https://hdl.handle.net/10481/15211 2026-04-05T12:35:01Z 2026-04-05T12:35:01Z Birch, Katherine Durán López, Alberto Bolaños Martinez, Daniel Pravin, Chandresh Bermúdez Edo, María del Campo Bauer, Roman De, Suparna https://hdl.handle.net/10481/107203 2026-03-20T09:18:57Z

Exploring Structural Brain Connectivity in Term and Preterm Infants with Explainable AI and Fuzzy Logic Birch, Katherine; Durán López, Alberto; Bolaños Martinez, Daniel; Pravin, Chandresh; Bermúdez Edo, María del Campo; Bauer, Roman; De, Suparna Preterm births have been associated with altered neurological development for neonatal infants; this has been implicated in certain neuro-developmental conditions in later life. Advances in brain imaging methods, such as Magnetic Resonance Imaging, have allowed for the analysis of physical connectivity of brain matter in infants shortly after birth. However, commonly used methods of investigating such data rely on a brain network analysis, traditionally based on graph-theoretical approaches, which may fail to capture complex patterns involving both local and global network structures and spatial information. Furthermore, many previous studies of infant brain data rely on a priori selection of specific graph connectivity measures. We propose employing machine learning models such as logistic regression and Graph Neural Networks (GNN) to provide a data-driven approach for classifying preterm and term brain networks at birth. We utilize fuzzy logic, and explainability methods including Shapley Additive Explanations (SHAP) to identify influential regions and connections in decision making. In our analysis, brain regions are represented as spatially embedded nodes, with edges representing strength of structural connections between areas. Using this setup, our model achieves a binary classification accuracy of 88.57%. This performance is further enhanced using a fuzzy boundary between preterm and term classes, achieving an accuracy of 96.19%. This demonstrates that the model can be assisted particularly by adding context to “near-term” born infant cases. These analyses highlight important connections and key nodes, including deep brain structures which are broadly consistent with biological literature.

Hourcade, J. P Bakala, E. Bonsignore, E. Currin, F. H. Fails, J. A. Gilhoi, A. Medina Medina, Nuria Norris, N. Onions, M. Pires, A.C. Walsh, G. Yarosh, S. Yip, J. https://hdl.handle.net/10481/106569 2025-09-23T12:07:29Z

Conducting Ethical Research on Emerging Technologies for Children Hourcade, J. P; Bakala, E.; Bonsignore, E.; Currin, F. H.; Fails, J. A.; Gilhoi, A.; Medina Medina, Nuria; Norris, N.; Onions, M.; Pires, A.C.; Walsh, G.; Yarosh, S.; Yip, J. T his SIG will provide child-computer interaction researchers and practitioners, as well as other interested CHI attendees, an opportunity to discuss topics related to conducting ethical research on emerging technologies for children. While the community has extensively debated on ethical issues, we have not had ample discussion of how to ethically manage research on emerging technologies, oftendesigned for adults, that may also be used by children or affect children. More specifically, we would like to discuss ethical aspects related to motivations for research, how research is conducted, and how it is reported.

Molina, Carlos Prados-Suárez, Belén Martinez-Sanchez, Beatriz https://hdl.handle.net/10481/101597 2025-01-31T10:29:02Z

Federated Mining of Interesting Association Rules over EHRs Molina, Carlos; Prados-Suárez, Belén; Martinez-Sanchez, Beatriz Federated learning has a great potential to create solutions working over different sources without data transfer. However current federated methods are not explainable nor auditable. In this paper we propose a Federated data mining method to discover association rules. More accurately, we define what we consider as interesting itemsets and propose an algorithm to obtain them. This approach facilitates the interoperability and reusability, and it is based on the accessibility to data. These properties are quite aligned with the FAIR principles.

Molina, Carlos Prados-Suárez, Belén https://hdl.handle.net/10481/101448 2025-01-31T08:07:28Z

Measuring the quality of data in electronic health records aggregators Molina, Carlos; Prados-Suárez, Belén There is an increasing work to integrate health related data for health care services and research purposes. Most of the current proposals adapt the schemas of the data sources to extract automatically the information, but they do not measure the quality of the resulted data. Even more, smart personal devices gather health related data into private non-standard compliant databases. Although these sources could be useful for health care systems and research, to consider the quality of the information they offer is essential. Electronic Health Records Aggregators (EHRagg ) are a new concept to integrate this kind of information, that considers the quality of the data. In this paper we present several factors that affect the quality (intrinsic to the data and related to the later use of it) and proposed a fuzzy quality measure to be used inside the EHRagg systems.

Martínez-Jiménez, Pedro Manuel Chamorro Martínez, Jesús Prados Suárez, María Belén https://hdl.handle.net/10481/101440 2025-01-31T08:05:03Z

A fuzzy approach for texture-based segmentation Martínez-Jiménez, Pedro Manuel; Chamorro Martínez, Jesús; Prados Suárez, María Belén In this paper, we propose a fuzzy approach for texture-based segmentation, where both the obtained regions and the texture features are fuzzy. This way, we have considered the imprecision related to the texture concept, as well as the imprecision associated with the boundaries between regions in the images. With regard to the texture descriptors, we propose to define them on the basis of the perceptual properties of texture. Since these properties are imprecise by nature, they are modelled by means of fuzzy sets defined on the domain of some of the most representative measures of each property. With regard to the segmentation technique, a fuzzy path-based segmentation is proposed. In this approach, fuzzy connectivity is used to measure the relationship between any pair of pixels. Thus, given a set of seed points, fuzzy regions are obtained on the basis of the connectivity between each seed point and the rest of pixels in the image. Finally, this proposal is applied to obtain fuzzy segmentations from real images