RU2587926C2 - Method of determining characteristics of circadian rhythm of subject - Google Patents

Abstract

FIELD: medicine.

SUBSTANCE: invention refers to medicine, namely cardiovascular diagnostics. Readings are determined of first input signal showing cardiac function of individual for period of time T to obtain, as result, row of values x′_n of first input signal. Determined samples of second input signal showing activity of individual for period of time, overlapping period T of time to obtain, as result, row of values x_n of second input signal. Then series of values y_n is calculated of output signal from values x_n of first input signal and values x_n of second input signal according to model of autoregressive moving average (ARMAX). Values x′_n of first input signal and values x_n of the second input signal are used as exogenous input data. Value y_n of output signal for any n is calculated by linear combination of values x_n of first input signal and values x_n of second input signal. Periodic output signal is removed from calculated row of values y_n of output signal by harmonic regression analysis. Vibration parameter is then determined as characteristics of circadian rhythm.

EFFECT: method enables providing reliable assessment of parameters of oscillations of central driver circadian rhythm through analysis of cardiac function of subject, eliminating masking effects of endogenous circadian rhythm.

9 cl, 1 dwg

Description

FIELD OF THE INVENTION

The invention relates to a method for determining a circadian rhythm characteristic of a subject.

BACKGROUND

Paper Scale-invariant Aspects of Cardiac Dynamics Across Sleep Stages and Circadian Phases, Plamen Ch. Invanov, 2006, describes the measurement of an input signal showing the cardiac function of a subject and the characterization of the resulting periodic output.

Document WO-A-98/46128 describes a measurement of a first input signal showing the cardiac function of a subject and a second input signal showing the activity of the subject to compare both signals by the nature of the change in order to suppress the effect of noise.

Document WO-A-2007/143535 contains a description of a method for controlling physiological features, for example, cardiac function and motor function, and combining these features in a mathematical model.

The fact of the presence of diurnal changes in the cardiovascular and metabolic functions is well known. However, it has only recently been shown that cardiovascular and metabolic processes are influenced not only by the behavioral sleep / wake cycle, but, in part, are directly regulated by the central circadian rhythm driver. This relationship between the central circadian rhythm driver and cardiac output, suggests that the latter can be used to reliably derive oscillation parameters of the former, for example, amplitude and acrophase, i.e. time maximum oscillations. These parameters have important applications for improving sleep and increasing the physical and mental performance of a person.

Knowledge of the phase of the human circadian cycle helps to optimize the time of small impacts that allow you to move the circadian rhythm. Such a shift in circadian rhythms is useful for treating circadian rhythm disturbances, facilitating the adaptation of shift workers to night work schedules, and bringing people with sleep disorders such as sleep phase delay syndrome or sleep phase advance syndrome to a more normal time chart of their sleep patterns / wakefulness. Optimization of the human circadian phase allows you to adjust the human biorhythm so that maximum care and performance take place at the right time. From solving this problem, one can expect improvements for everyday life, increased labor productivity, sports results, etc.

The main advantage of using cardiac output as a means of assessing the dynamic characteristics of the central circadian rhythm driver is that cardiac output can be reliably measured using sensors that do not take up much space. The data that these sensors generate can be used to collect signals that contain small endogenous circadian changes. Given the small amplitude of the mentioned changes, they are often masked by the influence of competing processes, for example, the influence of physical or psychological stress on the activity of the heart. Many of these masking effects create disturbances that are easy to filter out, since these effects introduce frequency components outside the normal circadian range. However, other masking effects are also circadian in nature and need more sophisticated filtering methods. Recent studies have suggested that circadian modulation of heart activity is influenced by previous periods of wakefulness. The sleep / wake cycle follows a circadian cycle that affects the activity of the heart in a frequency range similar to the frequency range of the central circadian rhythm driver.

SUMMARY OF THE INVENTION

In connection with the foregoing objective of the present invention is to provide a method for determining the characteristics of the circadian rhythm of the subject, which allows to obtain an estimate of the vibration parameters of the central driver of the circadian rhythm by analyzing the cardiac function of the subject, eliminating the masking effects of endogenous circadian rhythm present in the cardiac function.

This goal is achieved using a method containing the characteristics of claim 1 of the claims.

In accordance with the aforementioned method, the first input signal, which is associated with the cardiac function of the subject, is measured over a period of time. In addition, a second input signal is measured, which shows the physiological activity of the subject, over a period of time superimposed on said first period of time during which the first input signal is measured.

The first input signal and the second input signal are combined to produce a periodic output signal representing the circadian rhythm of the subject. From this output signal, the required circadian rhythm characteristic is derived.

The concept of combining the first input signal and the second input signal is that the latter contains information about effects that mask information about the endogenous circadian rhythm that is present in the first signal. Combination eliminates undesirable characteristics from the first input signal (representing the cardiac function), so that the oscillation parameters describing the dynamic characteristics of the circadian rhythm can be derived directly from the periodic output signal.

The combination of the first input signal and the second input signal can be performed by methods of autoregression, bleaching, analysis of independent components or (non-linear) analysis of the main components, which leads to an improved representation of the circadian rhythm with a periodic output signal. Then, the aforementioned more informative signal can be used to derive an estimate of the oscillation parameters, for example, the application of the harmonic regression method.

While the first input signal can be any sign of cardiac function, the second signal can be represented by physiological, behavioral data from the subject, which provide an indication of the previous period of wakefulness of the subject, for example, data recorded on the wrist activity, sleep logs, answers to questions from the questionnaire , skin temperature, light intensity, etc., for a period of time that is aligned with a period of time during which cardiac function is measured.

In a preferred embodiment, the circadian rhythm characteristic to be determined is an acrophase of a periodic output.

In accordance with another preferred embodiment of the method in accordance with the present invention, said characteristic is the amplitude of the output periodic signal.

In another preferred embodiment, the first input signal and the second input signal are sampled to obtain a series of values of the first input signal and the values of the second input signal, respectively, and a series of output signal values are calculated from the sampled values of the first input signal and the values of the second input signal.

In this embodiment, the first and second input signals are discrete values of the input signals obtained by sampling the parameters of the cardiac function and monitoring the activity of the subject, respectively. This approach results in a first set of values of the first input signal and a second set of values of the second input signal. From these sets, one can derive a set of values of the output signal, for example, by solving a system of equations that contains the values of the first and second input signals as input variables to obtain the values of the output signal as solutions.

In a preferred embodiment, the periodic output signal is derived from the calculated series of output signal values by regression analysis.

According to another preferred embodiment of the present invention, a series of output signal values are calculated from the values of the first input signal and the values of the second input signal in accordance with the autoregressive moving average (ARMAX) model, while the values of the first input signal and the values of the second input signal serve as exogenous input data .

Autoregressive moving average models are widely known in the field of signal processing in the processing of autocorrelated time series data. The model consists of two parts, namely the autoregressive part and the moving average part. The mathematical basis of the model is a system of linear equations that specify the relationship between the two time series represented in this case by the values of the first input signal and the values of the second input signal. The solution of the aforementioned system of linear equations results in a sequence of values of the output signal representing the circadian rhythm of the subject.

In a preferred embodiment, the values of the first input signal and the values of the second input signal are obtained by sampling with a frequency above 1 Hz.

According to another preferred embodiment of the present invention, the values of the first input signal or the values of the second input signal are obtained by sampling from the values of the original signal with a first frequency, wherein said values of the initial signal are obtained by sampling with a second frequency higher than the first frequency.

In practice, the foregoing means that one of the time series represented by the values of the first input signal or the values of the second input signal is obtained by sampling from the values of the original (unprocessed) signal, which, in turn, are obtained by sampling the measured parameter with a relatively high frequency. In only one example, the values of the original signal are the initial values obtained by sampling the parameter of cardiac activity with a frequency of 512 Hz. The mentioned values of the initial signal cannot be calculated directly together with the second time series of the values of the second input signal, which are obtained by sampling at a frequency of 60 Hz. Therefore, samples of the values of the original signal are selected with a relatively low frequency of 60 Hz to obtain the values of the first input signal, which can then be combined with the values of the second input signal, for example, using the aforementioned ARMAX model.

In a preferred embodiment, the first input signal is the interval between heartbeats (IBI) of the subject.

According to another preferred embodiment, the values of the initial ECG signal (electrocardiogram) are obtained by sampling the original ECG signal with a second frequency of 512 Hz, the IBI signal (interval between heartbeats) is derived from the obtained ECG signal values, and the IBI signal values representing the values of the first input signal, obtained by sampling the output IBI signal with a first frequency of 60 Hz.

In another preferred embodiment of the present invention, the second input signal is a wrist activity signal of a subject.

According to another preferred embodiment, the second input signal is a signal indicating the intensity of the light acting on the subject.

BRIEF DESCRIPTION OF THE DRAWINGS

The above and other aspects of the invention are clearly illustrated further on the example of the following embodiments.

The only figure in the drawings is a flowchart for illustrating one embodiment of a method for determining a circadian rhythm characteristic of a subject in accordance with the present invention.

DETAILED DESCRIPTION OF EMBODIMENTS

In the method represented by the flowchart of the figure, physiological signals are removed from the subject 10. ECG signals are obtained by a suitable sensor built into the wearable ECG device in the form of a strapping strap or similar device. The values of the original ECG signal are obtained by sampling at a frequency of 512 Hz (step 12). Samples are taken over a predetermined time period T to obtain a set of values of the original ECG signal. Second additional physiological data is collected from subject 10 through another suitable sensor. Said second data shows the physiological activity of the subject 10 over a period of time overlapping with the aforementioned period T of the sampling time of the original ECG signal. Said sample of the signal associated with the activity is performed at a reduced frequency of 60 Hz (step 14). A suitable sensor can be integrated into a bracelet or similar device for measuring an activity signal on the wrist of a subject 10. It should be noted that the second physiological data that indicate the activity of the subject 10 may also be other data based on sleep logs, skin temperature, light intensities, etc. It is assumed that the data contain information about the previous wakefulness of the subject 10. It will be shown hereinafter that the data can be combined with data related to the cardiac function of the subject 10 to exclude masking effects from the signals of the cardiac function, resulting in a signal a function representing the circadian activity of subject 10.

An IBI signal (interval between heartbeats) is derived from the original ECG signals by calculating the intervals of the R-R waves. At step 16, samples at a frequency of 60 Hz are additionally taken from the received IBI signal. The result of step 16 is a set of values of the first input signal, depending on the interval between heartbeats (IBI) of the subject 10 in the interval T time. On the other hand, the result of step 14 is a set of the same number of values of the second input signal showing the activity of the subject 10 over a period of time superimposed on a period T of time. Trends are removed from the first and second signals to exclude components of zero frequencies.

In the next step 18, the values of the first input signal and the values of the second input signal are mathematically combined using the autoregressive moving average (ARMAX) model, in which the values of the first input signal and the values of the second input signal represent exogenous input data. Let x _n represent the values of the second input signal and x ' _n represent the values of the first input signal, then in the framework of the ARMAX model it is possible to calculate a sequence of y _n values of the output signal that represent a linear function of x _n and x' _n . This function y _n is an extended representation of the circadian rhythm containing information about the acrophase of the circadian rhythm. The value of y _n for any n can be calculated from x _n , x ' _n through a linear combination of the preceding values.

Upon closer examination, the values x _n and x ′ _{n of the} first and second input signals can be combined to obtain an extended representation of y _n circadian rhythm through the following linear model:

(one)

where n is a non-negative integer, q means the backward shift operator, i.e. qy _n = y _n-1 , w _n means a Gaussian white noise process with an average value of 0 and a standard deviation of σ, and A (q), B (q) and C (q) are polynomials defined as follows:

A (q) = 1 + a ₁ q + a ₂ q ² + ... + a _r q ^r (2)

(3) C (q) = 1 + c ₁ q + c ₂ q ² + ... + c _u q ^u (four)

In the above expressions, the values of r, [s ₁ s ₂ ] ^T , u, a _i b _ij , c _i and σ are given model constants. Then the expanded representation of y _n rhythm is subjected to regression cosinor analysis (step 20) of the form:

μ + αcos (2πnΔt + Φ) (5)

Regression of the coefficient ϕ is the required evaluation function of the circadian rhythm.

Coefficient ϕ represents acrophase, i.e. the maximum oscillation time of the circadian rhythm function, which can be derived from the structure of the above equation (5) (step 22). After deriving the coefficient ϕ, it can be used to control the time of exposure to light, which allows you to shift the circadian rhythm of the subject 10 and helps to optimize the time of circadian activity. For example, closer to the time at which the internal body temperature is minimal (approximately two hours before spontaneous awakening on a non-working day), exposure to light gives a greater effect, and the applied exposure durations may be shorter.

It should be noted that the coefficient ϕ representing the acrophase of the circadian rhythm function is just one example of the characteristic of the circadian rhythm represented by the periodic output signal x _n . For example, the amplitude of a given signal may be another characteristic that can be used to control the exposure to light in order to optimize the circadian phase of subject 10. In general, another cardiac function can be used as the first input signal, except for the interval between heartbeats (IBI) of subject 10, and to represent the activity of subject 10, you can use another second input signal that is different from the signal recorded on the wrist of activity, as mentioned above.

The above invention is presented in the drawings and described in detail in the above description, however, the above image and description should be considered illustrative and approximate, and not limiting; the present invention is not limited to these embodiments. In the process of practical implementation of the claimed invention by specialists in this field of technology, after studying the drawings, description and appended claims, other changes to the above-described embodiments can be created and made. In the claims, the phrase “comprising” does not exclude other elements or steps, and the singular does not exclude a plural cycle. The obvious circumstance that some features are mentioned in mutually different dependent claims does not mean that, in appropriate cases, a combination of these features cannot be used. No position in the claims can be interpreted in the sense of limiting the scope of the invention.

Claims (9)

1. The method of determining the characteristics of the circadian rhythm of the subject (10), the method comprises the steps of:
- take samples of the first input signal showing the cardiac function of the subject (10) during a period T of time, to obtain, as a result, a series of values x ′ _{n of the} first input signal;
- take samples of at least one second input signal showing the activity of the subject (10) over a period of time superimposed on a period T of time to obtain, as a result, a series of values x _{n of the} second input signal;
- calculate a series of values y _{n of the} output signal from the values x _{n of the} first input signal and the values x _{n of the} second input signal in accordance with the autoregressive moving average model (ARMAX), while the x ′ _n values of the first input signal and the values x _{n of the} second input signal exogenous input, the value y _{n of the} output signal for any n being calculated by linearly combining the values x _{n of the} first input signal and the values x _{n of the} second input signal;
- deriving a periodic output signal from the calculated series of values of y _{n the} output signal by the method of harmonic regression analysis;
- determine the oscillation parameter as a characteristic of said periodic output signal. 2. A method according to Claim. 1, wherein said characteristic is acrophase periodic output signal y _n. 3. The method of claim 1 or 2, wherein said characteristic is the amplitude of the periodic output signal y _n . 4. The method according to one of paragraphs. 1-3, wherein the value of x _'n first input signal x _n and values of the second input signal obtained by sampling with a frequency above 1 Hz. 5. The method according to one of paragraphs. 1-4, in which the values x ′ _{n of the} first input signal or the values x _{n of the} second input signal are obtained by sampling from the values of the original signal with a first frequency,
wherein said initial signal values are obtained by sampling at a second frequency higher than the first frequency. 6. The method according to one of the preceding paragraphs, in which the first input signal is the interval between heartbeats (IBI) of the subject (10). 7. The method according to p. 6 in connection with p. 5, in which
the values of the original ECG signal (electrocardiogram) are obtained by sampling the original ECG signal with a second frequency of 512 Hz,
the IBI signal is derived from the obtained ECG signal values,
and IBI signal values representing the values of the first input signal are obtained by sampling the output IBI signal with a first frequency of 60 Hz. 8. The method according to p. 4, in which the second input signal is a signal of activity of the subject (10), recorded on the wrist. 9. The method according to one of paragraphs. 1-7, in which the second input signal is a signal showing the intensity of the light acting on the subject (10).

Images