Respiration pattern determination method and respiration pattern determination device

The breathing pattern determination method uses wavelet transformation to classify respiratory conditions, addressing the challenge of accurate classification beyond abnormal phase detection, enhancing the understanding of respiratory states.

JP2024166690APending Publication Date: 2024-11-29YAMAHA CORP +1
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
JP2023082978
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Filing Date
2023-05-19
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

Existing methods struggle to accurately classify respiratory conditions beyond the presence of abnormal respiratory phases, often requiring expert judgment.

Method used

A breathing pattern determination method utilizing wavelet transformation of lung expansion and contraction data to identify periodic breathing patterns through wavelet spectra analysis, including time domain division and similarity calculations.

Benefits of technology

Enables detailed classification of respiratory conditions, allowing for the detection of abnormal phases and understanding patient conditions, such as complications post-unilateral lung resection surgery.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2024166690000001_ABST
    Figure 2024166690000001_ABST
Patent Text Reader

Abstract

To provide a respiration pattern determination method that enables classification of a respiration state by grasping the respiration state in detail from expansion or contraction of a lung based on respiration.SOLUTION: A respiration pattern determination method of the present disclosure includes: a measurement step S1 of measuring a temporal variation value of expansion or contraction of a lung; a conversion step S2 of obtaining a wavelet spectrum by wavelet conversion of the temporal variation value; and a determination step S3 of determining a cyclic respiration pattern on the basis of the wavelet spectrum.SELECTED DRAWING: Figure 1
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present disclosure relates to a breathing pattern determination method and a breathing pattern determination device. [Background technology]

[0002] By detecting the expansion and contraction of the lungs due to breathing, it may be possible to obtain information on various biological activities associated with breathing. For example, the expansion and contraction of the lungs may enable early detection of complications of abnormal respiratory phases that occur after one-sided lung resection surgery. As a sensor unit capable of detecting the expansion and contraction of the surface of a living body, a sensor unit using multiple thread-like or band-like strain sensor elements that expand and contract in the longitudinal direction is known (see JP 2020-151294 A).

[0003] The sensor unit is held by a holding member so that the central axes of at least one pair of the strain sensor elements intersect with each other, making it possible to easily detect movement of the surface of the living body in a simple manner that places less strain on the subject. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2020-151294 A Summary of the Invention [Problem to be solved by the invention]

[0005] The detection of an abnormal respiratory phase using the sensor unit is performed, for example, according to the following procedure. First, a time-displacement curve of elapsed time and frequency displacement is obtained from the detection signal of the sensor unit. Next, the envelope of this time-displacement curve is obtained, and the envelope is transformed by a Fourier transform or the like in a predetermined time unit to obtain frequency characteristics per unit time. The presence or absence of an abnormal respiratory phase is determined from this frequency characteristic.

[0006] Although the above procedure can determine the presence or absence of an abnormal breathing phase, it is difficult to determine what type of abnormal breathing is occurring. Generally, cases that are extracted are those in which an abnormal breathing phase is suspected, and ultimately, a skilled physician's judgment is often required.

[0007] The present disclosure has been made in light of the above circumstances, and aims to provide a breathing pattern determination method and a breathing pattern determination device that are capable of grasping the breathing condition in detail from the expansion and contraction of the lungs due to breathing and classifying the breathing condition. [Means for solving the problem]

[0008] (1) A breathing pattern determination method according to one aspect of the present disclosure includes a measurement step of measuring a time variation value of lung expansion / contraction, a transformation step of obtaining a wavelet spectrum by wavelet transform of the time variation value, and a determination step of determining a periodic breathing pattern based on the wavelet spectrum.

[0009] (2) In the breathing pattern determination method of (1) above, the periodic breathing pattern may be determined using a time change in a peak frequency of a wavelet spectrum.

[0010] (3) In the breathing pattern determination method of (1) or (2), the determination step may include a time domain division step of dividing the wavelet spectrum into a plurality of regions at a constant time interval, and a similarity calculation step of calculating a similarity of the wavelet spectrum among the plurality of regions, and the similarity may be used to determine the periodic breathing pattern.

[0011] (4) In the breathing pattern determination method of (3) above, the similarity may be a correlation coefficient.

[0012] (5) In any of the breathing pattern determination methods (1) to (4) above, a strain sensor element that is placed on the surface of the living body and can detect the expansion and contraction of the living body surface in response to the expansion and contraction of the lungs may be used to measure the change over time of the lung expansion and contraction in the measurement step.

[0013] (6) Any one of the breathing pattern determination methods (1) to (5) may be used to determine an abnormal breathing phase.

[0014] (7) A breathing pattern determination device according to another aspect of the present disclosure includes a measurement unit that measures a time variation value of lung expansion and contraction, a transformation unit that obtains a wavelet spectrum by wavelet transform of the time variation value, and a determination unit that determines a periodic breathing pattern based on the wavelet spectrum. Effect of the Invention

[0015] The breathing pattern determination method and breathing pattern determination device disclosed herein focus on frequency changes over time in the respiratory waveform and extract repetitive patterns, thereby enabling detailed classification of breathing patterns and enabling a detailed understanding of the respiratory condition. [Brief description of the drawings]

[0016] [Figure 1] FIG. 1 is a flow diagram showing a flow of a breathing pattern determination method according to an embodiment of the present disclosure. [Diagram 2] FIG. 2 is a schematic configuration diagram showing a breathing pattern determination device according to an embodiment of the present disclosure, which is used in the breathing pattern determination method of FIG. [Diagram 3] FIG. 3 is a schematic plan view (outer surface view) showing the sensor unit of FIG. [Figure 4] FIG. 4 is a schematic diagram showing a state in which the sensor unit of FIG. 2 is attached to a living body. [Diagram 5] FIG. 5 is a graph showing an example of a time variation curve obtained by the measurement unit of FIG. [Figure 6] FIG. 6 is a graph illustrating the wavelet spectrum obtained in the transform unit of FIG. [Figure 7] FIG. 7 is a flow diagram showing the flow of the determination step in FIG. [Figure 8] FIG. 8 is an explanatory diagram for explaining the time domain division step of FIG. [Figure 9] FIG. 9 is an explanatory diagram for explaining the pattern extraction step in FIG. [Figure 10] FIG. 10 shows an example of a breathing pattern determined by the breathing pattern determination method of FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0017] Hereinafter, the embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. Note that, for the numerical values ​​described in this specification, only one of the upper limit value and the lower limit value described may be adopted, or the upper limit value and the lower limit value may be arbitrarily combined.

[0018] The breathing pattern determination method shown in Fig. 1 includes a measuring step S1, a converting step S2, and a determining step S3. The breathing pattern determination method can be performed using a breathing pattern determination device 1 shown in Fig. 2.

[0019] The breathing pattern determination device 1 includes a sensor unit 10, a measuring unit 20, a converting unit 30, a determining unit 40, and a display unit 50. The measuring unit 20, the converting unit 30, and the determining unit 40 are partially or entirely composed of a computer having a CPU, a ROM, a RAM, a large-capacity storage device, etc. The "computer" here includes not only a computer server and a personal computer, but also an embedded system such as a microcomputer board. The breathing pattern determination device 1 does not necessarily require a large amount of calculation, so it can also be realized by a microcomputer board, which is a lightweight application.

[0020] The breathing pattern determination method and the breathing pattern determination device 1 are preferably used for determining an abnormal breathing phase. By using the breathing pattern determination method, for example, complications of an abnormal breathing phase occurring after a unilateral lung resection can be found, and the pathological condition of the patient can be understood in detail. Hereinafter, the description will be continued by taking the determination of an abnormal breathing phase as an example, but this does not mean that the breathing pattern determination method is limited to the determination of an abnormal breathing phase.

[0021] Each step of the breathing pattern determination method will be described below together with the corresponding parts of the breathing pattern determination device 1.

[0022] <Measurement steps> In the measurement step S1, the time variation value of the expansion and contraction of the lungs is measured. This measurement step S1 is performed by the measurement section 20 using the sensor unit 10 of the breathing pattern determination device 1. That is, the measurement section 20 measures the time variation value of the expansion and contraction of the lungs.

[0023] (sensor unit) The sensor unit 10 has a strain sensor element 11 placed on the surface of the living body. In the breathing pattern determination method, the strain sensor element 11 capable of detecting the expansion and contraction of the living body surface according to the expansion and contraction of the lungs is used to measure the time change in the expansion and contraction of the lungs in the measurement step S1. Specifically, as shown in FIG. 3, the sensor unit 10 includes a pair of thread-like or band-like strain sensor elements 11 that expand and contract in the longitudinal direction, and a holding member 12 that is connected to the pair of strain sensor elements 11 and determines the positional relationship of the strain sensor elements 11 with respect to the living body surface. The central axes M of the pair of strain sensor elements 11 intersect with each other. Note that the term "band-like" means an elongated shape with a large width compared to the thickness, and includes a configuration in which the thickness and width are partially different.

[0024] The sensor unit 10 is configured to be capable of detecting the expansion and contraction of the biological surface that changes in response to the movement of the biological body by measuring the resistance value at both ends of the strain sensor element 11. The sensor unit 10 is unlikely to pressurize the biological surface of the subject X because the pair of strain sensor elements 11 expand and contract in the longitudinal direction in response to the expansion and contraction of the biological surface. Therefore, the sensor unit 10 is unlikely to cause the subject X a sense of strangeness or discomfort, and can detect natural movements of the biological surface of the subject X.

[0025] The strain sensor elements 11 are arranged in a straight line. Both ends of the strain sensor elements 11 are fixed to the holding member 12. The both ends of the strain sensor elements 11 may be fixed to the holding member 12 via a metal support member, or may be directly fixed to the holding member 12 with an adhesive or the like.

[0026] The strain sensor element 11 may be any element that has elasticity in the longitudinal direction and whose electrical characteristics change in response to elasticity, and is preferably a strain resistance element whose electrical resistance changes in response to elasticity. Among them, the strain sensor element 11 is particularly preferably a CNT strain sensor using carbon nanotubes (hereinafter also referred to as "CNT").

[0027] As shown in Fig. 4, a pair of strain sensor elements 11 are placed on the surface of the chest of a living body when in use. The pair of strain sensor elements 11 are capable of independently detecting the movements of the left and right rib cages, and are placed symmetrically apart from each other. The pair of strain sensor elements 11 are placed inclined outwardly in opposing directions from top to bottom. The pair of strain sensor elements 11 are preferably placed across the sixth and eighth ribs so as to detect the expansion and contraction of the area of ​​the rib cage that deforms most greatly.

[0028] The lower limit of the angle α (see FIG. 3) between the central axes M of the pair of strain sensor elements 11 is preferably 30°, more preferably 50°, and even more preferably 70°. On the other hand, the upper limit of the angle α is preferably 120°, more preferably 100°, and even more preferably 90°. If the angle α is outside the above range, it may be difficult to arrange the longitudinal direction of the pair of strain sensor elements 11 along the direction in which the thorax of the subject X deforms the most.

[0029] The holding member 12 can be composed of, for example, a stretchable fabric 13 , and an adhesive portion 14 and a reinforcing member 15 laminated on the fabric 13 .

[0030] The fabric 13 is configured to be capable of covering the surface of a living body. A pair of strain sensor elements 11 are arranged on the outer surface of the fabric 13 (the surface opposite to the side facing the surface of a living body). The adhesive portion 14 is configured to be capable of adhering the fabric 13 to the surface of a living body. The reinforcing member 15 is arranged at the end of the strain sensor element 11 in a plan view. The reinforcing member 15 can be fixed to the fabric 13 by bonding with an adhesive. When the reinforcing member 15 includes a woven fabric, it may be sewn on. The reinforcing member 15 may be layered on either the inner or outer surface of the fabric 13.

[0031] (Measuring section) The measurement unit 20 can be configured, for example, with an A / D converter that converts the detection signal by the pair of strain sensor elements 11 into an electric signal. The A / D converter can convert the resistance values ​​on both ends of the pair of strain sensor elements 11 into an electric signal. The A / D converter is controlled by the computer to obtain the electric signal.

[0032] The expansion and contraction of the strain sensor element 11 accompanying the expansion and contraction of the surface of the living body in response to the expansion and contraction of the lungs brings about a change in the resistance value on both ends of the strain sensor element 11, so that the electrical signal (e.g., voltage) obtained by this A / D converter represents the expansion and contraction of the strain sensor element 11. Therefore, by continuously measuring the strength of this electrical signal, it is possible to measure the time-varying value of the expansion and contraction of the strain sensor element 11 (e.g., FIG. 5).

[0033] In the breathing pattern determination method, as described later, a repetitive pattern of a breathing waveform is extracted, and when a change occurs in the breathing pattern, it is possible to determine that, for example, a change has occurred in the condition of the subject X (details will be described later). Therefore, the breathing pattern determination method requires a time variation value that can include at least two breathing patterns. Since it is generally known that breathing patterns are often repeated in a cycle of 40 seconds or more and 120 seconds or less, the lower limit of the measurement time of the time variation value is preferably 80 seconds, and more preferably 260 seconds. On the other hand, as long as two or more breathing patterns can be included, the upper limit of the measurement time of the time variation value is not particularly limited. For example, in the case of subject X after a lung resection operation, the time variation value may be continuously obtained at all times for a period until the subject X recovers from the operation. In addition, since the repetitive pattern of the most recent breathing waveform is highly important, processing such as deleting old data that has been used for a certain period of time or more may be performed from the viewpoint of reducing the storage capacity of data, for example.

[0034] <Conversion steps> In the conversion step S2, a wavelet spectrum is obtained by performing a wavelet transform on the time variation value. This conversion step S2 is performed by the conversion unit 30 of the breathing pattern determination device 1. That is, the conversion unit 30 obtains a wavelet spectrum by performing a wavelet transform on the time variation value.

[0035] The wavelet transform is one of the techniques for frequency analysis of signals. While the Fourier transform, which is a typical example of frequency analysis, loses information related to time, the wavelet transform retains the information related to time, and therefore it is possible to simultaneously extract information related to time and frequency. That is, by performing the wavelet transform, it is possible to know the intensity (represented by the wave height in FIG. 6) of a certain frequency (vertical axis in FIG. 6) at a certain time (horizontal axis in FIG. 6), for example, as shown in FIG. 6. In the breathing pattern determination device 1, the normalized complex wavelet transform (NCWT) can be suitably used as the wavelet transform technique.

[0036] The conversion unit 30 may include a high-pass filter. In other words, it is preferable not to perform wavelet conversion on frequency components below a certain frequency. In principle, if the strain sensor elements 11 are the same in terms of expansion and contraction, the same expansion and contraction signal should be obtained, but slight fluctuations may occur at low frequencies. In addition, for example, when the subject X turns sideways, the lower strain sensor element 11 tends to expand and the upper strain sensor element 11 tends to contract, and the tendency of the expansion and contraction of the strain sensor elements 11 may change depending on the behavior of the subject X. The high-pass filter can remove components caused by the fluctuations and changes in the tendency of expansion and contraction, thereby improving the judgment accuracy of the breathing pattern judgment device 1.

[0037] The wavelet transform is performed for a certain frequency range. The frequency range is preferably 0.1 Hz to 0.8 Hz, and more preferably 0.05 Hz to 1 Hz. A change in the breathing pattern during an abnormal breathing phase is likely to appear in a change in the wavelet spectrum within the frequency range. Therefore, by performing the wavelet transform within the frequency range, it is possible to obtain a large effect with a small amount of data. Furthermore, by performing the wavelet transform for a frequency equal to or higher than the lower limit of the frequency, the device is less susceptible to the influence of components caused by the fluctuations of the strain sensor element 11 and changes in the tendency of expansion and contraction, and the determination accuracy of the device 1 is improved. Furthermore, one cycle of the frequency is more likely to include one breathing motion, and from this viewpoint, the device 1 is more likely to determine the accuracy of the device 1. In addition, by performing the wavelet transform for a frequency equal to or lower than the upper limit of the frequency, it is easier to prevent other biological signals from being mixed in, and the device 1 is more likely to determine the accuracy of the device 1. Furthermore, the device 1 is more likely to determine the accuracy of the device 1 when the frequency range is wider. Therefore, from the viewpoint of a wide range, the frequency range may be selected to be 0.1 Hz or more and 1 Hz or less, or 0.05 Hz or more and 0.8 Hz or less.

[0038] <Judgment step> In the determination step S3, a periodic breathing pattern is determined based on the wavelet spectrum. This determination step S3 is performed by the determination unit 40 of the breathing pattern determination device 1. That is, the determination unit 40 determines a periodic breathing pattern based on the wavelet spectrum.

[0039] As shown in FIG. 7, the decision step S3 includes a maximum value extraction step S31, a time domain division step S32, a similarity calculation step S33, and a pattern extraction step S34.

[0040] (Maximum value extraction step) In this breathing pattern determination method, the periodic breathing pattern is determined using the time change of the peak frequency of the wavelet spectrum. Therefore, in the maximum value extraction step S31, the peak frequency at each time is extracted from the wavelet spectrum obtained in the conversion step S2. By using the time change of the peak frequency of the wavelet spectrum to determine the periodic breathing pattern, the breathing state can be classified more accurately.

[0041] The peak frequency is the frequency at which the intensity of the wavelet spectrum at a certain time is maximum. Figure 6 shows a schematic diagram of the wavelet spectrum, representing the wavelet spectrum at each time. For example, the maximum value P indicates the peak frequency of the rightmost wavelet spectrum. The peak frequency is defined for each time, and is therefore extracted as a quantity dependent on time.

[0042] For example, there are cases where a significant change in peak frequency over time cannot be extracted because multiple peaks are detected or no peak exists. In such cases, it is determined that no periodic breathing pattern exists (this pattern is referred to as "Pattern 1"). The cause of determining that it is Pattern 1 may be that the breathing of the subject X itself is abnormal, but it may also be a problem with the measurement environment, such as when the strain sensor element 11 is detached from the subject X.

[0043] If it is determined to be pattern 1, the following process is not performed and the determination ends.

[0044] (Time domain decomposition step) In the time domain division step S32, the wavelet spectrum is divided into a plurality of domains at regular time intervals.

[0045] In this breathing pattern determination method, the determination is made using the time change of the peak frequency in the wavelet spectrum. For example, when the time change of the peak frequency shown in Fig. 8 is obtained, this is divided into a plurality of regions R1 to R6 at intervals of, for example, 30 seconds (see Fig. 8). Note that this 30 second interval is merely an example, and since a plurality of correlation coefficients are extracted by changing the time interval as described later, it is possible to start from any value.

[0046] The range of the wavelet spectrum divided in the time domain division step S32 is preferably within a recent fixed time. By determining what is within the recent fixed time, the latest condition of the subject X can be determined. In addition, it is possible to determine an abnormal condition that occurs suddenly in the subject X without overlooking it. The above-mentioned fixed time is preferably 80 seconds or more, which is enough to include at least two breathing patterns.

[0047] (Similarity calculation step) In a similarity calculation step S33, the similarities of the wavelet spectra between the multiple regions R1 to R6 are calculated.

[0048] The similarity is not particularly limited as long as it is an index that can determine whether the wavelet spectra between the multiple regions R1 to R6 have similarity, but it is preferable to use a correlation coefficient. By using the correlation coefficient, the similarity can be easily and accurately determined. In the following, the explanation will be continued by taking the case where the similarity is a correlation coefficient as an example.

[0049] A known method can be used to calculate the correlation coefficient. For the regions R1 to R6, the average value of the correlation coefficients for all combinations of two regions may be calculated, or the average value of the correlation coefficients for two adjacent regions may be calculated, or the calculation may be performed by another method.

[0050] One correlation coefficient is calculated for the time interval set in the time domain division step S32. Therefore, the time interval in the time domain division step S32 is changed to calculate multiple correlation coefficients using the time interval as a parameter. That is, the time domain division step S32 and the correlation coefficient calculation step S33 are performed repeatedly. By performing the steps repeatedly, the result shown in FIG. 9 is calculated.

[0051] (Pattern extraction step) In the pattern extraction step S34, a periodic breathing pattern is determined. In this breathing pattern determination method, the similarity (correlation coefficient in this case) is used to determine the periodic breathing pattern. By using the similarity to determine the periodic breathing pattern, it is possible to efficiently determine whether the breathing pattern has periodicity and, if periodic, simultaneously extract one cycle of the breathing pattern.

[0052] Specifically, the time interval at which the correlation coefficient is maximized is obtained. In the example shown in Fig. 9, the correlation coefficient is maximized when the time interval is 63 seconds. By dividing the wavelet spectrum into a plurality of regions at a constant time interval using this time interval and obtaining the average value of the wavelet spectrum at corresponding times between the plurality of regions, for example, a periodic breathing pattern can be obtained.

[0053] As pattern 2 in Fig. 10, three examples of the change in peak frequency of the wavelet spectrum over time and the resulting periodic breathing patterns are shown. In this way, different periodic breathing patterns can be obtained by the wavelet spectrum.

[0054] These three examples are actually obtained changes in peak frequency over time and periodic breathing patterns, and are some of the typically seen patterns. More specifically, the upper part of pattern 2 is a pattern in which the peak frequency (vertical axis of the graph) constantly changes within a period. Specifically, in the upper part of pattern 2, the peak frequency rises in the first half of the period and falls in the second half, forming a triangular wave shape. The middle part of pattern 2 is a square wave shape that goes back and forth between two frequencies at regular time intervals. The lower part of pattern 2 attenuates from a specific frequency to a low frequency (around 0.05 Hz, the lower limit of analysis), and then recovers to the specific frequency again. In other words, it is a sawtooth wave shape. Thus, although pattern 2 has in common the fact that it exhibits periodic behavior, the changes in one period are various and have a unique pattern. It is considered that this unique pattern appears depending on the state of subject X.

[0055] Here, it is assumed that the time interval at which the correlation coefficient is maximized is not calculated, that is, there is no portion at which the correlation coefficient is maximized, and the correlation coefficient remains almost constant regardless of the time interval. In this case, the correlation coefficient may be constant at a high value (close to 1, e.g., greater than 0.5) (Pattern 3) or constant at a low value (e.g., 0.5 or less) (Pattern 4).

[0056] When the pattern belongs to pattern 3, the peak frequency is considered to be almost constant regardless of time. In this case, correlation is always obtained regardless of the time interval (see pattern 3 in Figure 10). This pattern is generally observed during normal breathing.

[0057] When the pattern belongs to pattern 4, it is considered that the change in the peak frequency occurs randomly, and the breathing pattern has no periodicity (see pattern 4 in FIG. 10). This pattern is generally observed during abnormal breathing.

[0058] (Display) The periodic breathing pattern determined by the determining unit 40 is displayed on the display unit 50, for example.

[0059] The display unit 50 is configured with, for example, a liquid crystal display. In addition to the periodic breathing pattern, the display unit 50 may display information such as the time variation value obtained by the measurement unit 20, the wavelet spectrum obtained by the conversion unit 30, and the time change of the peak frequency obtained by the judgment unit 40. The display of these information is controlled by the computer.

[0060] Fig. 10 summarizes the types of periodic breathing patterns determined by the breathing pattern determination method and the breathing pattern determination device 1. In the case of pattern 1, since the time change of the peak frequency is not extracted, it is considered that a problem has occurred in the measurement environment, for example, when the strain sensor element 11 is removed from the subject X. In the case of pattern 3, it can be determined that breathing is normal, and in the case of pattern 2 or pattern 4, it is determined that abnormal breathing is occurring. Abnormal breathing is divided into pattern 4, in which no periodicity is observed in breathing, and pattern 2, in which a periodic breathing pattern is observed.

[0061] In pattern 2, each unique pattern is determined as a periodic breathing pattern, and therefore, by observing the periodic breathing pattern of pattern 2 in particular, the severity and clinical condition of subject X can be judged. For this reason, an alarm that notifies by voice or the like that an abnormal breathing pattern has appeared may be provided in place of or together with the display unit 50. The operation of the alarm is also controlled by the computer.

[0062] In order to determine which unique pattern of pattern 2 the observed periodic breathing pattern corresponds to, a known pattern matching method such as machine learning may be used.

[0063] <Advantages> The breathing pattern determination method and breathing pattern determination device 1 disclosed herein focus on frequency changes over time in the respiratory waveform and extract the repetitive patterns, thereby classifying the breathing patterns in detail and enabling a detailed understanding of the respiratory condition.

[0064] [Other embodiments] The above-described embodiments do not limit the configuration of the present disclosure. Therefore, the above-described embodiments may include omissions, substitutions, or additions of components of each part of the embodiments based on the description in this specification and common technical knowledge, and all of these should be interpreted as belonging to the scope of the present disclosure.

[0065] In the above embodiment, the breathing pattern determination method and breathing pattern determination device are described as being used to determine an abnormal breathing phase. However, the breathing pattern determination method and breathing pattern determination device of the present disclosure can be applied to various biological activities in which insights can be obtained by determining whether an obtained breathing waveform has periodicity and, if periodicity is present, the type of breathing pattern.

[0066] An example of such biological activity is playing an orchestral instrument. By extracting the unique characteristics and psychological state of the player from the presence or absence of periodicity in the breathing pattern when playing an orchestral instrument and the periodic pattern, it is possible to use this information to improve the player's skill.

[0067] In the above embodiment, the pair of strain sensor elements for detecting the expansion and contraction of the lungs are arranged symmetrically and spaced apart from each other, but the arrangement of the strain sensor elements is not limited to this. For example, the strain sensor elements may be arranged in a circle around the chest to detect the change in chest circumference accompanying the expansion and contraction of the lungs.

[0068] In the above embodiment, the strain sensor element is installed on the surface of the living body via a holding member. However, the strain sensor element may be directly attached to the living body.

[0069] Furthermore, the means for detecting lung expansion / contraction is not limited to the strain sensor element, but may be, for example, a camera or the like that captures the movement of the chest and performs image analysis to detect the expansion / contraction of the lungs.

[0070] In the above embodiment, the determination step includes the averaging step, but the averaging step can be omitted. In other words, the present invention also intends to use the wavelet spectrum obtained in the transformation step as it is.

[0071] In the above embodiment, the periodic breathing pattern is determined using the time change of the peak frequency of the wavelet spectrum, but the amount used for determining the periodic breathing pattern is not limited to the time change of the peak frequency of the wavelet spectrum. For example, the periodic breathing pattern may be determined using the intensity of the wavelet spectrum.

[0072] In the above embodiment, the determination step includes a time domain division step and a similarity calculation step, and the correlation coefficient is used as an example of the similarity, but other methods can be used to calculate the correlation coefficient. Furthermore, the method of determining a periodic breathing pattern is not limited to the method using the correlation coefficient and thus the similarity.

[0073] The determination step of the embodiment may first include an averaging step. In this averaging step, the wavelet spectrum is averaged for each region divided two-dimensionally along the time axis and the frequency axis. That is, the frequency range in which the wavelet transform is performed along the frequency axis is divided into n regions, and the time range in which the wavelet transform is performed along the time axis is divided into m regions, and the average value of the wavelet spectrum intensity is calculated for each of the n×m regions. By performing such averaging, it is possible to easily extract a repeating pattern.

[0074] When performing the averaging process for each region, the intensity of the wavelet spectrum at any position within the region is required. If the wavelet spectrum is calculated discretely, the averaging process may be performed using the most recent discrete value as a representative value, or the representative value may be calculated by interpolation process from multiple nearby discrete values.

[0075] The number m of divisions of the time range is determined by the time width of one region, with the lower limit of the time width of one region being preferably 1 second, and more preferably 2 seconds. Meanwhile, the upper limit of the time width of one region being preferably 5 seconds, and more preferably 4 seconds. By setting the time width of one region to be equal to or greater than the lower limit, noise components superimposed on the wavelet spectrum can be removed, thereby improving the accuracy of the similarity calculated in similarity calculation step S33, which will be described later. Conversely, by setting the time width of one region to be equal to or less than the upper limit, it is possible to prevent the characteristics of the wavelet spectrum from disappearing. [Industrial Applicability]

[0076] As described above, the breathing pattern determination method and breathing pattern determination device disclosed herein focus on frequency changes over time in the breathing waveform and extract the repetitive patterns, thereby enabling detailed classification of breathing patterns and enabling a detailed understanding of the breathing condition. [Explanation of symbols]

[0077] 1. Breathing pattern determination device 10 Sensor unit 11 Strain sensor element 12 Retaining member 13 Fabric 14 Adhesive part 15 Reinforcement members 20 Measurement section 30 Conversion unit 40 Judgment section 50 Display M center axis X Subject P maximum value

Claims

1. A measuring step of measuring a time variation value of lung expansion and contraction; A transformation step of obtaining a wavelet spectrum by wavelet transforming the time variation value; determining a periodic breathing pattern based on the wavelet spectrum; A breathing pattern determination method comprising:

2. 3. The breathing pattern determination method according to claim 1, wherein the periodic breathing pattern is determined using a time change in a peak frequency of a wavelet spectrum.

3. The determining step comprises: a time domain division step of dividing the wavelet spectrum into a plurality of domains at regular time intervals; a similarity calculation step of calculating a similarity of the wavelet spectrum between the plurality of regions; having The breathing pattern determination method according to claim 1 or 2, wherein the similarity is used to determine whether the breathing pattern is a periodic breathing pattern.

4. The breathing pattern determination method according to claim 3 , wherein the similarity is a correlation coefficient.

5. 3. The method for determining a breathing pattern according to claim 1, wherein a strain sensor element is placed on the surface of the living body and capable of detecting expansion and contraction of the living body surface in response to expansion and contraction of the lungs in the measuring step.

6. The method for determining a breathing pattern according to claim 1 or 2, which is used for determining an abnormal breathing phase.

7. A measurement unit that measures the time variation of lung expansion and contraction; A transform unit that obtains a wavelet spectrum by performing a wavelet transform on the time variation value; A determination unit for determining a periodic breathing pattern based on the wavelet spectrum; A breathing pattern determination device comprising:

Citation Information

Patent Citations

  • Sensor unit and detection device

    JP2020151294A