20 Sleep-Related Modulations of Heart Rate Variability…
319
20.5 Detection of Sleep Apnea Through Changes in Heart
Rate and ECG Morphology
In 2000 at the Computers in Cardiology Conference in Boston, USA, public competition took place as part of a congress held by engineers of biomedical technology
who are involved with ECG analysis. This competition involved solving a problem
encountered in ECG analysis, i.e., recognition of sleep apnea by study of nocturnal
ECGs [29]. ECGs from healthy test subjects, patients with moderate sleep apnea, and
patients with severe degree of sleep apnea were made available on a server of PHYSIONET [13]. A total of 35 nocturnal ECG recordings were provided for training
purposes, and 35 recordings were provided to competition participants for purposes
of analysis. Two teams succeeded in correctly classifying all patients and were even
able to correctly identify sleep apnea epochs in 92 and 94% of cases.
In addition to data on the cyclical variation in heart rate, these teams had also
analyzed the corresponding ECG plots [29]. This indeed revealed that respiration
modulates the amplitudes of the R- and T-waves of the ECG—a phenomenon that
was observed earlier but had not been assessed in the context of apnea detection [19].
The so-called electrocardiographically derived respiration signal (EDR)—i.e., data
on respiration as derived from an ECG—enables evaluation of the respiration plot
and, in turn, detection of apnea and hypopnea events. Although when employed alone,
EDR also demonstrates weaknesses in recognition of such events, in combination
with data on the cyclical variation in heart rate, it offers an astonishingly high level
of certainty in detection of sleep-related breathing disorders from ECGs.
Morphological alterations of the ECG during sleep apnea arise from the fact that
the influence of breathing on an ECG is mechanical in nature, and it is therefore
independent of factors that affect the eloquence of the cyclical variation in heart
rate. As a result, combining the evaluation of autonomic influences on the ECG
(i.e., heart rate) and assessment of the mechanical, respiration-mediated influences
on the ECG (i.e., EDR) enables good detection of sleep apnea events. If additional
physiological measures are included such as oxygen saturation, snoring, and body
movement during sleep, a polygraph system can be applied that can achieve a high
degree of sensitivity and specificity for detection of sleep-related breathing disorders
even without direct registration of respiration [9].
20.6 Cardiopulmonary Coupling and Cardiorespiratory
Phase Synchronization
The regulation of breathing and heartbeat is coupled [18], and cardiopulmonary
coupling (CPC) is studied mostly in the context of respiratory sinus arrhythmia
(RSA), which leads to variations in the heart rate and constitutes a portion of HRV.
Respiratory sinus arrhythmia describes the respiratory-gated fluctuation of the heart
rate: during inhalation, heart rate increases and during exhalation, it again subsides.
319
20.5 Detection of Sleep Apnea Through Changes in Heart
Rate and ECG Morphology
In 2000 at the Computers in Cardiology Conference in Boston, USA, public competition took place as part of a congress held by engineers of biomedical technology
who are involved with ECG analysis. This competition involved solving a problem
encountered in ECG analysis, i.e., recognition of sleep apnea by study of nocturnal
ECGs [29]. ECGs from healthy test subjects, patients with moderate sleep apnea, and
patients with severe degree of sleep apnea were made available on a server of PHYSIONET [13]. A total of 35 nocturnal ECG recordings were provided for training
purposes, and 35 recordings were provided to competition participants for purposes
of analysis. Two teams succeeded in correctly classifying all patients and were even
able to correctly identify sleep apnea epochs in 92 and 94% of cases.
In addition to data on the cyclical variation in heart rate, these teams had also
analyzed the corresponding ECG plots [29]. This indeed revealed that respiration
modulates the amplitudes of the R- and T-waves of the ECG—a phenomenon that
was observed earlier but had not been assessed in the context of apnea detection [19].
The so-called electrocardiographically derived respiration signal (EDR)—i.e., data
on respiration as derived from an ECG—enables evaluation of the respiration plot
and, in turn, detection of apnea and hypopnea events. Although when employed alone,
EDR also demonstrates weaknesses in recognition of such events, in combination
with data on the cyclical variation in heart rate, it offers an astonishingly high level
of certainty in detection of sleep-related breathing disorders from ECGs.
Morphological alterations of the ECG during sleep apnea arise from the fact that
the influence of breathing on an ECG is mechanical in nature, and it is therefore
independent of factors that affect the eloquence of the cyclical variation in heart
rate. As a result, combining the evaluation of autonomic influences on the ECG
(i.e., heart rate) and assessment of the mechanical, respiration-mediated influences
on the ECG (i.e., EDR) enables good detection of sleep apnea events. If additional
physiological measures are included such as oxygen saturation, snoring, and body
movement during sleep, a polygraph system can be applied that can achieve a high
degree of sensitivity and specificity for detection of sleep-related breathing disorders
even without direct registration of respiration [9].
20.6 Cardiopulmonary Coupling and Cardiorespiratory
Phase Synchronization
The regulation of breathing and heartbeat is coupled [18], and cardiopulmonary
coupling (CPC) is studied mostly in the context of respiratory sinus arrhythmia
(RSA), which leads to variations in the heart rate and constitutes a portion of HRV.
Respiratory sinus arrhythmia describes the respiratory-gated fluctuation of the heart
rate: during inhalation, heart rate increases and during exhalation, it again subsides.