Abstract
Antibiotic resistance is a major health concern of the 21st century. The misuse of antibiotics over the years has led to their increasing presence in the environment, particularly in water resources, which can exacerbate the transmission of resistance genes and facilitate the emergence of resistant microorganisms. The objective of the present work is to develop a chemosensor for screening of sulfonamides in environmental waters, targeting sulfamethoxazole as the model analyte. The methodology was based on the retention of sulfamethoxazole in disks containing polystyrene divinylbenzene sulfonated sorbent particles and reaction with p-dimethylaminocinnamaldehyde, followed by colorimetric detection using a computer-vision algorithm. Several color spaces (RGB, HSV and CIELAB) were evaluated, with the coordinate a_star, from the CIELAB color space, providing the highest sensitivity. Moreover, in order to avoid possible errors due to variations in illumination, a color palette is included in the picture of the analytical disk, and a correction using the a_star value from one of the color patches is proposed. The methodology presented recoveries of 82-101% at 0.1 mu g and 0.5 mu g of sulfamethoxazole (25 mL), providing a detection limit of 0.08 mu g and a quantification limit of 0.26 mu g. As a proof of concept, application to in-field analysis was successfully implemented.

Abstract
Cardiac auscultation is one of the most cost-effective techniques used to detect and identify many heart conditions. Computer-assisted decision systems based on auscultation can support physicians in their decisions. Unfortunately, the application of such systems in clinical trials is still minimal since most of them only aim to detect the presence of extra or abnormal waves in the phonocardiogram signal, i.e., only a binary ground truth variable (normal vs abnormal) is provided. This is mainly due to the lack of large publicly available datasets, where a more detailed description of such abnormal waves (e.g., cardiac murmurs) exists. To pave the way to more effective research on healthcare recommendation systems based on auscultation, our team has prepared the currently largest pediatric heart sound dataset. A total of 5282 recordings have been collected from the four main auscultation locations of 1568 patients, in the process, 215780 heart sounds have been manually annotated. Furthermore, and for the first time, each cardiac murmur has been manually annotated by an expert annotator according to its timing, shape, pitch, grading, and quality. In addition, the auscultation locations where the murmur is present were identified as well as the auscultation location where the murmur is detected more intensively. Such detailed description for a relatively large number of heart sounds may pave the way for new machine learning algorithms with a real-world application for the detection and analysis of murmur waves for diagnostic purposes.

Abstract
Ordinal classification tasks are present in a large number of different domains. However, common losses for deep neural networks, such as cross-entropy, do not properly weight the relative ordering between classes. For that reason, many losses have been proposed in the literature, which model the output probabilities as following a unimodal distribution. This manuscript reviews many of these losses on three different datasets and suggests a potential improvement that focuses the unimodal constraint on the neighborhood around the true class, allowing for a more flexible distribution, aptly called quasi-unimodal loss. For this purpose, two constraints are proposed: A first constraint concerns the relative order of the top-three probabilities, and a second constraint ensures that the remaining output probabilities are not higher than the top three. Therefore, gradient descent focuses on improving the decision boundary around the true class in detriment to the more distant classes. The proposed loss is found to be competitive in several cases.

Abstract
Objective: Heart rate variability (HRV), an index of the autonomic cardiac activity, is decreased in patients with epilepsy, and a low HRV is associated with a higher risk of sudden death. Generalized tonic-clonic seizures are one of the most consistent risk factors for SUDEP, but the influence (and relative risk) of each type of seizure on cardiac function is still unknown. Our objective was to assess the impact of the type of seizure (focal to bilateral tonic-clonic seizure - FBTCS - versus non-FBTCS) on periictal HRV, in a group of patients with refractory epilepsy and both types of seizures. Methods: We performed a 48-hour Holter recording on 121 patients consecutively admitted to our Epilepsy Monitoring Unit. We only included patients with both FBTCS and non-FBTCS on the Holter recording and selected the first seizure of each type to analyze. To evaluate HRV parameters (AVNN, SDNN, RMSSD, pNN20, LF, HF, and LF/HF), we chose 5-min epochs pre-and postictally. Results: We included 14 patients, with a median age of 36 (min-max, 16-55) years and 64% were female. Thirty-six percent had cardiovascular risk factors, but no previously known cardiac disease. In the preictal period, there were no statistically significant differences in HRV parameters, between FBTCS and non-FBTCS. In the postictal period, AVNN, RMSSD, pNN20, LF, and HF were significantly lower, and LF/HF and HR were significantly higher in FBTCS. From preictal to postictal periods, FBTCS elicited a statistically significant rise in HR and LF/HF, and a statistically significant fall in AVNN, RMSSD, pNN20, and HF. Non-FBTCS only caused statistically significant changes in HR (decrease) and AVNN (increase). Significance/conclusion: This work emphasizes the greater effect of FBTCS in autonomic cardiac function in patients with refractory epilepsy, compared to other types of seizures, with a significant reduction in vagal tonus, which may be associated with an increased risk of SUDEP.

Abstract
Background: Physical exercise is a first-line treatment for peripheral arterial disease (PAD) and intermittent claudication (IC) reducing pain and increasing the distances walked. Home-based exercise therapy (HBET) has the advantage of reaching a higher number of patients and increasing adherence to physical exercise as it is performed in the patient's residential area and does not have the time, cost, and access restrictions of supervised exercise therapy (SET) implemented in a clinical setting. Even so, rates of adherence to physical exercise are relatively low, and therefore, m-health tools are promising in increasing motivation to behavior change and adherence to physical exercise. A built-in virtual assistant is a patient-focused tool available in a mobile interface, providing a variety of functions including health education, motivation, and implementation of behavior change techniques. Methods: This is a single-center, prospective, three-arm, single-blind, randomized, controlled, superior clinical trial with stratified and blocked random allocation. Three hundred participants with PAD and IC will be recruited from an Angiology and Vascular Surgery Department, Centro Hospitalar Universitario Porto (CHUPorto), Porto, Portugal. All patients will receive the same medical care recommended by current guidelines. Participants in all three groups will receive a personalized prescription for an HBET program and a behavioral change and motivational intervention. Participants in experimental groups 1 and 2 will receive a smartphone with the WalkingPad app to monitor exercise sessions. Experimental group 2 WalkingPad app will have a built-in virtual assistant that will promote behavioral change and provide motivational support. Participants allocated to the active control group will not receive the m-health tool, but a practice diary to encourage monitoring. The program will last for 6 months with three evaluation moments (baseline, 3, and 6 months). The primary outcome will be the change in distances walked (maximal and pain-free) from baseline to 3 and 6 months. Secondary outcomes will be changes in quality of life, patients' perception of resistance, and walking speed. Discussion: This study will allow measuring the effectiveness of an m-health tool in increasing motivation for behavior change and adherence to an HBET program in patients with PAD. The superiority of experimental group 2 in the primary and secondary outcomes will indicate that the virtual assistant is effective for motivating behavioral change and encouraging the practice and adherence to physical exercise. The use of m-health tools and virtual health assistants can potentially fill a gap in the access and quality of health services and information, reducing the burden on the health system and promoting self-management and self-care in chronic illness.

Abstract
Lower limb amputation is a condition affecting millions of people worldwide. Patients are often prescribed with lower limb prostheses to aid their mobility, but these prostheses require frequent adjustments through an iterative and manual process, which heavily depends on patient feedback and on the prosthetist's experience. New computer-aided design and manufacturing technologies have been emerging as ways to improve the fitting process by creating virtual models of the prosthesis' interface component with the limb, the socket. Using Adversarial Autoencoders, a generative model describing both transtibial and transfemoral sockets was created. Two strategies were tested to counteract the small size of the dataset: transfer learning using the ModelNet dataset and data augmentation through a previously validated socket statistical shape model. The minimum reconstruction error was 0.00124 mm and was obtained for the model which combined the two approaches. A single-blind assessment conducted with prosthetists showed that, while generated and real shapes are distinguishable, most generated ones assume plausible shapes. Our results show that the use of transfer learning allowed for a correct training and regularization of the latent space, inducing in the model generative abilities with potential clinical applications.