A Signal Segmentation‐Free Model for Electrocardiogram‐Based Obstructive Sleep Apnea Severity Classification

Abstract

Obstructive sleep apnea (OSA) has been a common sleep disorder for years, and polysomnography (PSG) remains the gold standard for diagnosing OSA. Nevertheless, PSG is a time and money consuming test, and patients have to wait long for arranging a PSG test in a hospital. In light of this, portable and wearable tools for OSA classification have been developed recently as a low‐cost and easy‐to‐use screening method before undergoing PSG. Using unsegmented electrocardiogram (ECG) signals, a deep neural network (DNN)‐based model is developed here to categorize OSA severity with the following features. First, the model takes unsegmented ECG signals recorded overnight as input, and then generates a four‐level scale as output. Since all the input ECG signals are unsegmented, the tremendous amount of effort spent on signal annotation can be fully saved. Second, the largest amount of data is used to test the model and consequently provide a high generalization ability, as compared with others in the literature. The overall outperformance of this work is highlighted at the end of this article, and this work is validated as an easy‐to‐use and effective screening tool for OSA accordingly.

Predicting Apnea‐Hypopnea Index in Patients with Obstructive Sleep Apnea Using Unsegmented ECG‐Signal‐Based Algorithms

“…In this manner, this work can provide advantages to users and doctors regarding more details of OSA assessment. As it turned out, a correlation coefficient of 0.8343, and an overall accuracy of 55.96% for fourlevel severity classification were given in the presented model, which is comparable with [5] and is superior to other counterparts in the literature.…”

“…Finally, this work gives an overall accuracy of 55.96%, which is comparable with Ref. [5] of 57.55% and is superior to Ref. [4] of 43.37%.…”

“…For comparison purposes, the presented model was trained and tested using the same data sets as that in Ref. [5], as listed in Table II. It comprises three public data sets: the SHHS data set, the MrOS data set and the MESA data set, all of which are provided by the National Sleep Research Resource (NSRR) [8].…”

“…However, AHI values and the correlation coefficient cannot be calculated from the approach of Ref. [5]. In the second experiment, all the AHI values are converted into a four-level severity classification as mentioned above, and then a 4 × 4 confusion matrix and its performance metrics can be acquired, as shown in Fig.…”

Predicting Apnea‐Hypopnea Index in Patients with Obstructive Sleep Apnea Using Unsegmented ECG‐Signal‐Based Algorithms

“…In this manner, this work can provide advantages to users and doctors regarding more details of OSA assessment. As it turned out, a correlation coefficient of 0.8343, and an overall accuracy of 55.96% for fourlevel severity classification were given in the presented model, which is comparable with [5] and is superior to other counterparts in the literature.…”

“…Finally, this work gives an overall accuracy of 55.96%, which is comparable with Ref. [5] of 57.55% and is superior to Ref. [4] of 43.37%.…”

“…For comparison purposes, the presented model was trained and tested using the same data sets as that in Ref. [5], as listed in Table II. It comprises three public data sets: the SHHS data set, the MrOS data set and the MESA data set, all of which are provided by the National Sleep Research Resource (NSRR) [8].…”

“…However, AHI values and the correlation coefficient cannot be calculated from the approach of Ref. [5]. In the second experiment, all the AHI values are converted into a four-level severity classification as mentioned above, and then a 4 × 4 confusion matrix and its performance metrics can be acquired, as shown in Fig.…”

Feature Extraction with Multiple Receptive Fields Conducted on ECG Signals for Performance Enhancement in OSA Severity Classification

“…Take a further observation, compared to Ref. [1], the major contribution of the overall accuracy improvement in this work is that the sensitivity of the normal category in the proposed model has been considerably raised by above 14%, however, at the cost of a decrease in the sensitivity of the mild category.…”

Predicting Apnea‐Hypopnea Index in Patients with Obstructive Sleep Apnea Using Unsegmented ECG‐Signal‐Based Algorithms

“…In this manner, this work can provide advantages to users and doctors regarding more details of OSA assessment. As it turned out, a correlation coefficient of 0.8343, and an overall accuracy of 55.96% for fourlevel severity classification were given in the presented model, which is comparable with [5] and is superior to other counterparts in the literature.…”

“…Finally, this work gives an overall accuracy of 55.96%, which is comparable with Ref. [5] of 57.55% and is superior to Ref. [4] of 43.37%.…”

“…For comparison purposes, the presented model was trained and tested using the same data sets as that in Ref. [5], as listed in Table II. It comprises three public data sets: the SHHS data set, the MrOS data set and the MESA data set, all of which are provided by the National Sleep Research Resource (NSRR) [8].…”

“…However, AHI values and the correlation coefficient cannot be calculated from the approach of Ref. [5]. In the second experiment, all the AHI values are converted into a four-level severity classification as mentioned above, and then a 4 × 4 confusion matrix and its performance metrics can be acquired, as shown in Fig.…”

Feature Extraction with Multiple Receptive Fields Conducted on ECG Signals for Performance Enhancement in OSA Severity Classification

“…Take a further observation, compared to Ref. [1], the major contribution of the overall accuracy improvement in this work is that the sensitivity of the normal category in the proposed model has been considerably raised by above 14%, however, at the cost of a decrease in the sensitivity of the mild category.…”

Predicting Apnea‐Hypopnea Index in Patients with Obstructive Sleep Apnea Using Unsegmented ECG‐Signal‐Based Algorithms

“…In this manner, this work can provide advantages to users and doctors regarding more details of OSA assessment. As it turned out, a correlation coefficient of 0.8343, and an overall accuracy of 55.96% for fourlevel severity classification were given in the presented model, which is comparable with [5] and is superior to other counterparts in the literature.…”

“…Finally, this work gives an overall accuracy of 55.96%, which is comparable with Ref. [5] of 57.55% and is superior to Ref. [4] of 43.37%.…”

“…For comparison purposes, the presented model was trained and tested using the same data sets as that in Ref. [5], as listed in Table II. It comprises three public data sets: the SHHS data set, the MrOS data set and the MESA data set, all of which are provided by the National Sleep Research Resource (NSRR) [8].…”

“…However, AHI values and the correlation coefficient cannot be calculated from the approach of Ref. [5]. In the second experiment, all the AHI values are converted into a four-level severity classification as mentioned above, and then a 4 × 4 confusion matrix and its performance metrics can be acquired, as shown in Fig.…”

Feature Extraction with Multiple Receptive Fields Conducted on ECG Signals for Performance Enhancement in OSA Severity Classification

“…Take a further observation, compared to Ref. [1], the major contribution of the overall accuracy improvement in this work is that the sensitivity of the normal category in the proposed model has been considerably raised by above 14%, however, at the cost of a decrease in the sensitivity of the mild category.…”