Leveraging IoT Devices for Atrial Fibrillation Detection: A Comprehensive Study of AI Techniques DOI Creative Commons

Alicia Pedrosa-Rodriguez,

Carmen Cámara, Pedro Peris‐Lopez

et al.

Applied Sciences, Journal Year: 2024, Volume and Issue: 14(19), P. 8945 - 8945

Published: Oct. 4, 2024

Internet of Things (IoT) devices play a crucial role in the real-time acquisition photoplethysmography (PPG) signals, facilitating seamless data transmission to cloud-based platforms for analysis. Atrial fibrillation (AF), affecting approximately 1–2% global population, requires accurate detection methods due its prevalence and health impact. This study employs IoT capture PPG signals implements comprehensive preprocessing steps, including windowing, filtering, artifact removal, extract relevant features classification. We explored broad range machine learning (ML) deep (DL) approaches. Our results demonstrate superior performance, achieving an accuracy 97.7%, surpassing state-of-the-art methods, those with FDA clearance. Key strengths our proposal include use shortened 15-second traces validation using publicly available datasets. research advances design cost-effective AF by leveraging diverse ML DL techniques enhance classification robustness.

Language: Английский

Accuracy of the Instantaneous Breathing and Heart Rates Estimated by Smartphone Inertial Units DOI Creative Commons

Eliana Cinotti,

Jessica Centracchio, Salvatore J. Parlato

et al.

Sensors, Journal Year: 2025, Volume and Issue: 25(4), P. 1094 - 1094

Published: Feb. 12, 2025

Seismocardiography (SCG) and Gyrocardiography (GCG) use lightweight, miniaturized accelerometers gyroscopes to record, respectively, cardiac-induced linear accelerations angular velocities of the chest wall. These inertial sensors are also sensitive thoracic movements with respiration, which cause baseline wanderings in SCG GCG signals. Nowadays, widely integrated into smartphones, thus increasing potential as cardiorespiratory monitoring tools. This study investigates accuracy smartphone simultaneously measuring instantaneous heart rates breathing rates. Smartphone-derived signals were acquired from 10 healthy subjects at rest. The performances heartbeats respiratory acts detection, well inter-beat intervals (IBIs) inter-breath (IBrIs) estimation, evaluated for both via comparison simultaneous electrocardiography respiration belt Heartbeats detected a sensitivity positive predictive value (PPV) 89.3% 93.3% 97.3% 97.9% Moreover, IBIs measurements reported strong relationships (R2 > 0.999), non-significant biases, Bland-Altman limits agreement (LoA) ±7.33 ms ±5.22 GCG. On other hand, detection scored PPV 95.6% 94.7% 95.7% 92.0% Furthermore, high R2 values (0.976 0.968, respectively), an LoA ±0.558 s ±0.749 achieved IBrIs estimates. results this confirm that can provide accurate rate without need additional devices.

Language: Английский

Citations

0
Leveraging IoT Devices for Atrial Fibrillation Detection: A Comprehensive Study of AI Techniques DOI Creative Commons

Alicia Pedrosa-Rodriguez,

Carmen Cámara, Pedro Peris‐Lopez

et al.

Applied Sciences, Journal Year: 2024, Volume and Issue: 14(19), P. 8945 - 8945

Published: Oct. 4, 2024

Internet of Things (IoT) devices play a crucial role in the real-time acquisition photoplethysmography (PPG) signals, facilitating seamless data transmission to cloud-based platforms for analysis. Atrial fibrillation (AF), affecting approximately 1–2% global population, requires accurate detection methods due its prevalence and health impact. This study employs IoT capture PPG signals implements comprehensive preprocessing steps, including windowing, filtering, artifact removal, extract relevant features classification. We explored broad range machine learning (ML) deep (DL) approaches. Our results demonstrate superior performance, achieving an accuracy 97.7%, surpassing state-of-the-art methods, those with FDA clearance. Key strengths our proposal include use shortened 15-second traces validation using publicly available datasets. research advances design cost-effective AF by leveraging diverse ML DL techniques enhance classification robustness.

Language: Английский

Citations

1