Device for determining breathing rate of subject

Abstract

The invention relates to a device for determining the breathing rate of a subject. The pressure signal providing unit provides a measured pressure signal of the subject over time, where the measured pressure signal is indicative of a blood pulsation and comprises a plurality of pressure pulses, and the processor determines a respective dicrotic incisura location for a respective one of the plurality of pressure pulses (9), and determining the respiration rate based on the dicrotic incisura locations determined for the plurality of pressure pulses (9). This allows a very accurate determination of the respiratory rate.

Description

Technical Field

[0001] The present invention relates to an apparatus, a method and a computer program for determining the breathing rate of a subject. Background Art

[0002] US 2013 / 0079657 A1 discloses a method for determining respiratory information. A photoplethysmography (PPG) signal is transformed based at least in part on a continuous wavelet transform to generate a wavelet energy spectrum, wherein a threshold energy level is calculated based at least in part on the wavelet energy spectrum. For a particular time, a scale is identified in the wavelet energy spectrum at an energy level that is greater than the threshold but less than a maximum energy level of the wavelet energy spectrum, and an estimate of respiratory rate is determined based at least in part on the scale.

[0003] WO 2011 / 098763 discloses an apparatus for determining a subject's respiratory rate. The apparatus includes an input for receiving an arterial pressure waveform and means for defining a starting point within the waveform at the beginning of the systolic arterial pressure increase during the ventricular systolic phase of a heartbeat. The apparatus also includes means for determining an end point of the heartbeat by identifying a portion or point of the waveform having an arterial pressure value substantially identical to the starting point, and, if the arterial pressure value at the starting point is not retrievable, extrapolating the waveform to a point having an arterial pressure value substantially identical to the starting point. The apparatus also includes means for calculating stroke volume from the waveform or the extrapolated waveform between the starting and ending points, and for determining the respiratory rate based on cyclical variations in the stroke volume. Summary of the Invention

[0004] It is an object of the present invention to provide a device, a method and a computer program which allow an improved determination of the breathing rate of a subject.

[0005] In a first aspect of the present invention, an apparatus for determining a breathing rate of a subject is proposed, wherein the apparatus comprises:

[0006] a pressure signal providing unit configured to provide a measured pressure signal of the subject over time, wherein the measured pressure signal is indicative of blood pulsation and comprises a plurality of pressure pulses,

[0007] - a processor configured to determine a respective dicrotic notch location for a respective pressure pulse of the plurality of pressure pulses, and to determine the respiratory rate based on the dicrotic notch locations determined for the plurality of pressure pulses.

[0008] Determining the respiratory rate based on the dicrotic notch position allows for very accurate determination of the respiratory rate. The processor can determine the respiratory rate directly or indirectly based on the dicrotic notch position. In the latter case, the dicrotic notch position is used to calculate another parameter related to the corresponding pressure pulse, and the respiratory rate is determined based on this other parameter. This will be further exemplarily described below.

[0009] Furthermore, the processor can determine the respective dicrotic notch position directly based on the respective measured pressure pulse or based on the respective processed measured pressure pulse. If the dicrotic notch position is determined based on the processed measured pressure pulse, the processing is preferably performed by the processor. However, it can also be performed at least in part by the pressure signal providing unit. The processor can be configured to determine, for a respective pressure pulse from the plurality of pressure pulses, the respective dicrotic notch position relative to the respective pressure pulse, in particular along the respective pressure pulse.

[0010] The measured pressure pulse can be, for example, a non-invasively measured pressure pulse or an invasively measured pressure pulse.The measured pressure pulse is an arterial pressure pulse measured for an artery.

[0011] The pressure signal providing unit can be a receiving unit configured to receive a measured pressure signal from, for example, a measuring device and provide the received pressure signal. However, the pressure signal providing unit can also be a storage device that stores previously measured pressure signals and can retrieve the pressure signal from the storage device. The pressure signal providing unit can also be or include a measuring device that measures the pressure signal.

[0012] In one example, the pressure signal providing unit is configured to provide a pressure signal measured using a measuring device as the pressure signal, the measuring device comprising: a) a housing configured to surround a portion of a subject through which blood flows; b) a pressure applicator configured to apply pressure to the housing and thereby to the enclosed portion of the subject; and c) a pressure sensor configured to measure the pressure signal on the skin of the enclosed portion of the subject, wherein the pressure applicator increases or decreases the applied pressure while measuring the pressure signal. Thus, the pressure signal can be a non-invasively measured pressure signal, while allowing for high-quality determination of the respiratory rate. However, as described above, the pressure signal providing unit can also be configured to provide an invasively measured pressure signal.

[0013] In examples where the measuring device includes a housing, a pressure applicator, and a pressure sensor, the pressure sensor can be arranged inside the housing. However, the pressure sensor can also be arranged in another manner to measure the pressure on the skin of the enclosed portion of the subject. For example, a fluid-filled pressure sensor pad can be arranged inside the housing and connected to a pressure sensor outside the housing via a fluid path (i.e., via a fluid-filled conduit) to measure the pressure on the skin of the enclosed portion of the subject.

[0014] The processor can be configured to determine a respiratory rate determination value for the corresponding pressure pulse based on the determined dicrotic notch position, so that for several pressure pulses present at different times, several respiratory rate determination values ​​are determined, wherein the several respiratory rate determination values ​​determined for the several pressure pulses and therefore for the several times form a respiratory rate determination curve, and the respiratory rate is determined based on the respiratory rate determination curve.

[0015] In particular, the processor can be configured to determine, for the respective pressure pulse, at least one feature that depends on the determined dicrotic notch position, and to determine, based on the at least one determined feature, a respective respiratory rate determination value, in particular, not based on other features that characterize the respective pressure pulse and do not depend on the determined dicrotic notch position. However, the processor can also be configured to determine, for the respective pressure pulse, at least one additional feature that characterizes the respective pressure pulse and does not necessarily depend on the dicrotic notch position, and also determine the respiratory rate based on the at least one additional feature. In particular, the processor can be configured to determine, for the respective pressure pulse, a respiratory rate determination value based on several features, wherein at least one of these features is based on the determined dicrotic notch position, and at least one additional feature also characterizes the respective pressure pulse, but not necessarily based on the dicrotic notch position, such that, for several pressure pulses present at different times, several respiratory rate determination values ​​are determined, wherein, for the several pressure pulses and therefore for the several respiratory rate determination values ​​determined at different times, a respiratory rate determination curve is formed. As described above, the processor can be configured to determine the respiratory rate based on the respiratory rate determination curve.

[0016] The processor can be adapted to process several respiration rate determination values ​​obtained for several pressure pulses to obtain a continuous respiration rate determination curve. For example, interpolation can be applied and optionally also smoothing can be applied.

[0017] In one example, the processor is configured to determine, for determining a corresponding respiratory rate determination value for a corresponding pressure pulse, at least one of: a) an area under the corresponding pressure pulse between i) a first position and ii) a determined dicrotic notch position, the first position being a starting position where the corresponding pressure pulse begins, or a position between the starting position and the determined dicrotic notch position, and b) a difference between the first position and the dicrotic notch position. It has been found that by using at least one of these features, namely the aforementioned area and / or the aforementioned difference, the respiratory rate determination can be further improved.

[0018] In particular, the processor can be configured to determine the corresponding respiratory rate determination value by applying a predefined function to at least one of: a) the area under the corresponding pressure pulse between the first location and the determined dicrotic notch location, and b) the difference between the first location and the dicrotic notch location. For example ... the following: a) the area under the corresponding pressure pulse between the first location and the determined dicrotic notch location, and b) the difference between the first location and the dicrotic notch location.

[0019] In one example, the processor is configured to determine the respiratory rate by determining the frequency of the respiratory rate determination curve (preferably determining the dominant frequency of the respiratory rate determination curve), in particular if the respiratory rate determination curve has been determined based on at least one of the area and difference features described in the previous paragraph.

[0020] As mentioned above, the corresponding pressure pulse for determining the dicrotic notch can be directly the measured pressure pulse or a processed measured pressure pulse, wherein the measured pressure pulse can be a pressure pulse among a plurality of pressure pulses of the measured pressure signal. The corresponding pressure pulse can be processed, for example, by subtracting the average measured pressure from the corresponding pressure pulse. However, the pressure pulse can also be processed in another way. In addition, the determination of the respiratory rate determination value, in particular the determination of the characteristic of the corresponding pressure pulse (for example, the area under the corresponding pressure pulse between the starting position of the corresponding pressure pulse and the determined dicrotic notch position), can be performed directly using the measured pressure pulse or using a processed measured pressure pulse.

[0021] In an example, the processor is configured to apply a frequency range preselection filter to the respiratory rate determination curve, which can also be considered a respiratory rate determination signal. The filter is preferably configured to leave a predetermined frequency range in which the respiratory rate can exist unaffected or substantially unaffected. For example, the filter can be a frequency filter that reduces or eliminates frequency components outside the predetermined frequency range. Preferably, the frequency range preselection filter is a bandpass filter, which can be a combination of a low-pass filter and a high-pass filter.

[0022] In an example, the processor is configured to determine the dicrotic notch location by determining at least one of: a) a notch determination intersection location, the notch determination intersection location being the location where a shifted tangent intersects the pressure pulse, wherein the shifted tangent is determined by determining a tangent to the pressure pulse at a location where a first derivative of the pressure pulse has a minimum value and by shifting the tangent by an offset distance in a direction of increasing time, b) a notch determination maximum location, the notch determination maximum location being the location where a derivative function has a maximum value, wherein the derivative function is determined by determining a first derivative of the pressure pulse, by determining a second derivative of the pressure pulse, and by combining the determined first and second derivatives, and c) a notch determination function location, the notch determination function location being determined by determining a characteristic of the pressure pulse and applying a position determination function to the characteristic, wherein the position determination function provides a relationship between the characteristic and the notch determination function location. A processor configured in this manner allows for improved determination of the dicrotic notch location and, therefore, improved determination of respiratory rate based on the dicrotic notch location.

[0023] According to the explanation given above, the pressure pulse determined by applying at least one of the notch-determined intersection position, the notch-determined maximum position, and the notch-determined function position can be directly the measured pressure pulse or a processed measured pressure pulse. The minimum value of the first-order derivative of the pressure pulse is preferably a global minimum value of the first-order derivative of the pressure pulse.

[0024] In one example, the processor is configured to provide an offset distance such that it depends on a characteristic of the pressure pulse. The characteristic can be any characteristic related to the pressure pulse, and the offset distance can depend on one or more characteristics. For example, the offset distance can depend on at least one of the following: a) pulse rate, which can also be referred to as heart rate, b) the location of a minimum in the first derivative of the pressure pulse between the maximum systolic pressure of the pressure pulse and the pressure at the end of the pressure pulse and / or between the time of the maximum systolic pressure of the pressure pulse and the time at the end of the pressure pulse, c) the minimum of the first derivative, and d) the width of the pressure pulse at a predefined percentage of the difference between the maximum systolic pressure of the pressure pulse and the pressure at the end of the pressure pulse. The predefined percentage is preferably in the range of 50% to 80%. In a preferred embodiment, the predefined percentage is 66%. In an example, the width of the pressure pulse at the predefined percentage of the difference between the maximum systolic pressure of the pressure pulse and the pressure at the end of the pressure pulse is divided by the full pulse length (i.e., divided by the pulse duration) to determine the offset distance. In particular, the resulting ratio, which may be referred to as the pulse width ratio, can be an input to an offset distance function that provides an offset distance as an output. For example, the offset distance can be calculated according to the following equation:

[0025] Offset distance = (pulse width ratio - m) * n, (1)

[0026] Here, m and n are parameters predetermined by calibration.

[0027] Thus, the dependence of the offset distance on one or more characteristics of the corresponding pulse can be determined during a calibration process, wherein the dependence of the offset distance on the one or more characteristics of the corresponding pulse is determined so that the determined intersection position corresponds to the position of the dicrotic notch known during the calibration process. It has been found that a dependence that defines a greater offset distance the higher the position of the minimum of the first derivative between the maximum systolic pressure of the pressure pulse and the pressure at the end of the diastole of the pressure pulse provides good results. It has also been found that a dependence that defines a smaller offset distance the greater the pulse rate, or that the offset distance decreases with increasing pulse rate, also provides good results. The offset distance can also be constant. It has been found that an already constant offset distance can lead to a high-quality determination of the dicrotic notch position, wherein the quality of the determination of the dicrotic notch position can be further improved if the offset distance depends on one or more characteristics of the pressure pulse.

[0028] The constant offset distance can be predetermined by statistical analysis of a training data set used during the calibration process and comprising the known dicrotic notch location, the tangent line, and the offset distance required to shift the tangent line so that the resulting intersection position coincides with the known dicrotic notch location. In one embodiment, the statistical analysis comprises calculating the average of the offset distances found while performing the calibration process. In a preferred embodiment, the constant offset distance is in the range of 10 ms to 50 ms. Preferably, the constant offset distance is 20 ms.

[0029] In an example, the processor is configured to select the temporally last intersection position as the notch determination intersection position if there are several intersection positions of the shift tangent line with the pressure pulse. It has been found that this allows determining the intersection position such that it corresponds to a further improved accuracy of the position of the dicrotic notch.

[0030] Furthermore, in an example, the processor is configured to determine the notch-determined maximum position by determining whether the highest maximum value of the derivative function is greater than the second highest maximum value multiplied by a predefined factor, wherein if this is the case, the notch-determined maximum position is the position of the highest maximum value, and if not, the notch-determined maximum position is the position of the first maximum value of the derivative function, wherein preferably, the processor is configured to only consider maxima of the derivative function that are greater than a first predefined derivative threshold value and / or within a provided search range. The first maximum value of the derivative function is the first in time, i.e., the earliest maximum value of the derivative function. This allows the notch-determined derivative position to be determined so that it corresponds even more accurately to the position of the dicrotic notch of the pressure pulse.

[0031] In one example, the processor is configured to determine that a pressure pulse has an artifact if the derivative function includes at least two maxima greater than a second predefined derivative threshold (particularly within a search range), wherein the second predefined derivative threshold is greater than the first predefined derivative threshold. If the derivative function of the corresponding pulse includes at least two maxima greater than the first predefined derivative threshold, then the derivative function has at least two very large maxima, which would not be the case if a normal pressure pulse were present. In other words, if there are at least two very large maxima, then an artifact is likely present.

[0032] The predefined derivative threshold values ​​can be determined again during the calibration process. During the calibration process, it is known whether the respective training pressure pulse comprises an artifact and, if the respective training pressure pulse does not comprise an artifact, the dicrotic notch position is known. These threshold values ​​are determined such that training pressure pulses with artifacts are identified as closely as possible and, for each respective training pressure pulse without artifacts, the respective determined dicrotic notch position is also likely to correspond to the respective known dicrotic notch position. It has been found that a first predefined derivative threshold value in the range from 0.003 to 0.070, in particular with a value of 0.015, and a second predefined derivative threshold value in the range from 0.05 to 1.00, in particular with a value of 0.25, provide good results, wherein these values ​​refer to the use of mmHg as the unit of pressure and s as the unit of time in the function (2) to be described below.

[0033] The processor can be configured to determine the derivative function by dividing a) a second-order derivative function that depends on the second-order derivative by b) a first-order derivative function that depends on the first-order derivative. Dividing by the first-order derivative function avoids overweighting of the derivative function in a range of positions where the first-order derivative has relatively large negative values. By avoiding overweighting, the notch determination maximum position can be determined so that it corresponds even better to the position of the dicrotic notch of the pressure pulse.

[0034] In a preferred embodiment, the second-order derivative function is a second-order derivative. In particular, the derivative function can be determined according to the following equation:

[0035] f_2deri=(d 2 P / dt 2 ) / (a+(dP / dt)^b), (2)

[0036] Among them, f_2deri represents the derivative function, d 2 P / dt 2Denotes the second derivative, dP / dt denotes the first derivative, and a and b denote predefined parameters. The parameters a and b can be predetermined during a calibration process, wherein the corresponding positions of the dicrotic notch are known for several training pressure pulses, and wherein the parameters a and b are determined so that the determined maximum position of the second derivative corresponds as well as possible to the known position of the dicrotic notch. In a preferred embodiment, a is in the range of 0 to 10 and b is in the range of 1 to 2. In particular, a is 1.0 and b is 1.3. Furthermore, these values ​​refer to the use of mmHg as the unit of pressure and s as the unit of time in equation (2). It has been found that by using the derivative function according to equation (2), the determined maximum position of the notch corresponds even more accurately to the position of the dicrotic notch of the pressure pulse.

[0037] In one example, the first-order derivative is filtered using a filter, preferably a low-pass filter, wherein, in one embodiment, the filter depends on the heart rate. For example, the processor can be configured such that within a normal heart rate range of, for example, 50 to 90 beats per minute, the first-order derivative is filtered several times, for example three times, using a moving average filter, wherein the moving average filter includes a window width of, for example, 28 ms. If the actual heart rate is outside this normal heart rate range, another filter can be used. The filtering is preferably performed to reduce noise, and the filter can also be constant, i.e., it can also be independent of the heart rate. The derivative function can also be filtered using a constant filter or a filter that depends on the heart rate. In particular, if the heart rate is within the normal heart rate range, for example, 50 to 90 beats per minute, the derivative function can also be filtered several times, for example three times, using a moving average filter, wherein the moving average filter includes a window width of, for example, 28 ms, and if the heart rate is outside this normal heart rate range, another filter is used. Furthermore, the derivative function is preferably filtered to reduce noise.

[0038] The processor can also be configured to align the derivative function with the pressure pulse. Misalignment may occur due to filter delay, wherein the delay can be corrected by aligning the derivative function with the pressure pulse.

[0039] In an example, the processor is configured to provide a position determination function such that it provides a relationship between the notch determination function position and at least one of the following characteristics of the pressure pulse: a maximum value of the pressure pulse, a maximum position along the pressure pulse, a first position along the pressure pulse before the maximum position (and at which the pressure pulse has a value that is a predefined first fraction of the maximum value), a second position along the pressure pulse after the maximum position (and at which the pressure pulse has a value that is a predefined second fraction of the maximum value), a maximum value of a first derivative of the pressure pulse, and a difference between the maximum systolic pressure of the pressure pulse and the pressure at the end-diastolic phase of the pressure pulse. In particular, any combination of two or more of the above characteristics can be used to determine the notch determination function position, i.e., the notch determination function can depend on one of these characteristics or on any combination of two or more of these characteristics. In one embodiment, the position determination function specifically provides a relationship between the notch determination function position and characteristics including i) a maximum position along the pressure pulse and ii) at least one of the following characteristics of the pressure pulse: the maximum value of the pressure pulse, a first position along the pressure pulse before the maximum position (and at this first position, the pressure pulse has a value that is a predefined first fraction of the maximum value), a second position along the pressure pulse after the maximum position (and at this second position, the pressure pulse has a value that is a predefined second fraction of the maximum value), a maximum value of a first derivative of the pressure pulse, and a difference between a maximum systolic pressure of the pressure pulse and a pressure at the end of diastole of the pressure pulse. In particular, the notch determination function position can be determined according to the following equation:

[0040] fp=c*t.dia.pre+d*t.max+e*t.dia.post+f*Pulse.max+

[0041] g*d(Pulse) / dt.max+h*PP+i*PR, (3)

[0042] Wherein, fp represents the position of the notch determination function, t.dia.pre represents the first position along the pressure pulse before the maximum position (and at which the pressure pulse has a value that is a predefined first fraction of the maximum value), t.max represents the maximum position along the pressure pulse, t.dia.post represents the second position along the pressure pulse after the maximum position (and at which the pressure pulse has a value that is a predefined second fraction of the maximum value), Pulse.max represents the maximum value of the pressure pulse, d(Pulse) / dt.max represents the maximum value of the first derivative of the pressure pulse, PP represents the difference between the maximum systolic pressure of the pressure pulse and the pressure at the end of diastole of the pressure pulse, and PR represents the pulse rate. Instead of the pulse rate, the pulse duration can also be used in equation (3). It has been found that if the position of the notch determination function fp is determined in this way, it can correspond very accurately to the position of the dicrotic notch of the pressure pulse.

[0043] Furthermore, the parameters c to i are predefined parameters that can be predetermined by calibration. In particular, the parameters c to i used in equation (3) can be predetermined during a calibration process during which the position of the dicrotic notch is known for several training pulses, and the parameters c to i are determined so that the position of the corresponding notch-determining function corresponds as well as possible to the corresponding position of the dicrotic notch. Furthermore, a predefined first fraction of the maximum value and a predefined second fraction of the maximum value can be determined in such a way that, during the calibration process, the resulting function position corresponds as well as possible to the known dicrotic notch position. The predefined first and second fractions are preferably in the range of 70% to 95%.

[0044] In one embodiment, the processor is configured to determine the dicrotic notch position only within a provided search range. In particular, the above-mentioned determination process for determining the notch determination intersection position, the notch determination maximum position, and / or the notch determination function position is applied only within the search range. In an example, the search range can be predetermined based on statistical analysis. In particular, for a large group of training pressure pulses with known dicrotic notch positions, the average position and standard deviation of the dicrotic notch position can be determined and used to determine the search range. For example, the center of the search range can be defined by the average value, and the width of the search range can be defined by the standard deviation.

[0045] Preferably, the processor is configured to provide a search range defined a) by a first and a second percentage of the difference PP between the maximum systolic pressure of the pressure pulse and the pressure at the end-diastolic phase of the pressure pulse and / or b) by a first and a second percentage of the pulse duration. For example, the starting and stopping points of the search range can be defined based on a) the first and a second percentage of the difference between the maximum systolic pressure of the pressure pulse and the pressure at the end-diastolic phase of the pressure pulse and / or b) the first and a second percentage of the pulse duration. If several starting and / or stopping points are determined, the search range is defined by the latest starting and / or earliest stopping points, respectively. One or more percentages can be static, i.e., have a predefined constant value, and / or one or more percentages can be dynamic, i.e., depend on the corresponding pulse or pulse rate. For example, the starting point can be defined by a static first percentage of the pulse duration (e.g., 18% of the pulse duration) and a static first percentage of the difference between the maximum systolic pressure of the pressure pulse and the pressure at the end-diastolic phase of the pressure pulse (e.g., 75% of the difference). The stopping point can be defined by a static second percentage of the pulse duration (e.g., 65% of the pulse duration), by a static second percentage of the difference between the maximum systolic pressure of the pressure pulse and the pressure at the end of diastole of the pressure pulse (e.g., 9% of that difference), and by a dynamic percentage of the pulse duration, where the dynamic percentage can depend on the pulse rate.

[0046] Furthermore, these percentages are predefined through calibration, i.e., during a training phase, by analyzing training pressure pulses with known dicrotic notches. The percentages can be determined to ensure a sufficient presence of the dicrotic notch so that it can actually be reliably detected. During the calibration process, the presence of the dicrotic notch is determined for a number of training pressure pulses, allowing its location to be determined. This determination can be performed manually by an experienced physician or technician, for example. Then, during this calibration process, the value range is determined by determining in which range relative to the corresponding difference PP the training pulses with a detectable presence of the dicrotic notch fall. It has been found that a static first percentage of the difference between the maximum systolic pressure of the pressure pulse and the pressure at the end-diastolic phase of the pressure pulse within the range of 75% to 85%, and a static second percentage of this difference within the range of 5% to 15%, allows for further improved dicrotic notch detection. Searching for the dicrotic notch location exclusively within the search range can be implemented, for example, by only determining the dicrotic notch if the pressure pulse has a value within the search range at the location where the first derivative of the pressure pulse has its minimum value. This ensures that the processor does not determine the unreliable so-called dicrotic notch position of the pressure pulse, for which the dicrotic notch position is actually undeterminable. This further increases the reliability of determining the dicrotic notch position and ultimately determining the respiratory rate based on the dicrotic notch position.

[0047] If the pressure pulse has been measured non-invasively, the processor is preferably configured to determine the dicrotic notch position by determining the notch determination function position. It has been found that, particularly in case of non-invasive measurement of the pressure pulse, the notch determination function position corresponds very well to the dicrotic notch position.

[0048] Furthermore, in a preferred embodiment, if a pressure pulse is non-invasively measured on the skin of the enclosed portion of the subject using a pressure sensor while the pressure applicator increases or decreases the applied pressure, then preferably only pressure pulses that were measured when the mean measured pressure (which can also be considered as mean tissue pressure, since it is measured on the skin of the subject's tissue) was below the systolic arterial blood pressure (SAP) and above the diastolic arterial blood pressure (DAP) by a predefined percentage are used to determine the dicrotic notch location. Preferably, the predefined percentage is in the range of 85% to 95%, and particularly preferably, it is 90%. The SAP and DAP values ​​here can be values ​​from a previous blood pressure measurement.

[0049] The processor can be configured to determine an area ratio TPA1.top / TPA2.top based on the partial areas TPA1.top and TPA2.top under the pressure pulse curve. To determine the partial areas TPA1.top and TPA2.top, a partial area TPA.top under the corresponding pressure pulse (i.e., below the corresponding pressure pulse curve) is formed, which is the area enclosed by the corresponding pressure pulse curve above a predefined percentage of the difference TPP between the maximum and minimum values ​​of the corresponding pressure pulse curve. This corresponds to the area of ​​the corresponding pressure pulse curve above a horizontal line arranged along the predefined percentage of the difference TPP. This predefined percentage is preferably in the range of 30% to 70%, and more preferably 50%. A vertical line is then arranged so that it intersects the vertex, i.e., the maximum value, of the corresponding pressure pulse curve. Furthermore, two straight lines are added, wherein the first of the two straight lines connects the vertex with the first of the intersection points of the horizontal line and the corresponding pressure pulse curve, and the second of the two straight lines connects the vertex with the second of the intersection points of the horizontal line and the corresponding pressure pulse curve. The first straight line together with the horizontal line and the vertical line encloses the first partial area TPA1.top, and the second straight line together with the horizontal line and the vertical line encloses the second partial area TPA2.top. These partial areas TPA1.top and TPA2.top are also Figure 6 A and described in WO 2018 / 210931 A1 on page 19, line 36 to page 20, line 8, which is incorporated herein by reference.

[0050] In one embodiment, the processor is configured to determine a moving average of the area ratio TPA1.top / TPA2.top over a predetermined number of pressure pulses. Then, for each respective pressure pulse, the difference between the moving average of the area ratio TPA1.top / TPA2.top and the respective area ratio TPA1.top / TPA2.top of the respective pressure pulse is calculated. Based on the obtained difference, a standard deviation function is calculated for the pressure pulse. The respective standard deviation function (TPA1.top / TPA2.top).sd is Figure 6 B and described in WO 2018 / 210931 A1, page 9, lines 19 to 26 and page 20, lines 10 to 23, which are incorporated herein by reference as described above. The processor can be configured to determine the dicrotic notch location using only pressure pulses that have been measured before the standard deviation function reaches its maximum value.

[0051] In one example, the processor is configured to determine, for each pressure pulse, at least one of the following features as at least one additional feature that characterizes the corresponding pressure pulse and does not necessarily depend on the location of the dicrotic notch: i) the difference between the maximum systolic pressure of the pressure pulse and the pressure at the end-diastolic phase of the pressure pulse, ii) the area enclosed by the upper portion of the pressure pulse, wherein a) the upper end of the upper portion is at the maximum systolic pressure of the pressure pulse and b) the lower end of the upper portion is between the maximum systolic pressure of the pressure pulse and a pressure value corresponding to the mean value of the pressure, iii) the duration of the corresponding pressure pulse, iv) the area of ​​the corresponding pressure pulse, and v) the half-peak width of the corresponding pressure pulse. It has been found that by using at least one of these additional features, the determination of the respiratory rate can be further improved.

[0052] The difference between the maximum systolic pressure of the pressure pulse and the pressure at the end of diastole of the pressure pulse is preferably the difference between the maximum and minimum measured pressures of the corresponding pressure pulse. The duration of the corresponding pressure pulse preferably corresponds to the time difference between the end of diastole and the subsequent end of diastole of the corresponding pressure pulse. The half-peak width corresponds to the width at 50% of the difference between the maximum and minimum pressures of the corresponding pressure pulse. Therefore, it is the width at 50% of the difference between the pressure at the maximum systolic point of the pressure pulse and the pressure at the end of diastole of the pressure pulse.

[0053] In another aspect of the present invention, a method for determining a respiration rate of a subject is provided, wherein the method comprises:

[0054] - providing a pressure signal of the subject over time by a pressure signal providing unit, wherein the pressure signal is indicative of blood pulsation and comprises a plurality of pressure pulses,

[0055] - determining, by a processor, a corresponding dicrotic notch position for a corresponding pressure pulse of the plurality of pressure pulses,

[0056] - determining the respiratory rate based on the dicrotic notch locations determined for the plurality of pressure pulses.

[0057] In one aspect of the present invention, a computer program for determining a breathing rate of a subject is proposed, wherein the computer program comprises program code means for causing the apparatus for determining the breathing rate according to claim 1 to perform the steps of the method according to claim 14.

[0058] The apparatus, method and computer program for determining the breathing rate can be adapted to determine the breathing rate continuously, ie performing several subsequent breathing rate measurements, or to determine the breathing rate discontinuously, ie for example once.

[0059] It shall be understood that the apparatus of claim 1, the method of claim 14 and the computer program of claim 15 have similar and / or identical preferred embodiments, in particular as defined in the dependent claims.

[0060] It shall be understood that a preferred embodiment of the invention can also be any combination of the dependent claims or the above embodiments with the respective independent claim.

[0061] These and other aspects of the invention will be apparent from and elucidated with reference to the embodiments described hereinafter. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In the following figures:

[0063] Figure 1 An embodiment of an apparatus for determining the respiratory rate of a subject is schematically and exemplarily shown, comprising a measuring device in the following circumstances: wherein a pressure cuff of the measuring device is inflated,

[0064] Figure 2 Schematically and exemplarily shown is a housing of a measuring device surrounding the upper arm of a subject,

[0065] Figure 3 The device is shown schematically and exemplarily with the pressure cuff deflated,

[0066] Figure 4 Schematically and exemplarily shown are measured tissue pressure and further values ​​derived from the tissue pressure measurement,

[0067] Figure 5 Schematically and exemplarily illustrating the determination of the dicrotic notch position as the intersection position of the displacement tangent and the corresponding pressure pulse,

[0068] Figure 6schematically and exemplarily illustrating the determination that the dicrotic notch position is the second derivative maximum position (ie the position where the second derivative of the corresponding pressure pulse has a maximum value),

[0069] Figure 7 The determination of the search range is schematically and exemplarily illustrated,

[0070] Figure 8 Schematically and exemplarily illustrating the determination of the dicrotic notch position as a function position,

[0071] Figure 9 illustrates the calculation of different characteristics of the pressure pulse measuring tissue pressure, depending on the determined location of the dicrotic notch,

[0072] Figures 10 to 13 illustrates the calculation of different additional features of the pressure pulse that measure tissue pressure, and

[0073] Figure 14 A flow chart exemplarily illustrating an embodiment of a method for determining the respiration rate of a subject is shown. DETAILED DESCRIPTION

[0074] Figure 1 Schematically and exemplarily, an apparatus 1 for determining the respiratory rate of a subject is shown. The apparatus 1 comprises a housing 4 which is capable of Figure 2 and is configured to surround a portion 5 of the subject through which blood flows. In this embodiment, the portion 5 of the subject is the subject's arm, wherein Figure 2 5 , wherein the brachial artery 11 in the arm 5 is shown, and wherein the arrow in the brachial artery 11 indicates the direction of blood flow away from the heart. The device 1 further comprises a pressure sensor 7, which is arranged inside the housing 4 and is configured to measure the pressure on the outer skin of the subject's enclosed arm 5. The measured pressure can also be referred to as tissue pressure (TP). Figure 2 As shown, the pulse wave in the brachial artery 11 causes a pressure wave 12 which is transmitted via the tissue of the arm 5 to the pressure sensor 7. For clarity, Figure 1 The housing 4 is not shown.

[0075] The device 1 further comprises a cuff 6 which surrounds the housing 4 and is inflatable using a pump 8 for applying pressure to the housing 4 and thereby to the subject's enclosed arm 5 from outside the housing 4. Since the cuff 6 and the pump 8 together enable pressure to be applied to the housing 4 and thereby to the subject's enclosed arm 5, they can be considered to form pressure applicators 6, 8. Furthermore, the housing 4, the pressure applicators 6, 8 and the pressure sensor 7 can be considered to be components of a measuring device controlled by the processor 3. The housing 4 with the cuff 6 is preferably an anti-kink housing cuff as described in WO 2014 / 121945 A1.

[0076] Processor 3 is configured to control the measurement device so that pressure applicators 6 and 8 increase applied pressure during a measurement period that extends to the end of the measurement time point and decrease applied pressure during a subsequent post-respiration rate measurement period, and so that pressure sensor 7 measures the pressure on the skin, i.e., tissue pressure TP, at least during the measurement period. Furthermore, processor 3 is configured to control pressure applicators 6 and 8 so that they increase applied pressure at a first rate during a preceding measurement period, followed by a measurement period in which applied pressure increases at a second rate, wherein the first rate is greater than the second rate. Figure 1 , the control of the measuring device is illustrated such that the cuff 6 is inflated, whereby the applied pressure increases, ie the bold arrow indicates the inflation.

[0077] The device 1 includes a valve 20 which, when open, allows compressed air within the system to leave the system and enter the surrounding atmosphere for deflation of the cuff. Figure 3 In FIG, this deflation situation with the pump 8 switched off is indicated by a bold arrow.

[0078] The processor 3 can be considered to comprise a control portion 10 for controlling the pump 8 and the valve 20 and a processing portion 11, the processing portion 11 being configured, inter alia, to perform some calculations as will be explained further below. The apparatus 1 can also comprise a display 22 for displaying the determined respiratory rate and optionally other physiological parameters, such as blood pressure.

[0079] In a first measurement period, which can also be considered a fast inflation period, processor 3 controls apparatus 1 so that valve 20 is closed and pump 8 inflates cuff 6 at a first, higher rate. In a subsequent measurement period, processor 3 also controls apparatus 1 so that valve 20 is closed, but controls pump 8 so that inflation of cuff 6 continues at a second, lower rate. Thus, the measurement period can also be considered a slow inflation period.

[0080] Figure 4The measured pressure TP is illustrated schematically and exemplarily with respect to the time t. The first inflation in the preceding measurement time period starts with a tissue pressure TP of an attachment pressure Patt, which is the measured tissue pressure in the uninflated cuff 6. This attachment pressure Patt can be in the range of, for example, 0 to 15 mmHg. An attachment pressure Patt of up to 15 mmHg has been shown not to cause venous congestion for more than 12 hours, making the assembly of the housing 4 and the cuff 6 optimally suitable for longer-term monitoring. For this reason, it is preferred that the attachment pressure Patt is not greater than 15 mmHg. Figure 4 In FIG. 1 , arrow 30 indicates the start of the pre-measurement period, i.e., the start of the rapid inflation period. During this pre-measurement period, the portion of the entire tissue pressure range that contains no or substantially no information for determining blood pressure should pass as quickly as possible. Therefore, during the pre-measurement period, the inflation rate is preferably as high as possible. For example, the inflation rate relative to the tissue pressure TP can be equal to or greater than 8 mmHg / s. This pre-measurement period with the rapid inflation rate is Figure 4 The test ends at the tissue pressure value indicated by “TPlow” in the figure.

[0081] The tissue pressure value TPlow also indicates the start of a measurement period having a second, slower inflation rate. Figure 4 In FIG. 3 , the start of the measuring period is indicated by arrow 31 , which can also be considered as a slow inflation period.

[0082] exist Figure 4 , time interval 40 indicates the inflation-deflation period from the start 30 of rapid inflation until tissue pressure TP has dropped below 20 mmHg during rapid deflation to allow venous return. Time period 41 indicates the cycle time, which is the time between the start of a measurement and the start of a subsequent measurement, and time period 42 indicates the break time period, which is the difference between inflation-deflation time period 40 and cycle time 41.

[0083] Figure 4 Also illustrated is a mean pressure TPcl, which can be considered as the tissue clamping pressure affecting the tissue when the cuff 6 is attached to the subject's arm 5 (e.g., upper arm). The mean pressure TPcl can be calculated by applying a low-pass filter to the tissue pressure TP, wherein the low-pass filter may be located within the processor 3. The processor 3 is preferably configured to determine the alternating component of the tissue pressure TPac by subtracting the mean pressure TPcl from the measured tissue pressure TP, i.e., TPac=TP-TPcl. Figure 4 In FIG, the TPac curve is shown magnified 2 times in order to improve the visibility of the curve.

[0084] During the slow inflation period, i.e. during the measurement period, a TP pulse curve and / or a TPac pulse curve is recorded which can be analyzed simultaneously, i.e. online, wherein the TP pulse curve and / or the TPac pulse curve has a tissue pressure waveform (TPW) including information allowing an accurate determination of the respiratory rate.

[0085] An important feature of the pressure pulse (i.e., the TP pulse curve and the TPac pulse curve) is the location of the dicrotic notch. Figure 5 、 Figure 6 and Figure 7 As explained, the processor 3 is thus configured to determine the location of the dicrotic notch. For example, the processor can be configured to determine the dicrotic notch location by determining at least one of: a) a notch determination intersection location, the notch determination intersection location being the location where the shift tangent 52 intersects the pressure pulse, wherein the shift tangent 52 is determined by determining a tangent 51 to the pressure pulse at the location where the first derivative of the pressure pulse has a minimum value and by shifting the tangent 51 in a direction of increasing time by an offset distance t.shift, b) a notch determination maximum location, the notch determination maximum location being the location where the derivative function f_2deri has a maximum value, wherein the derivative function is determined by determining a first derivative of the pressure pulse, by determining a second derivative of the pressure pulse, and by combining the determined first and second derivatives, and c) a notch determination function location, the notch determination function location being determined by determining a characteristic of the pressure pulse and by applying a position determination function to the characteristic, wherein the position determination function provides a relationship between the characteristic and the notch determination function location.

[0086] The processor 3 can be configured to determine one, two, or all of the notch-determined intersection position, the notch-determined maximum position, and the notch-determined function position. If the processor 3 determines two or three of the notch-determined intersection position, the notch-determined maximum position, and the notch-determined function position, the processor can be configured to combine the different positions to obtain a combined position corresponding to the dicrotic notch position. For example, the different positions can be averaged to determine the dicrotic notch position.

[0087] Before determining, for example, the intersection location of the notch, processor 3 preferably checks whether the pressure pulse, at the location where the first derivative of the pressure pulse has its minimum value, has a value within a search range defined by a first percentage and a second percentage of the difference PP between the maximum systolic pressure of the pressure pulse and the pressure at the end-diastolic phase of the pressure pulse. This search range can also be referred to as a first search range. Preferably, the static first percentage of the difference PP is in the range of 75% to 85%, and the static second percentage of the difference PP is in the range of 5% to 15%. For example, the search range can be from 75% of PP to 10% of PP.

[0088] If the pressure pulse has a value within a predefined search range depending on the difference PP at the location where the first derivative of the pressure pulse has its minimum, the processor 3 determines a tangent 51 at the location along the corresponding pulse 50 where the first derivative has its minimum. Figure 5 As shown, the tangent line 51 is shifted by an offset distance t.shift, wherein the intersection position of the shifted tangent line 52 and the pulse 50 corresponds to the dicrotic notch position t.notch. Figure 5 In the figure, the shift is a right shift because the direction of increasing time is from left to right. If there are several intersections of the shift tangent 52 and the pressure pulse, the processor 3 preferably selects the last intersection position in time as the intersection position for determining the dicrotic notch position.

[0089] Processor 3 can be configured to provide the offset distance t.shift such that it depends on a characteristic of the corresponding pressure pulse 50. The characteristic can be any characteristic related to the pressure pulse, and the offset distance can depend on one or more characteristics. For example, the offset distance can depend on at least one of the following: a) pulse rate (PR), which can also be referred to as heart rate; b) the location of the minimum of the first derivative of the pressure pulse between the maximum systolic pressure of the pressure pulse and the pressure at the end of the pressure pulse and / or between the time of the maximum systolic pressure of the pressure pulse and the time at which the pressure pulse ends; c) the minimum of the first derivative; and d) the width of the pressure pulse at a predefined percentage of the difference PP. The predefined percentage is preferably in the range of 50% to 80%. In a preferred embodiment, the predefined percentage is 66%. In one example, the width of the pressure pulse at the predefined percentage of the difference between the maximum systolic pressure of the pressure pulse and the pressure at the end of the pressure pulse is divided by the full pulse length, i.e., by the pulse duration, to determine the offset distance. However, the processor 3 can also be configured to provide a constant or static offset distance t.shift, ie an offset distance t.shift that is, for example, independent of any characteristics of the pressure pulse 50 .

[0090] Figure 5 Also shown is the position t.start where the pressure pulse 50 starts and the difference LVET (Left Ventricular Ejection Time) between the starting position t.start and the determined intersection position corresponding to the dicrotic notch position t.notch.

[0091] Also before determining the notch-determined maximum position, the processor 3 preferably checks whether the pressure pulse, at the position where the first derivative of the pressure pulse has its minimum value, has a value within the aforementioned search range defined by the percentage of the difference PP between the maximum systolic pressure of the pressure pulse and the pressure at the end of diastole of the pressure pulse. If this is the case, the notch-determined maximum position is determined. Otherwise, the notch-determined maximum position is not determined for the pressure pulse, and the process proceeds to the next pressure pulse of the pressure signal.

[0092] In order to determine the maximum position of the notch, the processor 3 can be configured to calculate the position where the derivative function has a maximum value, wherein the derivative function is determined by determining the first derivative of the pressure pulse, by determining the second derivative of the pressure pulse and by combining the determined first and second derivatives. The first derivative can be filtered in order to reduce noise. In particular, depending on in which heart rate range the actual heart rate is located, different noise reduction filters can be used. In an embodiment, if the heart rate is within the normal heart range of 50 to 90 beats per minute, the filter can be a moving average filter, wherein, for example, a filter with a width of 28 ms or with another width can be applied several times, for example three times. The filtered first derivative 1st_deriv is in Figure 6 It is shown as an example in FIG.

[0093] The processor 3 can also be configured to determine the derivative function by dividing a) a second-order derivative function that depends on the second-order derivative by b) a first-order derivative function that depends on the first-order derivative. Preferably, the second-order derivative function is a second-order derivative, and the derivative function is determined according to equation (2). Furthermore, the second-order derivative and the derivative function can be filtered to reduce noise, as explained above with respect to the first-order derivative. Furthermore, the filtered derivative function f_2deri is Figure 6 It is shown as an example in FIG.

[0094] The processor 3 can also be configured to define another search range within which the maximum value of the derivative function should be searched. In particular, the processor can be configured to use a predefined search range that has been predetermined, for example, by a statistical analysis as described above, or has been determined by a start point and a stop point, as will be explained below. This further search range can be regarded as a second search range.

[0095] The processor 3 can be configured to determine a starting point and a stopping point of the further search range based on a) the percentage of the difference PP and b) the percentage of the pulse duration. In particular, several starting points and several stopping points can be determined, wherein the further search range can be defined by the latest starting point and the earliest stopping point, and wherein the percentages can be static or depend on the respective pulse or pulse rate. For example, Figure 7As shown, a starting point start1 can be defined by a static first percentage of the pulse duration (e.g. 18%), another starting point start2 can be defined by a static first percentage of the difference PP (e.g. 85%), a first stopping point stop1 can be defined by a static second percentage of the pulse duration (e.g. 65%), a second stopping point stop2 can be defined by a static second percentage of the difference PP (e.g. 10%), and a third stopping point stop3 can be defined by a dynamic percentage of the pulse duration, which depends on the pulse rate and can be, for example, 43% for a pulse rate of 50 bpm. Figure 7 In , the search range starts at the latest starting point start1 and ends at the earliest stopping point stop3.

[0096] Preferably, the processor 3 is configured to determine all maxima of the derivative function f_2deri within the further search range that are greater than a first predefined derivative threshold value, so as to determine all maximum maxima of the second-order derivative within the search range. If the highest maximum of the determined maxima is greater than the product of a predefined factor and the second highest maximum in the search range, then the position of the highest maximum within the search range is considered to be the notch-determined maximum position, i.e., corresponding to the dicrotic notch position t.notch, for example. Otherwise, the position of the first temporal maximum of the derivative function f_2deri within the search range is considered to be the notch-determined maximum position, i.e., corresponding to the dicrotic notch position t.notch, for example.

[0097] If two maxima are found within the other search range that are greater than a second predefined derivative threshold, which is greater than the first predefined derivative threshold used to determine the large maximum value, i.e., if there are two or more very large maxima within the search range, the corresponding pressure pulse is not taken into account for determining the respiratory rate because it is considered to have an artifact.

[0098] Also before determining the notch determination function position, the processor 3 can check whether the pressure pulse has a value at the location where the first derivative of the pressure pulse has its minimum value that is within a first search range determined by the difference PP between the maximum systolic pressure of the pressure pulse and the pressure at the end of diastole of the pressure pulse. If this is the case, the notch determination function position is determined. Otherwise, the notch determination function position is not determined for the pressure pulse and the process continues with the next pressure pulse of the pressure signal.

[0099] In order to determine the notch determination function position, the processor 3 can be configured to provide a position determination function such that it defines a relationship between the notch determination function position and at least one of the following properties of the corresponding pressure pulse, which properties are also Figure 8, the maximum value Pulse.max of the corresponding pressure pulse, the maximum position t.max along the corresponding pressure pulse, a first position t.dia.pre along the corresponding pressure pulse before the maximum position t.max (and at this first position, the corresponding pressure pulse has a value that is a predefined first fraction p1 of the maximum value Pulse.max), a second position t.dia.post along the corresponding pressure pulse after the maximum position t.max (and at this second position, the corresponding pressure pulse has a value that is a predefined second fraction p2 of the maximum value Pulse.max), the maximum value d(Pulse) / dt.max of the first derivative of the corresponding pressure pulse, and the difference PP between the maximum systolic pressure of the pressure pulse and the pressure at the end of diastole of the pressure pulse. In particular, the processor 3 can be configured to calculate the notch determination function position fp according to the above equation (3), which should correspond to the dicrotic notch position t.notch.

[0100] The processor 3 is configured to determine, for the corresponding pressure pulse, at least one corresponding feature characterizing the corresponding pressure pulse based on the corresponding determined dicrotic notch position. Figure 9 Describe this.

[0101] from Figure 9 As can be seen in FIG, the pressure pulse 29 has a notch in its descending slope, wherein the time at which the notch occurs is indicated by "t.notch". In particular, a pressure pulse 29 that reaches t.stop from t.start and thus has a pulse curve duration of t.stop-t.start includes a contraction that begins with a systolic upslope at t.start and ends at the dicrotic notch at time t.notch in the subsequent downslope. In this example, the dicrotic notch is characterized by a tiny notch. The further descending portion of the pulse curve from t.notch to t.end can be referred to as "diastole".

[0102] The processor can be adapted to determine the difference between the start (t.start) of the pressure pulse 29 and the position of the notch (t.notch) as a feature for determining respiratory rate or another physiological parameter. This difference (i.e., this feature) can be called the "systolic interval" or "left ventricular ejection time" (LVET).

[0103] Furthermore, the processor 3 can be configured to determine the area TPA.sys under the pressure pulse 29 between the start position t.start and the notch position t.notch. The corresponding area is Figure 9In one embodiment, the processor 3 can be configured to additionally or alternatively determine another feature characterizing the ejection phase of the pressure pulse 29, in particular the left ventricular ejection phase, wherein the further feature also uses knowledge about the position of the dicrotic notch defining the transition between systole and diastole.

[0104] In one embodiment, the processor 3 is further configured to determine a respiratory rate determination value TPWP_R for each pressure pulse based on at least one feature that characterizes the corresponding pressure pulse and depends on the dicrotic notch position, such that a plurality of respiratory rate determination values ​​TPWP_R are determined for a plurality of pressure pulses present at different times. The plurality of respiratory rate determination values ​​TPWP_R determined for a plurality of pressure pulses and, therefore, for a plurality of times, form a respiratory rate determination curve. The respiratory rate determination curve may also be referred to as a respiratory pulse variation signal.

[0105] In a preferred embodiment, the respiratory rate determination value TPWP_R is determined based on at least one of TIP.sys and LVET. Furthermore, in a preferred embodiment, the respiratory rate determination value TPWP_R is identical to or linearly related to, in particular proportional to, TPA.sys. Therefore, the respiratory rate determination curve, which can also be referred to as the respiratory pulse variation signal, is preferably TCP.sys(t). That is, the respiratory rate determination curve or respiratory pulse variation signal can be formed using a sequence of TCP.sys values ​​determined over time. However, as previously noted, the respiratory rate determination value TPWP_R can also be determined by combining TCP.sys and LVET, for example, by a linear combination of TCP.sys and LVET.

[0106] Processor 3 can also be configured to apply a frequency range preselection filter to the respiratory pulse variation signal so that the respiratory rate is retained in the respiratory pulse variation signal. The filter can be, for example, a bandpass filter, which is configured to pass the portion of the respiratory pulse variation signal that is higher than a lower cutoff frequency and lower than a higher cutoff frequency through the filter. In one example, the bandpass filter is a combination of a high-pass filter having a lower cutoff frequency lower than the lowest expected respiratory rate and a low-pass filter having a higher cutoff frequency higher than the maximum expected respiratory rate. For example, the lower cutoff frequency of the bandpass filter (i.e., the cutoff frequency of the high-pass filter) can be in the range of 0.02Hz to 0.1Hz. In a preferred embodiment, the cutoff frequency is 0.083Hz, assuming that the lowest expected respiratory rate is greater than 5 breaths per minute, for example, 6 breaths per minute. The higher cutoff frequency of the bandpass filter (i.e., the cutoff frequency of the low-pass filter) can be in the range of 0.583Hz to 0.833Hz. In a preferred embodiment, the cut-off frequency is 0.667 Hz, assuming that the maximum expected respiratory rate is lower than 40 breaths per minute, for example 36 breaths per minute.

[0107] In this embodiment, the processor 3 is configured to determine the respiratory rate based on the respiratory pulse variation signal. Specifically, the processor 3 is configured to transform the signal from the time domain to the frequency domain and determine the frequency in the frequency domain. For the transformation from the time domain to the frequency domain, a corresponding transformation can be used, such as Fourier transform, in particular fast Fourier transform, wavelet transform, etc. For example, the corresponding frequency at the corresponding maximum value in the frequency domain (in particular within a predefined expected frequency range) can be determined as the corresponding respiratory rate. The predefined expected frequency range is the frequency range of the respiratory rate of the expected object.

[0108] The processor 3 can also be configured to determine additional features for the respective pressure pulse. In the following, the determination of additional features for the respective processed measured pressure pulse 29 will be described, wherein in this example the respective processed measured pressure pulse 29 is a respective TPac pulse curve.

[0109] For example, the difference TPP between the maximum measured pressure and the minimum measured pressure, ie the difference between the maximum systolic pressure of the pressure pulse (TPsys) and the pressure at the end diastolic phase of the pressure pulse (TPdia), can be determined, as Figure 10 As indicated.

[0110] The processor 3 can also be configured to determine the pulse duration (t(Pulse)) as the time difference between the end-diastolic phase of the pressure pulse 29 and the subsequent end-diastolic phase. This feature can also be defined as the time difference between the start of the corresponding pulse (t.start) and the end of the corresponding pulse (t.stop). This feature is Figure 11 Shown in.

[0111] The processor 3 can also be adapted to determine the pulse area (TPA) of the respective pulse 29, which is the area under the respective pulse curve over the time defined by t.start to t.stop and ranging from the pressure TPdia at the end diastole of the pressure pulse to the maximum systolic pressure TPsys of the pressure pulse. Preferably, the pulse area TPA is scaled to TPP=1, as Figure 11 The scaled pressure pulse area is named “TPA.norm”.

[0112] The processor 3 can also be adapted to determine the half-peak pulse width (W50) of the corresponding pulse 29, such as Figure 12 shown.

[0113] Furthermore, the processor 3 can be adapted to determine the area TPA+.top50 enclosed by the upper part of the pressure pulse 29, such as Figure 13Schematically shown in FIG. The upper end of the area TPA+.top50 enclosed by the upper portion is at the maximum systolic pressure TPsys of the pressure pulse 29, and the lower end of the upper area TPA+.top50 is between the maximum systolic pressure TPsys of the pressure pulse 29 and the pressure value of the average TPcl corresponding to the measured pressure TP. Because the pressure pulse 29 has been processed by subtracting the average TPcl from the measured pressure TP, the pressure value of the average TPcl corresponding to the measured pressure TP is zero. Preferably, the processor 3 is configured to determine the area TPA+.top50 enclosed by the upper portion of the pressure pulse so that the lower end of the upper portion is at a pressure value that is half the pressure distance TPP+ between the pressure values ​​corresponding to the maximum systolic pressure TPsys of the pressure pulse and the average TPcl corresponding to the measured pressure TP.

[0114] exist Figures 9 to 13 In the example, a reference numeral with a "T" as the first letter is used because, in this example, the measured pressure signal is a tissue pressure signal. However, as described above, instead of a tissue pressure signal, another pressure signal that indicates blood pulsation and includes multiple pressure pulses can also be measured. For example, the pressure signal can be measured invasively, and the above-described determination of the dicrotic notch position and physiological parameters can be applied to the pressure pulse of the invasively measured pressure signal.

[0115] The processor 3 can also be configured to determine a blood pressure determination value TPWP_M for the corresponding pressure pulse based on at least one determined feature, so that for several pressure pulses present at different times, several blood pressure determination values ​​TPWP_M are determined, wherein the several blood pressure determination values ​​TPWP_M determined for the several pressure pulses and therefore for the several times form a blood pressure determination curve TPW_M-curve. The processor 3 can be configured to determine the blood pressure based on the blood pressure determination curve TPW_M-curve, for example, as explained in WO 2018 / 210931A1. Specifically, the processor 3 can be configured to determine the position of the maximum value (TPW_M-curve.max) of the blood pressure determination curve TPW_M-curve, and determine the blood pressure based on the determined position and the measured pressure TP. For example, the processor 3 can be configured to determine the systolic arterial blood pressure (SAPni) based on the average TPcl of the measured pressure TP at the determined maximum position (TPW_M-curve.max) according to the following equation:

[0116] SAPni = α · (TPcl@TPW_M-curve.max), (4)

[0117] The parameter α can be predetermined through calibration.

[0118] In practice, the processor 3 can be configured to determine the position of the maximum of the blood pressure determination curve TPW_M-curve and / or the position of a derivative (such as a first-order derivative) of the blood pressure determination curve TPW_M-curve, and determine the blood pressure based on one or both of the determined position and the measured pressure TP. For example, the processor 3 can be configured to determine the systolic arterial blood pressure based on the average TPcl of the measured pressures TP at one or both of the determined positions according to SAPni=α'·(TPcl@TPW_M-curve.max)+β'·(TPcl@TPW_M-curve'.max), where TPW_M-curve' refers to the derivative of TPW_M-curve, and where the parameters α' and β' can be predetermined by calibration.

[0119] In one embodiment, the respiratory rate determination curve and / or the blood pressure determination curve are smooth curves, wherein the smoothing process for smoothing the respective physiological parameter determination curve can comprise, for example, filtering and / or fitting. To smooth the respective physiological parameter determination curve, a moving average filter, in particular a variable moving average filter, can be used, which is applied to the physiological parameter determination values, i.e., in this case the respiratory rate determination value and the blood pressure determination value for the pressure pulse determination. The window used for averaging can be fixed or variable, wherein in the latter case it preferably has a maximum duration of, for example, 8 seconds. The use of filters can result in filter delays, wherein the minimum filter delay is as long as the sum of the durations of several pulses.

[0120] The measurement of physiological parameters (ie, such as respiratory rate and blood pressure) is intended to be used in rapid succession in a series of measurements to allow for effective semi-continuous monitoring, minimizing stress on the monitored individual (ie, the monitored subject).

[0121] In the following, reference will be made to Figure 14 The shown flow chart exemplarily describes an embodiment of a method for determining the breathing rate of a subject.

[0122] In step 101, a measured pressure signal of the object over time is provided by a pressure signal providing unit, wherein, in this embodiment, the pressure signal has been measured by using a measuring device, the measuring device comprising: a) a housing configured to surround a portion of the object through which blood flows; b) a pressure applicator configured to apply pressure to the housing and thereby to the surrounded portion of the object; and c) a pressure sensor configured to measure a pressure signal on the skin of the surrounded portion of the object, wherein the pressure applicator increases or decreases the applied pressure when measuring the pressure signal, and wherein the measured pressure signal indicates blood pulsation and comprises a plurality of pressure pulses. In particular, by using the above reference Figures 1 to 3 The described measuring device is used to measure the pressure signal.

[0123] In step 102, for each pressure pulse from the plurality of pressure pulses, a corresponding feature characterizing the corresponding pressure pulse is determined, thereby determining a plurality of features for the plurality of pressure pulses. Furthermore, in step 102, a respiratory rate is determined based on the determined plurality of features. For each pressure pulse, one or more features, i.e., one or more different types of features, can be determined. The one or more features determined for each individual pressure pulse include at least one feature that depends on the location of the dicrotic notch.

[0124] The dicrotic notch position can be determined based on at least one of a notch determination intersection position, a notch determination maximum position, and a notch determination function position, which can be determined for the corresponding pressure pulse. In particular, the notch determination intersection position can be determined as the intersection of a shifted tangent on the corresponding pressure pulse, wherein the shifted tangent is determined by determining the tangent of the corresponding pressure pulse at the position where the first-order derivative of the corresponding pressure pulse has a minimum value and by shifting the tangent offset distance in the direction of increasing time. The notch determination maximum position can be determined as the position where the derivative function of the corresponding pressure pulse has a maximum value, wherein, in this regard, reference is made to the corresponding description provided above in more detail. Finally, the notch determination function position can be determined by determining the characteristics of the corresponding pressure pulse and by applying a position determination function to the characteristics, wherein the position determination function provides a relationship between the characteristics and the notch determination function position. For example, equation (3) can be used to determine the notch determination function position for the corresponding pressure pulse.

[0125] The notch-determined intersection position or the notch-determined maximum position or the notch-determined function position can directly be the dicrotic notch position to be determined, or at least one of the notch-determined intersection position, the notch-determined maximum position and the notch-determined function position can be combined, in particular averaged, to determine the dicrotic notch position. Furthermore, the dicrotic notch position determined for the corresponding pressure pulse can directly be a feature or can be processed to provide a corresponding feature. For example, as described above with respect to Figure 9 As explained, the determined dicrotic notch position t.notch can be used to determine the characteristic LVET and / or the characteristic TPA.sys.

[0126] In step 103 , the characteristics determined for several pressure pulses are used to determine a respiratory rate determination value TPWP_R forming a respiratory rate determination curve or a respiratory pulse variation signal, wherein the respiratory rate is determined based on the respiratory pulse variation signal as described above.

[0127] In step 104, it is determined whether a termination criterion is met. If so, the method stops in step 105, otherwise it continues with step 101. Thus, the method can be executed in a loop for continuously monitoring the respiratory rate and optionally one or more other physiological parameters (such as blood pressure) over time for several measurement cycles until the termination criterion is met. For example, the monitoring can be interrupted if a user, such as a physician, has entered a corresponding command into the device via an input unit, such as a keyboard, a computer mouse, a touchpad, or the like.

[0128] Although the features are determined using the TPac pulse as the processed pressure pulse in the above embodiment, the features can also be determined by directly using the measured pressure pulse (i.e., for example, the TP pulse). The pressure pulse can also be processed in a different manner, i.e., for example, other than by subtracting the average TP value used to determine the TPac pulse. For example, the pressure values ​​at t.start and t.stop for the respective measured pressure pulses can be connected by a straight line, and this straight line can be subtracted from the respective measured pressure pulses to determine the processed pressure pulse.

[0129] Although in the above embodiment, the pressure signal is a tissue pressure signal that has been measured by using a shell cuff, that is, although in the above embodiment, the pressure signal has been measured non-invasively, the pressure signal can also be measured invasively. In order to invasively measure the pressure signal, a known corresponding measuring device can be used, such as the measuring device described in the article "How to measure blood pressure using an arterial catheter: a systematic 5-step approach" by B. Saugel et al. (Critical Care, 24: 172 (2020)), which is incorporated herein by reference. Other known techniques can also be used to invasively measure the pressure signal. After the dicrotic notch position has been determined for the invasive measurement of the pressure pulse, the dicrotic notch position can be used to calculate the respiratory rate and optionally other physiological parameters. For example, as also explained above with respect to the non-invasive measurement of the pulse, LVET and / or TPA.sys can be calculated as physiological parameters and can also be used to calculate the respiratory rate. In addition, pulse contour stroke volume (PCSV) can be determined based on invasive pressure pulse and dicrotic notch location, for example as described in WO 2019 / 211210 A1, which is incorporated herein by reference.

[0130] The measured pressure pulse can also be a synthetic measured pressure pulse, for example, a pressure pulse composed of a non-invasively measured tissue pressure pulse or an invasively measured tissue pressure pulse. For example, a synthetic pressure pulse can be generated from a non-invasive pressure pulse with the aim of generating an equivalent of an invasively measured arterial pressure pulse. Therefore, all parameters that can be determined based on the invasively measured arterial pressure pulse can also be determined from the synthetic pressure pulse. In one example, the synthetic pressure pulse can be generated by weighting the non-invasive pressure pulses and by adding the weighted non-invasive pressure pulses. In particular, the synthetic pressure pulse can be generated by weighted averaging two or more consecutive non-invasive pressure pulses. In one embodiment, the synthetic pressure pulse is generated as described in EP2759258A1, in particular as described in claim 1 of EP 2759258A1, which is incorporated herein by reference.

[0131] Although in the embodiments described above the respiratory rate and optionally the blood pressure have been determined as physiological parameters, the device may also determine other physiological parameters, such as a fluid responsiveness parameter.

[0132] Although in the embodiments described above for determining the respiratory rate, the area under the corresponding pressure pulse between the starting position where the corresponding pressure pulse begins and the determined dicrotic notch position is used, for determining the respiratory rate, other dicrotic notch-related parameters can also be used, such as the area under the corresponding pressure pulse between i) the position between the starting position and the determined dicrotic notch position (for example, the position where the corresponding pressure pulse reaches its maximum value or the position where the increase (i.e., the first-order derivative of the pressure pulse) has its maximum value) and ii) the dicrotic notch position.

[0133] Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims.

[0134] In the claims, the word "comprising" does not exclude other elements or steps, and the word "a" or "an" does not exclude a plurality.

[0135] A single unit or device may fulfill the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.

[0136] Calculations like determining pulse characteristics, certain curves, respiratory rate, etc., which are performed by one or several units or devices, can be performed by any other number of units or devices. The calculations and determination and / or control of the apparatus for determining the respiratory rate of a subject according to the method for determining the respiratory rate of a subject can be implemented as program code modules of a computer and / or as dedicated hardware.

[0137] The computer program can be stored / distributed on a suitable medium, such as an optical storage medium or a solid-state medium, provided together with or as a part of other hardware, but can also be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems.

[0138] Any reference signs in the claims such as TPWP_R, t.start, t.notch or t.shift shall not be construed as limiting the scope.

[0139] The present invention relates to an apparatus for determining a subject's respiratory rate. A pressure signal providing unit provides a measured pressure signal of the subject over time, wherein the measured pressure signal indicates blood pulsation and includes a plurality of pressure pulses. A processor determines a corresponding dicrotic notch position for a corresponding pressure pulse in the plurality of pressure pulses, and determines the respiratory rate based on the dicrotic notch positions determined for the plurality of pressure pulses. This allows for highly accurate determination of the respiratory rate.

Claims

1. An apparatus for determining a subject's respiratory rate, the apparatus comprising: - a pressure signal providing unit configured to provide a measured pressure signal of the subject over time, wherein the measured pressure signal is indicative of blood pulsation and comprises a plurality of pressure pulses (9), - a processor configured to determine a respective dicrotic notch position for a respective pressure pulse of the plurality of pressure pulses (9) and to determine the respiratory rate based on the dicrotic notch positions determined for the plurality of pressure pulses (9).

2. The device according to claim 1, wherein The processor is configured to determine a respiratory rate determination value (TPWP_R) for the corresponding pressure pulse based on the determined dicrotic notch position, so that several respiratory rate determination values ​​(TPWP_R) are determined for several pressure pulses present at different times, wherein the several respiratory rate determination values ​​(TPWP_R) determined for the several pressure pulses and therefore for several times form a respiratory rate determination curve, and the respiratory rate is determined based on the respiratory rate determination curve.

3. The device according to claim 2, wherein The processor is configured to determine, in order to determine a corresponding respiratory rate determination value (TPWP_R) for a corresponding pressure pulse, at least one of: - an area under the respective pressure pulse between i) a first position, being a starting position (t.start) where the respective pressure pulse begins, or a position between the starting position (t.start) and the determined dicrotic notch position (t.notch), and ii) a determined dicrotic notch position (t.notch), and - The difference between the first position and the dicrotic notch position (t. notch).

4. The device according to claim 3, wherein The processor is configured such that the corresponding respiratory rate determination value (TPWP_R) is determined by applying a predefined function to at least one of: a) the area under the corresponding pressure pulse between the first position and the determined dicrotic notch position (t.notch), and b) the difference between the first position (t.start) and the dicrotic notch position (t.notch).

5. The device according to claim 3, wherein The processor is configured such that the respective respiratory rate determination value (TPWP_R) is the same as or linearly related to an area under the respective pressure pulse between the first location where the respective pressure pulse begins and a determined dicrotic notch location (t.notch).

6. The device according to any one of claims 2 to 5, wherein: The processor is configured to determine the respiratory rate based on a frequency of the respiratory rate determination curve.

7. The device according to any one of claims 2 to 6, wherein: The processor is configured to apply a frequency range pre-selection filter to the respiration rate determination curve.

8. A device according to any one of the preceding claims, wherein The processor is configured to determine the corresponding dicrotic notch location for the corresponding pressure pulse by determining at least one of: a) a notch-determined intersection position, the notch-determined intersection position being the position where a shifted tangent line (52) intersects the pressure pulse, wherein the shifted tangent line (52) is determined by determining a tangent line (51) to the pressure pulse at a position where a first derivative of the pressure pulse has a minimum value, and by shifting the tangent line (51) by an offset distance (t.shift) in a direction of increasing time, b) a notch-determined maximum position, the notch-determined maximum position being the position at which a derivative function has a maximum, wherein the derivative function is determined by determining a first derivative of the pressure pulse, by determining a second derivative of the pressure pulse and by combining the determined first and second derivatives, and c) a notch determination function position, the notch determination function position being determined by determining a characteristic of the pressure pulse and by applying a position determination function to the characteristic, wherein the position determination function provides a relationship between the characteristic and the notch determination function position.

9. The device according to claim 8, wherein The processor is configured to provide the shift distance (t.shift) such that the shift distance depends on a characteristic of the pressure pulse (50).

10. The device according to any one of claims 8 and 9, wherein The processor is configured to: if there are several intersection positions of the displacement tangent line and the pressure pulse, select the temporally latest intersection position as the notch determination intersection position.

11. The device according to any one of claims 8 to 10, wherein The processor is configured to determine the notch-determined maximum position by determining whether the highest maximum value of the derivative function is greater than the second highest maximum value of the derivative function multiplied by a predefined factor, wherein, if this is the case, the notch-determined maximum position is the position of the highest maximum value, and if this is not the case, the notch-determined maximum position is the position of the first maximum value of the derivative function.

12. The device according to any one of claims 8 to 11, wherein The processor is configured to determine the derivative function by dividing a) a second-order derivative function that depends on the second-order derivative by b) a first-order derivative function that depends on the first-order derivative.

13. The device according to any one of claims 8 to 12, wherein The processor is configured to provide the position determination function such that it provides a relationship between the position of the notch determination function and at least one of the following characteristics of the pressure pulse: a maximum value (Pulse.max) of the pressure pulse (50), a maximum position (t.max) along the pressure pulse (50), a first position (t.dia.pre) along the pressure pulse (50) before the maximum position (t.max) and at which the pressure pulse (50) has a value that is a predefined first fraction (p1) of the maximum value (Pulse.max), a second position (t.dia.post) along the pressure pulse (50) after the maximum position (Pulse.max) and at which the pressure pulse (50) has a value that is a predefined second fraction (p2) of the maximum value (Pulse.max), a maximum value of the first derivative of the pressure pulse (50) (d(Pulse) / dt.max), a difference (PP) between the maximum systolic pressure of the pressure pulse and the pressure at the end diastolic phase of the pressure pulse.

14. A method for determining a subject's respiratory rate, the method comprising: - providing a measured pressure signal of the subject over time by a pressure signal providing unit, wherein the measured pressure signal is indicative of blood pulsation and comprises a plurality of pressure pulses (9), - determining, by a processor, a corresponding dicrotic notch position for a corresponding pressure pulse of the plurality of pressure pulses (9), - determining the respiratory rate based on the dicrotic notch locations determined for the plurality of pressure pulses (9).

15. A computer program for determining a breathing rate of a subject, the computer program comprising program code means for causing the apparatus for determining the breathing rate according to claim 1 to perform the steps of the method according to claim 14.

Images