WO2014101913A1 - Method and device for quantifying a respiratory sinus arrhythmia and use of said type of method or said type of device - Google Patents

Abstract

The invention relates to a method for quantifying the respiratory sinus arrhythmia, in order to minimize computing complexity. Said method is characterized in that a heart rate curve is divided into two different heart rate partial curves by at least two filters. At least one of said heart rate partial curves is tangent to the frequency band of the respiratory sinus arrhythmia and by analyzing both of the heart rate partial curves, the respiratory sinus arrhythmia is quantified.

Description

 Method and apparatus for quantifying a respiratory sinus arrhythmia Use of such a method or device

[01] The invention relates to a method and a device for quantifying a respiratory sinus arrhythmia and to the use of such a method or device.

[02] Heart rate variability, HRV for short or heart rate variability, is the term used to describe heart rate fluctuations from heartbeat to heartbeat. The HRV can be quantified by suitable mathematical methods, e.g. The heart rate variability is an expression of the continuous adjustment of the heart rate to changing requirements in the human organism and therefore allows conclusions on the neurovegetative regulatory ability of humans, in particular the function of the parasympathetic nervous system with its main nerve (vagus nerve) attributed a particular importance (Clin Sei (Lond.) 2012 Apr; 122 (7): 323-8 You may need the vagus nerve to understand pathophysiology and treat diseases; De Couck M, Mravec B, Gidron Y).

An insufficient function of the parasympathetic nervous system or a predominance of the sympathetic nervous system can be manifested in a low HRV. In severe cases, the picture of a heart rate stiffness is present. The association between low HRV and increased morbidity and mortality has been widely acknowledged in many studies and is now considered to be reliable (Int J Cardiol 2010 May 28; 141 (2): 122-31.) Epub 2009 Nov 11 The relationship of autonomic imbalance, heart rate variability and cardiovascular disease risk factors; Thayer JF, Yamamoto SS, Brosschot JF).

The determination of HRV by means of heart rate analysis at rest and under stimulation is considered today as the most important method of investigation in the diagnosis of the autonomic nervous system (Eur Heart J. 1996 Mar; 17 (3): 354-81 Heart rate variability. Physiological interpretation, and clinical use, Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology).

| Confirmation copy | [05] The influence of the parasympathetic nervous system is particularly evident in the size of the respiratory sinus arrhythmia (RSA). This refers to the oscillation of the heart rate in synchrony with the breathing, whereby the inhalation increases the heart rate and drops again when exhaling. Figures la and b show an example of a pronounced respiratory sinus arrhythmia. The respiratory sinus arrhythmia is mediated exclusively by the parasympathetic nervous system, the sympathetic has no part in it. This is due to the speed of the two control systems sympathetic and parasympathetic, whereby the parasympathetic nervous system can regulate much faster than the sympathetic nervous system. This property of the parasympathetic nervous system is used to differentiate between the activity of the sympathetic and the parasympathetic nervous system by subjecting the heart rate profile to spectral analysis, eg by means of a Fast Fourier Transformation (FFT) (see FIG. 1b).

FIG. 2a shows a typical heart rate curve in humans, FIG. 2b shows the associated spectral analysis. The area of the spectral analysis, designated as "HF" (high frequency), designates a control range of the heart rate, which is exclusively due to the influence of the parasympathetic nervous system (0.15 Hz to 0.4 Hz) .The "JLF" (low frequency) The area identified comprises a frequency range in which the sympathetic and parasympathetic nerves overlap (0.04 to 0.15 Hz). The frequency range referred to as "VLF" involves very slow acting influences such as thermoregulation Because of the required long recording time, the "VLF" range does not matter in the short term HRV analysis and may be neglected in the further consideration.

[07] It has long been the state of the art to quantify the different activities of the sympathetic and parasympathetic nervous system by parameters obtained by spectral analysis. For example, in many scientific studies the LF / HF ratio is used to indicate the balance of sympathetic and parasympathetic nervous system. The ratio of the frequency distribution in the LF range to the integral of the frequency curve in the HF range is set into proportion. It is assumed that the HF-range reflects only the parasympathetic nervous system and the LF-range approximately reflects the activity of the sympathetic nervous system (Eur Heart J. 1996 Mar; 17 (3): 354-81; Heart rate variability , Physiological Interpretation, and Clinical Use, Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology). [08] Another way to determine the parasympathetic activity is to quantify respiratory sinus arrhythmia. In the condition of respiratory sinus arrhythmia, the synchronization of the rhythms of respiration, blood pressure and heart rate leads to an approximately sinusoidal oscillation of the heart rate, which has a characteristic pattern in spectral analysis and is characterized by a large peak above the respiratory rate (FIG. 1b). Due to the optimal resonance frequency of the respiratory sinus arrhythmia of about 0.1 Hz, the peak of the respiratory sinus arrhythmia is almost always in the LF (low frequency) range of the spectral analysis, but is still an expression of parasympathetic activity and not to be confused with sympathetic activity by the absence of a characteristic peak and a broad distribution of frequencies occurring in the LF range is characterized (see Figure 2b).

[09] It is therefore possible to determine the amplitude of the respiratory sinus arrhythmia by spectral analysis. For example, DE 10 2006 039 957 A1 describes a spectral-analytical method for determining the respiratory sinus arrhythmia, which is based on a ratio of the integral component of two frequency ranges similar to the LF / HF quotient. EP 1 156 851 B1 or DE 600 32 581 T2, in turn, describes a spectral-analytical method for determining the respiratory sinus arrhythmia, in which the frequency of the peak is determined by means of a peak detection and subsequently an integral which comprises the peak, is set in relation to two integrals, each comprising an area below and above the peak frequency. A practical implementation of the methods mentioned in the two patents are, for example, devices for carrying out the so-called HRV biofeedback process. The respiratory sinus arrhythmia of the user is measured in real time and visualized by different colored LEDs. Through the so-called biofeedback effect, the user can now specifically train his respiratory sinus arrhythmia and thus the influence of the health-promoting parasympathetic nervous system. DE 10 2008 030 956 AI shows an embodiment of such a device.

The determination of the respiratory sinus arrhythmia by means of the spectral analysis of the prior art is due to the required mathematical methods such as a Fourier analysis, as disclosed for example in US 2007/0208266 AI, an FFT analysis, as in particular, for example EP 1 156 851 B1, WO 2008/028912 A2 of US Pat. No. 6,358,201, US Pat. No. 7,163,512 Bl, US Pat. No. 7,462,151 B2, US Pat. No. 8,066,637 B2 or US Pat US 8,123,696 B2 is disclosed, or a wavelet analysis before a high computing power, as it is quite common in the PC field today. Also, US 2006/0074333 AI discloses complex computational steps to infer from biometric parameters that are measured to the state of the nervous system. In the realization of simple battery powered

₅ devices for the determination of respiratory sinus arrhythmia but less compute-intensive analysis methods are desirable because then can be dispensed with the use of expensive microprocessors with high power consumption. Thus, for example, the standard deviation can be used as a statement for a variance of the corresponding signal with relatively little computation effort in the time domain (Proceedings of the 17th IEEE Symposium on Computer-Based Medical Systems (CBMS'04) 2004, pages 1 to 4; Raghavan V, Vikas Lath, Ashish Patil Sreejit Pillai), which in itself is not very meaningful.

[1 1] It is therefore an object of the present invention to avoid these disadvantages, with corresponding simplifications of course also for devices with larger computing power to better performance and the ability to entrust the corresponding devices with other i ₅ calculations and other tasks at the same time, leads.

[12] As a solution, methods and devices for quantifying the respiratory sinus arrhythmia with the features of the independent claims have been proposed. Further advantageous embodiments can be found in the subclaims and the following description.

[13] The LF / HF quotient mentioned in the prior art as well as the quotients mentioned in the patents DE 10 2006 039 957 A1 and EP 1 156 851 B1 ultimately correspond to a signal-to-noise ratio, the amplitude of the respiratory sinusoidal signal being used as signal component. Arrhythmia is used and are used as noise all other frequencies occurring outside the respiratory rate. The noise component is therefore usually

₂₅ , because with high activity of the parasympathetic nervous system, the noise is greatly reduced. This is also called coherence between heart rate and respiration. For this reason, the magnitude of the noise, in addition to the amplitude of the respiratory sinus arrhythmia, provides valuable information about parasympathetic activity. Due to the resonance characteristics of the physiological reflex arcs, the respiratory sinus arrhythmia is more pronounced in slow breathing with a maximum at about 0.1 Hz. The higher the respiratory rate, the lower its amplitude becomes. The determination of respiratory sinus arrhythmia is therefore usually performed at slower breathing. For this reason, it is usually sufficient to use the frequency components which occur above the respiratory frequency as the noise component. It is now possible to produce such a signal-to-noise ratio without spectral analysis in a simple manner:

mit Aßandpass als die Fläche unter der Bandpass-gefilterten Herzfrequenzkurve, wobei der Frequenzbereich, der den Filter möglichst ungedämpft passiert, den physiologischen Bereich der respiratorischen Sinusarrhythmie umfassen sollte. Mögliche Frequenzbereiche wären z.B. der aus dem Stand der Technik bekannte LF-Bereich 0,04 bis 0,15 Hz. Alternativ könnte auch ein Frequenzbereich gewählt werden, der eine physiologische Atemfrequenz umfasst, z.B. zwischen 5 und 10 Atemzügen pro Minute, was einem Frequenzbereich von 0,083 bis 0,166 Hz entspricht. A respiratory sinus arrhythmia corresponds to the occurrence of a sinusoidal oscillation of a specific frequency in the heart rate curve, the frequency of the sinusoidal oscillation coinciding with the respiratory rate. The amplitude of the sinusoidal oscillation corresponds to the area swept by the sinusoidal curve, see FIG. 3. If respiratory sinus arrhythmia overlaps with frequencies of other genesis (FIG. 4a), the respiratory sinus arrhythmia can be set by the use of a digital bandpass or lowpass filter Separate from the interfering frequency components (Figure 4b) and determine the area under the curve. By using a digital high-pass filter on the heart rate curve, for example, the noise component can now also be separated and likewise the area under the curve can be determined (FIG. 4c). The respiratory sinus arrhythmia can then be quantified as the ratio of the areas of the filtered curves:

with passband as the area under the bandpass-filtered heart rate curve, where the frequency range that passes through the filter as undamped as possible should include the physiological range of the respiratory sinus arrhythmia. Possible frequency ranges would be, for example, the LF range known from the prior art 0.04 to 0.15 Hz. Alternatively, a frequency range could be selected which includes a physiological respiratory rate, eg between 5 and 10 breaths per minute, which corresponds to a frequency range of 0.083 to 0.166 Hz.

and with A _H and _P ass as the area under the high pass filtered heart rate curve, where the frequency range that passes through the filter as undamped as possible should be above the frequency range of the bandpass filter. Possible frequency ranges would be For example, known from the prior art RF range> 0.15 Hz, or any other suitable frequency range, which is above the frequency range of the bandpass filter.

By suitable selection of the filter coefficients and frequency ranges, a signal-to-noise ratio can be achieved which quantifies the magnitude of the respiratory sinus arrhythmia with sufficient accuracy. Due to the simple realization of digital but also analogue filters by means of fewer additions and multiplications, a complex spectral analysis can now be dispensed with, as a result of which smaller, more energy-efficient microprocessors can be used in devices for the determination of respiratory sinus arrhythmia.

[17] It is understood that the filters act in the period or that the heart rate curve is recorded in the period and separated accordingly. Any complex transitions in the frequency domain, for example by a Fourier transform, can then be avoided. On the other hand, it is conceivable that the filters themselves possibly resort to the Fourier space in order to exercise their filter function, which, however, in view of the limited bandwidth in the bandwidths in which the filters then operate, leads to a reasonably low computational effort, since not the entire , occurring during the period heart rate curve must be transformed accordingly.

[18] The heart rate curve is preferably recorded as a heartbeat rate over time or as a number of heartbeats per unit of time over time.

Thus, in particular, a method for quantifying the respiratory sinus arrhythmia can be characterized in that a heart rate curve is separated by at least two filters in two different heart rate subcurves, wherein at least one of these heart rate sub-curves affects the frequency band of the respiratory sinus arrhythmia - or at least partially cuts or completely and by analyzing the two heart rate subcultures the respiratory sinus arrhythmia is quantified. Such a filter arrangement can be implemented on the one hand even by analog electronic modules and on the other hand by the simplest software engineering measures and allows a quick analysis of the heart rate curve. As a frequency band of respiratory sinus arrhythmia, a range between 0.06 heart and 0.166 heart, which corresponds to approximately 4 to 10 breaths per minute, is preferably selected or narrowed - the latter at the risk of breathing frequencies of individuals, such as athletes , no longer applicable. Accordingly, a device for quantifying the respiratory sinus arrhythmia is advantageous, which is characterized by measuring means for measuring a heart rate curve, by at least two filters for separating the heart rate curve into two different heart rate subcurves, by at least one analyzer for analyzing juxtaposition of the two heart rate subcurves and by at least one output device for Output of the result of the analyzer distinguished.

[21] Preferably, the other of the heart rate sub-curves does not include the respiratory sinus arrhythmia frequency band. In this way structurally very easy separation and thus good signal sharpness can be achieved.

[22] The latter also applies if the two heart rate subcurves do not overlap.

Preferably for analysis, the areas A and B are determined under the two heart rate sub-curves and set in relation to each other, wherein the ratio G quantifies the respiratory sinus arrhythmia.

[24] It is understood that, before the heart rate component curves are set in relation to each other, a weighting, for example via a factor c, d or an exponent p, q, may be possible. Such weighting may be readily experimentally chosen and, for example, tuned to output a binary signal G, such as "good" and "bad", as quantification, if the corresponding factors c, d and exponents p, q are chosen such that that a sufficiently significant transition is available at an output of a corresponding device. In particular, the heart rate curve can be separated by more than two filters into a corresponding number of different heart rate subcurves. Preferably, the heart rate sub-curves which affect the frequency band of the respiratory sinus arrhythmia are first evaluated and the rating A summarized and the heart rate sub-curves which do not contain or affect the frequency band of the respiratory sinus arrhythmia are evaluated and the rating evaluated B summarized. Subsequently, the summaries for the quantification G of the respiratory sinus arrhythmia can be analyzed.

[26] For the evaluations A, B, the areas under the two heart rate subcurves are preferably determined and added to the summary, wherein in particular the combined heart rate subcurves can be set in relation to one another for analysis. This ratio can then be used to quantify G respiratory sinus arrhythmia.

[27] An appropriate summary can be done, for example, by simple summation.

 Preferably, at least one of the scores, preferably in each case via a factor c, d, is weighted prior to combining the respective heart rate subcurves. As a result, the significance of the output signal can be specifically adapted to the respective requirements.

_ l _k .c _k A _k

Σ _? ά i B _\

[29] In particular, at least one of the summaries may be weighted prior to analysis. This can also be done via a factor or the like. Preferably, this is done via an exponent.

[30] It goes without saying that even more complex weighting functions can be used if necessary.

[31] The analysis device may have at least one integrator for determining the area under a heart rate sub-curve, whereby a corresponding area calculation can be carried out quickly and easily. Preferably, an integrator is provided per heart rate sub-curve.

[32] A corresponding analysis device can have at least one amplifier for amplifying one of the two heart rate subcurves, wherein in the present context the term "amplification" includes a multiplication by a factor which can certainly also take place with a value below 1, which leads to a corresponding reduction Likewise, the amplifier may have an exponential or logarithmic or a multiplying characteristic curve, Preferably an amplifier is provided for each heart rate component curve.

Cumulatively or alternatively, the analysis means may comprise at least one adder for adding heart rate subcurves modified by the analyzer, for example heart rate subcircuits modified by the enhancers or integrators. Also, the analyzing means may have at least one divider for setting heart rate part curves modified by the analyzing means, whereby a corresponding analysis can be performed quickly and precisely. Likewise, the analysis device can cumulatively or alternatively also comprise a multiplier, possibly a plurality of multipliers with different characteristic curves, in particular in order, for example, to weight individual or groups of high-frequency partial curves with a factor or exponent.

[34] In particular, one of the filters can also be a digital filter, so that the overall arrangement can also be implemented in digital form. In this regard, it is accordingly advantageous if the other components of the analysis device, such as the integrators, amplifiers, multipliers or adders, are also configured correspondingly digital. [35] The corresponding method or the device is particularly suitable for biofeedback by taking a measurement and evaluation in real time and displaying the result of the evaluation in real time. A corresponding display can be made for example by "good" or "bad" but also by a finer resolution.

[36] It is understood that the features of the solutions described above or in the claims can optionally also be combined in order to be able to implement the advantages in a cumulative manner.

[37] Further advantages, objects and characteristics of the present invention will be explained with reference to the following description of exemplary embodiments, which are also illustrated in particular in the appended drawing. In the drawing show:

 1 shows an example of a respiratory sinus arrhythmia with the individual recorded data (a) and the resulting spectral analysis (b);

 FIG. 2 shows a ten-minute recording of the heart rate curve (a) and the associated spectral analysis (b);

 FIG. 3 shows an exemplary heart rate variability only with consideration of

 Respiratory rate;

 FIG. 4 shows an exemplary heart rate variability taking into account further

 Recreate (a) using a filter (b) matched to the respiratory rate and the remaining parts (c); and

 Figure 5 shows an exemplary construction of the filter and analysis device.

According to the arrangement according to FIG. 5, a heart rate curve (FIG. 1 a) via an input 1 via three filters 11, 12, 13 on the one hand frequencies below the respiratory sinus arrhythmia via the filter 11, which very low frequencies (very low frequencies VLF ) as well as above the respiratory sinus arrhythmia via the filter 13, which filters high frequencies (high frequencies HF), from the frequency band of respiratory sinus arrhythmia between 0.06 heart and 0.166 heart (low frequencies LF) passing through the filter 12 into one third branch to be filtered out, separated.

[39] Integrators 21, 22, 23 are used to integrate the areas under the respective variations, as shown by way of example in FIGS. 4b and 4c and then weighted via suitable amplifiers 31, 32, 33. In the present embodiment, the amplifiers 31, 33 are attenuating amplifiers with a factor below one, while the amplifier 32, which amplifies the respiratory sinus arrhythmia, sets an over unity factor to the integrated values.

[40] Following this, the values correspondingly reduced via the amplifiers 31, 33 are added up in an adder 41. Optionally, further adders may be provided, in particular if a plurality of filters are also provided in the area of the respiratory sinus arrhythmia, the outputs of which are optionally integrated separately and weighted with different factors. After that, the accumulated values from the adder 41, or the amplified values from the amplifier 32, are supplied to an exponential weighting in the multipliers 51, 52, in order subsequently to divide them in a divisor 61 according to the above-mentioned formula. From this follows then a corresponding output signal at an output 2, which represents a measure of the respiratory sinus arrhythmia.

[41] As can readily be understood, all of the aforementioned operators are easily implemented by simple conventional electronic components but also by appropriate software. The latter is in particular an advantageous implementation of the present invention if the correspondingly measured values are available digitally and a corresponding digital heart rate curve has been determined from the digital data. The linear arrangement requires immediate results that can be achieved quickly and with little effort.

LIST OF REFERENCE NUMBERS

 1 input 10 31 amplifiers

2 output 32 amplifiers

11 VLF filter 33 Amplifier 12 LF filter 41 Adder

13 RF filter 51 multiplier

21 Integrator 15 52 Multiplier

22 Integrator 61 Divisor

23 integrator

Claims

claims: 1. A method for quantifying a respiratory sinus arrhythmia, characterized in that a heart rate curve is separated by at least two filters in a corresponding number of different heart rate subcurves, wherein at least a first of these heart rate subcurves affecting a respiratory sinus arrhythmia comprehensive frequency band or this at least partially cuts or completely covers and wherein respiratory sinus arrhythmia is quantified by analysis of the two heart rate subcurves. 2. The method according to claim 1, characterized in that a second of the heart rate sub-curves does not include the frequency band of the respiratory sinus arrhythmia. 3. The method according to claim 1 or 2, characterized in that the two heart rate subcurves do not overlap. 4. The method according to any one of claims 1 to 3, characterized in that the areas under the two heart rate sub-curves are determined and set in relation to each other, wherein the ratio quantifies the respiratory sinus arrhythmia. 5. The method according to any one of claims 1 to 4, characterized in that the heart rate curve is separated by more than two filters in a corresponding number of different heart rate subcurves, wherein initially the heart rate subcurves which affect the frequency band of the respiratory sinus arrhythmia, evaluated and summarized the evaluation and the heart rate sub-curves, which do not include or affect the frequency band of the respiratory sinus arrhythmia, are evaluated and the score summarized and then the summaries for the quantification of the respiratory sinus arrhythmia are analyzed. 6. The method according to claim 5, characterized in that prior to combining the respective heart rate subcurves at least one of the ratings, preferably by a factor, is weighted. 7. The method according to claim 5 or 6, characterized in that at least one of the summaries before the analysis, preferably via an exponent weighted. 8. The method according to any one of claims 5 to 7, characterized in that for the evaluation, the areas under the heart rate sub-curves are determined and added to the summary. 9. The method according to claim 8, characterized in that the summarized heart rate sub-curves are set in relation to each other for analysis, wherein the ratio quantifies the respiratory sinus arrhythmia. 10. A device for quantifying a respiratory sinus arrhythmia, characterized by measuring means for measuring a heart rate curve, by at least two filters for separating the heart rate curve in two different heart rate subcurves, by at least one analyzer for analyzing juxtaposition of the two heart rate subcurves and by at least one output device for outputting the result of analysis device. 11. The device according to claim 10, characterized in that the analysis device comprises at least one amplifier for amplifying one of the two heart rate subcurves, preferably per heart rate subcircuit an amplifier. 12. The apparatus of claim 10 or 1 1, characterized in that the analysis device curve at least one integrator for determining the area under one of the heart rate subcircuits, preferably per heart rate subculture an integrator for determining the area under the respective heart rate. 13. Device according to one of claims 10 to 12, characterized in that the analysis device at least one adder for the addition of modified by the analysis device heart rate subcircuits and / or at least one Divider for setting by the analyzer modified heart rate sub-curves in a ratio. Method or device according to one of the preceding claims, characterized in that at least one of the filters is a digital filter. 15. Use of the method or the device according to one of the preceding claims for biofeedback by making a real-time measurement and evaluation and the result of the evaluation is displayed in real time.