Estimation of cerebral blood flow velocity during breath-hold challenge using artificial neural networks

Mohammad A. Al-Abed, Areen K. Al-Bashir, Ayman Al-Rawashdeh, Raichel M. Alex, Rong Zhang, Donald E. Watenpaugh, Khosrow Behbehani

Research output: Contribution to journalArticle

Abstract

The effect of untreated Obstructive Sleep Apnoea (OSA) on cerebral haemodynamics and CA impairment is an active field of research interest. A breath-hold challenge is usually used in clinical and research settings to simulate cardiovascular and cerebrovascular changes that mimic OSA events. This work utilises temporal arterial oxygen saturation (SpO2) and photoplethysmography (PPG) signals to estimate the temporal cerebral blood flow velocity (CBFv) waveform. Measurements of CBFv, SpO2, and PPG, were acquired concurrently from volunteers performing two different protocols of breath-hold challenge in the supine position. Past values of the SpO2 and PPG signals were used to estimate the current values of CBFv using different permutations and topologies of supervised learning with shallow artificial neural networks (ANNs). The measurements from one protocol were used to train the ANNs and find the optimum topologies, which in turn were tested using the other protocol. Data collected from 10 normotensive, healthy subjects (four females, age 28.5 ± 6.1 years, Body Mass Index (BMI) 24.0 ± 4.7 kg/m2) were used in this study. The results show that different subjects have different optimum topologies for ANNs, thus indicating the effects of inter-subject variability on ANNs. Successfully reconstructed blind waveforms for the same subject group in the second protocol showed a reasonable accuracy of 60–80% estimation compared to the measured waveforms. Hypothesis: Temporal waveforms for SpO2 and PPG contain adequate information to estimate the temporal CBFv waveform using ANNs. Methodology: Concurrent measurements of SpO2 and PPG using pulse oximetry from the forehead and CBFv from the middle cerebral artery (MCA) using transcranial Doppler (TCD) were recorded from healthy, normotensive subjects performing a breath-hold challenge. The breath-hold challenge mimicked the cerebrovascular response to apnoea, and was recorded by measuring CBFv in MCA. Two protocols were used, each consisting of five breath-holding manoeuvres and differing in terms of the time between the five successive breath-holds. Using data from one protocol, several permutations of the temporal values of SpO2 and PPG signals were used as inputs to different ANN topologies, in order to train and find the optimum model. The optimum model was evaluated using the data from the other protocol as a blind dataset. Results: Using the first protocol for training, optimum ANN configurations were found to be different for each subject, and accuracy of 75–87% was achieved. When these optimum ANN models were tested using the second protocol as a blind dataset, the accuracy achieved was around 60–80%. Conclusions: A novel approach employing temporal records of SpO2 and PPG can be used to estimate the CBFv waveform using ANNs with acceptable accuracy. Increases in the size and diversity of the population dataset and the use of features extracted from SpO2 and PPG signals are needed for generalisation of the method and potential future clinical applications.

Original languageEnglish (US)
Article number103508
JournalComputers in Biology and Medicine
Volume115
DOIs
StatePublished - Dec 2019
Externally publishedYes

Fingerprint

Cerebrovascular Circulation
Photoplethysmography
Blood Flow Velocity
Flow velocity
Blood
Neural networks
Topology
Middle Cerebral Artery
Obstructive Sleep Apnea
Healthy Volunteers
Breath Holding
Neural Networks (Computer)
Oximetry
Forehead
Supine Position
Apnea
Population Density
Supervised learning
Research
Hemodynamics

Keywords

  • Biomedical signal processing
  • Blood oxygen saturation
  • Breath-hold challenge
  • Cerebral blood flow velocity
  • Machine learning
  • Obstructive sleep apnea
  • Photoplethysmography
  • Time series estimation

ASJC Scopus subject areas

  • Computer Science Applications
  • Health Informatics

Cite this

Estimation of cerebral blood flow velocity during breath-hold challenge using artificial neural networks. / Al-Abed, Mohammad A.; Al-Bashir, Areen K.; Al-Rawashdeh, Ayman; Alex, Raichel M.; Zhang, Rong; Watenpaugh, Donald E.; Behbehani, Khosrow.

In: Computers in Biology and Medicine, Vol. 115, 103508, 12.2019.

Research output: Contribution to journalArticle

Al-Abed, Mohammad A. ; Al-Bashir, Areen K. ; Al-Rawashdeh, Ayman ; Alex, Raichel M. ; Zhang, Rong ; Watenpaugh, Donald E. ; Behbehani, Khosrow. / Estimation of cerebral blood flow velocity during breath-hold challenge using artificial neural networks. In: Computers in Biology and Medicine. 2019 ; Vol. 115.
@article{afcdfa61ba414268b4434328ee5beb11,
title = "Estimation of cerebral blood flow velocity during breath-hold challenge using artificial neural networks",
abstract = "The effect of untreated Obstructive Sleep Apnoea (OSA) on cerebral haemodynamics and CA impairment is an active field of research interest. A breath-hold challenge is usually used in clinical and research settings to simulate cardiovascular and cerebrovascular changes that mimic OSA events. This work utilises temporal arterial oxygen saturation (SpO2) and photoplethysmography (PPG) signals to estimate the temporal cerebral blood flow velocity (CBFv) waveform. Measurements of CBFv, SpO2, and PPG, were acquired concurrently from volunteers performing two different protocols of breath-hold challenge in the supine position. Past values of the SpO2 and PPG signals were used to estimate the current values of CBFv using different permutations and topologies of supervised learning with shallow artificial neural networks (ANNs). The measurements from one protocol were used to train the ANNs and find the optimum topologies, which in turn were tested using the other protocol. Data collected from 10 normotensive, healthy subjects (four females, age 28.5 ± 6.1 years, Body Mass Index (BMI) 24.0 ± 4.7 kg/m2) were used in this study. The results show that different subjects have different optimum topologies for ANNs, thus indicating the effects of inter-subject variability on ANNs. Successfully reconstructed blind waveforms for the same subject group in the second protocol showed a reasonable accuracy of 60–80{\%} estimation compared to the measured waveforms. Hypothesis: Temporal waveforms for SpO2 and PPG contain adequate information to estimate the temporal CBFv waveform using ANNs. Methodology: Concurrent measurements of SpO2 and PPG using pulse oximetry from the forehead and CBFv from the middle cerebral artery (MCA) using transcranial Doppler (TCD) were recorded from healthy, normotensive subjects performing a breath-hold challenge. The breath-hold challenge mimicked the cerebrovascular response to apnoea, and was recorded by measuring CBFv in MCA. Two protocols were used, each consisting of five breath-holding manoeuvres and differing in terms of the time between the five successive breath-holds. Using data from one protocol, several permutations of the temporal values of SpO2 and PPG signals were used as inputs to different ANN topologies, in order to train and find the optimum model. The optimum model was evaluated using the data from the other protocol as a blind dataset. Results: Using the first protocol for training, optimum ANN configurations were found to be different for each subject, and accuracy of 75–87{\%} was achieved. When these optimum ANN models were tested using the second protocol as a blind dataset, the accuracy achieved was around 60–80{\%}. Conclusions: A novel approach employing temporal records of SpO2 and PPG can be used to estimate the CBFv waveform using ANNs with acceptable accuracy. Increases in the size and diversity of the population dataset and the use of features extracted from SpO2 and PPG signals are needed for generalisation of the method and potential future clinical applications.",
keywords = "Biomedical signal processing, Blood oxygen saturation, Breath-hold challenge, Cerebral blood flow velocity, Machine learning, Obstructive sleep apnea, Photoplethysmography, Time series estimation",
author = "Al-Abed, {Mohammad A.} and Al-Bashir, {Areen K.} and Ayman Al-Rawashdeh and Alex, {Raichel M.} and Rong Zhang and Watenpaugh, {Donald E.} and Khosrow Behbehani",
year = "2019",
month = "12",
doi = "10.1016/j.compbiomed.2019.103508",
language = "English (US)",
volume = "115",
journal = "Computers in Biology and Medicine",
issn = "0010-4825",
publisher = "Elsevier Limited",

}

TY - JOUR

T1 - Estimation of cerebral blood flow velocity during breath-hold challenge using artificial neural networks

AU - Al-Abed, Mohammad A.

AU - Al-Bashir, Areen K.

AU - Al-Rawashdeh, Ayman

AU - Alex, Raichel M.

AU - Zhang, Rong

AU - Watenpaugh, Donald E.

AU - Behbehani, Khosrow

PY - 2019/12

Y1 - 2019/12

N2 - The effect of untreated Obstructive Sleep Apnoea (OSA) on cerebral haemodynamics and CA impairment is an active field of research interest. A breath-hold challenge is usually used in clinical and research settings to simulate cardiovascular and cerebrovascular changes that mimic OSA events. This work utilises temporal arterial oxygen saturation (SpO2) and photoplethysmography (PPG) signals to estimate the temporal cerebral blood flow velocity (CBFv) waveform. Measurements of CBFv, SpO2, and PPG, were acquired concurrently from volunteers performing two different protocols of breath-hold challenge in the supine position. Past values of the SpO2 and PPG signals were used to estimate the current values of CBFv using different permutations and topologies of supervised learning with shallow artificial neural networks (ANNs). The measurements from one protocol were used to train the ANNs and find the optimum topologies, which in turn were tested using the other protocol. Data collected from 10 normotensive, healthy subjects (four females, age 28.5 ± 6.1 years, Body Mass Index (BMI) 24.0 ± 4.7 kg/m2) were used in this study. The results show that different subjects have different optimum topologies for ANNs, thus indicating the effects of inter-subject variability on ANNs. Successfully reconstructed blind waveforms for the same subject group in the second protocol showed a reasonable accuracy of 60–80% estimation compared to the measured waveforms. Hypothesis: Temporal waveforms for SpO2 and PPG contain adequate information to estimate the temporal CBFv waveform using ANNs. Methodology: Concurrent measurements of SpO2 and PPG using pulse oximetry from the forehead and CBFv from the middle cerebral artery (MCA) using transcranial Doppler (TCD) were recorded from healthy, normotensive subjects performing a breath-hold challenge. The breath-hold challenge mimicked the cerebrovascular response to apnoea, and was recorded by measuring CBFv in MCA. Two protocols were used, each consisting of five breath-holding manoeuvres and differing in terms of the time between the five successive breath-holds. Using data from one protocol, several permutations of the temporal values of SpO2 and PPG signals were used as inputs to different ANN topologies, in order to train and find the optimum model. The optimum model was evaluated using the data from the other protocol as a blind dataset. Results: Using the first protocol for training, optimum ANN configurations were found to be different for each subject, and accuracy of 75–87% was achieved. When these optimum ANN models were tested using the second protocol as a blind dataset, the accuracy achieved was around 60–80%. Conclusions: A novel approach employing temporal records of SpO2 and PPG can be used to estimate the CBFv waveform using ANNs with acceptable accuracy. Increases in the size and diversity of the population dataset and the use of features extracted from SpO2 and PPG signals are needed for generalisation of the method and potential future clinical applications.

AB - The effect of untreated Obstructive Sleep Apnoea (OSA) on cerebral haemodynamics and CA impairment is an active field of research interest. A breath-hold challenge is usually used in clinical and research settings to simulate cardiovascular and cerebrovascular changes that mimic OSA events. This work utilises temporal arterial oxygen saturation (SpO2) and photoplethysmography (PPG) signals to estimate the temporal cerebral blood flow velocity (CBFv) waveform. Measurements of CBFv, SpO2, and PPG, were acquired concurrently from volunteers performing two different protocols of breath-hold challenge in the supine position. Past values of the SpO2 and PPG signals were used to estimate the current values of CBFv using different permutations and topologies of supervised learning with shallow artificial neural networks (ANNs). The measurements from one protocol were used to train the ANNs and find the optimum topologies, which in turn were tested using the other protocol. Data collected from 10 normotensive, healthy subjects (four females, age 28.5 ± 6.1 years, Body Mass Index (BMI) 24.0 ± 4.7 kg/m2) were used in this study. The results show that different subjects have different optimum topologies for ANNs, thus indicating the effects of inter-subject variability on ANNs. Successfully reconstructed blind waveforms for the same subject group in the second protocol showed a reasonable accuracy of 60–80% estimation compared to the measured waveforms. Hypothesis: Temporal waveforms for SpO2 and PPG contain adequate information to estimate the temporal CBFv waveform using ANNs. Methodology: Concurrent measurements of SpO2 and PPG using pulse oximetry from the forehead and CBFv from the middle cerebral artery (MCA) using transcranial Doppler (TCD) were recorded from healthy, normotensive subjects performing a breath-hold challenge. The breath-hold challenge mimicked the cerebrovascular response to apnoea, and was recorded by measuring CBFv in MCA. Two protocols were used, each consisting of five breath-holding manoeuvres and differing in terms of the time between the five successive breath-holds. Using data from one protocol, several permutations of the temporal values of SpO2 and PPG signals were used as inputs to different ANN topologies, in order to train and find the optimum model. The optimum model was evaluated using the data from the other protocol as a blind dataset. Results: Using the first protocol for training, optimum ANN configurations were found to be different for each subject, and accuracy of 75–87% was achieved. When these optimum ANN models were tested using the second protocol as a blind dataset, the accuracy achieved was around 60–80%. Conclusions: A novel approach employing temporal records of SpO2 and PPG can be used to estimate the CBFv waveform using ANNs with acceptable accuracy. Increases in the size and diversity of the population dataset and the use of features extracted from SpO2 and PPG signals are needed for generalisation of the method and potential future clinical applications.

KW - Biomedical signal processing

KW - Blood oxygen saturation

KW - Breath-hold challenge

KW - Cerebral blood flow velocity

KW - Machine learning

KW - Obstructive sleep apnea

KW - Photoplethysmography

KW - Time series estimation

UR - http://www.scopus.com/inward/record.url?scp=85073924309&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85073924309&partnerID=8YFLogxK

U2 - 10.1016/j.compbiomed.2019.103508

DO - 10.1016/j.compbiomed.2019.103508

M3 - Article

C2 - 31698237

AN - SCOPUS:85073924309

VL - 115

JO - Computers in Biology and Medicine

JF - Computers in Biology and Medicine

SN - 0010-4825

M1 - 103508

ER -