Journal Publications
Article |
Link |
Abstract |
[1] Sanchez-Morillo, D., Sales-Lerida, D., Priego-Torres, B., & León-Jiménez, A. (2024). Cough Detection Using Acceleration Signals and Deep Learning Techniques. Electronics, 13(12), 2410.
|
|
Cough is a frequent symptom in many common respiratory diseases and is considered a predictor of early exacerbation or even disease progression. Continuous cough monitoring offers valuable insights into treatment effectiveness, aiding healthcare providers in timely intervention to prevent exacerbations and hospitalizations. Objective cough monitoring methods have emerged as superior alternatives to subjective methods like questionnaires. In recent years, cough has been monitored using wearable devices equipped with microphones. However, the discrimination of cough sounds from background noise has been shown a particular challenge. This study aimed to demonstrate the effectiveness of single-axis acceleration signals combined with state-of-the-art deep learning (DL) algorithms to distinguish intentional coughing from sounds like speech, laugh, or throat noises. Various DL methods (recurrent, convolutional, and deep convolutional neural networks) combined with one- and two-dimensional time and time–frequency representations, such as the signal envelope, kurtogram, wavelet scalogram, mel, Bark, and the equivalent rectangular bandwidth spectrum (ERB) spectrograms, were employed to identify the most effective approach. The optimal strategy, which involved the SqueezeNet model in conjunction with wavelet scalograms, yielded an accuracy and precision of 92.21% and 95.59%, respectively. The proposed method demonstrated its potential for cough monitoring. Future research will focus on validating the system in spontaneous coughing of subjects with respiratory diseases under natural ambulatory conditions. |
[2] Camacho-Magriñán, P., Sales-Lerida, D., León-Jiménez, A., & Sanchez-Morillo, D. (2025). Indoor Environmental Monitoring and Chronic Respiratory Diseases: A Systematic Review. Technologies, 13(3), 122.
|
Chronic respiratory diseases (CRD), which include Chronic Obstructive Pulmonary Disease (COPD) and asthma, are significant global health issues, with air quality playing a vital role in exacerbating these conditions. This systematic review explores how monitoring indoor air quality (IAQ) can help manage and reduce respiratory exacerbations in CRD patients. A search of the Web of Science database, yielding 301 articles, was conducted following PRISMA guidelines. Of these, 60 met the inclusion criteria, and after screening, 21 articles were analyzed. The review identified substantial gaps in current research: the lack of standardization in IAQ monitoring; the need for considering geographic variability and for long-term longitudinal studies; and the importance of linking monitored air quality data with respiratory health indicators. It also stressed the importance of considering the heterogeneity of patients in the methodological study design, as well as the convenience of introducing recommendation systems to assess the true impact of corrective measures on indoor air quality in the homes of chronic respiratory patients. The integration of home-based IAQ monitoring with machine learning techniques to enhance our understanding of the relationship between IAQ and respiratory health is emerging as a key area for future research. Addressing all these challenges has the potential to mitigate the impact of CRD and improve the quality of life for patients. | |
[3] Khalili Khalili, E., Sánchez Morillo, D., Priego Torres, B. M., & León Jiménez, A. (2025). Localization and classification of abnormalities on chest X-ray images using a Mamba-YOLOvX model. Expert Systems with Applications, 284, 127929.
|
Chest X-rays (CXR) are critical diagnostic tools for detecting thoracic abnormalities. However, challenges such as overlapping anatomical structures, class imbalance, and dataset heterogeneity hinder accurate interpretation and limit model generalizability. To address these issues, a Mamba-YOLOvX model is presented in this study. It was aimed to integrate global and local lesion information to improve the detection and localization of thoracic abnormalities. The model incorporates novel architectural improvements, including combined spatial and channel attention mechanisms and selective scanning blocks, to capture fine-grained features and enhance multi-scale detection. In addition, a projection-based data augmentation strategy, leveraging rib segmentation and keypoint alignment was developed to improve the anatomical consistency and the intensity normalization across datasets. Extensive experiments were conducted on three large-scale datasets (VinDr-CXR, ChestX-ray8, and CXR-AL14), achieving state-of-the-art performance in detecting abnormalities of varying sizes. The proposed method reached an average precision at 50 % intersection over union of 0.366, 0.153, and 0.615 on the VinDr-CXR, ChestX-ray8, and CXR-AL14 datasets, respectively. Results demonstrated significant improvements in precision, recall, and mean average precision, particularly for small lesions. Cross-dataset validation confirmed the model’s robustness and generalizability. This study highlights the potential of integrating advanced deep learning techniques with domain-specific augmentations to enhance clinical decision support systems for thoracic disease detection. By addressing critical challenges such as class imbalance, annotation inconsistencies, and scale variations, the enhanced Mamba-YOLOvX model is shown as a scalable, accurate, and generalizable solution for automated CXR analysis. | |
[4] Camacho-Magriñán, P., Sales-Lerida, D., Lara-Doña, A., & Sanchez-Morillo, D. (2025). Leveraging Low-Cost Sensor Data and Predictive Modelling for IoT-Driven Indoor Air Quality Monitoring. Smart Cities, 8(6), 200.
|
Indoor air quality (IAQ) in residential settings is often dominated by high-concentration pollutant events from activities such as cooking and occupancy, which are overlooked by traditional 24 h average assessments. In this, we have designed and implemented a low-cost unit for remote IAQ monitoring. We deployed these units for high-resolution remote monitoring of CO2, particulate matter (PM), and volatile organic compounds (VOCs) in three different domestic environments: a kitchen, a living room, and a bedroom. The monitoring campaign confirmed that, while daily averages frequently remained below guideline limits, transient peaks (e.g., CO2 exceeding 2800 ppm in bedrooms and significant increases in PM during cooking) posed acute exposure risks. This dataset was used to train and evaluate machine learning models for 10 min ahead pollutant forecasting. Ensemble tree-based methods (Random Forest) and gradient boosting algorithms (XGBoost, LGBM, and CatBoost) were effective and robust. The predictability of the models correlated with room dynamics: performance improved under clear cyclical patterns (bedroom) and remained stable under stochastic events (kitchen). This work shows that integrating low-cost IoT sensing with machine learning enables proactive IAQ management, supporting health interventions driven by predictive risk rather than static averages. | |
[5] Sanchez-Morillo, D.; Camacho-Magriñán, P.; León-Jiménez, A. Multimodal Prodromal Patterns Associated with COPD Exacerbations in Home Telemonitoring: An Exploratory Observational Study. Bioengineering 2026, 13, 1195.
|
DOI | Characterizing multi-parameter physiological, acoustic, and environmental changes before acute chronic obstructive pulmonary disease exacerbations (AECOPD) is crucial for improving home telemonitoring and enabling timely clinical intervention. In a 12-month prospective observational exploratory pilot study (𝑁=16 ), a multimodal Internet of Things (IoT) ecosystem collected daily data on symptoms, lung function, cardiorespiratory dynamics, actigraphy, sleep, fine motor skills, bioacoustics, and indoor air quality. From 427 daily features across 4738 patient-days, collinearity mitigation retained 245 independent features. Linear mixed-effects models (LMM), nonparametric paired tests, and multi-timescale exponential moving average smoothing quantified prodromal trajectories and univariate discrimination during the 7-day pre-exacerbation window. Engagement analysis showed a dual hierarchy: passive background monitoring had high retention (>72 %), while demanding active tasks like peak flow spirometry dropped to 26.9 % compliance. Patients with frequent exacerbations showed higher reactive adherence. LMM frameworks identified a multisystem prodromal footprint marked by coordinated increases in cough frequency, sputum purulence, dyspnea, and resting heart rate. Acoustic measures showed notable pre-event changes in higher-order speech cepstral coefficients, voice-intensity variability, and cough spectral dispersion, providing objective proxies for airway limitation. Indoor environmental signals suggested domestic confinement but weakened after adjustment for macro-seasonal confounders. Multi-temporal kinetics demonstrated that acute physiological symptoms peaked within 72 h, whereas acoustic features required 3-to-7-day temporal smoothing (EMA3–EMA7) to optimize signal-to-noise ratios. In conclusion, these preliminary observations in the exploratory small cohort suggest that the transition from a stable baseline to the prodromal AECOPD phase is marked by a quantifiable, multi-systemic digital footprint across cardiorespiratory, bioacoustic, and symptom domains. Although requiring validation in broader populations, findings highlight the potential of multi-tiered digital health architectures that pair continuous passive sensing with low-friction active reporting to overcome task attrition and secure the uninterrupted longitudinal baselines needed for automated early-warning systems. |
Grant PID2021-126810OB-I00 funded by:
