As a technique that can compactly represent complex patterns, machine learning has significant potential for predictive inference. K-fold cross-validation (CV) is the most common approach to ascertaining the likelihood that a machine learning outcome is generated by chance, and it frequently outperforms conventional hypothesis testing. This improvement uses measures directly obtained from machine learning classifications, such as accuracy, that do not have a parametric description. To approach a frequentist analysis within machine learning pipelines, a permutation test or simple statistics from data partitions (i.e., folds) can be added to estimate confidence intervals. Unfortunately, neither parametric nor non-parametric tests solve the inherent problems of partitioning small sample-size datasets and learning from heterogeneous data sources. The fact that machine learning strongly depends on the learning parameters and the distribution of data across folds recapitulates familiar difficulties around excess false positives and replication. A novel statistical test based on K-fold CV and the Upper Bound of the actual risk (K-fold CUBV) is proposed, where uncertain predictions of machine learning with CV are bounded by the worst case through the evaluation of concentration inequalities. Probably Approximately Correct-Bayesian upper bounds for linear classifiers in combination with K-fold CV are derived and used to estimate the actual risk. The performance with simulated and neuroimaging datasets suggests that K-fold CUBV is a robust criterion for detecting effects and validating accuracy values obtained from machine learning and classical CV schemes, while avoiding excess false positives.

Autism Spectrum Condition (ASC) is a neurodevelopmental condition characterized by impairments in communication, social interaction and restricted or repetitive behaviors. Extensive research has been conducted to identify distinctions between individuals with ASC and neurotypical individuals. However, limited attention has been given to comprehensively evaluating how variations in image acquisition protocols across different centers influence these observed differences. This analysis focuses on structural magnetic resonance imaging (sMRI) data from the Autism Brain Imaging Data Exchange I (ABIDE I) database, evaluating subjects' condition and individual centers to identify disparities between ASC and control groups. Statistical analysis, employing permutation tests, utilizes two distinct statistical mapping methods: Statistical Agnostic Mapping (SAM) and Statistical Parametric Mapping (SPM). Results reveal the absence of statistically significant differences in any brain region, attributed to factors such as limited sample sizes within certain centers, noise effects and the problem of multicentrism in a heterogeneous condition such as autism. This study indicates limitations in using the ABIDE I database to detect structural differences in the brain between neurotypical individuals and those diagnosed with ASC. Furthermore, results from the SAM mapping method show greater consistency with existing literature.

Alzheimer's disease (AD) is a progressive neurodegenerative disorder characterized by cognitive decline and substantial brain atrophy. Early and accurate prediction of disease progression and staging is crucial for timely intervention and effective treatment planning. Previous studies, including those based on artificial intelligence techniques, have employed neuroimaging, biomarkers and clinical data to model AD progression; however, many of these approaches rely on strong parametric assumptions or lack robust statistical guarantees regarding model validity. To bridge this gap, this study proposes a novel framework for validating predictive and staging models of disease using a statistically agnostic methodology. The objective is to take the advantages of an unconventional method for robust validation of ML models related to AD. Validation is performed using the Statistical Agnostic Regression (SAR) methodology applied to the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset. The method tests for a linear relationship by resampling and estimating an upper bound on the expected risk (R) via a Bayesian bound under the worst-case scenario. The SAR power assesses the likelihood of detecting a true linear relationship using the test statistic R, via Monte Carlo simulations under the null distribution. Three predictive models related to structural neuroimaging are assessed: one for the Mini Mental State Examination (MMSE) score, another for the concentration of amyloid beta 1-42 protein in the cerebrospinal fluid, and a third for age. In addition, a model for staging based on Alzheimer's-related clinical groups is explored through the joint analysis of segmented gray matter and white matter images. The findings indicate that the SAR methodology not only facilitates robust validation of predictive ML models related to neuroimaging and AD but also enables an effective staging of the AD continuum. This SAR-proposed framework opens new perspectives for the validation of ML models for early diagnosis and provides a solid foundation for future research in computational neuroscience.

Medical imaging fusion combines complementary information from multiple modalities to enhance diagnostic accuracy. However, evaluating the quality of fused images remains challenging, with many studies relying solely on classification performance, which may lead to incorrect conclusions. We introduce a novel framework for improving image fusion, focusing on preserving fine-grained details. Our model uses a siamese autoencoder to process T1-MRI and FDG-PET images in the context of Alzheimer’s disease (AD). The framework optimizes fusion by minimizing reconstruction error between generated and input images, while maximizing differences between modalities through cosine distance. Additionally, we propose a supervised variant, incorporating binary cross-entropy loss between diagnostic labels and probabilities. Fusion quality is rigorously assessed through three tests: 1) classification of AD patients and controls using fused images; 2) an atlas-based occlusion test for identifying regions relevant to cognitive decline; and 3) analysis of structural-functional relationships via Euclidean distance. Results show an AUC of 0.92 for AD detection, reveal the involvement of brain regions linked to preclinical AD stages, and demonstrate preserved structural-functional brain networks, indicating that subtle differences are successfully captured through our fusion approach.

Brain tumor resection is a complex procedure with significant implications for patient survival and quality of life. Predictions of patient outcomes provide clinicians and patients the opportunity to select the most suitable onco-functional balance. In this study, global features derived from structural magnetic resonance imaging in a clinical dataset of 49 pre- and post-surgery patients identified potential biomarkers associated with survival outcomes. We propose a framework that integrates Explainable AI (XAI) with neuroimaging-based feature engineering for survival assessment, offering guidance for surgical decision-making. In this study, we introduce a global explanation optimizer that refines survival-related feature attribution in deep learning models, enhancing interpretability and reliability. Our findings suggest that survival is influenced by alterations in regions associated with cognitive and sensory functions, indicating the importance of preserving areas involved in decision-making and emotional regulation during surgery to improve outcomes. The global explanation optimizer improves both fidelity and comprehensibility of explanations compared to state-of-the-art XAI methods. It effectively identifies survival-related variability, underscoring its relevance in precision medicine for brain tumor treatment.

INTRODUCTION:Regression analysis is a central topic in statistical modeling, aimed at estimating the relationships between a dependent variable, commonly referred to as the response variable, and one or more independent variables, i.e., explanatory variables. Linear regression is by far the most popular method for performing this task in various fields of research, such as data integration and predictive modeling when combining information from multiple sources. OBJECTIVES:Classical methods for solving linear regression problems, such as Ordinary Least Squares (OLS), Ridge, or Lasso regressions, often form the foundation for more advanced machine learning (ML) techniques, which have been successfully applied, though without a formal definition of statistical significance. At most, permutation or analyses based on empirical measures (e.g., residuals or accuracy) have been conducted, leveraging the greater sensitivity of ML estimations for detection. METHODS:In this paper, we introduce Statistical Agnostic Regression (SAR) for evaluating the statistical significance of ML-based linear regression models. This is achieved by analyzing concentration inequalities of the actual risk (expected loss) and considering the worst-case scenario. To this end, we define a threshold that ensures there is sufficient evidence, with a probability of at least 1-η, to conclude the existence of a linear relationship in the population between the explanatory (feature) and the response (label) variables. CONCLUSIONS:Simulations demonstrate that the proposed agnostic (non-parametric) test can perform an analysis of variance comparable to the classical multivariate F-test for the slope parameter, without relying on the underlying assumptions of classical methods. A power analysis on a putative regression task revealed an overinflated false positive rate in standard ML methods, whereas the SAR test exhibited excellent control. Moreover, the residuals computed using this method represent a trade-off between those obtained from ML approaches and classical OLS.

The analysis of neuroimaging data by means of computer systems has become a general practice in the diagnosis and monitoring of Alzheimer’s disease. In recent years, different systems based on neural networks have been proposed to aid in the diagnosis of this disorder. These systems usually contain millions of parameters that must be adjusted during the training process. This requires large datasets, often not available in most studies. The use of pre-trained systems (also known as transfer learning) would help to alleviate this problem, however these systems are often pre-trained with data of a very different nature than neuroimaging data, so their use could be counterproductive. In this work we evaluate the use of transfer learning for the development of a computer-aided diagnosis system for Alzheimer’s disease. To this end, we compared the performance obtained by different systems with and without transfer learning. The results show that transfer learning improves the results in some cases, reaching higher performances than other systems, however, the improvement depends largely on the optimizer used. In addition, we propose a novel method to convert neuroimaging data from 3D volumes to the 2D images used as input for most pre-trained systems.

The RESISTO project, a pioneer innovation initiative in Europe, endeavors to enhance the resilience of electrical networks against extreme weather events and associated risks. Emphasizing intelligence and flexibility within distribution networks, RESISTO aims to address climatic and physical incidents comprehensively, fostering resilience across planning, response, recovery, and adaptation phases. Leveraging advanced technologies including AI, IoT sensors, and aerial robots, RESISTO integrates prediction, detection, and mitigation strategies to optimize network operation. This article summarizes the main technical aspects of the proposed solutions to meet the aforementioned objectives, including the development of a climate risk detection platform, an IoT-based monitoring and anomaly detection network, and a fleet of intelligent aerial robots. Each contributing to the project’s overarching objectives of enhancing network resilience and operational efficiency.

Neurodegenerative diseases pose a formidable challenge to medical research, demanding a nuanced understanding of their progressive nature. In this regard, latent generative models can effectively be used in a data-driven modeling of different dimensions of neurodegeneration, framed within the context of the manifold hypothesis. This paper proposes a joint framework for a multi-modal, common latent generative model to address the need for a more comprehensive understanding of the neurodegenerative landscape in the context of Parkinson’s disease (PD). The proposed architecture uses coupled variational autoencoders (VAEs) to joint model a common latent space to both neuroimaging and clinical data from the Parkinson’s Progression Markers Initiative (PPMI). Alternative loss functions, different normalization procedures, and the interpretability and explainability of latent generative models are addressed, leading to a model that was able to predict clinical symptomatology in the test set, as measured by the unified Parkinson’s disease rating scale (UPDRS), with R2 up to 0.86 for same-modality and 0.441 cross-modality (using solely neuroimaging). The findings provide a foundation for further advancements in the field of clinical research and practice, with potential applications in decision-making processes for PD. The study also highlights the limitations and capabilities of the proposed model, emphasizing its direct interpretability and potential impact on understanding and interpreting neuroimaging patterns associated with PD symptomatology.

Medical imaging plays a pivotal role in understanding neurodegenerative diseases like Parkinson’s, aiding in early diagnosis and treatment monitoring. Despite its importance, obtaining comprehensive imaging datasets remains challenging. In response, we introduce a new database comprising brain images from Parkinson’s patients and healthy controls, addressing the scarcity of such resources in the field. The database currently houses around 3000 subjects, offering a diverse and extensive collection for research purposes. Leveraging this dataset, we conduct experiments employing classical models to delineate neuroanatomical disparities between Parkinson’s patients and controls. Our findings not only underscore the potential of this database in advancing Parkinson’s research but also highlight its significance in facilitating the translation of findings into clinical practice, ultimately enhancing patient care and outcomes.

Cancer is one of the leading causes of death in the world, with radiotherapy as one of the treatment options. Radiotherapy planning starts with delineating the affected area from healthy organs, called organs at risk (OAR). A new approach to automatic OAR segmentation in the chest cavity in Computed Tomography (CT) images is presented. The proposed approach is based on the modified U-Net architecture with the ResNet-34 encoder, which is the baseline adopted in this work. The new two-branch CS-SA U-Net architecture is proposed, which consists of two parallel U-Net models in which self-attention blocks with cosine similarity as query-key similarity function (CS-SA) blocks are inserted between the encoder and decoder, which enabled the use of consistency regularisation. The proposed solution demonstrates state-of-the-art performance for the problem of OAR segmentation in CT images on the publicly available SegTHOR benchmark dataset in terms of a Dice coefficient (oesophagus-0.8714, heart-0.9516, trachea-0.9286, aorta-0.9510) and Hausdorff distance (oesophagus-0.2541, heart-0.1514, trachea-0.1722, aorta-0.1114) and significantly outperforms the baseline. The current approach is demonstrated to be viable for improving the quality of OAR segmentation for radiotherapy planning.

Artificial Intelligence (AI) has improved our ability to process large amounts of data. These tools are particularly interesting in medical contexts because they evaluate the variables from patients’ screening evaluation and disentangle the information that they contain. In this study, we propose a novel method for detecting developmental dyslexia by extracting heart signals from NIRS. Features in terms of different domains based on heart rate variability (HRV) are computed from the extracted signal, and dimensionality of the resulting data is reduced through Principal Component Analysis (PCA). To evaluate the discriminability of the information patterns associated with normal controls and dyslexic patients, the resulting components are entered into a linear classifier to evaluate the discriminability of the information patterns associated with normal controls and dyslexic patients, leading to an area under the ROC curve of 0.79. The explanatory nature of our framework, based on Shapley Additive Explanations (SHAP), yields a deeper understanding of the evaluated phenomenon, revealing the presence of behavioral variables highly correlated with the model’s features. These findings demonstrate that heart information can be extracted from a different equipment than electrocardiogram tools, and that cardiac signal variables can be used to detect dyslexia in an early stage.

The global prevalence of dementia is on the rise, posing a challenge to healthcare systems worldwide. The disease leads to irreversible deterioration of cognitive function, which underlines the importance of early detection to mitigate its impact. The Clock Drawing Test (CDT) is a widely used tool in cognitive assessment, as it involves manually drawing a clock on a piece of paper. Despite its widespread use, CDT scoring methods often rely on subjective expert assessments. Thus, machine learning and deep learning-based models are recently being proposed for the automated evaluation of CDT drawings. In this study, we compare two state-of-the-art models, a simple CNN and API-Net, as cognitive state classification systems. Two databases were used, one from Spanish clinical centers (7009 samples) and the other from a hospital in Thailand (3108 samples). The obtained results align with expected accuracy rates in such scenarios (around 80 % ) and are similar in both models. Specifically, the accuracy rates obtained with the Spanish database are 75.65 % and 72.42 % , and with the Thai database, 86.42 % and 86.90 % . This reflects that the implementation of an excessively complex model is not necessary given the available sample size and the binary classification scenario. Therefore, although both models could be useful in the clinical domain, opting for models with lower computational costs is advisable to make them more cost-effective and easily accessible.

Parkinson's disease (PD) is a neurodegenerative disorder that affects millions of people worldwide. The diagnosis of PD is based on clinical and neuroimaging data. This work proposes a novel approach that jointly models several Variational Autoencoder (VAE) architectures in order to maximize cross-modality prediction. We hypothesize that 123I-ioflupane SPECT could be related to motor symptomatology and other dopaminergic deficits. We propose a joint modelling of several VAE architectures for maximizing cross-modality prediction of the PD Clinical and Neuroimaging Data. The final model, with 5 common latents and 2 neuroimaging and data specific latents achieve R2 values up to 0.8 for scores related to PD, including well known PD symptomatology scales such as UPDRS (R2 = 0.545), at the same time that provides tools for interpreting the results and the common latent distribution for both clinical data and neuroimaging, paving the way for interpretable machine learning tools in neurodegeneration.

Multiple sclerosis is a neurodegenerative disease characterized by the presence of lesions in brain tissues due to the inflammatory processes. The analysis of these lesions, visible through magnetic resonance imaging, is of great utility for the diagnosis and monitoring of the disease. In this work we propose a computerized procedure based on isolines to automatically segment the lesions caused by the disease, thus reducing the burden of physicians and removing human factor from the process. The proposed method calculates the isolines of all axial slices of the magnetic resonance image corresponding to a subject, selects those corresponding to points of higher intensity and triangulates the points of selected isolines, forming solid surfaces covering the proximities of the isolines. The process concludes by returning a 3D map containing all the lesions of a given subject in a standardized space that can be used to determine the volume of lesioned tissues or to perform quantitative analysis of molecular neuroimaging data to improve patient monitoring. This method has been evaluated using neuroimaging data from 16 patients and the results, similar to those obtained by manual segmentation, have been visually validated by experienced clinicians.

Novel features derived from imaging and artificial intelligence systems are commonly coupled to construct computer-aided diagnosis (CAD) systems that are intended as clinical support tools or for investigation of complex biological patterns. This study used sulcal patterns from structural images of the brain as the basis for classifying patients with schizophrenia from unaffected controls. Statistical, machine learning and deep learning techniques were sequentially applied as a demonstration of how a CAD system might be comprehensively evaluated in the absence of prior empirical work or extant literature to guide development, and the availability of only small sample datasets. Sulcal features of the entire cerebral cortex were derived from 58 schizophrenia patients and 56 healthy controls. No similar CAD systems has been reported that uses sulcal features from the entire cortex. We considered all the stages in a CAD system workflow: preprocessing, feature selection and extraction, and classification. The explainable AI techniques Local Interpretable Model-agnostic Explanations and SHapley Additive exPlanations were applied to detect the relevance of features to classification. At each stage, alternatives were compared in terms of their performance in the context of a small sample. Differentiating sulcal patterns were located in temporal and precentral areas, as well as the collateral fissure. We also verified the benefits of applying dimensionality reduction techniques and validation methods, such as resubstitution with upper bound correction, to optimize performance.

This chapter provides an in-depth exploration of artificial intelligence (AI)-based organ segmentation in medical imaging, a crucial task with significant implications for disease diagnosis, treatment planning, and patient care. The chapter delves into the core components of AI-based segmentation, including the utilization of large annotated datasets, the training process of artificial neural networks (ANNs), and the adoption of advanced architectures like convolutional neural networks and transformer-based models. It also discusses the different architectural configurations, such as 2D, 3D, 2.5D, and cascaded 3D approaches, highlighting their strengths and limitations. State-of-the-art approaches are explored to showcase the current advancements in the field. Furthermore, the chapter addresses the challenges and limitations of AI-based segmentation, such as generalization issues and computational resources limitations, as well as the need for diverse and representative datasets. By providing a comprehensive overview, this chapter aims to equip readers with a solid understanding of AI-based segmentation in medical imaging and its potential for enhancing patient outcomes.

Developmental dyslexia (DLX) hinders the reading learning process of 5

The RESISTO project represents a pioneering initiative in Europe aimed at enhancing the resilience of the power grid through the integration of advanced technologies. This includes artificial intelligence and thermal surveillance systems to mitigate the impact of extreme meteorological phenomena. RESISTO endeavors to predict, prevent, detect, and recover from weather-related incidents, ultimately enhancing the quality of service provided and ensuring grid stability and efficiency in the face of evolving climate challenges. In this study, we introduce one of the fundamental pillars of the project: a monitoring system for the operating temperature of different regions within power transformers, aiming to detect and alert early on potential thermal anomalies. To achieve this, a distributed system of thermal cameras for real-time temperature monitoring has been deployed in The Doñana National Park, alongside servers responsible for the storing, analyzing, and alerting of any potential thermal anomalies. An adaptive prediction model was developed for temperature forecasting, which learns online from the newly available data. In order to test the long-term performance of the proposed solution, we generated a synthetic temperature database for the whole of the year 2022. Overall, the proposed system exhibits promising capabilities in predicting and detecting thermal anomalies in power electric transformers, showcasing potential applications in enhancing grid reliability and preventing equipment failures.

The postyield rheological regime is investigated in sheared magnetic field-responsive composites (i.e. carbonyl iron based magnetorheological fluids). When subjected to uniaxial DC fields, high-speed videomicroscopy techniques and dedicated image analysis tools demonstrate that dispersed magnetic microparticles self-assemble to form concentric layered patterns above a particular shear rate ( γ ˙ R ,c ). This critical shear rate for layer formation is dictated by a critical Mason number M n c ∼ 1 that is associated to the destruction of the last doublet in the chain-like aggregates. The number of layers, mean width, percentage of occupation and mean period are found to be very weakly dependent on the shear rate in start-up shearing flow tests. Experimental data for the mean period are in good agreement with an energy minimization theory.