Search

Background: Deep-learning applications in cardiology typically perform trivial binary classification and are able to discriminate between subjects affected or not affected by a specific cardiac disease. However, this working scenario is very different from the real one, where clinicians are required to recognize the occurrence of one cardiac disease among the several possible ones, performing a multiclass classification. The present work aims to create a new interpretable deep-learning tool able to perform a multiclass classification and, thus, discriminate among several different cardiac diseases. Methods: The “China Physiological Signal Challenge in 2018” Physionet database was used to develop a multiclass neural network, constructed by the Advanced Repeated Structuring & Learning Procedure (AdvRS&LP). Data, consisting of 6877 12-lead 10-second electrocardiograms, was processed to obtain 252 electrocardiographic and vectorcardiographic input features, used to classify the data into eight classes (normal sinus rhythm, atrial fibrillation, first-degree atrioventricular block, left bundle branch block, right bundle branch block, premature atrial contraction, premature ventricular contraction, and unknown). Classification performance was evaluated by the area under the curve of the receiver operating characteristics. Clinical interpretability was assessed by standard statistical analysis and the local interpretable model-agnostic explainer algorithm. Results: Performance ranged from 89.88% to 90.10% (95.98 ± 3.32%) in the learning dataset and from 69.15% to 91.14% (83.65 ± 8.24%) in the testing dataset. These results are good considering the difficult, realistic multiclass classification task. Conclusions: The proposed multiclass neural network constructed using the AdvRS&LP represents a promising deep-learning tool for discriminating several cardiac diseases while ensuring clinical interpretability.

The volitional control of powered assistive devices is commonly performed by mapping the electromyographic (EMG) activity of the lower limb to joints’ angular kinematics, which are then used as the input for regulation. However, during walking, the ground reaction force (GRF) plays a central role in the modulation of the gait, providing dynamic stability and propulsion during the stance phase. Including this information within the control loop of prosthetic devices can improve the quality of the final output, providing more physiological walking dynamics that enhances the usability and patient comfort. In this work, we explored the feasibility of the estimation of the ground reaction force vertical component (VGRF) by using only the EMG activities of the thigh and shank muscles. We compared two deep learning models in three experiments that involved different muscular configurations. Overall, the outcomes show that the EMG signals could be leveraged to obtain a reliable estimation of the VGRF during walking, and the shank muscles alone represent a viable solution if a reduced recording setup is needed. On the other hand, the thigh muscles failed in providing performance enhancements, either when used alone or together with the shank muscles. The results outline the feasibility of including GRF information within an EMG-driven control scheme for prosthetic and assistive devices.

The database is constituted by 50 datasets containing cardiorespiratory signals acquired from 50 healthy volunteer subjects (one dataset for each subject; 23 males and 27 females; age: 23±5 years) while performing normal breathing, deep breathing, and breath holding, and two spreadsheet files, namely the “SubjectsInfo.xlsx” and “DBInfo.xlsx” containing the metadata of subjects (including demographic data) and of acquired signals, respectively.Cardiorespiratory signals consisted in simultaneously recorded 12-lead electrocardiograms acquired by the clinical M12 Global InstrumentationⓇ digital Holter ECG recorder, and single-lead electrocardiograms and respiration signals acquired by the wearable chest strap BioHarness 3.0 by Zephyr.The database may be useful to: (1) validate the use of wearable sensors in the acquisition of cardiorespiratory data during different respiration kinds, including apnea; (2) investigate the physiological association between cardiovascular and respiratory systems; (3) validate algorithms able to indirectly extract the respiration signal from the electrocardiogram; (4) study the fatigue level induced by a series ofcontrolled respiration patterns; and (5) investigate the effect of COVID-19 infection on the cardiorespiratory system.

The limited mobility of lower limb amputees highlights the need for advancements in prosthetic control strategies to restore natural locomotion. This paper proposes an information fusion approach for gait phase recognition using surface electromyography (sEMG) and kinematics data. Time-domain (TD) features were extracted from the myoelectric data and three data-driven models, specifically Support Vector Machine (SVM), K-Nearest Neighbours (KNN), and Artificial Neural Network (ANN), were compared in three different input conditions i.e. sEMG features, hip angle, and their fusion. Gait phase estimation results averaged from 40 healthy participants during normal walking with 10 strides per each demonstrated that the proposed fusion approach has consistently outperformed (p

In this study, a minimal setup for the ankle joint kinematics estimation is proposed relying only on proximal information of the lower-limb, i.e. thigh muscles activity and joint kinematics. To this purpose, myoelectric activity of Rectus Femoris (RF), Biceps Femoris (BF), and Vastus Medialis (VM) were recorded by surface electromyography (sEMG) from six healthy subjects during unconstrained walking task. For each subject, the angular kinematics of hip and ankle joints were synchronously recorded with sEMG signal for a total of 288 gait cycles. Two feature sets were extracted from sEMG signals, i.e. time domain (TD) and wavelet (WT) and compared to have a compromise between the reliability and computational capacity, they were used for feeding three regression models, i.e. Artificial Neural Networks, Random Forest, and Least Squares - Support Vector Machine (LS-SVM). BF together with LS-SVM provided the best ankle angle estimation in both TD and WT domains (RMSE < 5.6 deg). The inclusion of Hip joint trajectory significantly enhanced the regression performances of the model (RMSE < 4.5 deg). Results showed the feasibility of estimating the ankle trajectory using only proximal and limited information from the lower limb which would maximize a potential transfemoral amputee user’s comfortability while facing the challenge of having a small amount of information thus requiring robust data-driven models. These findings represent a significant step towards the development of a minimal setup useful for the control design of ankle active prosthetics and rehabilitative solutions.

The early diagnosis of diabetic neuropathy (DN) is fundamental in order to enact timely therapeutic strategies for limiting disease progression. In this work, we explored the suitability of standing balance task for identifying the presence of DN. Further, we proposed two diagnosis pathways in order to succeed in distinguishing between different stages of the disease. We considered a cohort of non-neuropathic (NN), asymptomatic neuropathic (AN), and symptomatic neuropathic (SN) diabetic patients. From the center of pressure (COP), a series of features belonging to different description domains were extracted. In order to exploit the whole information retrievable from COP, a majority voting ensemble was applied to the output of classifiers trained separately on different COP components. The ensemble of kNN classifiers provided over 86% accuracy for the first diagnosis pathway, made by a 3-class classification task for distinguishing between NN, AN, and SN patients. The second pathway offered higher performances, with over 97% accuracy in identifying patients with symptomatic and asymptomatic neuropathy. Notably, in the last case, no asymptomatic patient went undetected. This work showed that properly leveraging all the information that can be mined from COP trajectory recorded during standing balance is effective for achieving reliable DN identification. This work is a step toward a clinical tool for neuropathy diagnosis, also in the early stages of the disease.

Changes in cardiac function and morphology are reflected in variations in the electrocardiogram (ECG) and, in turn, in the cardiac risk indices derived from it. These variations have led to the introduction of normalization as a step to compensate for possible biasing factors responsible for inter- and intra-subject differences, which can affect the accuracy of ECG-derived risk indices in assessing cardiac risk. The aim of this work is to perform a scoping review to provide a comprehensive collection of open-access published research that examines normalized ECG-derived parameters used as markers of cardiac anomalies or instabilities. The literature search was conducted from February to July 2024 in the major global electronic bibliographic repositories. Overall, 39 studies were selected. Results suggest extensive use of normalization on heart rate variability-related indices (49% of included studies), QT-related indices (18% of included studies), and T-wave alternans (5% of included studies), underscoring their recognized importance and suggesting that normalization may enhance their role as clinically useful risk markers. However, the primary objective of the included studies was not to evaluate the effect of normalization itself; thus, further research is needed to definitively assess the impact and advantages of normalization across various ECG-derived parameters.

Gait phase recognition systems based on surface electromyographic signals (EMGs) are crucial for developing advanced myoelectric control schemes that enhance the interaction between humans and lower limb assistive devices. However, machine learning models used in this context, such as Linear Discriminant Analysis (LDA) and Support Vector Machine (SVM), typically experience performance degradation when modeling the gait cycle with more than just stance and swing phases. This study introduces a generalized phasor-based feature extraction approach (PHASOR) that captures spatial myoelectric features to improve the performance of LDA and SVM in gait phase recognition. A publicly available dataset of 40 subjects was used to evaluate PHASOR against state-of-the-art feature sets in a five-phase gait recognition problem. Additionally, fully data-driven deep learning architectures, such as Rocket and Mini-Rocket, were included for comparison. The separability index (SI) and mean semi-principal axis (MSA) analyses showed mean SI and MSA metrics of 7.7 and 0.5, respectively, indicating the proposed approach’s ability to effectively decode gait phases through EMG activity. The SVM classifier demonstrated the highest accuracy of 82% using a five-fold leave-one-trial-out testing approach, outperforming Rocket and Mini-Rocket. This study confirms that in gait phase recognition based on EMG signals, novel and efficient muscle synergy information feature extraction schemes, such as PHASOR, can compete with deep learning approaches that require greater processing time for feature extraction and classification.

IntroductionFunctional connectivity (FC) is defined in terms of temporal correlations between physiological signals, which mainly depend upon structural (axonal) connectivity; it is commonly studied using functional magnetic resonance imaging (fMRI). Interhemispheric FC appears mostly supported by the corpus callosum (CC), although several studies investigating this aspect have not provided conclusive evidence. In this context, patients in whom the CC was resected for therapeutic reasons (split-brain patients) provide a unique opportunity for research into this issue. The present study was aimed at investigating with resting-state fMRI the interhemispheric FC in six epileptic patients who have undergone surgical resection of the CC.MethodsThe analysis was performed using fMRI of the Brain Software Library; the evaluation of interhemispheric FC and the recognition of the resting-state networks (RSNs) were performed using probabilistic independent component analysis.ResultsGenerally, bilateral brain activation was often observed in primary sensory RSNs, while in the associative areas, such as those composing the default mode and fronto-parietal networks, the activation was often unilateral.DiscussionThese results suggest that even in the absence of the CC, some interhemispheric communication is still present. This residual FC might be supported through extra-callosal pathways that are likely subcortical, making it possible for some interhemispheric integration. Further studies are needed to confirm these conclusions.

Electroencephalography (EEG) wearable devices are particularly suitable for monitoring a subject’s engagement while performing daily cognitive tasks. EEG information provided by wearable devices varies with the location of the electrodes, the suitable location of which can be obtained using standard multi-channel EEG recorders. Cognitive engagement can be assessed during working memory (WM) tasks, testing the mental ability to process information over a short period of time. WM could be impaired in patients with epilepsy. This study aims to evaluate the cognitive engagement of nine patients with epilepsy, coming from a public dataset by Boran et al., during a verbal WM task and to identify the most suitable location of the electrodes for this purpose. Cognitive engagement was evaluated by computing 37 engagement indexes based on the ratio of two or more EEG rhythms assessed by their spectral power. Results show that involvement index trends follow changes in cognitive engagement elicited by the WM task, and, overall, most changes appear most pronounced in the frontal regions, as observed in healthy subjects. Therefore, involvement indexes can reflect cognitive status changes, and frontal regions seem to be the ones to focus on when designing a wearable mental involvement monitoring EEG system, both in physiological and epileptic conditions.

Despite hand gesture recognition is a widely investigated field, the design of myoelectric architectures for detecting finer motor task, like the handwriting, is less studied. However, writing tasks involving cognitive loads represent an important aspect toward the generalization of myoelectric-based human-machine interfaces (HMI), and also for many rehabilitative tasks. In this study, the handwriting recognition of the ten digits was faced under the myoelectric control perspective, considering the probes setup and the feature extraction step. Time and frequency domain features were extracted from surface electromyography (sEMG) signals of 11 subjects who wrote the ten digits following a standardized template and 8 sEMG probes were equally distributed between forearm and wrist. Feature class separability was investigated and an aggregated feature set was built to train pattern recognition architectures, i.e. linear discriminant analysis (LDA) and quadratic support vector machine (QSVM). Also, four reduced probes setups were investigated. LDA and QSVM showed mean accuracy of about 97%, with all the forearm and wrist sEMG information. A significant reduction of performances was observed considering the wrist or the forearm only ( $\leq 92$ %) and when LDA and QSVM were trained with two electrodes information ( $\leq 90$ %). For the reliable classification performances in a motor task involving high cognitive demands, like the handwriting, it is required the use of probes fully covering forearm and wrist. Outcomes support the methodological transfer from myoelectric hand gesture to the handwriting recognition, which represents a key aspect in the development of new HMI for rehabilitation tasks.

Goal: To evaluate suitability of respiratory signals derived from clinical 12-lead electrocardiograms (ECGs) and wearable 1-lead ECG to identify different respiration types. Methods: ECGs were simultaneously acquired through the M12R ECG Holter by Global Instrumentation and the chest strap BioHarness 3.0 by Zephyr from 42 healthy subjects alternating normal breathing, breath holding, and deep breathing. Respiration signals were derived from the ECGs through the Segmented-Beat Modulation Method (SBMM)-based algorithm and the algorithms by Van Gent, Charlton, Soni and Sarkar, and characterized in terms of breathing rate and amplitude. Respiration classification was performed through a linear support vector machine and evaluated by F1 score. Results: Best F1 scores were 86.59%(lead V2) and 80.57%, when considering 12-lead and 1-lead ECGs, respectively, and using SBMM-based algorithm. Conclusion: ECG-derived respiratory signals allow reliable identification of different respiration types even when acquired through wearable sensors, if associated to appropriate processing algorithms, such as the SBMM-based algorithm.

Recent studies have highlighted the possibility of using surface electromyographic (EMG) signals to develop human–computer interfaces that are also able to recognize complex motor tasks involving the hand as the handwriting of digits. However, the automatic recognition of words from EMG information has not yet been studied. The aim of this study is to investigate the feasibility of using combined forearm and wrist EMG probes for solving the handwriting recognition problem of 30 words with consolidated machine-learning techniques and aggregating state-of-the-art features extracted in the time and frequency domains. Six healthy subjects, three females and three males aged between 25 and 40 years, were recruited for the study. Two tests in pattern recognition were conducted to assess the possibility of classifying fine hand movements through EMG signals. The first test was designed to assess the feasibility of using consolidated myoelectric control technology with shallow machine-learning methods in the field of handwriting detection. The second test was implemented to assess if specific feature extraction schemes can guarantee high performances with limited complexity of the processing pipeline. Among support vector machine, linear discriminant analysis, and K-nearest neighbours (KNN), the last one showed the best classification performances in the 30-word classification problem, with a mean accuracy of 95% and 85% when using all the features and a specific feature set known as TDAR, respectively. The obtained results confirmed the validity of using combined wrist and forearm EMG data for intelligent handwriting recognition through pattern recognition approaches in real scenarios.

Timely and reliable fetal monitoring is crucial to prevent adverse events during pregnancy and delivery. Fetal phonocardiography, i.e., the recording of fetal heart sounds, is emerging as a novel possibility to monitor fetal health status. Indeed, due to its passive nature and its noninvasiveness, the technique is suitable for long-term monitoring and for telemonitoring applications. Despite the high share of literature focusing on signal processing, no previous work has reviewed the technological hardware solutions devoted to the recording of fetal heart sounds. Thus, the aim of this scoping review is to collect information regarding the acquisition devices for fetal phonocardiography (FPCG), focusing on technical specifications and clinical use. Overall, PRISMA-guidelines-based analysis selected 57 studies that described 26 research prototypes and eight commercial devices for FPCG acquisition. Results of our review study reveal that no commercial devices were designed for fetal-specific purposes, that the latest advances involve the use of multiple microphones and sensors, and that no quantitative validation was usually performed. By highlighting the past and future trends and the most relevant innovations from both a technical and clinical perspective, this review will represent a useful reference for the evaluation of different acquisition devices and for the development of new FPCG-based systems for fetal monitoring.

Abstract Background The triglyceride-glucose index (TyG) has been proposed as a surrogate marker of insulin resistance, which is a typical trait of pregnancy. However, very few studies analyzed TyG performance as marker of insulin resistance in pregnancy, and they were limited to insulin resistance assessment at fasting rather than in dynamic conditions, i.e., during an oral glucose tolerance test (OGTT), which allows more reliable assessment of the actual insulin sensitivity impairment. Thus, first aim of the study was exploring in pregnancy the relationships between TyG and OGTT-derived insulin sensitivity. In addition, we developed a new version of TyG, for improved performance as marker of insulin resistance in pregnancy. Methods At early pregnancy, a cohort of 109 women underwent assessment of maternal biometry and blood tests at fasting, for measurements of several variables (visit 1). Subsequently (26 weeks of gestation) all visit 1 analyses were repeated (visit 2), and a subgroup of women (84 selected) received a 2 h-75 g OGTT (30, 60, 90, and 120 min sampling) with measurement of blood glucose, insulin and C-peptide for reliable assessment of insulin sensitivity (PREDIM index) and insulin secretion/beta-cell function. The dataset was randomly split into 70% training set and 30% test set, and by machine learning approach we identified the optimal model, with TyG included, showing the best relationship with PREDIM. For inclusion in the model, we considered only fasting variables, in agreement with TyG definition. Results The relationship of TyG with PREDIM was weak. Conversely, the improved TyG, called TyGIS, (linear function of TyG, body weight, lean body mass percentage and fasting insulin) resulted much strongly related to PREDIM, in both training and test sets (R2 > 0.64, p

COVID-19 is an infectious disease that has greatly affected worldwide healthcare systems, due to the high number of cases and deaths. As COVID-19 patients may develop cardiac comorbidities that can be potentially fatal, electrocardiographic monitoring can be crucial. This work aims to identify electrocardiographic and vectorcardiographic patterns that may be related to mortality in COVID-19, with the application of the Advanced Repeated Structuring and Learning Procedure (AdvRS&LP). The procedure was applied to data from the “automatic computation of cardiovascular arrhythmic risk from electrocardiographic data of COVID-19 patients” (COVIDSQUARED) project to obtain neural networks (NNs) that, through 254 electrocardiographic and vectorcardiographic features, could discriminate between COVID-19 survivors and deaths. The NNs were validated by a five-fold cross-validation procedure and assessed in terms of the area under the curve (AUC) of the receiver operating characteristic. The features’ contribution to the classification was evaluated through the Local-Interpretable Model-Agnostic Explanations (LIME) algorithm. The obtained NNs properly discriminated between COVID-19 survivors and deaths (AUC = 84.31 ± 2.58% on hold-out testing datasets); the classification was mainly affected by the electrocardiographic-interval-related features, thus suggesting that changes in the duration of cardiac electrical activity might be related to mortality in COVID-19 cases.

Abstract Despite early repolarization (ER) syndrome being usually considered benign, its association with severe/malignant ventricular arrhythmias (VA) was also reported. Microvolt T‐wave alternans (MTWA) is an electrocardiographic marker for the development of VA, but its role in ER syndrome remains unknown. A 90‐second 6‐lead electrocardiogram from an ER syndrome patient, acquired with the Kardia recorder, was analyzed by the enhanced adaptive matched filter for MTWA quantification. On average, MTWA was 50 μV, higher than what was previously observed on healthy subjects using the same method. In our ER syndrome patient, MTWA plays a potential role in VA development in ER syndrome.

Amino acids (AAs) are well known to be involved in the regulation of glucose metabolism and, in particular, of insulin secretion. However, the effects of different AAs on insulin release and kinetics have not been completely elucidated. The aim of this study was to propose a mathematical model that includes the effect of AAs on insulin kinetics during a mixed meal tolerance test. To this aim, five different models were proposed and compared. Validation was performed using average data, derived from the scientific literature, regarding subjects with normal glucose tolerance (CNT) and with type 2 diabetes (T2D). From the average data of the CNT and T2D people, data for two virtual populations (100 for each group) were generated for further model validation. Among the five proposed models, a simple model including one first-order differential equation showed the best results in terms of model performance (best compromise between model structure parsimony, estimated parameters plausibility, and data fit accuracy). With regard to the contribution of AAs to insulin appearance/disappearance (kAA model parameter), model analysis of the average data from the literature yielded 0.0247 (confidence interval, CI: 0.0168 – 0.0325) and -0.0048 (CI: -0.0281 – 0.0185) μU·ml-1/(μmol·l-1·min), for CNT and T2D, respectively. This suggests a positive effect of AAs on insulin secretion in CNT, and negligible effect in T2D. In conclusion, a simple model, including single first-order differential equation, may help to describe the possible AAs effects on insulin kinetics during a physiological metabolic test, and provide parameters that can be assessed in the single individuals.

Background: This review systematically examined the scientific literature about electroencephalogram-derived ratio indexes used to assess human mental involvement, in order to deduce what they are, how they are defined and used, and what their best fields of application are. (2) Methods: The review was carried out according to the Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) guidelines. (3) Results: From the search query, 82 documents resulted. The majority (82%) were classified as related to mental strain, while 12% were classified as related to sensory and emotion aspects, and 6% to movement. The electroencephalographic electrode montage used was low-density in 13%, high-density in 6% and very-low-density in 81% of documents. The most used electrode positions for computation of involvement indexes were in the frontal and prefrontal cortex. Overall, 37 different formulations of involvement indexes were found. None of them could be directly related to a specific field of application. (4) Conclusions: Standardization in the definition of these indexes is missing, both in the considered frequency bands and in the exploited electrodes. Future research may focus on the development of indexes with a unique definition to monitor and characterize mental involvement.

Knowledge about the anatomical structures of the left heart, specifically the atrium (LA) and ventricle (i.e., endocardium—Vendo—and epicardium—LVepi) is essential for the evaluation of cardiac functionality. Manual segmentation of cardiac structures from echocardiography is the baseline reference, but results are user-dependent and time-consuming. With the aim of supporting clinical practice, this paper presents a new deep-learning (DL)-based tool for segmenting anatomical structures of the left heart from echocardiographic images. Specifically, it was designed as a combination of two convolutional neural networks, the YOLOv7 algorithm and a U-Net, and it aims to automatically segment an echocardiographic image into LVendo, LVepi and LA. The DL-based tool was trained and tested on the Cardiac Acquisitions for Multi-Structure Ultrasound Segmentation (CAMUS) dataset of the University Hospital of St. Etienne, which consists of echocardiographic images from 450 patients. For each patient, apical two- and four-chamber views at end-systole and end-diastole were acquired and annotated by clinicians. Globally, our DL-based tool was able to segment LVendo, LVepi and LA, providing Dice similarity coefficients equal to 92.63%, 85.59%, and 87.57%, respectively. In conclusion, the presented DL-based tool proved to be reliable in automatically segmenting the anatomical structures of the left heart and supporting the cardiological clinical practice.

The brain is the organ most studied using Magnetic Resonance (MR). The emergence of 7T scanners has increased MR imaging resolution to a sub-millimeter level. However, there is a lack of automatic segmentation techniques for 7T MR volumes. This research aims to develop a novel deep learning-based algorithm for on-cloud brain extraction and multi-structure segmentation from unenhanced 7T MR volumes. To this aim, a double-stage 3D U-Net was implemented in a cloud service, directing its first stage to the automatic extraction of the brain and its second stage to the automatic segmentation of the grey matter, basal ganglia, white matter, ventricles, cerebellum, and brain stem. The training was performed on the 90% (the 10% of which served for validation) and the test on the 10% of the Glasgow database. A mean test Dice Similarity Coefficient (DSC) of 96.33% was achieved for the brain class. Mean test DSCs of 90.24%, 87.55%, 93.82%, 85.77%, 91.53%, and 89.95% were achieved for the brain structure classes, respectively. Therefore, the proposed double-stage 3D U-Net is effective in brain extraction and multi-structure segmentation from 7T MR volumes without any preprocessing and training data augmentation strategy while ensuring its machine-independent reproducibility.

Wearable and portable devices capable of acquiring cardiac signals are at the frontier of the sport industry. They are becoming increasingly popular for monitoring physiological parameters while practicing sport, given the advances in miniaturized technologies, powerful data, and signal processing applications. Data and signals acquired by these devices are increasingly used to monitor athletes’ performances and thus to define risk indices for sport-related cardiac diseases, such as sudden cardiac death. This scoping review investigated commercial wearable and portable devices employed for cardiac signal monitoring during sport activity. A systematic search of the literature was conducted on PubMed, Scopus, and Web of Science. After study selection, a total of 35 studies were included in the review. The studies were categorized based on the application of wearable or portable devices in (1) validation studies, (2) clinical studies, and (3) development studies. The analysis revealed that standardized protocols for validating these technologies are necessary. Indeed, results obtained from the validation studies turned out to be heterogeneous and scarcely comparable, since the metrological characteristics reported were different. Moreover, the validation of several devices was carried out during different sport activities. Finally, results from clinical studies highlighted that wearable devices are crucial to improve athletes’ performance and to prevent adverse cardiovascular events.

American football is the sport with the highest rates of concussion injuries. Biomedical engineering applications may support athletes in monitoring their injuries, evaluating the effectiveness of their equipment, and leading industrial research in this sport. This literature review aims to report on the applications of biomedical engineering research in American football, highlighting the main trends and gaps. The review followed the PRISMA guidelines and gathered a total of 1629 records from PubMed (n = 368), Web of Science (n = 665), and Scopus (n = 596). The records were analyzed, tabulated, and clustered in topics. In total, 112 studies were selected and divided by topic in the biomechanics of concussion (n = 55), biomechanics of footwear (n = 6), biomechanics of sport-related movements (n = 6), the aerodynamics of football and catch (n = 3), injury prediction (n = 8), heat monitoring of physiological parameters (n = 8), and monitoring of the training load (n = 25). The safety of players has fueled most of the research that has led to innovations in helmet and footwear design, as well as improvements in the understanding and prevention of injuries and heat monitoring. The other important motivator for research is the improvement of performance, which has led to the monitoring of training loads and catches, and studies on the aerodynamics of football. The main gaps found in the literature were regarding the monitoring of internal loads and the innovation of shoulder pads.

Gestational diabetes mellitus (GDM) is a type of diabetes that usually resolves at the end of the pregnancy but exposes to a higher risk of developing type 2 diabetes mellitus (T2DM). This study aimed to unravel the factors, among those that quantify specific metabolic processes, which determine progression to T2DM by using machine-learning techniques. Classification of women who did progress to T2DM (labeled as PROG, n = 19) vs. those who did not (labeled as NON-PROG, n = 59) progress to T2DM has been performed by using Orange software through a data analysis procedure on a generated data set including anthropometric data and a total of 34 features, extracted through mathematical modeling/methods procedures. Feature selection has been performed through decision tree algorithm and then Naïve Bayes and penalized (L2) logistic regression were used to evaluate the ability of the selected features to solve the classification problem. Performance has been evaluated in terms of area under the operating receiver characteristics (AUC), classification accuracy (CA), precision, sensitivity, specificity, and F1. Feature selection provided six features, and based on them, classification was performed as follows: AUC of 0.795, 0.831, and 0.884; CA of 0.827, 0.813, and 0.840; precision of 0.830, 0.854, and 0.834; sensitivity of 0.827, 0.813, and 0.840; specificity of 0.700, 0.821, and 0.662; and F1 of 0.828, 0.824, and 0.836 for tree algorithm, Naïve Bayes, and penalized logistic regression, respectively. Fasting glucose, age, and body mass index together with features describing insulin action and secretion may predict the development of T2DM in women with a history of GDM.

Tidal volume (TV), defined as the amount of air that moves in or out of the lungs with each respiratory cycle, is important in evaluating the respiratory function. Although TV can be reliably measured in laboratory settings, this information is hardly obtainable under everyday living conditions. Under such conditions, wearable devices could provide valuable support to monitor vital signs, such as heart rate (HR) and breathing rate (BR). The aim of this study was to develop a model to estimate TV from wearable-device measures of HR and BR during exercise. HR and BR were acquired through the Zephyr Bioharness 3.0 wearable device in nine subjects performing incremental cycling tests. For each subject, TV during exercise was obtained with a metabolic cart (Cosmed). A stepwise regression algorithm was used to create the model using as possible predictors HR, BR, age, and body mass index; the model was then validated using a leave-one-subject-out cross-validation procedure. The performance of the model was evaluated using the explained variance (R2), obtaining values ranging from 0.65 to 0.72. The proposed model is a valid method for TV estimation with wearable devices and can be considered not subject-specific and not instrumentation-specific.