Multiparameteric estimation of cardiorespiratory fitness in seismocardiography

1 . A method for quantifying cardiorespiratory fitness of a person, the method comprising:

(a) obtaining a seismocardiogram (SCG) recorded with an accelerometer placed on the chest of the person and configured to measure accelerations and vibrations of the chest wall of the person caused by myocardial movement;

(b) determining properties of a first signal feature in the SCG, wherein the first signal feature corresponds to an aortic valve closure of a heartbeat, wherein determining properties of the first signal feature comprises:

(b)(i) determining a frequency measure of the first signal feature; and

(b)(ii) determining a morphology measure of the first signal feature by (1) determining a first signal window of the SCG that covers the first signal feature, (2) providing a first machine learning model trained to determine the morphology measure based on a time-domain waveform representation of the corresponding first signal window, and (3) inputting the time-domain waveform representation of the first signal window to the first machine learning model, wherein (4) the morphology measure indicates the shape, contour and/or outline of the first signal feature, wherein the frequency measure and the morphology measure are properties of the first signal feature; and

(c) determining a measure indicating the cardiorespiratory fitness of the person by (5) providing a second machine learning model trained to determine the measure indicating cardiorespiratory fitness based at least on corresponding properties of the first signal feature, and (6) inputting at least the properties of the first signal feature to the second machine learning model.

2 . The method according to claim 1 , wherein determining the properties of the first signal feature further comprises:

(b)(iii) determining a plurality of diastolic segments of the SCG; and

(b)(iv) discarding noisy diastolic segments, wherein the non-discarded noisy diastolic segments comprise the first signal feature.

3 . The method according to claim 2 , wherein determining the properties of the first signal feature further comprises:

(b)(v) identifying one or more fiducial points of the first signal feature;

wherein the measure indicating cardiorespiratory fitness is further determined based on the one or more fiducial points;

wherein the fiducial points comprise at least one of a local maximum of the first signal feature, and a first local minimum immediately before the local maximum.

4 . The method according to claim 1 , wherein determining properties of the first signal feature further comprises:

(b)(iii) determining an amplitude difference between a local maximum of the first signal feature and a first local minimum immediately before the local maximum of the first signal feature;

wherein determining the measure indicating cardiorespiratory fitness is further based on the amplitude difference.

5 . The method according to claim 1 , wherein determining properties of the first signal feature further comprises:

(b)(iii) determining a first time separation between a local maximum of the first signal feature and a first local minimum immediately before the local maximum of the first signal feature; and

wherein determining the measure indicating cardiorespiratory fitness is based on the first time separation.

6 . The method according to claim 1 , wherein determining the morphology measure comprises:

determining an SCG window having a time domain, wherein the SCG covers the first signal feature;

wherein determining the morphology measure is based on the time domain of the SCG window.

7 . The method according to claim 6 , wherein the SCG window has a width in the range 200 ms to 500 ms.

8 . The method according to claim 6 , wherein determining the frequency measure comprises:

determining an SCG window having a frequency domain, wherein the SCG window covers the first signal feature; and

determining the frequency measure based on the frequency domain of the SCG window.

9 . The method according to claim 1 , further comprising:

(d) determining a heart rate variability;

wherein determining the measure indicating cardiorespiratory fitness is further based on the heart rate variability.

10 . The method according to claim 1 , further comprising:

(d) determining properties of a second signal feature in the SCG, wherein the second signal feature is different from the first signal feature;

wherein determining the measure indicating cardiorespiratory fitness is further based on the properties of the second signal feature; and

wherein the second signal feature corresponds to at least one of a mitral valve closure and an aortic valve opening of a heart cycle.

11 . The method according to claim 10 , wherein determining properties of the second signal feature further comprises:

determining a second time separation between a zero crossing of the mitral valve closure and a first local maximum of the aortic valve opening after the mitral valve closure; and

wherein determining the measure indicating cardiorespiratory fitness is further based on the second time separation.

12 . The method according to claim 1 , wherein determining the first signal window of the SCG that covers the first signal feature comprises identifying a fiducial point of the first signal feature and positioning the first signal window relative to the fiducial point.

13 . A system for quantifying or determining an indication of cardiorespiratory fitness of a person, the system comprising:

an accelerometer configured to be placed on the chest of the person for measuring accelerations and vibrations of the chest wall of the person caused by myocardial movement; and

a processor operatively connected to the accelerometer, wherein the processor is configured to perform any of the steps of the method according to any of claims 1 to 11 .

14 . A non-transient memory storing program code instructions for use in a system for quantifying or determining an indication of cardiorespiratory fitness, wherein the system comprises (A) an accelerometer configured to be placed on the chest of a person for measuring accelerations and vibrations of the chest wall of the person caused by myocardial movement, and (B) a processor operatively connected to the accelerometer, wherein the program code instructions are configured, when executed by the processor, to perform the method according to any of claims 1 to 11 .