Umberto Mosca

Mr. Umberto Mosca

orcid

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
8 November 2025 in Sensors
Verified
Journal Article
Open access
PDF file
MCMichele Antonio Gazzanti Pugliese di Cotrone
Michele Antonio Gazzanti Pugliese di Cotrone
NANidà Farooq Akhtar
Nidà Farooq Akhtar
MPMARTINA PATERA
MARTINA PATERA
SGSilvia Gallo
Silvia Gallo
Umberto MoscaUmberto Mosca
Umberto Mosca
Fernanda IrreraFernanda Irrera
Fernanda Irrera
The present study evaluates the effectiveness of a non-invasive wearable sensor system, combining accelerometers, surface electromyography, and artificial intelligence, to objectively characterize swallowing in elderly individuals affected by Parkinson’s Disease, without clinically manifested dysphagia. A cohort of patients and healthy control subjects performed the same swallowing test protocol, including tasks with different viscosity boluses, positioning a commercial adhesive grid of High-Density surface Electromyography (HD-sEMG) electrodes on the submental muscle and a triaxial accelerometer over the thyroid cartilage. Relevant temporal and spectral features were extracted from electromyography data. Proper filtering and processing by machine learning and Principal Component Analysis allowed identification of two distinct clusters of subjects, one predominantly composed of controls with just a few patients, the other mostly crowded by patients. Excellent classification performances were achieved (accuracy = 83.3%, precision = 79.0%, recall = 90.7%, F1-score = 84.5%, Cohen’s kappa = 0.67), revealing consistent differences in muscle activation patterns among subjects, even in the absence of clinically diagnosed dysphagia. These results support the feasibility of wearable sensor-based assessment as a reliable and non-invasive tool for the early detection of subclinical swallowing dysfunction in Parkinson’s Disease.


11 November 2025 in TechRxiv
Verified
Preprint
Open access
Umberto MoscaUmberto Mosca
Umberto Mosca
IRIrene Rechichi
Irene Rechichi
ACAlessandro Cicolin
Alessandro Cicolin
GOGabriella Olmo
Gabriella Olmo
REM Sleep Behavior Disorder (RBD) is a reliable prodromal marker of synucleinopathies, including Parkinson's Disease, yet it often remains underdiagnosed due to the need for specialized and intrusive assessments. Heart Rate Variability (HRV), reflecting autonomic dysfunction, often associated with RBD, offers a non-intrusive alternative that can be collected via wearable devices, thus enabling scalable early screening. This study investigates the feasibility of automatically detecting RBD with Machine Learning (ML), using HRV features, including linear and nonlinear metrics. Electrocardiographic (ECG)recordings from four datasets (three public, one private) were employed, comprising healthy controls, RBD subjects, and subjects with Obstructive Sleep Apnea (OSAS). OSAS cases were deliberately included to evaluate model robustness under realistic, real-world screening scenarios with confounding sleep disturbances. Among tested feature selection and classifier combinations, the best models achieved an average Accuracy of 88.0 ± 3.5% and an F1-score of 89.3 ± 2.6% for the Healthy versus RBD classification. Including OSAS subjects (identified as non-RBD) slightly reduced performance (Accuracy: 85.9 ± 4.7%, F1score: 79.4 ± 8.7%), but maintained promising results. These findings demonstrate the potential of HRVbased ML models to support early RBD detection, facilitating timely intervention and monitoring in highrisk populations, through non-intrusive, scalable, and wearable-compatible data collection.

1 July 2026 in Algorithms
Verified
Journal Article
Open access
PDF file
Ilaria CiampaIlaria Ciampa
Ilaria Ciampa
Benedetta PerroneBenedetta Perrone
Benedetta Perrone
Umberto MoscaUmberto Mosca
Umberto Mosca
EFElisa Fattori
Elisa Fattori
SSSerena Sinagra
Serena Sinagra
Gabriella OlmoGabriella Olmo
Gabriella Olmo
Polysomnography (PSG) is the gold standard for diagnosing sleep disorders, but its complexity and cost limit widespread use. Heart rate variability (HRV) is traditionally assessed from electrocardiography (ECG), while photoplethysmography (PPG), widely available in wearable devices, offers a more accessible alternative. However, its reliability over full-night recordings remains underexplored. This study analyzes data from 50 subjects across five groups (healthy controls, rapid eye movement sleep behavior disorder, obstructive sleep apnea, periodic limb movements, and mixed comorbidities) to assess agreement between ECG-derived HRV and PPG-derived pulse rate variability (PRV), considering time-, frequency-, and nonlinear-domain features. Correlation and equivalence analyses were performed, with and without removal of artifactual segments. Correlation coefficients exceeded 0.6 for most features and improved to above 0.7 after artifact removal. Consistent improvements were observed across all subject groups. Equivalence testing further identified a subset of features showing high agreement and low bias. The results indicate that, with appropriate pre-processing, PPG can approximate ECG-derived variability in full-night sleep recordings. The identification of robust features for screening purposes supports the use of PRV for wearable-based screening and monitoring in heterogeneous sleep disorder populations.

3 September 2025 in Research Square
Verified
Preprint
Open access
PDF file
Umberto MoscaUmberto Mosca
Umberto Mosca
IRENE RECHICHIIRENE RECHICHI
IRENE RECHICHI
Alessandro CicolinAlessandro Cicolin
Alessandro Cicolin
Gabriella OlmoGabriella Olmo
Gabriella Olmo
Heart rate variability (HRV) is a promising biomarker for detecting subtle autonomic alterations in sleep disorders such as REM Sleep Behavior Disorder (RBD). However, the relevance of HRV-derived features depends heavily on methodological choices, including data segmentation, preprocessing, and dataset composition. This study systematically investigates how four key factors (epoch duration, inclusion of wake segments, dataset heterogeneity, and the presence of comorbidities such as Obstructive Sleep Apnea Syndrome (OSAS)), affect HRV feature discriminative power. Using multiple polysomnography datasets, we extract a comprehensive suite of 528 features per subject and evaluate their importance using a combination of statistical and machine learningbased techniques. Results show that longer epochs typically yield more robust features, though shorter windows enhance the prominence of nonlinear metrics. Including wake segments improves feature relevance in heterogeneous datasets, supporting their use in real-world conditions. Furthermore, dataset aggregation enhances generalizability for full-night recordings, but may reduce performance when analyzing sleep-only data. Finally, including OSAS subjects had minimal effect on top-ranked features, indicating the robustness of HRV-based RBD markers across varying clinical contexts. These findings highlight the importance of thoughtful experimental design in HRV-based classification models in sleep research.