with shallow neural network methods demonstrated the advantages of the LSTM algorithm when coupled with secondary decomposition techniques for processing time series datasets. Furthermore, the interpretable analysis derived by the SHAP approach revealed that precipitation followed by the total phosphorus had strong effects on the predicted concentration of OCPs in the given water. The data presented herein shows the effectiveness of decomposition technique-based deep learning algorithms in capturing the dynamic characteristics of pollutants in surface water.

RESUMEN

Reliably detecting potentially misleading patterns in automated diagnostic assistance systems, such as those powered by artificial intelligence (AI), is crucial for instilling user trust and ensuring reliability. Current techniques fall short in visualizing such confounding factors. We propose DiffChest, a self-conditioned diffusion model trained on 515,704 chest radiographs from 194,956 patients across the US and Europe. DiffChest provides patient-specific explanations and visualizes confounding factors that might mislead the model. The high inter-reader agreement, with Fleiss' kappa values of 0.8 or higher, validates its capability to identify treatment-related confounders. Confounders are accurately detected with 10%-100% prevalence rates. The pretraining process optimizes the model for relevant imaging information, resulting in excellent diagnostic accuracy for 11 chest conditions, including pleural effusion and heart insufficiency. Our findings highlight the potential of diffusion models in medical image classification, providing insights into confounding factors and enhancing model robustness and reliability.

Asunto(s)

RESUMEN

Background: Follicular thyroid carcinoma (FTC) and follicular thyroid adenoma (FTA) present diagnostic challenges due to overlapping clinical and ultrasound features. Improving the diagnosis of FTC can enhance patient prognosis and effectiveness in clinical management. This study seeks to develop a predictive model for FTC based on ultrasound features using machine learning (ML) algorithms and assess its diagnostic effectiveness. Methods: Patients diagnosed with FTA or FTC based on surgical pathology between January 2009 and February 2023 at Zhejiang Provincial Cancer Hospital and Zhejiang Provincial People's Hospital were retrospectively included. A total of 562 patients from Zhejiang Provincial Cancer Hospital comprised the training set, and 218 patients from Zhejiang Provincial People's Hospital constituted the validation set. Subsequently, clinical parameters and ultrasound characteristics of the patients were collected. The diagnostic parameters were analyzed using the least absolute shrinkage and selection operator and multivariate logistic regression screening methods. Next, a comparative analysis was performed using seven ML models. The area under the receiver operating characteristic (ROC) curve (AUC), accuracy, sensitivity, specificity, positive predicted value (PPV), negative predicted value (NPV), precision, recall, and comprehensive evaluation index (F-score) were calculated to compare the diagnostic efficacy among the seven models and determine the optimal model. Further, the optimal model was validated, and the SHapley Additive ExPlanations (SHAP) approach was applied to explain the significance of the model variables. Finally, an individualized risk assessment was conducted. Results: Age, echogenicity, thyroglobulin antibody (TGAb), echotexture, composition, triiodothyronine (T3), thyroglobulin (TG), margin, thyroid-stimulating hormone (TSH), calcification, and halo thickness >2 mm were influential factors for diagnosing FTC. The XGBoost model was identified as the optimal model after a comprehensive evaluation. The AUC of this model in the validation set was 0.969 [95% confidence interval (CI), 0.946-0.992], while its precision sensitivity, specificity, and accuracy were 0.791, 0.930, 0.913 and 0.917, respectively. Conclusions: XGBoost model based on ultrasound features was constructed and interpreted using the SHAP method, providing evidence for the diagnosis of FTC and guidance for the personalized treatment of patients.

RESUMEN

In recent years, we have witnessed both artificial intelligence obtaining remarkable results in clinical decision support systems (CDSSs) and explainable artificial intelligence (XAI) improving the interpretability of these models. In turn, this fosters the adoption by medical personnel and improves trustworthiness of CDSSs. Among others, counterfactual explanations prove to be one such XAI technique particularly suitable for the healthcare domain due to its ease of interpretation, even for less technically proficient staff. However, the generation of high-quality counterfactuals relies on generative models for guidance. Unfortunately, training such models requires a huge amount of data that is beyond the means of ordinary hospitals. In this paper, we therefore propose to use federated learning to allow multiple hospitals to jointly train such generative models while maintaining full data privacy. We demonstrate the superiority of our approach compared to locally generated counterfactuals. Moreover, we prove that generative models for counterfactual generation that are trained using federated learning in a suitable environment perform only marginally worse compared to centrally trained ones while offering the benefit of data privacy preservation. Finally, we integrate our method into a prototypical CDSS for treatment recommendation for sepsis patients, thus providing a proof of concept for real-world application as well as insights and sanity checks from clinical application.

RESUMEN

How does the act of explaining influence learning? Prior work has studied effects of explaining through a predominantly proximal lens, measuring short-term outcomes or manipulations within lab settings. Here, we ask whether the benefits of explaining extend to academic performance over time. Specifically, does the quality and frequency of student explanations predict students' later performance on standardized tests of math and English? In Study 1 (N = 127 5th-6th graders), participants completed a causal learning activity during which their explanation quality was evaluated. Controlling for prior test scores, explanation quality directly predicted both math and English standardized test scores the following year. In Study 2 (N = 20,384 10th graders), participants reported aspects of teachers' explanations and their own. Controlling for prior test scores, students' own explanations predicted both math and English state standardized test scores, and teacher explanations were linked to test performance through students' own explanations. Taken together, these findings suggest that benefits of explaining may result in part from the development of a metacognitive explanatory skill that transfers across domains and over time. Implications for cognitive science, pedagogy, and education are discussed.

Asunto(s)

RESUMEN

Deep learning applications in taxonomic classification for animals and plants from images have become popular, while those for microorganisms are still lagging behind. Our study investigated the potential of deep learning for the taxonomic classification of hundreds of filamentous fungi from colony images, which is typically a task that requires specialized knowledge. We isolated soil fungi, annotated their taxonomy using standard molecular barcode techniques, and took images of the fungal colonies grown in petri dishes (n = 606). We applied a convolutional neural network with multiple training approaches and model architectures to deal with some common issues in ecological datasets: small amounts of data, class imbalance, and hierarchically structured grouping. Model performance was overall low, mainly due to the relatively small dataset, class imbalance, and the high morphological plasticity exhibited by fungal colonies. However, our approach indicates that morphological features like color, patchiness, and colony extension rate could be used for the recognition of fungal colonies at higher taxonomic ranks (i.e. phylum, class, and order). Model explanation implies that image recognition characters appear at different positions within the colony (e.g. outer or inner hyphae) depending on the taxonomic resolution. Our study suggests the potential of deep learning applications for a better understanding of the taxonomy and ecology of filamentous fungi amenable to axenic culturing. Meanwhile, our study also highlights some technical challenges in deep learning image analysis in ecology, highlighting that the domain of applicability of these methods needs to be carefully considered.

RESUMEN

The evaluation of creep rupture life is complex due to its variable formation mechanism. In this paper, machine learning algorithms are applied to explore the creep rupture life span as a function of 27 physical properties to address this issue. By training several classical machine learning models and comparing their prediction performance, XGBoost is finally selected as the predictive model for creep rupture life. Moreover, we introduce an interpretable method, Shapley additive explanations (SHAP), to explain the creep rupture life predicted by the XGBoost model. The SHAP values are then calculated, and the feature importance of the creep rupture life yielded by the XGBoost model is discussed. Finally, the creep fracture life is optimized by using the chaotic sparrow optimization algorithm. We then show that our proposed method can accurately predict and optimize creep properties in a cheaper and faster way than other approaches in the experiments. The proposed method can also be used to optimize the material design across various engineering domains.

RESUMEN

Alzheimer's disease (AD) is an important research topic. While amyloid plaques and neurofibrillary tangles are hallmark pathological features of AD, cognitive resilience (CR) is a phenomenon where cognitive function remains preserved despite the presence of these pathological features. This study aimed to construct and compare predictive machine learning models for CR scores using RNA-seq data from the Religious Orders Study and Memory and Aging Project (ROSMAP) and Mount Sinai Brain Bank (MSBB) cohorts. We evaluated support vector regression (SVR), random forest, XGBoost, linear, and transformer-based models. The SVR model exhibited the best performance, with contributing genes identified using Shapley additive explanations (SHAP) scores, providing insights into biological pathways associated with CR. Finally, we developed a tool called the resilience gene analyzer (REGA), which visualizes SHAP scores to interpret the contributions of individual genes to CR. REGA is available at https://igcore.cloud/GerOmics/REsilienceGeneAnalyzer/.

RESUMEN

BACKGROUND: Understanding the factors contributing to mental well-being in youth is a public health priority. Self-reported enthusiasm for the future may be a useful indicator of well-being and has been shown to forecast social and educational success. Typically, cross-domain measures of ecological and health-related factors with relevance to public policy and programming are analyzed either in isolation or in targeted models assessing bivariate interactions. Here, we capitalize on a large provincial data set and machine learning to identify the sociodemographic, experiential, behavioral, and other health-related factors most strongly associated with levels of subjective enthusiasm for the future in a large sample of elementary and secondary school students. OBJECTIVE: The aim of this study was to identify the sociodemographic, experiential, behavioral, and other health-related factors associated with enthusiasm for the future in elementary and secondary school students using machine learning. METHODS: We analyzed data from 13,661 participants in the 2019 Ontario Student Drug Use and Health Survey (OSDUHS) (grades 7-12) with complete data for our primary outcome: self-reported levels of enthusiasm for the future. We used 50 variables as model predictors, including demographics, perception of school experience (i.e., school connectedness and academic performance), physical activity and quantity of sleep, substance use, and physical and mental health indicators. Models were built using a nonlinear decision tree-based machine learning algorithm called extreme gradient boosting to classify students as indicating either high or low levels of enthusiasm. Shapley additive explanations (SHAP) values were used to interpret the generated models, providing a ranking of feature importance and revealing any nonlinear or interactive effects of the input variables. RESULTS: The top 3 contributors to higher self-rated enthusiasm for the future were higher self-rated physical health (SHAP value=0.62), feeling that one is able to discuss problems or feelings with their parents (SHAP value=0.49), and school belonging (SHAP value=0.32). Additionally, subjective social status at school was a top feature and showed nonlinear effects, with benefits to predicted enthusiasm present in the mid-to-high range of values. CONCLUSIONS: Using machine learning, we identified key factors related to self-reported enthusiasm for the future in a large sample of young students: perceived physical health, subjective school social status and connectedness, and quality of relationship with parents. A focus on perceptions of physical health and school connectedness should be considered central to improving the well-being of youth at the population level.

Asunto(s)

RESUMEN

Accurately mapping ground-level ozone concentrations at high spatiotemporal resolution (daily, 1 km) is essential for evaluating human exposure and conducting public health assessments. This requires identifying and understanding a proxy that is well-correlated with ground-level ozone variation and available with spatiotemporal high-resolution data. This study introduces a high-resolution ozone modeling method utilizing the XGBoost algorithm with satellite-derived land surface temperature (LST) as the primary predictor. Focusing on China in 2019, our model achieved a cross-validation R2 of 0.91 and a root-mean-square error (RMSE) of 13.51 µg/m3. We provide detailed maps highlighting ground-level ozone concentrations in urban areas, uncovering spatial variations previously unresolved, along with time series aligning with established understandings of ozone dynamics. Our local interpretation of the machine learning model underscores the significant contribution of LST to spatiotemporal ozone variations, surpassing other meteorological, pollutant, and geographical predictors in its influence. Validation results indicate that model performance decreases as spatial resolution becomes coarser, with R2 decreasing from 0.91 for the 1 km model to 0.85 for the 25 km model. The methodology and data sets generated by this study offer new insights into ground-level ozone variability and mapping and can significantly aid in exposure assessment and epidemiological research related to this critical environmental challenge.

Asunto(s)

RESUMEN

Reservoir nearshore areas are influenced by both terrestrial and aquatic ecosystems, making them sensitive regions to water quality changes. The analysis of basin landscape hydrological features provides limited insight into the spatial heterogeneity of eutrophication in these areas. The complex characteristics of shoreline morphology and their impact on eutrophication are often overlooked. To comprehensively analyze the complex relationships between shoreline morphology and landscape hydrological features, with eutrophication, this study uses Danjiangkou Reservoir as a case study. Utilizing Landsat 8 OLI remote sensing data from 2013 to 2022, combined with a semi-analytical approach, the spatial distribution of the Trophic State Index (TSI) during flood discharge periods (FDPs) and water storage periods (WSPs) was obtained. Using Extreme Gradient Boosting (XGBoost) and SHapley Additive exPlanations (SHAP), explained the relationships between landscape composition, landscape configuration, hydrological topography, shoreline morphology, and TSI, identified key factors at different spatial scales and validated their reliability. The results showed that: (1) There is significant spatial heterogeneity in the TSI distribution of Danjiangkou Reservoir. The eutrophication levels are significant in the shoreline and bay areas, with a tendency to extend inward only during the WSPs. (2) The importance of landscape composition, landscape configuration, hydrological topography, and shoreline morphology to TSI variations during the FDPs are 25.12 %, 29.6 %, 23.09 %, and 22.19 % respectively. Besides shoreline distance, the Landscape Shape Index (LSI) and Hypsometric Integral (HI) are the two most significant environmental variables overall during the FDPs. Forest and grassland areas become the most influential factors during the WSPs. The influence of landscape patterns and hydrological topography on TSI varies at different spatial scales. At the 200 m riparian buffer zone, the increase in cropland and impervious areas significantly elevates eutrophication levels. (3) Morphology complexity, shows a noticeable threshold effect on TSI, with complex shoreline morphology increasing the risk of eutrophication.

RESUMEN

Out-of-hospital cardiac arrest (OHCA) is a critical condition with low survival rates. In patients with a return of spontaneous circulation, brain injury is a leading cause of death. In this study, we propose an interpretable machine learning approach for predicting neurologic outcome after OHCA, using information available at the time of hospital admission. METHODS: The study population were 55 615 OHCA cases registered in the Swedish Cardiopulmonary Resuscitation Registry between 2010 and 2020. The dataset was split to training and validation sets (for model development) and test set (for evaluation of the final model). We used an XGBoost algorithm with stratified, repeated 10-fold cross-validation along with Optuna framework for hyperparameters tuning. The final model was trained on 10 features selected based on the importance scores and evaluated on the test set in terms of discrimination, calibration and bias-variance tradeoff. We used SHapley Additive exPlanations to address the 'black-box' model and align with eXplainable artificial intelligence. RESULTS: The final model achieved: area under the receiver operating characteristic value 0.964 (95% confidence interval (CI) [0.960-0.968]), sensitivity 0.606 (95% CI [0.573-0.634]), specificity 0.975 (95% CI [0.972-0.978]), positive predictive value (PPV) 0.664 (95% CI [0.625-0.696]), negative predictive value (NPV) 0.969 (95% CI [0.966-0.972]), macro F1 0.803 (95% CI [0.788-0.816]), and showed a very good calibration. SHAP features with the highest impact on the model's output were:'ROSC on arrival to hospital', 'Initial rhythm asystole' and 'Conscious on arrival to hospital'. CONCLUSIONS: The XGBoost machine learning model with 10 features available at the time of hospital admission showed good performance for predicting neurologic outcome after OHCA, with no apparent signs of overfitting.

Asunto(s)

RESUMEN

This study explores the molecular alterations and disease progression in COVID-19 patients using ATR-FTIR spectroscopy combined with spectrochemical and explainable artificial intelligence (XAI) approaches. Blood serum samples from intubated patients (IC), those receiving hospital services (SC), and recovered patients (PC) were analyzed to identify potential spectrochemical serum biomarkers. Spectrochemical parameters such as lipid, protein, nucleic acid concentrations, and IgG glycosylation were quantified, revealing significant alterations indicative of disease severity. Notably, increased lipid content, altered protein concentrations, and enhanced protein phosphorylation were observed in IC patients compared to SC and PC groups. The serum AGR (Albumin/Globulin Ratio) index demonstrated a distinct shift among patient groups, suggesting its potential as a rapid biochemical marker for COVID-19 severity. Additionally, alterations in IgG glycosylation and glucose concentrations were associated with disease severity. Spectral analysis highlighted specific bands indicative of nucleic acid concentrations, with notable changes observed in IC patients. XAI techniques further elucidated the importance of various spectral features in predicting disease severity across patient categories, emphasizing the heterogeneity of COVID-19's impact. Overall, this comprehensive approach provides insights into the molecular mechanisms underlying COVID-19 pathogenesis and offers a transparent and interpretable prediction algorithm to aid decision-making and patient management.

Asunto(s)

RESUMEN

This paper establishes requirements for assessing the usability of Explainable Artificial Intelligence (XAI) methods, focusing on non-AI experts like healthcare professionals. Through a synthesis of literature and empirical findings, it emphasizes achieving optimal cognitive load, task performance, and task time in XAI explanations. Key components include tailoring explanations to user expertise, integrating domain knowledge, and using non-propositional representations for comprehension. The paper highlights the critical role of relevance, accuracy, and truthfulness in fostering user trust. Practical guidelines are provided for designing transparent and user-friendly XAI explanations, especially in high-stakes contexts like healthcare. Overall, the paper's primary contribution lies in delineating clear requirements for effective XAI explanations, facilitating human-AI collaboration across diverse domains.

Asunto(s)

RESUMEN

This study leverages data from a Canadian database of primary care Electronic Medical Records to develop machine learning models predicting type 2 diabetes mellitus (T2D), prediabetes, or normoglycemia. These models are used as a basis for extracting counterfactual explanations and derive personalized changes in biomarkers to prevent T2D onset, particularly in the still reversible prediabetic state. The models achieve satisfactory performance. Furthermore, feature importance analysis underscores the significance of fasting blood sugar and glycated hemoglobin, while counterfactuals explanations emphasize the centrality of keeping body mass index and cholesterol indicators within or close to the clinically desirable ranges. This research highlights the potential of machine learning and counterfactual explanations in guiding preventive interventions that may help slow down the progression from prediabetes to T2D on an individual basis, eventually fostering a recovery from prediabetes to a normoglycemic state.

Asunto(s)

RESUMEN

Deep learning models have achieved state-of-the-art performance for text classification in the last two decades. However, this has come at the expense of models becoming less understandable, limiting their application scope in high-stakes domains. The increased interest in explainability has resulted in many proposed forms of explanation. Nevertheless, recent studies have shown that rationales, or language explanations, are more intuitive and human-understandable, especially for non-technical stakeholders. This survey provides an overview of the progress the community has achieved thus far in rationalization approaches for text classification. We first describe and compare techniques for producing extractive and abstractive rationales. Next, we present various rationale-annotated data sets that facilitate the training and evaluation of rationalization models. Then, we detail proxy-based and human-grounded metrics to evaluate machine-generated rationales. Finally, we outline current challenges and encourage directions for future work.

RESUMEN

OBJECTIVE: Percutaneous transluminal angioplasty (PTA) is the primary method for treatment in peripheral arterial disease. However, some patients experience flow-limiting dissection (FLD) after PTA. We utilized machine learning and SHapley Additive exPlanations to identify and optimize a classification system to predict FLD after PTA. METHODS: This was a multi-center, retrospective, cohort study. The cohort comprised 407 patients who underwent treatment of the femoropopliteal (FP) arteries in 3 institutions between January 2021 and June 2023. Preoperative computed tomography angiography images were evaluated to identify FP artery grading, chronic total occlusion (CTO), and vessel calcification (peripheral artery calcium scoring system [PACSS]). After PTA, FLD was identified by angiography. We trained and validated 6 machine-learning models to estimate FLD occurrence after PTA, and the best model was selected. Then, the sum of the Shapley values for each of CTO, FP, and PACSS was calculated for each patient to produce the CTO-FP-PACSS value. The CTO-FP-PACSS classification system was used to classify the patients into classes 1 to 4. Univariate and multivariate analyses were performed to validate the effectiveness of the CTO-FP-PACSS classification system for predicting FLD. RESULTS: Overall, 407 patients were analyzed, comprising 189 patients with FLD and 218 patients without FLD. Differences in sex (71% males vs 54% males, p<0.001), CTO (72% vs 43%, p<0.001), FP (3.26±0.94 vs 2.66±1.06, p<0.001), and PACSS (2.39±1.40 vs 1.74±1.35, p<0.001) were observed between patients with and without FLD, respectively. The random forest model demonstrated the best performance (validation set area under the curve: 0.82). SHapley Additive exPlanations revealed CTO, PACSS, and FP as the 3 most influential FLD predictors, and the univariate and multivariate analyses confirmed CTO-FP-PACSS classification as an independent FLD predictor (multivariate hazard ratio 4.13; p<0.001). CONCLUSION: The CTO-FP-PACSS classification system accurately predicted FLD after PTA. This user-friendly system may guide surgical decision-making, helping choose between PTA and additional devices to reduce FLD in FP artery treatment. CLINICAL IMPACT: We utilised machine-learning techniques in conjunction with SHapley Additive exPlanations to develop a clinical classification system that predicts the probability of flow-limiting dissection (FLD) after plain old balloon angioplasty. This classification system categorises lesions into Classes 1-4 based on three factors: chronic total occlusion, femoropopliteal grading, and peripheral artery calcium scoring. Each class demonstrated a different probability of developing FLD. This classification system may be valuable for surgeons in their clinical practice, as well as serving as a source of inspiration for other researchers.

RESUMEN

Analyses of complex behaviors of Cerebrospinal Fluid (CSF) have become increasingly important in diseases diagnosis. The changes of the phase-contrast magnetic resonance imaging (PC-MRI) signal formed by the velocity of flowing CSF are represented as a set of velocity-encoded images or maps, which can be thought of as signal data in the context of medical imaging, enabling the evaluation of pulsatile patterns throughout a cardiac cycle. However, automatic segmentation of the CSF region in a PC-MRI image is challenging, and implementing an explained ML method using pulsatile data as a feature remains unexplored. This paper presents lightweight machine learning (ML) algorithms to perform CSF lumen segmentation in spinal, utilizing sets of velocity-encoded images or maps as a feature. The Dataset contains 57 PC-MRI slabs by 3T MRI scanner from control and idiopathic scoliosis participants are involved to collect data. The ML models are trained with 2176 time series images. Different cardiac periods image (frame) numbers of PC-MRIs are interpolated in the preprocessing step to align to features of equal size. The fivefold cross-validation procedure is used to estimate the success of the ML models. Additionally, the study focusses on enhancing the interpretability of the highest-accuracy eXtreme gradient boosting (XGB) model by applying the shapley additive explanations (SHAP) technique. The XGB algorithm presented its highest accuracy, with an average fivefold accuracy of 0.99% precision, 0.95% recall, and 0.97% F1 score. We evaluated the significance of each pulsatile feature's contribution to predictions, offering a more profound understanding of the model's behavior in distinguishing CSF lumen pixels with SHAP. Introducing a novel approach in the field, develop ML models offer comprehension into feature extraction and selection from PC-MRI pulsatile data. Moreover, the explained ML model offers novel and valuable insights to domain experts, contributing to an enhanced scholarly understanding of CSF dynamics.

Asunto(s)