Biosignal measurement apparatus, method, and system

Abstract

The objective is to provide a means for stably acquiring biological information optically even when skin displacement is occurring. [Solution] The system includes an optical receiving unit 101 that acquires optical receiving signals of multiple wavelengths containing biological information, a ΔSR1 processor 103 that determines an estimated value 102 of the first skin surface reflected light change amount using the optical receiving signals of multiple wavelengths, a sensitivity analyzer 104 that determines the sensitivity coefficients of the optical receiving signals of multiple wavelengths for each of the first skin surface reflected light change amount 102, and a ΔHb processor 106 that determines an estimated value 105 of the second surface reflected light change amount using the optical receiving signals of multiple wavelengths to determine the hemoglobin change amount of the living organism.

Description

Technical Field

[0001] The present invention relates to the field of medical devices, particularly to an apparatus, method, and system for optically acquiring and processing physiological signals of a living body.

Background Art

[0002] Conventionally, pulse oximeters and heart rate monitors that are attached to a finger or earlobe with a clip are widely used. Alternatively, near-infrared spectroscopy (NIRS) that measures brain activity by attaching a near-infrared sensor to the head is used for research purposes and medical purposes. Pulse oximeters and heart rate monitors are fixed to a finger or earlobe with a clip or the like. NIRS fixes the near-infrared sensor with a belt, a headgear, or the like.

[0003] Non-Patent Document 1 discloses a technique for measuring the change amount of hemoglobin (ΔHb) in a living body by applying NIRS technology to a PPG (Photoplethysmography) sensor, which is an optical transceiver element used in a pulse oximeter or the like. Furthermore, Non-Patent Document 2 and Patent Document 1 disclose techniques for quantifying subcutaneous blood flow dynamics during exercise using a PPG sensor. These techniques also fix the PPG sensor using a load or the like.

[0004] In the example of the blood pressure sensor shown in Patent Document 2, a two-dimensional flexible reflecting surface is arranged near a location where the blood pressure data of a patient is to be acquired, light from a light source is reflected by the reflecting surface, and a configuration in which the light travels toward a two-dimensional photodetector is shown.

[0005] In the example of the pulse wave measurement sensor shown in Patent Document 3, a configuration having an insertion layer that transmits the first light and reflects the second light and that measures by distinguishing between a skin surface wave and a deep wave is shown.

[0006] In the example of the device shown in Patent Document 4 that enables remote determination of vascular health parameters, a configuration for remotely observing the displacement of blood vessels is shown. A time-varying signal indicating a movement synchronized with the heart is combined according to their polarities to obtain a combined signal.

[0007] Non-patent document 3 describes a configuration in which the optical path length ratio at each measurement wavelength is calculated in NIRS to improve the accuracy of hemoglobin change measurement. [Prior art documents] [Patent Documents]

[0008] [Patent Document 1] :WO2023 / 068313 [Patent Document 2] : Special Publication 2003-532478 [Patent Document 3] : JP 2017-018569 [Patent Document 4] : Special Publication 2020-510487 [Non-patent literature]

[0009] [Non-Patent Document 1] :TY Abay and PA Kyriacou, "Photoplethysmography for blood volumes and oxygenation changes during intermittent vascular occlusions", Journal of Clinical Monitoring and Computing, 2018, Vol. 32, pages 447 to 455. [Non-Patent Document 2] :A Tanaka, "Analysis of a microcirculatory windkessel model using photoplethysmography with green light: A pilot study", IEICE Electronics Express, 2022, Vol. 19, No. 21, pages 20220371-1 to 20220371-6. [Non-Patent Document 3] :S. Umeyama and T. Yamada, "New method of estimating wavelength-dependent optical path length ratios for oxy- and deoxyhemoglobin measurement using near-infrared spectroscopy", Journal of Biomedical Optics 2009, Vol. 14, No. 5, pages 054038-1 to 054038-6. [Overview of the Initiative] [Problems that the invention aims to solve]

[0010] The problem that this invention aims to solve is to provide a means for stably acquiring biological information optically even when skin displacement is occurring. [Means for solving the problem]

[0011] A biosignal measurement device comprising: an optical receiving unit that acquires optical receiving signals of multiple wavelengths containing biological information; a ΔSR1 processor that acquires an estimated value of a first skin surface reflected light change from at least a portion of the optical receiving signals of multiple wavelengths; a sensitivity analyzer that determines the sensitivity coefficients of at least a portion of the optical receiving signals of multiple wavelengths for the first skin surface reflected light change; and a ΔHb processor that acquires an estimated value of a second surface reflected light change from at least a portion of the optical receiving signals of multiple wavelengths to determine the hemoglobin change of a living organism.

[0012] The ΔHb processor is a biosignal measuring device comprising a multiplier that multiplies the sensitivity coefficient and the second surface reflected light change amount, a subtractor that subtracts the multiplication results from at least a portion of the multiple wavelength optical received signals to determine the absorbance change at each wavelength, and a ΔHb analyzer that determines the amount of hemoglobin change in the living organism from the absorbance change and the absorption coefficient at each wavelength. A biosignal measuring device comprising a time interval determination device that acquires an estimated value of the first skin surface reflected light change amount by setting a time interval in at least a portion of the multiple wavelength light received signal according to the heart rate of the living organism or the DC change amount of at least a portion of the multiple wavelength light received signal.

[0013] A biosignal measurement device further comprising: an optical receiving unit that acquires an optical receiving signal of three or more wavelengths consisting of Hb-sensitive light having high sensitivity to hemoglobin; a ΔSR1 processor that acquires an estimated value of the first skin surface reflected light change amount from the optical receiving signal of three or more wavelengths consisting of Hb-sensitive light; and a ΔSR analyzer that obtains multiple solutions of the hemoglobin change amount from the optical receiving signal of three or more wavelengths consisting of Hb-sensitive light, and optimizes the second surface reflected light change amount so that the multiple solutions fall within a predetermined error.

[0014] The ΔSR1 processor is a biosignal measuring device that obtains an estimated value of the first skin surface reflected light change amount from the average of one or at least some of the three or more wavelengths of received light signals.

[0015] A biosignal measurement device that determines the hemoglobin change at a given wavelength by applying the optimized second surface reflected light change to optical received signals of wavelengths other than the three or more optical received signals consisting of the Hb high-sensitivity light used by the ΔHb processor to determine the solution for the hemoglobin change.

[0016] A biosignal measurement device in which the light receiving unit acquires a light receiving signal consisting of at least one Hb low-sensitivity light having low sensitivity to hemoglobin and a plurality of Hb high-sensitivity lights having high sensitivity to hemoglobin, the ΔSR1 processor acquires an estimated value of the first skin surface reflected light change amount from the light receiving signal of the Hb low-sensitivity light, and the ΔHb processor acquires an estimated value of the second skin surface reflected light change amount from the light receiving signal of the Hb low-sensitivity light.

[0017] The Hb processing device is a biosignal measuring device that uses the second amount of change in reflected light from the skin surface and the plurality of Hb-sensitive lights to determine the amount of change in hemoglobin in the living organism.

[0018] A biological signal measurement device, wherein the ΔHb processor further includes an optical path length ratio analyzer that obtains a plurality of solutions for the amount of change in hemoglobin from optical reception signals of three or more wavelengths composed of the Hb-high-sensitivity light, and optimizes the estimated value of the optical path length ratio so that the plurality of solutions fall within a predetermined error.

[0019] A biological signal measurement device comprising the optical reception unit fixed with a space from an observation region of a living body, and a structure that holds the optical reception unit and contacts a biological site separated from the observation region by 10 mm or more, preferably 15 mm or more.

[0020] A biological signal measurement device comprising an optical transmission unit arranged at a position close to the optical reception unit, and a control unit including a power source arranged in the structure.

[0021] A biological signal measurement method comprising an optical reception step of acquiring optical reception signals of a plurality of wavelengths, a ΔSR1 processing step of acquiring an estimated value of a first skin surface reflection light change amount from at least a part of the optical reception signals of the plurality of wavelengths, a sensitivity analysis step of respectively obtaining sensitivity coefficients of at least a part of the optical reception signals of the plurality of wavelengths with respect to the first skin surface reflection light change amount, and a ΔHb processing step of acquiring an estimated value of a second surface reflection light change amount from at least a part of the optical reception signals of the plurality of wavelengths to obtain the amount of change in hemoglobin of a living body. The ΔHb processing step includes a multiplication step of respectively multiplying the sensitivity coefficient and the second surface reflection light change amount, a subtraction step of respectively subtracting the multiplication results from at least a part of the optical reception signals of the plurality of wavelengths to obtain the absorbance change of each wavelength, and a ΔHb analysis step of obtaining the amount of change in hemoglobin of the living body by using the absorbance change and the absorption coefficient of each wavelength.

[0022] The light reception step obtains a light reception signal of three or more wavelengths consisting of Hb highly sensitive light having high sensitivity to hemoglobin, the ΔSR1 processing step obtains an estimated value of the first skin surface reflection light change amount from the light reception signal of three or more wavelengths consisting of the Hb highly sensitive light, and the ΔHb analysis step obtains a plurality of solutions of the hemoglobin change amount from the light reception signal of three or more wavelengths consisting of the Hb highly sensitive light, and a biological signal measurement method further comprising a ΔSR analysis step of optimizing the second surface reflection light change amount so that the plurality of solutions fall within a predetermined error.

[0023] The light reception step obtains a light reception signal consisting of at least one Hb low sensitive light having low sensitivity to hemoglobin and a plurality of Hb highly sensitive lights having high sensitivity to hemoglobin, the ΔSR1 processing step obtains an estimated value of the first skin surface reflection light change amount from the light reception signal of the Hb low sensitive light, and the ΔHb processing step obtains an estimated value of the second skin surface reflection light change amount from the light reception signal of the Hb low sensitive light.

[0024] A biological measurement system comprising a biological measurement device, an edge processing device that processes the biological information acquired by the biological measurement device near the living body, and a cloud processing device that processes the biological information far from the living body.

Advantages of the Invention

[0025] The biological signal measurement device, method and system of the present invention have the advantage of being able to optically acquire biological information while minimizing the influence of skin displacement.

Brief Description of the Drawings

[0026] [Figure 1] It is a block diagram of an example of Embodiment 1 of the biological signal measurement device of the present invention. [Figure 2] It is an example of a ΔSR1 processor and a sensitivity analyzer. [Figure 3] It is an example of a ΔSR1 processor and a ΔHb processor. [Figure 4] It is an example of using Hb highly sensitive light. [Figure 5]This is an example of using Hb-low sensitivity light. [Figure 6] This is an example of a ΔHb processor that utilizes Hb-high sensitivity light. [Figure 7] This is an example of a ΔHb processor that utilizes Hb-low sensitivity light. [Figure 8] This is an example of a ΔHb processor that performs optical path length ratio estimation. [Figure 9] This is an example of a ΔSR1 processor and a ΔHb processor. [Figure 10] This is an example of data acquisition in the implemented example. [Figure 11] This is an example of data acquisition in the implemented example. [Figure 12] This is a flowchart illustrating an example of the biosignal measurement method of the present invention (using Hb high-sensitivity light). [Figure 13] This is a flowchart of an example of the biosignal measurement method of the present invention (using low-sensitivity Hb light). [Figure 14] This is an overhead view of an example of a biosignal measurement device according to Embodiment 3 of the present invention. [Figure 15] This is an example of a perspective view when the device is worn by a user (observed from the upper left). [Figure 16] This is an example of a perspective view when the device is worn by a user (observed from the rear). [Figure 17] This is an overhead view of an example where the entire sensor is placed in eyeglasses. [Figure 18] This is a block diagram of an example sensor configuration. [Figure 19] This is an example of a circuit that adds an LED to the outside of a PPG (Power Programming Device). [Figure 20] This is a perspective view of an example of a biosignal measuring device according to Embodiment 4 of the present invention. [Figure 21] This is a block diagram of an example of a biosignal measurement system according to Embodiment 5 of the present invention. [Figure 22] This diagram shows the relationship between equations 1 through 11 and each unit. [Modes for carrying out the invention]

[0027] (First embodiment) In experiments measuring blood flow in living organisms using light, the inventors revealed that in both the method of attaching the light sensor in contact with the living organism (hereinafter referred to as the contact type) and the method of placing the light sensor away from the living organism (hereinafter referred to as the remote type), there are problems with the optical measurement due to the displacement of blood vessels.

[0028] In contact-type sensors, for example, a light sensor (light-emitting element + light-receiving element) is placed on the skin. Inevitably, some stress is applied to the skin. When a person exercises and their blood pressure rises, displacement occurs in the blood vessels, including the peripheral system, causing expansion and contraction, which in turn causes expansion and contraction of the skin. In many cases, this skin displacement causes modulation of the light sensor stress. This is because the skin often has some curvature. The curvature of the skin can be the shape of the area or the shape of large blood vessels pushing up against the skin. For example, the distance between two points A and B on the skin changes due to the displacement of the curved skin (if skin expansion occurs, the distance between A and B increases). Considering points A and B as the contact surface between the light sensor and the skin, the rigid structure of the light sensor often exerts stress to suppress the skin displacement. This modulation of skin stress is transmitted to the blood vessels, causing modulation of the optical signal observing the blood vessels. This phenomenon is a type of artifact that is introduced into the true blood flow in the absence of compression, with time variations. However, because it changes in accordance with fluctuations in the body's blood pressure, it differs from the motion artifacts that are known to change in accordance with the body's movement (acceleration).

[0029] In remote applications, for example, a light sensor is fixed at a certain distance from the skin when the body is at rest, and light is emitted and received. Taking a reflective light sensor as an example, it receives not only light reflected from beneath the skin but also light reflected from the skin surface. This amount of light reflected from the skin surface is modulated by skin displacement due to blood vessel displacement. This phenomenon can be understood through a thought experiment of illuminating a flat plate with a light. As the distance between the plate and the light increases, the spot of the light becomes larger, the amount of light escaping from the field of view of the light-receiving part of the light sensor increases, and the light signal decreases. In other words, this phenomenon also applies to transmissive light sensors (light incident on a living body is scattered within the body, but as the spot on the incident surface widens, the light that passes through the body also widens). Actual living bodies have curvature, and reflective and transmissive light can take different and complex paths, but here we define the change in light amount based on skin displacement in both reflective and transmissive types as the change in surface reflected light ΔSR. This ΔSR is a type of artifact because it is superimposed on the true signal in relation to the measurement system, but its phenomenon differs from conventionally known motion artifacts. Directly measuring this ΔSR is difficult.

[0030] Figure 1 is a block diagram of a biosignal measurement device according to a first embodiment of the present invention. It comprises an optical reception signal 101 composed of optical reception signals of multiple wavelengths, a ΔSR1 processor 103 that holds an estimated ΔSR value 1(102) which is an estimated value of the first surface reflected light change, a sensitivity analyzer 104 that determines the sensitivity coefficient of each optical reception signal to the ΔSR estimate 1(102), and a ΔHb processor 106 that holds an estimated ΔSR value 2(105) which is an estimated value of the second surface reflected light change to determine the hemoglobin change of the living organism. The ΔSR in the remote type is estimated and removed to determine the hemoglobin change.

[0031] The optical reception signal 101 can use, for example, a received signal of light with wavelengths ranging from ultraviolet to visible light and then to near-infrared light (for example, wavelengths from 300 to 2000 nm, considering applications to various absorbing species). A PPG sensor can be used, but signals from NIRS equipment or remote optical measurement devices can be used as appropriate. An inexpensive silicon (Si) photodiode (PD) may be used as the photodetector. In that case, based on the sensitivity characteristics, wavelengths from 300 to 1200 nm, and more preferably from 400 to 1100 nm where a certain degree of quantum efficiency can be expected, can be used. In addition to using an LED with a built-in PPG as the photoluminescent element (optical transmission unit), near-infrared lasers in NIRS equipment and other light sources with arbitrary wavelengths ranging from ultraviolet to visible light and then to near-infrared light can be used.

[0032] Since photodetectors are sensitive to a wide range of wavelengths, it is necessary to synchronize them with the light source to identify the wavelength of light intensity. However, they can also be used in combination with passive light sources such as ambient light, optical filters, and optical shutters as appropriate. A unit that acquires the signals measured by these devices and elements from an external source can also be used within this biosignal measurement device.

[0033] It is preferable to position the light-receiving element and the light-emitting element relatively close together to prevent the observation area from expanding too much. If the observation area expands, the contact area with the skin will be determined further away in remote-type systems, which would require longer arms to support the light-receiving element and the light-emitting element. In this embodiment, hemoglobin is used as an example of the physiological signal to be measured, but the effects of this embodiment are not limited to hemoglobin measurement (it can also be applied to myoglobin, bilirubin, porphyrin, as well as alcohols that absorb around 1160 nm, glucose that absorbs around 1600 nm, etc.).

[0034] The ΔSR estimate 1(102) held by the ΔSR1 processor 103 can be determined by two main methods. One is 1) a method using light of a wavelength that has high sensitivity (absorption) to hemoglobin (hereafter referred to as Hb high-sensitivity light), and the other is 2) a method using light of a wavelength that has relatively low sensitivity to hemoglobin (hereafter referred to as Hb low-sensitivity light).

[0035] 1) When using Hb high-sensitivity light, it is necessary to extract the ΔSR component from the superimposed ΔHb and ΔSR components. This can be done in two ways: a) by applying a stimulus to the living body that causes only a ΔSR change without causing a ΔHb change, and b) by selecting a wavelength from among multiple wavelengths of Hb high-sensitivity light that causes little ΔHb change, or by selecting a combination from among multiple wavelengths that cancels out the ΔHb change.

[0036] As an example of method a), we attempted to give a living organism a light exercise task such as walking (instructing it to exercise). In the inventor's experiments, for example, when a subject with a resting heart rate of approximately 60 was given a walking exercise task for about 100 seconds, resulting in a heart rate of approximately 90, a change in ΔSR, i.e., a change in the DC of the optically received signal, could be induced with almost no change in ΔHb. By increasing the rate of breathing or giving other light exercise tasks, it is possible to suppress the change in ΔHb and induce a change in ΔSR.

[0037] As an example of method b), consider the case where the changes in oxygenated hemoglobin (HbO) and deoxygenated hemoglobin (HbR) are measured using red light with a wavelength around 660 nm and near-infrared light with a wavelength around 880 nm. In the inventor's observations around the ear, HbO and HbR change due to exercise, etc., but the rate of change in HbR is smaller. This is related to the fact that HbO normally accounts for more than 90% of blood. Since red light has relatively high sensitivity to HbR and low sensitivity to HbO, light around 660 nm can be used as the ΔSR estimate value 1(10²). Furthermore, in some cases, the polarity of change in HbO (ΔHbO) and HbR (ΔHbR) can be reversed. Since near-infrared light has relatively high sensitivity to HbO and low sensitivity to HbR, the average of the time-series signals of infrared and near-infrared light can be used as the ΔSR estimate value 1(10²). Furthermore, by using three or more wavelengths, it is possible to take the average or weighted average of all wavelength combinations or any combination of wavelengths.

[0038] Furthermore, by appropriately combining a) and b), it is possible to give a living organism a task that minimizes the change in ΔHb while causing a change in ΔSR, and use the signal obtained by averaging the time-series signals of infrared and near-infrared light as the estimated ΔSR value 1(10²).

[0039] 2) When using low-sensitivity Hb light, select a wavelength of light that is insensitive to both HbO and HbR. Such wavelengths can be around 1100 nm or higher. This is because the molar extinction coefficient of HbO at 1100 nm is reduced to less than half of that at 880 nm, the molar extinction coefficient of HbR is sufficiently low, and water absorption increases, reducing the photoresponse from the subcutaneous tissue. Alternatively, light with a wavelength of around 400 nm or lower can be used. Although the extinction coefficients of HbO and HbR are high, the penetration depth into the subcutaneous tissue is shallower, reducing the photoresponse from the subcutaneous tissue. When using the aforementioned Si PD, wavelengths from 400 to 1100 nm are preferable, but there is also the option of using light with wavelengths around 400 nm or 1100 nm as low-sensitivity Hb light, or using a photodetector other than Si PD.

[0040] The sensitivity analyzer 104 determines the sensitivity coefficient of each received optical signal with respect to the ΔSR estimate 1(102). Each received optical signal is an optical signal necessary for calculating the hemoglobin change amount in the ΔHb processor 106 described later, and is Hb-sensitive light that has high sensitivity to hemoglobin. In addition to the red light and near-infrared light mentioned above, green light around 540 nm and wavelengths around 740 nm between red light and near-infrared light can also be used. When the ΔSR estimate 1(102) changes, the ratio of the changes occurring in each of these received optical signals necessary for hemoglobin calculation is determined as the sensitivity coefficient for each.

[0041] The reason why sensitivity to ΔSR differs with wavelength is due to factors such as the different positions of light sources like LEDs, the different penetration depth into the skin depending on the wavelength, and the different reflectivity in the skin depending on the wavelength. Since reflection occurs at the boundaries of substances with different refractive indices, it occurs not only at the outermost surface of the skin but also at the boundaries of the five layers that make up the epidermis.

[0042] In experiments conducted by the inventor, for example, when blood pressure increased due to walking, and the estimated ΔSR value 1(10²) increased by approximately 10% compared to the resting state due to vasodilation, the values ​​for each light received signal increased by approximately 12% for green light, 10% for red light, 8% for 740nm light, and 10% for near-infrared light compared to the resting state. In other words, the sensitivity coefficients for the green, red, 740nm, and near-infrared light received signals with respect to the estimated ΔSR value 1(10²) were 1.2:1:0.8:1. In this case, the estimated ΔSR value 1(10²) was obtained by averaging the time-series signals of infrared and near-infrared light. The deviation of green light from 1 is thought to be due to its shallow penetration depth into the subcutaneous tissue and differences in reflectivity. The reason for the discrepancy at 740nm is thought to be that the LED was placed outside the PPG (which contains three LEDs: green, red, and near-infrared) (they are located somewhat far apart) (the photodiode used is the one inside the PPG).

[0043] The ΔHb processor 106 stores the ΔSR estimate 2(105) and also determines the hemoglobin change ΔHb in the living organism. The difference between ΔSR estimate 1(102) and ΔSR estimate 2(105) is that the former is an estimate obtained by minimizing ΔHb, such as during walking, in order to determine the sensitivity coefficient, while the latter needs to be an estimate in any state of the living organism in order to determine ΔHb. If the Hb low-sensitivity light described in 2) above does not cause a significant change even in response to large Hb changes in the living organism, and only a ΔSR change occurs, then both ΔSR estimate 1(102) and ΔSR estimate 2(105) can use the entire interval of the Hb low-sensitivity light signal. On the other hand, if a large Hb change significantly mixes with the ΔSR, then ΔSR estimate 1(102) can use the Hb low-sensitivity light signal while excluding the interval in which the Hb change is mixed, but ΔSR estimate 2(105) needs to be determined by a different method. Furthermore, if it is difficult to use Hb low-sensitivity light due to hardware limitations, the ΔSR estimate 1(102) should be obtained by the method described in 1) above, and the ΔSR estimate 2(105) should be obtained using another method 3).

[0044] 3) As a method for determining the estimated ΔSR value 2(10⁵) from Hb-sensitive light, the inventors have developed a method to derive it from a system of simultaneous equations of optical received signals with three or more wavelengths. The background technology is the technique for determining the optical path length ratio in NIRS described in Non-Patent Document 3. Normally, in NIRS, the unknowns ΔHbO and ΔHbR are determined from signals of two wavelengths: red light and near-infrared light. If a third wavelength is added and measurements are performed with three wavelengths, two ΔHbO values ​​(ΔHbOa and ΔHbOb) and two ΔHbR values ​​(ΔHbRa and ΔHbRb) can be obtained from a system of three simultaneous equations. When light of three wavelengths passes through a living organism, there is a difference in optical path length, which causes errors in the calculated ΔHb. However, by determining the optical path length ratio (only the ratio can be determined, not the absolute value of the optical path length) from a system of three simultaneous equations, the accuracy of ΔHb measurement is improved.

[0045] In Non-Patent Literature 3, the unknowns in the system of equations are ΔHbO, ΔHbR, and the optical path length ratio. However, by changing the unknowns to ΔHbO, ΔHbR, and the estimated ΔSR value 2(105), it becomes possible to determine ΔHbO and ΔHbR while simultaneously determining the estimated ΔSR value 2(105). Experiments by the inventors show that changes in ΔSR cause changes in the received optical signal that are several times larger than changes in ΔHb. While errors will remain due to the lack of correction for the optical path length ratio, it becomes possible to extract the ΔHb signal that exists within ΔSR, which is normally difficult to estimate.

[0046] For example, consider the case where we use three wavelengths of Hb-sensitive light: red light, 740 nm light, and near-infrared light. In the Modified Lambert-Beer (MLB) law, the change in the received light signal is expressed as the product of the molar extinction coefficient of each absorbing substance (HbO, HbR) and the change in hemoglobin (ΔHbO, ΔHbR). In a typical two-wavelength NIRS, the absorbance changes of red light and near-infrared light (ΔAr, ΔAir) can be used to determine the two unknowns ΔHbO and ΔHbR from a system of two MLB equations. If we use the absorbance changes of three wavelengths (ΔAr, ΔA740, ΔAir), we can solve three MLB equations simultaneously and determine the three unknowns: the estimated ΔSR value 2(10⁵), ΔHbO, and ΔHbR.

[0047] Figure 2 shows an example in which the ΔSR1 processor is equipped with a time interval determination device 201 that determines time intervals including exercise tasks, etc., in order to carry out method a) above. It determines time intervals such as light exercise intervals in which ΔSR changes predominantly while ΔHb changes to a negligible degree. For this purpose, the heart rate of the living body and the DC change amount of the optical received signal 101 are compared with, for example, the resting level to make the determination. For example, it determines time intervals that include the interval in which the heart rate changes from 60 to 90, or time intervals that include the interval in which the DC change amount changes from a few percent to several tens of percent. Based on this determination, the time interval of the optical received signal 101 to be used as the ΔSR estimate value 1 (102) can be determined.

[0048] Furthermore, Figure 2 shows an example configuration of the sensitivity analyzer 104. It can include units for processing covariance 202 and variance 203. Covariance 202 processes the covariance between the optically received signal and the ΔSR estimate 1(102). Variance 203 processes the variance of the ΔSR estimate 1(102). By dividing the result of covariance 202 by the result of variance 203, the sensitivity of the optically received signal to ΔSR can be determined at each wavelength.

[0049] Equation 3 is the formula that shows these calculations. The signal used in Equation 3 is calculated using Equations 1 and 2.

[0050]

number

[0051]

number

[0052]

number

[0053] In Equation 1, rawDCλ is the received optical signal at each wavelength λ, and normDCλ is the received optical signal after normalization by a reference received optical signal level, rawDC0λ. rawDC0λ can be the received optical signal level at the start of the experiment or the resting optical signal level. This allows the received optical signal to be expressed as a change relative to 1. The calculation in Equation 1 can be performed, for example, with 101 units of received optical signal.

[0054] Equation 2 is an example of an equation for calculating the ΔSR estimate 1(102) (denoted as ΔSR1). For example, ΔSR1 can be calculated from the average of two Hb high-sensitivity light sources normDCλ1 and normDCλ2. ΔSR1 is subtracted by 1 to make it a change relative to zero for subsequent processing. Similarly, ΔSR1 can be calculated from the average of two Hb low-sensitivity light sources. Averaging can also be omitted by using a single Hb low-sensitivity light source where the Hb change is negligible. The purpose of averaging is to suppress the ΔHb change within the ΔSR change and to suppress noise contained in the optically received signal. The latter meaning of averaging can be replaced with a moving average or a filter. The calculation of Equation 2 can be performed, for example, by a ΔSR1 processor 103 that holds the ΔSR estimate 1(102).

[0055] Equation 3 processes the covariance of 202 in the numerator and the variance of 203 in the denominator. The covariance of 202 in the numerator can be obtained by subtracting 1 from normDCλ (based on 1) to obtain a zero-referenced signal, multiplying it by the ΔSR estimate 1(102), and taking the sum over a certain interval. The variance of 203 in the denominator can be obtained by squaring the ΔSR estimate 1(102) and taking the sum over a certain interval. The sensitivity coefficient Cλ obtained by dividing the covariance of 202 by the variance of 203 is the sensitivity coefficient to ΔSR for an optical received signal of a certain wavelength λ. By multiplying this Cλ by the ΔSR estimate, the ΔSR component contained in the optical received signal of wavelength λ can be obtained.

[0056] For the interval over which the sum is taken, it is preferable to use an interval where the change in ΔHb is negligible and the change in ΔSR is significant. As explained in the interval determination device 201 above, this can be done by giving a living organism a light exercise task such as walking and using that interval as the fixed interval. This Cλ is the proportion or correlation amount of a certain ΔSR estimated waveform mixed into each optical received signal, and it can be determined by methods other than Equation 3, such as using differential coefficients.

[0057] Figure 3 shows an example configuration of the ΔSR1 processor 103 and the ΔHb processor 105. The ΔSR1 processor 103 has a ΔSR estimate value 1(102) that utilizes an optically received signal selected from the group consisting of Hb high-sensitivity light 301 and Hb low-sensitivity light 302. The ΔHb processor 105 has a ΔSR estimate value 2(105) that utilizes an optically received signal selected from the group consisting of Hb high-sensitivity light 303 and Hb low-sensitivity light 304. Note that Figure 3 is a diagram to explain the characteristic signal flow, and the units in which the time-series data of the optically received signal exists (passes through or is temporarily stored) can be set as appropriate.

[0058] As mentioned above, it is preferable that the ΔSR estimate 1 (102) of the ΔSR1 processor 103 utilizes the optically received signal from an interval in which the ΔHb change is suppressed and the ΔSR change is significant, such as an interval including light exercise.

[0059] The ΔSR estimate 1(102) of the ΔSR1 processor 103 utilizes Hb high-sensitivity light 301 and Hb low-sensitivity light 302, as explained in Figure 1. Both can be used as needed. To obtain the sensitivity coefficient Cλ, it is important that the ΔHb change is suppressed and the ΔSR change is significant (if the true change in ΔHb remains, it will result in an error in Cλ), and it is also preferable that electrical and optical noise is suppressed. In light of this objective, depending on the required accuracy, if the accuracy is permitted by the specifications, a single Hb high-sensitivity light 301 or a single Hb low-sensitivity light 302 can be used, or to improve accuracy, the average of multiple Hb high-sensitivity light 301s in which the ΔHb change is canceled can be used, or the average of multiple Hb low-sensitivity light 302s can be used. To reduce noise, the average of the Hb high-sensitivity light 301 and the average of the Hb low-sensitivity light 302 can be further averaged. To suppress ΔHb changes and significantly induce ΔSR changes, tasks that alter heart rate and cause peripheral vascular displacement can be imposed on the body, such as light exercise like walking or stretching, standing, respiratory changes, and external stimuli.

[0060] The ΔSR estimate 2(105) of the ΔHb processor 105 is, for example, the optical received signal for the section over which ΔHb is to be determined, or it may be the entire measured section. The ΔSR estimate 2(105) utilizes Hb high-sensitivity light 303 or Hb low-sensitivity light 304, as explained in Figure 1. In order to derive ΔHb, it is necessary to clarify the ΔSR that is mixed in. In the method using Hb high-sensitivity light 303, a method has already been explained in which ΔHb and ΔSR are obtained from a system of equations of the MLB law using Hb high-sensitivity light 303 composed of optical received signals of three or more wavelengths. In the method using Hb low-sensitivity light 304 as the ΔSR estimate 2(105), the ΔHb change within the Hb low-sensitivity light 304 must be small enough to be negligible (if the true ΔHb is included in the ΔSR estimate, it will result in an error). The optical wavelength is selected and used to satisfy this condition.

[0061] For the ΔSR estimate 1(102), it is sufficient to have a signal in an interval where ΔHb is sufficiently small and ΔSR is sufficiently large for the calculation of Cλ. Here, "sufficient" can be determined, for example, by whether the error (correction error) from the true ΔHb when ΔHb is obtained using the calculated sensitivity coefficient Cλ with the ΔHb processor 105 is acceptable with respect to the expected accuracy. Similarly, for the ΔSR estimate 2(105), it is important that the ΔSR is corrected appropriately in the ΔHb processor 105, and that the error from the true ΔHb is acceptable with respect to the expected accuracy. In light of this purpose, it is possible to determine which optical received signal to use as the ΔSR estimate 1(102), that is, to select from the group consisting of Hb high-sensitivity light 301 and Hb low-sensitivity light 302, and to determine which interval to use. The same applies to the optical received signal selected from the group consisting of Hb high-sensitivity light 303 and Hb low-sensitivity light 304, which will be used as the ΔSR estimate 2(105).

[0062] However, the selection from the group consisting of Hb high-sensitivity light and Hb low-sensitivity light can be appropriately determined according to the desired error, available materials, cost, power consumption, etc. For example, in order to pursue low cost and to find the fewest possible number of LEDs (number of light wavelengths used), only Hb high-sensitivity light is used for both ΔSR estimate 1 (10²) and ΔSR estimate 2 (10⁵). For example, to improve accuracy, Hb low-sensitivity light consisting of wavelengths that minimize ΔHb contamination is used for ΔSR estimate 1 (10²) and ΔSR estimate 2 (10⁵). For example, to reduce power consumption, Hb low-sensitivity light is used when accuracy is required, and Hb high-sensitivity light is used by stopping the emission of Hb low-sensitivity light otherwise. It is also possible to dynamically switch the selection from the group during measurement.

[0063] Figure 4 shows an example of a configuration that utilizes an optically received signal consisting of Hb high-sensitivity light 301 for ΔSR estimate 1(102) and Hb high-sensitivity light 303 for ΔSR estimate 2(105). The configuration uses method 1) for ΔSR estimate 1(102) and method 3) for ΔSR estimate 2(105). As mentioned above, this can be used when pursuing low cost or when it is difficult to use Hb low-sensitivity light due to the availability of materials. When using Hb high-sensitivity light 301 for ΔSR estimate 1(102), a) Hb high-sensitivity light of any wavelength in a section where Hb changes can be ignored, such as a light motion section, can be used. Furthermore, b) one wavelength of high-sensitivity light with little Hb change can be selected, or the average of multiple high-sensitivity signals where Hb changes are canceled out can be used. When using Hb high-sensitivity light 303 for ΔSR estimate 2(105), ΔSR estimate 2(105) can be obtained from the simultaneous equations of the MLB law explained in 3).

[0064] Figure 6 shows an example configuration of the ΔHb processor 106 when using method 3), which utilizes an optically received signal consisting of Hb-high sensitivity light 303 as the ΔSR estimate 2(105). The ΔHb processor 106 comprises Hb-high sensitivity light 303, a ΔSR analyzer 601, a ΔSR estimate 2(105), a first multiplier 602, a subtractor 603, and a ΔHb analyzer 606, and outputs ΔHb (ΔHbO and ΔHbR). The ΔHb analyzer 606 can be realized, for example, by having an absorption coefficient 604 and a second multiplier 605.

[0065] Equation 4 is the formula for calculating calDCλ, the DC level of the optically received signal corrected for ΔSR, from the normalized optically received signal level normDCλ at the aforementioned wavelength λ, the sensitivity coefficient Cλ, and the estimated ΔSR value 2 (10⁵, hereafter referred to as ΔSR²). For each wavelength, it is obtained by subtracting the product of Cλ and ΔSR² from normDCλ. Here, ΔSR² is an unknown variable, but the three unknowns, including ΔHbO and ΔHbR which are also unknowns, are obtained from three simultaneous equations using the MLB law.

[0066]

number

[0067] For example, the Hb high-sensitivity light 303 in Figure 6 outputs normDCλ. The ΔSR analyzer 601 analyzes three unknowns ΔSR2, ΔHbO, and ΔHbR. The ΔSR estimate 2(105) holds the analysis result ΔSR2. The first multiplier 602 multiplies Cλ and ΔSR2. The subtractor 603 subtracts this multiplication result from normDCλ to calculate calDCλ. The second multiplier 605 multiplies calDCλ by the inverse matrix of the absorption coefficient 604 (illustration of the inverse matrix operation is omitted) and outputs ΔHb (ΔHbO and ΔHbR).

[0068] The following explains how to find the three unknowns using the MLB rule, with the help of mathematical formulas. We will consider the case where Hb high-sensitivity light 303 has the aforementioned green light, red light, 740nm light, and near-infrared light as an example.

[0069] In the MLB law, the absorbance change ΔA is expressed as the product of the molar extinction coefficient ε of a certain absorbing substance and the change in the concentration of the absorbing substance, and the products obtained for each changing absorbing substance are added together. For example, when measuring the changes in HbO and HbR with light of wavelength λ, the absorbance change ΔAλ is the sum of the HbO term, εHbO·ΔHbO, and the HbR term, εHbR·ΔHbR. ΔAλ is expressed by equation 5 and is the logarithm of the reciprocal of the DC level of the received optical signal, reflecting the DC change (absorbance increases as the DC level decreases). The DC level of the received optical signal used for ΔAλ is calDCλ, which is corrected for ΔSR. By setting up the relationship between ΔAλ and ε·ΔHb for each measurement wavelength, a system of simultaneous equations can be constructed (if there are N measurement wavelengths, a system of simultaneous equations consisting of N equations will be formed).

[0070]

number

[0071] When measuring HbO or HbR, the multiple optical signals used in this system of equations are Hb-sensitive light. Since the penetration depth into the subcutaneous tissue changes depending on the wavelength of the light used, it is preferable to select wavelengths that are as close as possible to each other. On the other hand, if wavelengths that are too close are selected, the difference in molar extinction coefficients becomes small, and the influence of noise contained in the optical signals becomes more noticeable when solving the system of equations.

[0072] The wavelength range from the aforementioned red light through 740nm light to near-infrared light penetrates more than 650um into the subcutaneous tissue, allowing observation of microcirculation consisting of arterioles and venules. On the other hand, focusing on the molar extinction coefficient, when comparing 660nm red light and 740nm light, the difference in εHbR is large, but the difference in εHbO is small. When comparing 740nm light and 880nm near-infrared light, conversely, the difference in εHbO is large, but the difference in εHbR is small. When comparing 660nm red light and 880nm near-infrared light, the difference in both εHbO and εHbR is large. From the system of three equations for the three wavelengths of red, 740nm, and near-infrared, three values ​​can be obtained for both ΔHbO and ΔHbR. For ΔHbO, we focus on two solutions: ΔHbOrir, which is the solution obtained by the system of red light and near-infrared light where the difference in εHbO is large, and ΔHbOz, which is the solution obtained by 740nm light and near-infrared light. For ΔHbR, we focus on two solutions: ΔHbRrir, which is a simultaneous solution of red light and near-infrared light with a large difference in εHbR, and ΔHbRy, which is a solution of red light and 740nm light. The above explanation is expressed mathematically in equations 6 to 11.

[0073]

number

[0074]

number

[0075]

number

[0076]

number

[0077]

number

[0078]

number

[0079] Equation 6 is the equation for the absorbance vector a for three wavelengths λ1, λ2, and λ3. For example, consider λ1 as red light, λ2 as 740 nm light, and λ3 as near-infrared light.

[0080] In equation 7, the inverse matrix of ε is defined for solving the system of equations.

[0081] Equation 8 defines matrices U, V, and W, whose elements are the inverse of ε. Matrix U is defined to find solutions for combinations of λ1 and λ2. Matrix V is defined to find solutions for combinations of λ2 and λ3. Matrix W is defined to find solutions for combinations of λ1 and λ3.

[0082] In equation 9, y is defined as the solution to ΔHb for the combination of λ1 and λ2. z is defined as the solution to ΔHb for the combination of λ2 and λ3. rir is defined as the solution to ΔHb for the combination of λ1 and λ3. In this example, y is the solution using red light and 740 ray light, z is the solution using 740 ray and near-infrared light, and rir is the solution using red light and near-infrared light.

[0083] In equation 10, the three vectors y, z, and rir, which are the solutions to ΔHb, are obtained from the matrix product of the inverse matrix of ε and the absorbance vector a.

[0084] Focusing on the combination that minimizes the noise mentioned above, we can define the evaluation formula shown in Equation 11. Resid is the sum of the squares of the difference between ΔHbOrir and ΔHbOz, and the sum of the squares of the difference between ΔHbRrir and ΔHbRy, over a certain interval T. ΔSR2, which minimizes this resist, is the solution for the unknowns ΔSR2 and ΔHb. By setting the interval T to, for example, about 10 seconds and calculating while sequentially shifting it over the entire measurement, a continuous ΔSR2 can be calculated. ΔSR is due to skin displacement caused by vascular displacement and is a relatively slow change due to the autonomic nervous system, etc. The interval T can be appropriately determined according to the required accuracy and temporal resolution.

[0085] The evaluation function `resid` can be appropriately determined depending on the required accuracy, the absorbing material of interest, and the computational resources available. For example, it is possible to use a different combination from that in Equation 11, including ΔHbOy and ΔHbRz, which are not used in Equation 11, or to weight each combination to reduce the influence of noise. Furthermore, algorithms such as quasi-Newtonian optimization algorithms, successive least-squares programming, Powell's conjugate direction method, and the simplex method can be used as appropriate to minimize `resid`.

[0086] Although green light is not used in the ΔSR2 search in the above example, the ΔHb (ΔHbOg, ΔHbRg) observed by green light can be calculated using the ΔSR2 obtained above. According to Non-Patent Literature 2, green light observes capillaries in the shallow subcutaneous region, while red light and near-infrared light observe deeper arterioles and venules. It is thought that the behavior of ΔHb differs between green light and red light and wavelengths greater than that, and although green light is not used in the ΔSR2 search in the above example, the obtained ΔSR2 is thought to be applicable to green light.

[0087] The procedure for determining ΔHb for green light involves finding Cλ using equation 3, similar to other wavelengths, then finding calDCλ from equation 4, and finally finding ΔAλ using equation 5. According to the MLB law, the absorbance change ΔAg for green light is the product of the molar extinction coefficients (εHbOg, εHbRg) and the ΔHb vector (ΔHbOg, ΔHbRg) for green light. In the assumed green light wavelength of around 540 nm, εHbOg and εHbRg are almost the same value εHbg, and ΔHbOg and ΔHbRg cannot be separated. However, the total hemoglobin change ΔHbTg, which is the sum of ΔHbOg and ΔHbRg, can be obtained by dividing ΔAg by εHbg.

[0088] In other words, by using ΔSR2 optimized with red, 740nm, and near-infrared light, we can correct for ΔSR contained in green light (a wavelength not used for optimization) to obtain the absorbance change ΔA, and then calculate ΔHb from ΔA and the absorption coefficient of green light. When εHbO and εHbR are approximately equal, the total hemoglobin change can be determined from a single wavelength. Applying the optimized ΔSR2 to wavelengths other than those used for optimization can be applied not only to a single wavelength but also to multiple wavelengths, and ΔHbO and ΔHbR can be calculated from the simultaneous equations for those multiple wavelengths.

[0089] Minimizing equation 11 yields ΔSR2 along with ΔHbO and ΔHbR. While ΔSR1 represents changes during, for example, light exercise, ΔSR2 represents the ΔSR over the entire analysis interval. ΔSR reflects the expansion and contraction of blood vessels and reflects information about the blood vessels in the observed area. Furthermore, it is possible to evaluate the state of the circulatory system of the living body from the acquired ΔSR.

[0090] Figure 5 shows an example of using an optically received signal consisting of Hb-low sensitivity light 302 as ΔSR estimate 1(102) and Hb-low sensitivity light 304 as ΔSR estimate 2(105). This can be used to suppress the inclusion of ΔHb within the ΔSR estimate and improve measurement accuracy. When using ΔSR estimate 1(102), it may be limited to sections of light motion, such as when Hb changes remain in the Hb-low sensitivity light 302 due to strong motion (since it is sufficient to extract Cλ). When using ΔSR estimate 2(105), it is used in the section where ΔHb is to be determined to calculate ΔHb. The accuracy can be further improved by performing the optical path length ratio correction described above.

[0091] Figure 7 shows an example configuration of the ΔHb processor 106 when the ΔSR estimate 2(105) utilizes an optically received signal consisting of Hb-low sensitivity light 304. The ΔHb processor 106 has Hb-low sensitivity light 304, the ΔSR estimate 2(105), a first multiplier 602, a subtractor 603, and a ΔHb analyzer 606 (for example, the absorption coefficient 604 and the second multiplier 605), and outputs ΔHb (ΔHbO and ΔHbR).

[0092] The difference from Figure 6 is that in Figure 6, the ΔSR estimate 2(105) is determined from a search using Hb high-sensitivity light 303, whereas in Figure 7, the ΔSR estimate 2(105) is determined by directly using Hb low-sensitivity light 304.

[0093] In the configuration shown in Figure 7, calDCλ, the DC level of the optically received signal corrected for ΔSR, is calculated using Equation 4. The ΔSR2 used here directly utilizes the Hb low-sensitivity light 304. One or multiple waves of Hb low-sensitivity light 304 can be used. If multiple waves are used, the average can be taken, or a weighted average can be taken with different weights depending on the noise. normDCλ uses the Hb high-sensitivity light necessary for calculating ΔHb, but two wavelengths necessary to determine the two unknowns ΔHbO and ΔHbR can also be used. When calculating Cλ, it is preferable to use the Hb low-sensitivity light for the ΔSR estimate 1 if available, but it is also possible to switch to Hb high-sensitivity light as appropriate, or to use the average of the Hb low-sensitivity and Hb high-sensitivity light.

[0094] For these two wavelengths, for example, red light and near-infrared light can be used. Red light has a large εHbR and a small εHbO. Conversely, near-infrared light has a large εHbO and a small εHbR, so the influence of noise can be reduced when solving simultaneous equations using the MLB law. In the case of two wavelengths, ΔHbO and ΔHbR can be directly obtained using, for example, the rir equation in Equation 10. Equations 6 to 10 are for the case of three wavelengths, so there are redundant parts when using them for two wavelengths, but they can be modified as appropriate for calculation. Basically, ΔHb can be calculated by solving the MLB law, ΔAλ = ελ·ΔHb, for two wavelengths simultaneously.

[0095] In Figure 7, as in Figure 6, the solution for ΔHb can be obtained using the inverse matrix of ελ. The first multiplier 602 multiplies ΔSR2 by the sensitivity coefficient Cλ, and the subtractor 603 subtracts the result of the multiplication Cλ·ΔSR2 from the received optical signals of two wavelengths of Hb-sensitive light, such as red light and near-infrared light. The value of Cλ is used for each wavelength. The subtraction result is calDCλ, and ΔAλ is obtained by taking the logarithm of the reciprocal of calDCλ (Equation 5, not shown in the figure). Then, by taking the matrix product of ΔAλ and the inverse matrix of the molar absorption coefficient ελ (not shown in the figure) in the second multiplier 605, ΔHb (ΔHbO and ΔHbR) can be obtained.

[0096] Figure 8 shows an example of a configuration in which the ΔHb processor 106 corrects the optical path length ratio in a configuration that utilizes the Hb low-sensitivity light of Figure 5 for ΔSR1 and ΔSR2. Compared to the configuration of Figure 7, it further includes an optical path length ratio analyzer 801, an optical path length ratio estimater 802, and a third multiplier 803.

[0097] To calculate the unknowns ΔHbO and ΔHbR, including the optical path length ratio L, a light received signal of three or more wavelengths is required as Hb-sensitive light (input to subtractor 603, normDCλ in equation 4). Using the three-wavelength absorbance vector a in equation 6, the unknowns, including the optical path length ratio L, are determined from a system of equations based on the MLB law. The processes in equations 1 to 5 are performed to derive the absorbance vector a, but the same argument as in Figure 7 applies to obtaining the sensitivity coefficient Cλ.

[0098] The ΔSR estimate 2(105) in Figure 8 directly utilizes the Hb low-sensitivity light 304. The Hb low-sensitivity light 304 can be a single wave or multiple waves. If multiple waves are used, the average can be taken, or a weighted average can be taken with varying weights depending on the noise. The first multiplier 602 multiplies Cλ and ΔSR2. The subtractor 603 subtracts the result of the first multiplier 602 from the Hb high-sensitivity light (normDCλ) to calculate calDCλ. The third multiplier 803 multiplies the inverse matrix of the optical path length estimate 802 (illustration of the inverse matrix operation is omitted) by calDCλ. The second multiplier 605 multiplies the output of the third multiplier 803 by the inverse matrix of the absorption coefficient 604 (illustration of the inverse matrix operation is omitted) to output ΔHb (ΔHbO and ΔHbR). The optical path length ratio analyzer 801 performs optimization processing on ΔHb and the estimated optical path length ratio 802. The estimated optical path length ratio 802, the third multiplier 803, the absorption coefficient 604, and the second multiplier 605 constitute the ΔHb analyzer 606, which finds the solution for ΔHb.

[0099] Equation 12 is the definition of the diagonal matrix L, which represents the ratios of the optical path lengths (L1, L2, L3) for the three wavelengths. The three wavelengths are Hb-sensitive light (normDCλ), and for example, the aforementioned red light, 740nm light, and near-infrared light can be used.

[0100]

number

[0101] Equation 13 shows the relationship between the three vectors (y, z, rir) that are solutions to ΔHb, the optical path length ratio matrix L, the reciprocal of the molar extinction coefficient (U, V, W), and the absorbance vector a corrected for ΔSR1. This can be derived from the MLB law, which states that the absorbance vector is expressed as the product of the optical path length ratio matrix, the molar extinction coefficient matrix, and the ΔHb vector.

[0102]

number

[0103] The unknowns in equation 13 are the optical path length ratios for each of the three wavelengths (L1, L2, L3), ΔHbO, and ΔHbR. As with equation 10, the solution can be optimized using, for example, the evaluation formula resid in equation 11. In the optimization of equation 10, ΔSR2, ΔHbO, and ΔHbR were calculated sequentially for each interval T (e.g., T=10 seconds), but in equation 13, the optical path length ratio can be optimized under the assumption that the optical path length ratio is constant during the measurement (in this case, ΔHbO and ΔHbR are assumed to change with each time step of the measurement). The absolute values ​​of the optical path length ratios (L1, L2, L3) cannot be calculated, but the optical path length ratios can be determined by optimization. This optimization process can be performed by the optical path length ratio analyzer 801.

[0104] As an example of the ΔHb analyzer 606, we showed how to multiply the absorbance vector a by its inverse matrix. However, to solve simultaneous equations involving multiple wavelengths, it is possible to use other methods such as addition / subtraction or substitution as appropriate.

[0105] Figure 9 shows an example of a configuration for obtaining an estimated ΔSR value by combining high-sensitivity Hb light and low-sensitivity Hb light. There are two cases: (a) for ΔSR estimate 1 and (b) for ΔSR estimate 2.

[0106] In Figure 9(a), the ΔSR1 processor 103 has Hb high-sensitivity light 301 and Hb low-sensitivity light 302, a synthesizer 901 that combines them, and a ΔSR estimate value 1(102) that holds the combination result. The synthesizer 901 can perform selective synthesis by appropriately switching between Hb high-sensitivity light 301 and Hb low-sensitivity light 302, simple average synthesis by taking a simple average with equal gain, and weighted average synthesis by changing the weighting according to the error. Errors include the inclusion of the true component ΔHb, as well as optical and electrical noise. The timing of switching the selection can be determined by using an acceleration sensor or the like to detect light or high-intensity movement of the body. This can reduce the inclusion of errors and improve the accuracy of Cλ calculation.

[0107] In Figure 9(b), the ΔHb processor 103 has Hb high-sensitivity light 303 and Hb low-sensitivity light 304, a synthesizer 902 that combines them, and a ΔSR estimate 2(105) that holds the synthesis result. The synthesizer 902 can either combine the Hb high-sensitivity light 301 and Hb low-sensitivity light 302 as described above, or it can first calculate ΔSR2 from the Hb high-sensitivity light 303 and then combine the calculated result with the Hb low-sensitivity light 302. Since the Hb high-sensitivity light 301 changes significantly during high-intensity exercise, the former can only be used during light exercise intervals, etc. However, in the latter synthesis method where ΔSR2 is first calculated, switching, simple averaging, or weighted averaging can function advantageously depending on the Hb contamination error contained in the Hb low-sensitivity light 302.

[0108] When synthesizing Hb-high sensitivity light and ΔSR2 with Hb-low sensitivity light, the sensitivity to ΔSR usually differs, so calibration that takes this difference in sensitivity into account is typically required. Protocols that minimize ΔHb contamination, such as light exercise, can be incorporated into the calibration procedure. The measurement system can also be equipped with a mechanism that slightly changes the distance to the living body to alter ΔSR, which can be used in the calibration procedure. This mechanism can be used in various processes other than the synthesis process. Piezoelectric actuators or motors can be provided to create these slight changes. It is also possible to control the distance between the skin and the light transceiver to keep it constant so that ΔSR does not change.

[0109] Figure 22 shows the relationship between equations 1 to 11 and each unit. The ΔSR1 processor 103 can perform the processing of equations 1 to 2, for example. The sensitivity analyzer 104 can perform the processing of equation 3, for example. The ΔHb processor 106 can perform the processing of equations 4 to 11, for example. [Examples]

[0110] Figure 10 shows the results of an experiment conducted with the configurations shown in Figures 4 and 6. The signal from the PPG is used as the optically received signal. The aforementioned wavelengths of green, red, 740 nm, and near-infrared are used. The PPG is mounted on the temple of the glasses, and the back of the auricle is observed non-contact (remotely). (a) to (d) show the measurement results of normDCλ, ΔHb, ΔSR, and heart rate HR, respectively. Although heart rate HR can be extracted from the optically received signal from the PPG, the signal-to-noise ratio (SNR) was low, so the ECG (Electrocardiogram) signal measured simultaneously in the chest is also displayed.

[0111] In the experiment, participants sat quietly in a chair for the first 100 seconds, stood up at 100 seconds, walked in place from 200 to 300 seconds, and performed leg-raising walking from 300 to 500 seconds, with their thighs almost horizontal. The resting heart rate was approximately 75 bpm, the average heart rate during walking was approximately 90 bpm, and the average heart rate during leg-raising walking was approximately 100 bpm.

[0112] The normalized optical reception signal normDCλ increases at each wavelength after standing, changing by slightly less than 10% during walking compared to the resting state, and by about 10 to 20% during leg-raising walking. The PPG is held in place by glasses, which are held on the head at the nose pads and the tips of the temples, and this change is thought to be due to the displacement of the skin behind the auricle at the observation site. The distance between the PPG and the skin is several millimeters, and the skin displacement is thought to be several hundred microns. The waveforms of red and near-infrared light (R and IR in the figure) are almost identical, but the green light (G in the figure) changes about 20% more, and the 740nm light (740 in the figure) changes slightly less by a little over 10%. The reason for the large change in green light is thought to be that it penetrates to a shallow depth into the subcutaneous tissue. The reason for the small change in 740nm light is thought to be that the 740nm LED is placed outside the PPG, resulting in a large distance to the PD.

[0113] ΔHb in Figure 10(b) and ΔSR2 in Figure 10(c) are the results obtained by the optimization process of equations 10 and 11. The interval T is calculated to be 10 seconds. Optimization was performed using successive least squares programming (SLSQP) in the Python programming language. ΔSR1 is not shown, but the average of both red light and near-infrared light over the entire interval (up to 700 seconds) is used (Cλ is extracted using this ΔSR1). Since there is no significant difference in the results even when using ΔSR1 for the interval up to 400 seconds at the end of walking, it is considered that no large Hb changes occur even during leg-raising walking. ΔHbO, ΔHbR, and ΔHbTg in Figure 9 (magnified 5 times in Figure 9 to make the changes easier to see) show small changes, but these are not considered significant changes. ΔSR2 in Figure 10(c) increases by about 10% during walking and about 15% during leg-raising walking from the resting state. No large change occurs in ΔHb during this time. Possible reasons for this include the possibility that vasodilation occurred in the arteries upstream of the arterioles in the observation area, increasing ΔSR, but blood flow in the venules of the observation area did not increase, and the possibility that the auricle itself underwent displacement relative to the glasses (for example, displacement of a large blood vessel near the base of the auricle may have changed the tilt of the auricle relative to the head).

[0114] In another experiment where the exercise intensity during leg-raising walking was increased, a phenomenon was observed in which the difference in DC levels between infrared and near-infrared light reception signals widened (diverged). In this case, it is thought that the true component ΔHb is included in the light reception signal during leg-raising walking, and it is preferable to use the interval from leg-raising walking to the light exercise before the above-mentioned divergence phenomenon occurs as ΔSR1.

[0115] Figure 11 shows the section from 600 seconds onward in the same experiment. Cλ extraction was performed using Hb-high sensitivity light from 0 to 700 seconds as ΔSR1. The Valsalva maneuver was performed at 900 seconds and 1100 seconds. The Valsalva maneuver increases intrathoracic pressure by holding one's breath and straining, inducing sympathetic nerve stimulation and vasoconstriction, and parasympathetic nerve stimulation and vasodilation upon release of breath-holding. The room temperature was 19°C, and the auricle may have been vasoconstricted in the section prior to the Valsalva maneuver. After the increase and decrease in HR caused by the Valsalva maneuver, increases in ΔHbO and ΔHbTg occurred at points A and B in Figure 11(b). During this time, ΔHbR remained almost unchanged. This suggests that the Valsalva maneuver caused vasodilation in the peripheral system, with the increase in ΔHbO mainly occurring in the venules and capillaries. This phenomenon is normally buried within ΔSR and goes unobserved, but the configuration of this embodiment makes it possible to observe it for the first time.

[0116] A typical pulse oximeter observes the AC component of the light received from the PPG, which is the pulsation of arterial blood flow. This is because the pulsation disappears beyond the arteries. On the other hand, changes in the DC components, ΔHbO and ΔHbR, mainly observe the veins, which have a larger volume. Green light penetrates shallowly into the subcutaneous tissue, so it mainly observes capillaries. The changes observed in the Valsalva maneuver are thought to be in arterioles, which are called resistance vessels, but the ΔHb changes observed are thought to be in venules and capillaries. Since more than 90% of blood is normally ΔHbO, it is thought that ΔHbO was mainly observed as a change in the Valsalva maneuver (an increase in total ΔHbTg was observed with green light).

[0117] (Second embodiment) Figure 12 shows an example of a biosignal measurement method according to a second embodiment of the present invention, and is an example of a flowchart that mainly uses Hb high-sensitivity light as the optically received signal. It includes a step 1201 for acquiring the optically received signal, a step 1204 for acquiring a first ΔSR estimate 1203 from the optically received signal, a step 1207 for analyzing the sensitivity coefficient Cλ, which is the sensitivity of each optically received signal to the first ΔSR estimate 1203, and a ΔHb processing step 1215 for acquiring a second ΔSR estimate 1208 and determining the change in hemoglobin in the living organism. The execution order of the above steps and the steps described below can be changed as appropriate.

[0118] As the first ΔSR estimate 1203, the optically received signal when a specific task is given to the living organism can be used. Preferably, the specific task is one in which the change in ΔHb is small and the change in ΔSR is large. A section determination step 1202 can be included to determine if such a state exists by observing the living organism's heart rate and the DC level of the optically received signal. Alternatively, the average of multiple optically received signals may be used as the first ΔSR estimate 1203.

[0119] The step 1207 for analyzing the sensitivity coefficient Cλ may include, for example, a step 1205 for calculating the covariance between the optically received signal and the first ΔSR estimate 1203, a step 1206 for calculating the variance of the first ΔSR estimate 1203, and a step for dividing the covariance by the variance (not shown in the diagram) (e.g., Equation 3). This process can be performed for each wavelength of the optically received signal to obtain Cλ for each wavelength.

[0120] The ΔHb processing step 1215 may include, for example, a first multiplication step 1209 that calculates the product of a second ΔSR estimate 1208 and the sensitivity coefficient Cλ for each wavelength; a subtraction step 1211 (e.g., Equation 4) that subtracts the output of the first multiplication step 1209 from the optically received signal; a ΔHb analysis step 1213 that obtains multiple ΔHb values ​​from the output of the subtraction step 1211; and a ΔSR analysis step 1214 that analyzes the second ΔSR estimate 1208 while comparing these multiple ΔHb values. The ΔHb analysis step 1213 may include, for example, a second multiplication step 1212 (e.g., Equation 10) that obtains multiple ΔHb values ​​by multiplying the output of the subtraction step 1211 (ΔAλ) by the inverse matrix of the absorption coefficient 1210 (not shown in the figure). The ΔSR analysis step 1214 determines a second ΔSR estimate 1208 such that, for example, the difference between ΔHbO values ​​among multiple ΔHb values ​​falls within a predetermined error, and furthermore, the difference between ΔHbR values ​​falls within a predetermined error. For example, the process from setting the second ΔSR estimate 1208 to the second multiplication step 1212 is repeated until the resist in equation 11 falls within a predetermined error.

[0121] Figure 13 is an example of a flowchart that primarily uses Hb-low sensitivity light as the optical received signal for ΔSR estimation. It includes a step 1201 for acquiring the optical received signal, a step 1204 for acquiring a first ΔSR estimate 1203 from the Hb-low sensitivity light 1301 in the optical received signal, a step 1207 for calculating the sensitivity coefficient Cλ, which is the sensitivity of each optical received signal to the first ΔSR estimate 1203, and a ΔHb processing step 1215 for setting the Hb-low sensitivity light 1301 in the optical received signal as a second ΔSR estimate 1208 and calculating ΔHb from the second ΔSR estimate 1208 and the Hb-high sensitivity light 1302 in the optical received signal.

[0122] If the ΔHb change in the Hb-low sensitivity light 1301 is negligible, the Hb-low sensitivity light 1301 can be used as is as the first ΔSR estimate 1203. It is preferable that there is a significant ΔSR change in order to determine Cλ, and the optically received signal when a specific task is given to the living organism can be used. As a specific task, it is preferable that the change in ΔHb is small and the change in ΔSR is large. For this reason, the interval determination step 1202 described above can also be included in this example.

[0123] The ΔHb processing step 1215 can perform the same processing as in Figure 12 from the second ΔSR estimate 1208 to the ΔHb analysis step 1213. If the ΔHb change contained in the Hb low-sensitivity light 1301 is negligible, the Hb low-sensitivity light 1301 can be used as is as the second ΔSR estimate 1208. In this case, the first ΔSR estimate 1203 and the second ΔSR estimate 1208 may be the same.

[0124] If a light receiving signal with three or more wavelengths can be used as the Hb high-sensitivity light 1302, the accuracy of ΔHb calculation can be improved by performing optical path length correction. Similar to the case in Figure 8, there is a third multiplication step (not shown) between the subtraction step 1211 and the second multiplication step 1212, and an optical path length ratio analysis step (not shown) that compares multiple ΔHb outputs from the second multiplication step 1212. The inverse matrix of the optical path length ratio estimate (not shown) output by the optical path length ratio analysis step is input to the third multiplication step.

[0125] (Third embodiment) Figure 14 shows an example of a biosignal measurement device according to a third embodiment of the present invention, and is an overhead view of a configuration in which an optical transceiver is arranged in eyeglasses (for clarity, the eyeglasses are shown in an oblique view, and the user is shown from the back).

[0126] A sensor head 1403, which has an optical transceiver, is positioned on the end of the eyeglasses 1400, the temple 1402, and is connected to a control unit 1405 via wiring 1404. When a user wears the eyeglasses 1400, the sensor head 1403 is positioned in the space surrounded by the surface behind the ear 1407, the bottom of the area behind the ear 1408, and the temporal surface 1409. In this embodiment, the sensor head 1403 does not come into contact with the body in order to function as a remote sensor.

[0127] As already mentioned, with contact-type sensors, even when the sensor is lightly placed, there is a problem in that the curvature and expansion of the skin induce stress changes from the sensor to the skin. The eyeglasses 1400 are usually held in contact with the skin at the nose pads 1401 and the tips of the temples 1402. The temples of the eyeglasses are usually springy, and a clamping force acts on the head to prevent the eyeglasses from moving. This clamping force not only prevents lateral displacement of the body, but also fixes the eyeglasses in the front-to-back direction due to the frictional force from the clamping force and the support from the nose pads. Regarding the vertical direction of the eyeglasses, the weight of the eyeglasses and the frictional force from the clamping force of the temples act as forces to prevent displacement. In other words, the eyeglasses and the head are mechanically fixed up to a certain critical point. In this state, if the sensor comes into contact with somewhere behind the ear near the temple, the stress on the skin from the sensor changes when skin displacement occurs due to the displacement of blood vessels in the body. Normally, the skin is compressed by atmospheric pressure, which determines the subcutaneous blood flow and blood pressure. If we consider the biological signal in this free state as the true signal, the modulation that occurs when the sensor is in contact with the skin and the stress from the glasses changes due to skin displacement becomes an artifact that introduces an error into the true biological signal.

[0128] Normally, the tip of the temple 1402 of the eyeglasses 1400 contacts the temporal surface 1409 behind the ear, called the mastoid process. The distance from this contact point to the observation area at which the above artifact can be ignored may vary depending on the course of blood vessels, the condition of the hair on the mastoid process, and the load of the eyeglasses. Experiments by the inventors showed that the surface behind the ear 1407 and the base of the ear 1408, which are about 10 to 15 mm away from the mastoid process, are less affected by the stress changes of the mastoid process. On the other hand, there are somewhat thick blood vessels running along the base of the ear 1408, and moving the field of view of the optical transceiver away from this area tends to reduce the occurrence of the above artifact. Also, if the curved part of the temple 1402 (see Figure 14) makes strong contact with the auricle, the above artifact may occur even if the optical transceiver is more than 10 mm away from the contact point. Due to the course of blood vessels in the auricle, the blood flow in the auricle is interconnected, and compression of a part of the auricle may have an effect.

[0129] Figure 18 shows an example of the overall sensor configuration 1804, which includes a sensor head 1403 and a control unit 1405.

[0130] The sensor head 1403 mechanically holds the optical transceiver 1801, which is composed of, for example, a PPG or an external LED. In the remote type, the sensor head 1403 is fixed in the air across space. The sensor head 1403 is mechanically coupled to the glasses 1400, which are fixed to the skin by contact with the nose pads 1401 and the tips of the temples 1402, for example. The sensor head 1403 can also be fixed to headgear, headbands, headsets, etc., in addition to glasses. In situations where the living body is considered to be fixed, the sensor head 1403 can also be installed on a bed, arm, or other fixing device.

[0131] The sensor head 1403 can be incorporated into a fixed device, and the fixed device can also be equipped with only an optical receiving unit such as a PD that receives light. A rigid structure (rigid body) can be used that holds the optical receiving unit spaced apart in the observation area and is mechanically coupled to a biological part located 10 mm or more, preferably 15 mm or more, away from the observation area. The length, cross-sectional area, and Young's modulus of the structure must be considered to prevent artifacts from occurring due to body movement, etc. While it is necessary to separate the observation area from the biological contact area, it is preferable to minimize the separation distance and increase the Young's modulus of the structure.

[0132] The control unit 1405 can have a control circuit 1802 and a power supply 1803. The control circuit 1802 can control the timing of light emission from the optical transceiver 1801 and process the optical signals received from the optical transceiver 1801. Other sensors such as acceleration sensors and pressure sensors, and actuators such as piezoelectric elements can also be included and controlled by the control circuit 1802.

[0133] Power supply 1803 can use lithium-ion batteries or the like. To drive the optical transceiver 1801 and control circuit 1802, it typically has a mass of several grams and a length, width, and thickness of several millimeters. This mass and volume increase the volume and mass of the eyeglasses 1400 when power supply 1803 is placed within the eyeglasses 1400. The increased mass poses a risk of increased pressure on the skin at the nose pads and mastoid process. The increased volume poses a risk of increasing the number of contact points. Power supply 1803 only needs to have the ability to supply power, and wireless power transfer or solar cells can also be used.

[0134] In the configuration shown in Figure 14, the control unit 1405, including the power supply 1803, is placed outside the glasses 1400 to mitigate the above risks. The control circuit 1802 may be placed on the sensor head 1403 side if its volume and mass are small. If the wiring 1404 comes into contact with the auricle, it may cause artifacts due to pressure. It is preferable to attach the control unit 1405 to the back of the neck, for example, and route the wiring 1404 to avoid contact with the auricle or the area around the observation site.

[0135] Figure 15 is a perspective view of the user wearing glasses 1400 with the sensor head 1403 positioned on the front lens 1402. The view is observed from the upper left of the user. The sensor head 1403 is positioned in the air surrounded by the auricle 1406, the base of the earlobe 1408, and the temporal surface 1409 (1503 is the boundary between the base of the earlobe and the temporal surface). The sensor head 1403 is set in a position where the relatively large blood vessel 1501, which runs between the temporal region and the auricle, does not enter the observation area 1502 (the approximate bright area). In the perspective view, part of the observation area 1502 is hidden by the auricle 1406. The observation area 1502 is located in the region from the base of the earlobe 1407 (not visible on the back of the auricle in the perspective view) to the base of the earlobe 1408. The bottom of the earlobe 1408 is generally a surface parallel to the tip cell, but it has undulations. The front part a of the sensor head 1403 has an upward slope, while the rear part b has a hill with a triangular mark as its peak.

[0136] A PPG (hidden in the perspective view of Figure 15) is located on the underside of the sensor head 1403. The distance from the PPG to the skin is approximately 2 mm. The distance from the center of the PPG to the posterior end of the tip cell 1402 is approximately 20 mm. The posterior end of the tip cell 1402 is located at the mastoid process (position c). The light emission position 1504 of the optical transceiver 1801 can be confirmed from the user's profile or front view. The mounting position can be checked by examining the cartilage pattern of the auricle 1406, the distance from the outer circumference of the auricle 1406, and the relationship between the light emission position 1504 and the auricle.

[0137] Figure 16 is a similar perspective view seen from behind the user. The PPG1601 is visible on the underside of the sensor head 1403. An external LED is positioned so as to be in contact with the rear of the PPG1601 (not shown in the illustration). There is a distance of approximately 2 mm from the glass surface of the PPG1601 to the observation area 1502.

[0138] Figure 17 shows an example where the entire sensor 1804 is placed on the glasses 1400, eliminating external components and wiring. The sensor head 1403 can be placed, for example, on the end cell 1402. The control circuit 1802 and power supply 1803 can be placed in various locations on the glasses 1400. In particular, the power supply 1803, which has a large mass and volume, can be distributed across the glasses 1400 to avoid the concentration of stress and volume in a specific location.

[0139] Figure 19 shows an example of a circuit that adds an LED to the outside of the PPG. Outside the PPG1905, there is, for example, an LED1904, a first transistor 1903 that supplies current to the LED1904, a second transistor 1902 that controls the on / off state of the first transistor, and a resistor 1901.

[0140] LED1904 allows you to select LEDs that emit wavelengths around 740nm, for example. This wavelength lies somewhere between red and near-infrared, and it allows for a large difference in molar extinction coefficient compared to red and near-infrared LEDs. Many general-purpose PPGs use red and near-infrared wavelengths. Some also have a green wavelength. Green LEDs have a higher forward bias voltage Vf than longer wavelength LEDs. Therefore, when a 740nm LED and a green LED are connected in parallel, almost all of the current flows to the 740nm LED.

[0141] In the example in Figure 19, a PPG1905 with an externally exposed cathode terminal is used to connect the cathode of a green LED1906 to the cathode of an externally placed 740nm LED1904. When the P-type MOSFET1903 connected to the anode of the external LED1904 is on, current flows when the green light emission of the PPG is selected. The voltage between the power supply voltage 1909 and the cathode is the Vf at 740nm and will not exceed the Vf of the green LED. The N-type transistor 1902 and resistor 1901 form a level conversion circuit, and the logic signal input to the gate of the N-type transistor 1902 controls the on / off state of the P-type transistor 1903. Depending on the PPG circuit configuration, it is possible to replace the connection between cathodes with anodes, reverse the polarity of P-type and N-type transistors, replace transistors and level conversion circuits with equivalent ones as appropriate, or omit the level conversion circuit.

[0142] The PPG has multiple wavelength LEDs, such as green 1906, red 1907, and near-infrared 1908, a photodiode (PD) 1910, and an ADC 1911 that converts the received light signal digitally. For example, the light emission of multiple LEDs is sequentially switched and received by a single PD. For example, during the time interval when the green LED is emitting light, a green light received signal is obtained by the PD. During the interval when the second transistor 1903 is turned on, a 740nm light received signal is obtained instead of green light. Although a green light received signal is not obtained during this time, the changes in ΔHb and ΔSR are relatively slow, so there is little problem.

[0143] (Fourth embodiment) Figure 20(a) shows an example of a contact type. As mentioned above, in the contact type, particularly on uneven skin surfaces, strain occurs between the displacement of the skin due to blood vessel displacement and the sensor in contact, generating stress. To minimize this, it is preferable to 1) reduce the load on the sensor to reduce the stress itself, and 2) increase the contact area to distribute the stress. In case 2), if the mass of the elastic material or other component in contact with the skin is dominant, the load (stress) per unit area does not change even if the area is increased. It is preferable to make the mass of the optical transceiver and the substrate that holds it as dominant as possible and to distribute it.

[0144] When an adhesive is used on the contact surface between the skin and the sensor, a spring stress acts between the skin and the solid sensor when the skin is displaced. This is because the distance between points A and B on the curvature changes due to the change in curvature. For example, if the distance between A and B on the skin increases, the distance between A and B on the sensor remains unchanged, so the adhesive acts to pull back the skin's displacement. The skin, including the subcutaneous blood vessels, is pulled and deformed. In experiments conducted by the inventor, when a sensor was attached to the back of the auricle with medical-grade double-sided tape (e.g., 3M 2477P) and light exercise was performed, artifacts due to skin deformation were observed. After the experiment, redness was observed on the skin, which appeared to be due to localized vascular compression or occlusion.

[0145] In the example shown in Figure 20(a), there is a skin contact portion 2001, an opening 2003 provided in the skin contact portion 2001, an optical transceiver 2002 provided on the side opposite to the side of the skin contact portion 2001 that contacts the skin, and wiring 1404.

[0146] An elastic material such as silicone or alginate can be used as the skin contact portion 2001. To reduce stress on the skin, it is preferable that the specific gravity of this elastic material is low. Furthermore, it is preferable that it has a hardness sufficient to distribute the mass of the optical transceiver 2002. To distribute stress on the skin, it is preferable that it conforms to the contours of the skin at the observation site. An impression material can be used to transfer the contours of the skin. Alternatively, an impression material such as silicone can be used, and the transferred impression material itself can be used as the skin contact portion 2001.

[0147] The skin contact area 2001 in Figure 20(a) is formed by taking an impression of the area from the surface 1407 to the bottom 1408 of the back of the ear using silicone impression material. The black triangular marks in the figure represent the vertices of the convex parts of the ridges (concave parts on the skin), and the white triangular marks represent the vertices of the concave parts of the ridges (convex parts on the skin). The right side of Figure 20(a) is the front of the user (direction from the ear towards the eye). The opening 2003 is an opening for shining light from the optical transceiver 2002 onto the skin. A transparent film can be attached to the skin-facing side of this part as well to contribute to stress distribution.

[0148] The optical transceiver 2002 is composed of, for example, a PPG, an external LED, and a PWB (printed circuit board) that holds them. A film-like material with low mass can be used as the PWB. The wiring 1404 is electrically connected to the control unit 1405, which is positioned, for example, at the back of the neck. To avoid pressure on the skin, it is preferable to use a lightweight and flexible wire for the wiring 1404. For example, a bundle of polyurethane wires with a diameter of about 0.1 mm can be used.

[0149] Figure 20(b) is a perspective view (observed from above the user's head) of the sensor shown in Figure 20(a) when worn by a user. The left side of the figure is the front of the user. It is attached to the area from the bottom of the earlobe 1408 to the surface of the earlobe 1407 (not visible in the perspective view). Part of the sensor is hidden behind the auricle 1405. The back surface of the PWB is visible on the optical transceiver 2002, and the PPG1601 and external LEDs are positioned on the surface of the PWB. The relatively large blood vessel 1501 is a blood vessel running between the temporal region 1408 and the auricle 1405, and is outside the observation area. When attached to the skin, applying a material with weak adhesion and fluidity, such as petroleum jelly, can reduce the risk of skin compression while reducing the risk of the sensor falling off.

[0150] (Fifth embodiment) Figure 21 shows an example of a biosignal measurement system according to a fifth embodiment of the present invention. The biological body 2101 can have a remote sensor 2102, a contact sensor 2103, an actuator 2104, an edge processing device 2105, and a cloud processing device 2106.

[0151] The remote sensor 2102 has an optical sensor and other sensors and acquires biological signals without compressing the skin in the observation area. The configurations described in Embodiments 1 to 3 can be used as appropriate. The contact sensor 2103 is attached at a location away from the observation area and acquires biological signals using an optical sensor and other sensors. It is preferable to attach it in a location that does not affect the blood flow, etc., in the observation area of ​​the remote sensor 2102. For example, heart rate information can be acquired by ECG to supplement missing biological information. The actuator 2104 can stimulate the living body with heat, stress, vibration, electromagnetic waves including light, etc. Biological information that changes depending on the type and magnitude of the stimulation can be measured by the remote sensor 2102 or the contact sensor 2103.

[0152] The edge processing unit 2105 is placed in or near a living organism to perform processing with demanding latency and transmission rate requirements. The cloud processing unit 2106 is placed on a server with sufficient computing resources to analyze biological signals, create models from information of multiple users, and infer relationships between multiple models. It can also infer which model a user's biological information is closest to. These models include, for example, models of organs called receptors that monitor blood pressure and body temperature, models of the central nervous system such as the circulatory center and thermoregulatory center, models of the autonomic nervous system, models of blood vessels and blood flow called windkessel models, and models of vasomotor processes that modulate blood flow. Furthermore, there are models related to healthy individuals, disease cases, gender, age, and other attributes. These models can be built on a computer to reproduce or simulate living organisms on the computer and infer past, present, and future biological conditions. Measurement values ​​can also be calibrated and validated using catheters, MRI, laser Doppler flowmeters, and other biological information in combination. [Explanation of Symbols]

[0153] 101 Optical received signal 102 ΔSR Estimate 1 103 ΔSR1 processor 104 Sensitivity analyzer 105 ΔSR Estimate 2 106 ΔHb treatment unit 201 Interval Detector 202 Covariance 203 Dispersion 301 Hb high sensitivity light 302 Hb low sensitivity light

Claims

1. A light receiving unit that acquires optical receiving signals of multiple wavelengths containing biological information, and a ΔSR1 processor that acquires an estimated value of a first skin surface reflected light change amount from at least a portion of the optical receiving signals of multiple wavelengths, A sensitivity analyzer that determines the sensitivity coefficients of at least some of the multiple wavelengths of light received signals with respect to the first amount of change in reflected light from the skin surface, A biosignal measurement device comprising a ΔHb processor that obtains an estimated value of a second surface reflected light change from at least a portion of the aforementioned multi-wavelength optical received signals to determine the hemoglobin change of a living organism.

2. The ΔHb processor, A multiplier that multiplies the aforementioned sensitivity coefficient by the second surface reflected light change amount, A subtractor that subtracts the multiplication result from at least a portion of the optical received signals of multiple wavelengths to determine the absorbance change for each wavelength, The biological signal measurement device according to claim 1, further comprising a ΔHb analyzer that determines the amount of change in the hemoglobin of the living organism from the absorbance change and the absorption coefficient of each wavelength.

3. The biosignal measuring device according to claim 2, further comprising a time interval determination device that provides a time interval in at least a portion of the multiple wavelength light reception signal according to the heart rate of the living organism or the DC change amount of at least a portion of the multiple wavelength light reception signal, and obtains an estimated value of the first skin surface reflected light change amount.

4. The aforementioned optical receiving unit acquires optical receiving signals consisting of three or more wavelengths of Hb-sensitive light that has high sensitivity to hemoglobin. The ΔSR1 processor obtains an estimated value of the first skin surface reflected light change amount from the light received signal consisting of three or more wavelengths of Hb high-sensitivity light. The biosignal measurement device according to any one of claims 1 to 3, further comprising a ΔSR analyzer which the ΔHb processor obtains multiple solutions for the hemoglobin change amount from a light reception signal of three or more wavelengths consisting of the Hb high-sensitivity light, and optimizes the second surface reflected light change amount so that the multiple solutions fall within a predetermined error.

5. The biosignal measuring device according to claim 4, wherein the ΔSR1 processor obtains an estimated value of the first skin surface reflected light change amount from the average of one or at least a portion of the three or more wavelengths of received light signals.

6. The biosignal measurement device according to claim 4, wherein the ΔHb processor applies the optimized second surface reflected light change amount to optical received signals of wavelengths other than the three or more wavelength optical received signals consisting of the Hb high-sensitivity light used to find the solution for the hemoglobin change amount, thereby determining the hemoglobin change amount at that wavelength.

7. The optical receiving unit acquires an optical receiving signal consisting of at least one Hb low-sensitivity light having low sensitivity to hemoglobin and a plurality of Hb high-sensitivity light having high sensitivity to hemoglobin. The ΔSR1 processor obtains an estimated value of the first skin surface reflected light change amount from the light received signal of the Hb low sensitivity light, The biosignal measuring device according to any one of claims 1 to 3, wherein the ΔHb processor obtains an estimated value of the second skin surface reflected light change amount from the light received signal of the Hb low sensitivity light.

8. The biosignal measuring device according to claim 7, wherein the Hb processor determines the amount of change in the hemoglobin of the living organism using the second amount of change in reflected light from the skin surface and the plurality of Hb-sensitive lights.

9. The biosignal measurement device according to claim 8, further comprising an optical path length ratio analyzer that obtains multiple solutions for the hemoglobin change amount from a light received signal of three or more wavelengths consisting of the Hb high-sensitivity light, and optimizes the optical path length ratio estimate so that the multiple solutions fall within a predetermined error.

10. The optical receiving unit is fixed in a space separated from the observation area of ​​the living organism, The biosignal measuring device according to any one of claims 1 to 3, comprising a structure that holds the light receiving unit and contacts a biological site located 10 mm or more, preferably 15 mm or more, from the observation area.

11. An optical transmitting unit positioned in close proximity to the optical receiving unit, The biosignal measuring device according to claim 10, further comprising a control unit including a power supply arranged in the aforementioned structure.

12. A light reception step in which multiple wavelengths of light received signals are acquired, A ΔSR1 processing step is performed to obtain an estimated value of the first skin surface reflected light change amount from at least a portion of the optical reception signals of multiple wavelengths, A sensitivity analysis step to determine the sensitivity coefficients of at least some of the multiple wavelengths of light received signals with respect to the first amount of change in reflected light from the skin surface, The system includes a ΔHb processing step to obtain an estimated value of a second surface reflected light change from at least a portion of the multiple wavelength light received signals to determine the hemoglobin change of the living organism. The ΔHb processing step is, A multiplication step in which the sensitivity coefficient and the second surface reflected light change amount are multiplied, A subtraction step to obtain the change in absorbance for each wavelength by subtracting the multiplication result from at least a portion of the optical received signals of multiple wavelengths, A biosignal measurement method comprising a ΔHb analysis step for determining the amount of hemoglobin change in the living organism using the absorbance change and the absorption coefficient of each wavelength.

13. The aforementioned light reception step acquires a light reception signal consisting of three or more wavelengths of Hb-sensitive light that has high sensitivity to hemoglobin. The ΔSR1 processing step obtains an estimated value of the first skin surface reflected light change amount from the light received signal of three or more wavelengths consisting of the Hb high-sensitivity light, The method for measuring a biological signal according to claim 12, further comprising a ΔSR analysis step in which the ΔHb analysis step obtains multiple solutions for the hemoglobin change amount from a light reception signal of three or more wavelengths consisting of the Hb high-sensitivity light, and optimizes the second surface reflected light change amount so that the multiple solutions fall within a predetermined error.

14. The light reception step acquires a light reception signal consisting of at least one Hb low-sensitivity light having low sensitivity to hemoglobin and a plurality of Hb high-sensitivity light having high sensitivity to hemoglobin. The ΔSR1 processing step obtains an estimated value of the first skin surface reflected light change amount from the light received signal of the Hb low sensitivity light, The biosignal measurement method according to claim 12, wherein the ΔHb processing step obtains an estimated value of the second skin surface reflected light change amount from the light received signal of the Hb low sensitivity light.

15. A biological measurement system comprising a biological measurement device according to claims 1 to 3, an edge processing device for processing the biological information acquired by the biological measurement device near the living body, and a cloud processing device for processing the biological information far from the living body.

Images