Method and system for monitoring sleep

1. A method comprising:

recording, with one or more processors, electrocardiogram (ECG) data and photoplethysmogram (PPG) data;

deriving, with the one or more processors, a first parameter from the recorded ECG data independently of the PPG data, wherein the first parameter is (a) an RR interval or (b) a pulse arrival time;

extracting, with the one or more processors, a first feature from the first parameter, wherein the first feature is (a) a value produced by a detrended fluctuation analysis, (b) a time or frequency domain descriptor of variability, or (c) a multi-scale entropy;

deriving, with the one or more processors, a second parameter from the recorded PPG data independently of the ECG data, wherein the second parameter is (a) a pulse amplitude or (b) a respiratory rate;

extracting, with the one or more processors, a second feature from the second parameter, wherein the second feature is (a) a value produced by a detrended fluctuation analysis, (b) a time or frequency domain descriptor of variability, (c) a multi-scale entropy, or (d) a difference between an instantaneous respiratory frequency and a baseline respiratory frequency; and

providing, with the one or more processors, an output indicative of a sleep stage based upon the first and second features.

2. The method of claim 1 , wherein the ECG and PPG data are acquired with a recording apparatus wearable in ambulatory manner.

3. The method of claim 2 , wherein the recording apparatus comprises a Holter monitor and a pulse oximeter.

4. The method of claim 1 , wherein the first or second feature is a value produced by a detrended fluctuation analysis.

5. The method of claim 1 , wherein the second parameter is a respiratory rate, and wherein the second feature is a difference between an instantaneous respiratory frequency and a baseline respiratory frequency.

6. The method of claim 1 , wherein the first or second feature is a time or frequency domain descriptor of variability.

7. The method of claim 1 , wherein the first or second feature is a multi-scale entropy.

8. The method of claim 1 , wherein the output indicative of the sleep stage is provided based upon a classification of the first and second features.

9. The method of claim 8 , wherein the classification of the first and second features is obtained with a trained classifier.

10. The method of claim 9 , wherein the trained classifier is configured to produce a probability of an epoch being associated with wakefulness, REM sleep, or non-REM sleep.

11. The method of claim 10 , wherein the probability is adjusted based on a length of time for which the ECG and PPG data have been simultaneously recorded.

12. The method of claim 1 further comprising:

deriving, with the one or more processors, a third parameter from the recorded ECG data independently of the PPG data; and

extracting, with the one or more processors, a third feature from the third parameter,

wherein the output indicative of the sleep stage is provided based upon the first, second, and third features, and

wherein the first parameter is an RR interval, wherein the second parameter is a pulse amplitude or a respiratory rate, and wherein the third parameter is a pulse arrival time.

13. The method of claim 1 further comprising:

deriving, with the one or more processors, a third parameter from the recorded PPG data independently of the ECG data; and

extracting, with the one or more processors, a third feature from the third parameter,

wherein the output indicative of the sleep stage is provided based upon the first, second, and third features, and

wherein the first parameter is an RR interval or a pulse arrival time, wherein the second parameter is a pulse amplitude, and wherein the third parameter is a respiratory rate.

14. A system comprising:

a sensing apparatus configured to acquire electrocardiogram (ECG) data and photoplethysmogram (PPG) data; and

one or more processors configured to:

derive a first parameter from the acquired ECG data independently of the PPG data, wherein the first parameter is (a) an RR interval or (b) a pulse arrival time;

extract a first feature from the first parameter, wherein the first feature is (a) a value produced by a detrended fluctuation analysis, (b) a time or frequency domain descriptor of variability, or (c) a multi-scale entropy;

derive a second parameter from the acquired PPG data independently of the ECG data, wherein the second parameter is (a) a pulse amplitude or (b) a respiratory rate;

extract a second feature from the second parameter, wherein the second feature is (a) a value produced by a detrended fluctuation analysis, (b) a time or frequency domain descriptor of variability, (c) a multi-scale entropy, or (d) a difference between an instantaneous respiratory frequency and a baseline respiratory frequency; and

provide an output indicative of a sleep stage based upon the first and second features.

15. The system of claim 14 , wherein the sensing apparatus comprises a Holter monitor and a pulse oximeter.

16. The system of claim 14 , wherein the first or second feature is a value produced by a detrended fluctuation analysis.

17. The system of claim 14 , wherein the second parameter is a respiratory rate, and wherein the second feature is a difference between an instantaneous respiratory frequency and a baseline respiratory frequency.

18. The system of claim 14 , wherein the first or second feature is a time or frequency domain descriptor of variability.

19. The system of claim 14 , wherein the first or second feature is a multi-scale entropy.

20. The system of claim 14 , wherein the one or more processors are further configured to (a) derive a third parameter from the acquired ECG or PPG data and (b) extract a third feature from the third parameter,

wherein the third parameter is (a) an RR interval, (b) a pulse arrival time, (c) a pulse amplitude, or (d) a respiratory rate,

wherein the third feature is (a) a value produced by a detrended fluctuation analysis, (b) a time or frequency domain descriptor of variability, (c) a multi-scale entropy, (d) a difference between an instantaneous respiratory frequency and a baseline respiratory frequency, (e) a PPG spectral spread, or (f) a mean of a pulse amplitude, and

wherein the output indicative of the sleep stage is provided based upon the first, second, and third features.