Automatic detection of breathing disorder from ballistocardiography signals

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F62690094%3A18470%2F20%3A50016011" target="_blank" >RIV/62690094:18470/20:50016011 - isvavai.cz</a>

Výsledek na webu

<a href="https://www.sciencedirect.com/science/article/pii/S0950705119304009" target="_blank" >https://www.sciencedirect.com/science/article/pii/S0950705119304009</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1016/j.knosys.2019.104973" target="_blank" >10.1016/j.knosys.2019.104973</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Automatic detection of breathing disorder from ballistocardiography signals

Popis výsledku v původním jazyce

Ballistocardiography (BCG) is a common method, wherein sensory information is used to identify blood-flow cardiac activity by measuring the mechanical micromovements of the human body generated by heart movements and blood eviction to the large arteries. BCG signals can be used to detect non-standard vital functions or predict likely health problems. However, the analysis of BCG signal is challenging because it contains various mechanical noises made by human body movements. This study is aimed at extracting information regarding the pulse arrival time from BCG signals and then establishing a connection with changes in breathing disorders, such as simulated apnoea, using convolutional neural networks. We present a novel approach toward recognizing the form of breathing which is independent of the body position while data are being collected from tensometers measuring the mechanical micromovements (motion) of the individual. The mechanical motions are caused by cardiac activity with multivariate time series output, which is processed to obtain the source data for breath detection. The signals are first processed by Cartan curvature. This is a differential geometric invariant, which enables the detection of marginal variations in the signals. Conditional dependency and short-term fluctuations are eliminated in longer measuring-periods. By these means, the breathing anomalies of individuals are subsequently detected between heartbeats using the time delay between the R-wave from the electrocardiogram (ECG) and the pulse arrival times. Moreover, ECG signals are included in the system for data sampling. In addition, the values of the time delay are used as the inputs to train a convolutional neural network classifier with two outputs (regular and disordered breathing) to validate the experiment. We achieved an average accuracy of 89.35%, sensitivity of 86.35%, and specificity of 91.22% on 828 regular and 1332 disordered breathing states from eight human subjects. The conclusion is that our novel method can detect disordered breathing from processed BCG signal, i.e. from the pulse arrival time, in a manner not previously used elsewhere.

Název v anglickém jazyce

Automatic detection of breathing disorder from ballistocardiography signals

Popis výsledku anglicky

Ballistocardiography (BCG) is a common method, wherein sensory information is used to identify blood-flow cardiac activity by measuring the mechanical micromovements of the human body generated by heart movements and blood eviction to the large arteries. BCG signals can be used to detect non-standard vital functions or predict likely health problems. However, the analysis of BCG signal is challenging because it contains various mechanical noises made by human body movements. This study is aimed at extracting information regarding the pulse arrival time from BCG signals and then establishing a connection with changes in breathing disorders, such as simulated apnoea, using convolutional neural networks. We present a novel approach toward recognizing the form of breathing which is independent of the body position while data are being collected from tensometers measuring the mechanical micromovements (motion) of the individual. The mechanical motions are caused by cardiac activity with multivariate time series output, which is processed to obtain the source data for breath detection. The signals are first processed by Cartan curvature. This is a differential geometric invariant, which enables the detection of marginal variations in the signals. Conditional dependency and short-term fluctuations are eliminated in longer measuring-periods. By these means, the breathing anomalies of individuals are subsequently detected between heartbeats using the time delay between the R-wave from the electrocardiogram (ECG) and the pulse arrival times. Moreover, ECG signals are included in the system for data sampling. In addition, the values of the time delay are used as the inputs to train a convolutional neural network classifier with two outputs (regular and disordered breathing) to validate the experiment. We achieved an average accuracy of 89.35%, sensitivity of 86.35%, and specificity of 91.22% on 828 regular and 1332 disordered breathing states from eight human subjects. The conclusion is that our novel method can detect disordered breathing from processed BCG signal, i.e. from the pulse arrival time, in a manner not previously used elsewhere.

Druh

J_imp - Článek v periodiku v databázi Web of Science

CEP obor

—

OECD FORD obor

10102 - Applied mathematics

Projekt

<a href="/cs/project/EF17_048%2F0007441" target="_blank" >EF17_048/0007441: PERSONMED - Centrum rozvoje personalizované medicíny u věkem podmíněných onemocnění</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)<br>S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2020

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název periodika

Knowledge-based systems

ISSN

0950-7051

e-ISSN

—

Svazek periodika

188

Číslo periodika v rámci svazku

January

Stát vydavatele periodika

NL - Nizozemsko

Počet stran výsledku

11

Strana od-do

"Article number: 104973"

Kód UT WoS článku

000513295000011

EID výsledku v databázi Scopus

2-s2.0-85071566714