Biological signal processing system, signal processing device, computer program, and biological signal generation method

JP2024129505A5Pending Publication Date: 2026-03-13KOBE UNIV +1
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Filing Date
2023-03-13
Publication Date
2026-03-13

AI Technical Summary

Technical Problem

Existing biological signal measurement systems require precise placement of a predetermined number of electrodes at specific positions, which can be inconvenient and difficult to achieve, limiting the ability to obtain necessary multi-channel biosignals with non-standard electrode arrangements.

Method used

A biological signal processing system using a flexible electrode device with integrated electrodes on a single sheet, combined with a generative model that can convert signals from one electrode arrangement to another, allowing for accurate signal reconstruction even with different electrode placements.

Benefits of technology

Enables convenient and accurate acquisition of multi-channel biological signals, such as a standard 12-lead electrocardiogram, without the need for precise electrode placement, facilitating easier use and potential for telemedicine applications.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 00000000_0000_ABST
    Figure 00000000_0000_ABST
Patent Text Reader

Abstract

To acquire necessary biological signals of a plurality of channels even in another electrode arrangement different from a predetermined electrode arrangement.SOLUTION: A biological signal processing system includes: m pieces of electrodes mounted on a living body in a first arrangement; and a signal processing device for executing processing to first biological signals of p channels acquired from the m pieces of electrodes in a generative model and output second biological signals of q channels from the generative model. The second biological signals of q channels are signals equivalent to signals obtained from n pieces of electrodes mounted on the living body in a second arrangement (n being the integer of three or more). The second arrangement is an electrode arrangement different from the first arrangement in at least any one of the number of electrodes and the mounting position. The generative model is configured to output the second biological signals of the q channels when the first biological signals of the p channels are input.SELECTED DRAWING: Figure 5
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present disclosure relates to a biological signal processing system, a signal processing device, a computer program, and a method for generating a biological signal. [Background technology]

[0002] Patent Document 1 discloses a biosignal measuring device. The device in Patent Document 1 uses electrodes to measure changes in potential over time, such as electrical activity in a living body. Patent Document 1 discloses that the device can be applied to an electroencephalograph, as well as an electrocardiograph, etc. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] International Publication No. 2017 / 122379 Summary of the Invention

[0004] An example of the measurement of biological signals is measurement to obtain a standard 12-lead electrocardiogram. A standard 12-lead electrocardiogram is composed of 12 electrocardiograms (12-channel electrocardiogram). A total of 10 electrodes are used to record a standard 12-lead electrocardiogram: four limb electrodes and six chest electrodes. To obtain an appropriate standard 12-lead electrocardiogram, a specified number (10) of electrodes must be attached accurately to specified positions on the living body.

[0005] Thus, when recording a standard 12-lead ECG or other required multi-channel biosignals, a certain number of electrodes must be placed at certain locations on the body, and the required multi-channel biosignals can be obtained with certain electrode configurations but not previously with other electrode configurations.

[0006] However, if a predetermined electrode arrangement is required in which a predetermined number of electrodes are arranged at predetermined positions on the living body, this may lead to a decrease in convenience. In addition, it may be difficult to obtain the predetermined electrode arrangement. Therefore, it is desirable to be able to obtain the necessary multi-channel biosignals even with an electrode arrangement other than the predetermined electrode arrangement.

[0007] An aspect of the present disclosure is a biosignal processing system. The disclosed biosignal processing system includes m (m is an integer of 3 or more) electrodes attached to a living body in a first arrangement, and a signal processing device that executes a process of inputting a first biosignal of p (p is an integer of 2 or more) channels acquired from the m electrodes to a generation model and outputting a second biosignal of q (q is an integer of 2 or more) channels from the generation model, the second biosignal of the q channels being a signal equivalent to a signal obtained from n (n is an integer of 3 or more) electrodes attached to the living body in a second arrangement, the second arrangement being an electrode arrangement different from the first arrangement in at least one of the number of electrodes and the attachment positions, and the generation model being configured to output the second biosignal of the q channel when the first biosignal of the p channel is input.

[0008] Other aspects of the present disclosure include a signal processing device, a method for generating a biological signal, or a computer program, which will be described in further detail in the following embodiments. [Brief description of the drawings]

[0009] [Figure 1] FIG. 1 is a configuration diagram of a biological signal processing system. [Diagram 2] FIG. 2 is a cross-sectional view of the electrode device taken along line AA. [Diagram 3] FIG. 3 is an explanatory diagram showing a state in which electrodes are attached using the electrode device according to the embodiment. [Figure 4] FIG. 4 is an explanatory diagram showing the electrode positions for a standard 12-lead electrocardiogram. [Diagram 5] FIG. 5 is a flowchart showing a processing procedure in the signal processing device. [Figure 6]FIG. 6 is a waveform showing the first biological signal. [Figure 7] FIG. 7 is an explanatory diagram of the learning phase and inference phase of a generative model. [Figure 8] FIG. 8 is a waveform showing the second biological signal. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] <1. Overview of the biosignal processing system, the signal processing device, the biosignal generating method, and the computer program>

[0011] (1) The system according to the embodiment may be a biosignal processing system. The biosignal processing system may include m (m is an integer of 3 or more) electrodes attached to a living body in a first arrangement, and a signal processing device that executes a process of inputting a first biosignal of p (p is an integer of 2 or more) channels acquired from the m electrodes to a generative model and outputting a second biosignal of q (q is an integer of 2 or more) channels from the generative model. The second biosignal of the q channel may be a signal equivalent to a signal obtained from n (n is an integer of 3 or more) electrodes attached to a living body in a second arrangement. The second arrangement may be an electrode arrangement different from the first arrangement in at least one of the number of electrodes and the attachment positions. The generative model may be configured to output the second biosignal of the q channel when the first biosignal of the p channel is input.

[0012] According to the biosignal processing system of the embodiment, a second biosignal corresponding to a signal obtained from electrodes in a second arrangement can be generated from a first biosignal acquired from electrodes in a first arrangement different from the second arrangement. In other words, even if the electrode arrangement in the first arrangement is different from a predetermined second arrangement, a necessary second biosignal can be obtained.

[0013] (2) The m electrodes attached to the living body in the first arrangement are preferably provided on a single base with the relative positions of the m electrodes fixed. Fixing the relative positions of the electrodes makes it easy to attach the electrodes.

[0014] (3) It is preferable that the generative model is a trained model that has been machine-learned to output the second biological signal of the q channel when the first biological signal of the p channel is input.

[0015] (4) The second biological signal of the q channel is preferably a signal of 12 channels in a standard 12-lead electrocardiogram or a signal of less than 12 channels included in a standard 12-lead electrocardiogram.

[0016] (5) It is preferable that all of the m electrodes from which the first biological signals of the p channels are acquired are attached to the center of the chest of the living body.

[0017] (6) The device according to the embodiment may be a signal processing device that executes signal processing. The signal processing may include acquiring a first biosignal of p (p is an integer of 2 or more) channels from m (m is an integer of 3 or more) electrodes attached to a living body in a first arrangement, inputting the first biosignal of the p channels into a generation model, and outputting a second biosignal of q (q is an integer of 2 or more) channels from the generation model. The second biosignal of the q channels may be a signal equivalent to a signal obtained from n (n is an integer of 3 or more) electrodes attached to a living body in a second arrangement. The second arrangement may be an electrode arrangement different from the first arrangement in at least one of the number of electrodes and the attachment positions. The generation model may be configured to output the second biosignal of the q channel when the first biosignal of the p channels is input.

[0018] (7) The method according to the embodiment may be a method for generating a biosignal. The generation method may include acquiring a first biosignal of p (p is an integer of 2 or more) channels from m (m is an integer of 3 or more) electrodes attached to a living body in a first arrangement, inputting the first biosignal of the p channels into a generation model, and outputting a second biosignal of q (q is an integer of 2 or more) channels from the generation model. The second biosignal of the q channels may be a signal equivalent to a signal obtained from n (n is an integer of 3 or more) electrodes attached to a living body in a second arrangement. The second arrangement may be an electrode arrangement different from the first arrangement in at least one of the number of electrodes and the attachment positions. The generation model may be configured to output the second biosignal of the q channel when the first biosignal of the p channels is input.

[0019] (8) A computer program according to an embodiment may be a computer program that causes a computer to execute signal processing. The signal processing may include acquiring a first biosignal of p (p is an integer of 2 or more) channels from m (m is an integer of 3 or more) electrodes attached to a living body in a first arrangement, inputting the first biosignal of the p channels into a generation model, and outputting a second biosignal of q (q is an integer of 2 or more) channels from the generation model. The second biosignal of the q channels may be a signal equivalent to a signal obtained from n (n is an integer of 3 or more) electrodes attached to a living body in a second arrangement. The second arrangement may be an electrode arrangement different from the first arrangement in at least one of the number of electrodes and the attachment positions. The generation model may be configured to output the second biosignal of the q channel when the first biosignal of the p channels is input.

[0020] The computer program may be stored in a non-transitory computer readable storage medium.

[0021] <2. Examples of overviews of biosignal processing systems, signal processing devices, computer programs, and biosignal generating methods>

[0022] 1 shows a biosignal processing system 1 according to an embodiment. The system 1 includes an electrode device 10 and a signal processing device 20. The electrode device 10 measures a first biosignal. The signal processing device 20 reconstructs a second biosignal, which is different from the first biosignal, from the first biosignal.

[0023] The electrode device 10 is attached to the surface of a living body to measure a first biosignal. The first biosignal is a signal measured by the electrode device 10 according to the embodiment. The living body is, for example, a human. The biosignal is, for example, the body surface potential of the living body. The body surface potential is, for example, a weak potential (tens of microvolts to tens of millivolts) generated by activity of the heart and other muscles.

[0024] The electrode device 10 includes a plurality of electrodes E0 to E15. Here, the "plurality of electrodes E0 to E15" is also referred to as "m electrodes E0 to E15". It is preferable that m is an integer equal to or greater than 3. Each of the electrodes E0 to E15 is attached so as to be in contact with the body surface in order to measure the body surface potential. The electrode device 10 is attached to the body surface so that each of the electrodes E0 to E15 is in contact with the body surface.

[0025] The electrode device 10 includes a sheet 11 as a base supporting the electrodes E0 to E15. The sheet 11 is made of, for example, a thin film of synthetic resin or other material. The sheet 11 is preferably flexible. The flexibility of the sheet 11 allows the sheet 11 to be attached in accordance with the shape of the body surface. The sheet 11 is preferably stretchable in the planar direction. The sheet 11 is preferably transparent or translucent.

[0026] As shown in Fig. 2, the sheet 11 has a first surface 11a and a second surface 11b opposite to the first surface 11a. The first surface 11a is a surface that is attached to the body surface of a living organism. The first surface 11a has adhesiveness so that the sheet 11 can be attached to the body surface. The adhesiveness may be inherent to the sheet 11 itself, or may be obtained by applying an adhesive to the first surface 11a or by carrying out other surface treatments to generate adhesiveness.

[0027] The m electrodes E0 to E15 are provided on the first surface 11a. The m electrodes E0 to E15 are provided so as to be exposed on the first surface 11a. Therefore, when the sheet 11 is attached to the body surface so that the first surface 11a is in contact with the body surface, the electrodes E0 to E15 come into contact with the skin, which is the body surface. Each of the electrodes E0 to E15 measures the body surface potential at the position where it is attached.

[0028] In the electrode device 10 of the embodiment, all of the electrodes E0 to E15 are provided in a concentrated manner on a single sheet 11, so that the electrodes E0 to E15 can be worn by a person in a single step of attaching the sheet 11. This makes it easy to wear the electrode device 10.

[0029] In the electrode device 10 of the embodiment, all of the electrodes E0 to E15 are fixedly provided on the sheet 11. As a result, the relative positional relationship of the electrodes E0 to E15 is regulated and fixed by the single sheet 11 (base). The fact that the relative positional relationship of the electrodes E0 to E15 is fixed is advantageous in terms of accurately reconstructing the second biological signal.

[0030] Wiring 12 is also provided on first surface 11a of sheet 11. Wiring 12 extends from each of electrodes E0 to E15 to communication device 15 attached to sheet 11. Wiring 12 is connected to communication device 15 via a connector (not shown).

[0031] The communication device 15 transmits the first biosignal measured by the electrodes E0 to E15 to the signal processing device 20. The communication device 15 includes, for example, a signal processing circuit and a wireless circuit. The signal processing circuit includes, for example, an amplifier for amplifying the first biosignal measured by each of the electrodes E0 to E15, and an AD converter for converting the amplified signal into a digital signal. The wireless circuit wirelessly transmits a wireless signal obtained by modulating the digital signal to the signal processing device 20. The communication device 15 may transmit the first biosignal to the signal processing device 20 by wired transmission. The signal processing device 20 receives the first biosignal transmitted from the communication device 15, and generates a second biosignal by using the first biosignal to execute processing described below.

[0032] 3(A), as an example, the electrode device 10 is attached to a person so that the electrodes E0 to E15 are located at the center of the person's chest. Attaching the electrodes E0 to E15 to the center of the chest is advantageous because many of the electrodes E0 to E15 are located at positions on the body surface that overlap the heart H. Since the body surface potential changes greatly on the body surface near the heart H (near the center of the chest) as described below, by arranging a large number of integrated electrodes E0 to E15 on the body surface near the heart H (near the center of the chest), it is possible to obtain a lot of information related to the body surface potential, which is advantageous.

[0033] Here, the human chest is the area of ​​the front body surface where the sternum and ribs are present. The center of the chest is the approximate center position in the left-right direction of the chest. Since the heart is located in the center of the chest (slightly to the left), the center of the chest is the position near the heart.

[0034] FIG. 3(B) shows an example of the body surface potential distribution of a person's upper body including the chest. In FIG. 3(B), 16 white circles represent 16 electrodes E0 to E15 attached to the center of the chest. In FIG. 3(B), white lines drawn on the human body indicate that the positions of the white lines are at the same potential. Also, in FIG. 3(B), A represents either the position with the highest or lowest potential, and B represents the other. The potentials of A and B are generated at both ends of the heart. As shown in FIG. 3(B), the white lines are densely packed between A and B, and the change in potential is large.

[0035] As shown in FIG. 3(B), most of the potential values ​​present on the surface of a person's body are distributed in the center of the chest (roughly corresponding to the area where 16 white circles exist in FIG. 3(B)). That is, almost all of the white lines showing the same potential in FIG. 3(B) pass through the center of the chest. The distribution of potential in areas other than the center of the chest is determined by the distribution of the white lines that pass near the center of the chest. Therefore, information on the potential at multiple points within the center of the chest is useful for estimating the potential at body surface positions other than the center of the chest. Therefore, by using the first biosignal obtained by attaching electrode device 10 to the center of the chest, it is possible to obtain the second biosignal at other positions.

[0036] In the embodiment, as an example, the signal processing device 20 generates a second biosignal corresponding to a standard 12-lead electrocardiogram using the first biosignal. In general, when obtaining a standard 12-lead electrocardiogram, potentials at the attachment positions of the four limb electrodes R, L, F, and N and the six chest electrodes C1, C2, C3, C4, C5, and C6, totaling ten electrodes, are used to measure potentials at the attachment positions of the electrodes.

[0037] FIG. 4 shows the attachment positions of ten electrodes (n=10) R, L, F, N, C1, C2, C3, C4, C5, and C6 for a standard 12-lead electrocardiogram. As shown in FIG. 4(A), electrode R is attached to the right hand. Electrode L is attached to the left hand. Electrode F is attached to the left foot. Electrode N is attached to the right foot. As shown in FIG. 4(A), the four limb electrodes R, L, F, and N are used to obtain leads I, II, III, and aV. R Induction, aVL Induction, aV F A total of six channels of electrocardiogram (second biosignal) can be obtained through leads.

[0038] As shown in FIG. 4(B), electrodes C1, C2, C3, C4, C5, and C6 are attached to the chest. Electrode C1 is attached to the right edge of the sternum at the fourth intercostal space, and is used to obtain an electrocardiogram in lead V1. Electrode C2 is attached to the left edge of the sternum at the fourth intercostal space, and is used to obtain an electrocardiogram in lead V2. Electrode C3 is attached to the midpoint between electrodes C2 and C4, and is used to obtain an electrocardiogram in lead V3. Electrode C4 is attached to the intersection of the fourth intercostal space and the left midclavicular line, and is used to obtain an electrocardiogram in lead V4. Electrode C5 is attached to the intersection of a horizontal line at the same height as electrode C4 and the left anterior axillary line, and is used to obtain an electrocardiogram in lead V5. Electrode C6 is attached to the intersection of a horizontal line at the same height as electrode C4 and the left mid-axillary line, and is used to obtain an electrocardiogram in lead V6.

[0039] In this way, using the six electrodes C1, C2, C2, C3, C4, C5, and C6, a total of six channels of electrocardiograms (second biological signals) can be obtained, including leads V1, V2, V3, V4, V5, and V6.

[0040] Therefore, as a standard 12-lead electrocardiogram, a total of 12 channels of electrocardiograms (second biological signals) can be obtained: 6 channels using the four limb electrodes R, L, F, and N, and 6 channels using the six chest electrodes C1, C2, C2, C3, C4, C5, and C6.

[0041] The 10 electrodes (n=10) R, L, F, N, C1, C2, C3, C4, C5, and C6 for a standard 12-lead electrocardiogram are separate from each other, and each electrode must be attached accurately in the designated position. If each electrode is not attached accurately in the designated position, the electrocardiogram obtained will not be useful for diagnosis. However, it can be difficult even for medical professionals to attach these 10 electrodes R, L, F, N, C1, C2, C3, C4, C5, and C6 in the correct positions. Also, if there is no medical professional who is familiar with attaching electrodes, it may not be possible to measure the electrocardiogram.

[0042] However, when the system 1 of the embodiment is used, a second biosignal corresponding to a standard 12-lead electrocardiogram can be obtained from the first biosignal obtained by the electrode device 10. This reduces the burden of wearing electrodes.

[0043] Returning to Fig. 3(A), in the electrode device 10 of the embodiment, at least one electrode, preferably multiple electrodes E0 to E15, are attached to the center of the chest so as to overlap the position of the heart. Here, the arrangement of the electrodes E0 to E15 when the electrode device 10 is attached to a living body is referred to as the "first arrangement". Fig. 3(A) shows an example of the first arrangement. The first arrangement shown in Fig. 3(A) is an electrode arrangement in which 16 integrated electrodes E0 to E15 are attached to the center of the chest.

[0044] An electrode arrangement different from the first arrangement is referred to as a "second arrangement." It is sufficient that the second arrangement differs from the first arrangement in at least one of the number of electrodes and the attachment positions. The electrode arrangement for a standard 12-lead electrocardiogram shown in Figure 4 is an example of the second arrangement. The second arrangement shown in Figure 4 differs from the first arrangement shown in Figure 3 in both the number of electrodes and the attachment positions. In the second arrangement, the electrodes are attached over a wider area compared to the first arrangement, in which the electrodes are concentrated in the center of the chest.

[0045] In the electrode device 10 shown in Figs. 1 and 3, m indicating the number of electrodes E0 to E15 is 16. That is, the electrode device 10 shown in Fig. 1 includes 16 electrodes E0 to E15. These electrodes E0 may include one or more reference electrodes E15. The reference electrode E15 is an electrode for obtaining a reference biosurface potential (reference potential). In the electrode device 10 shown in the figure, the electrode E15 at the left corner is the reference electrode. Of the m electrodes E0 to E15, the electrodes E0 to E14 other than the reference electrode E15 are measurement electrodes for measuring a p-channel biosignal (biosurface potential) provided to the signal processing device 20. Here, the 15 measurement electrodes E0 to E14 are used to measure 15-channel body surface potentials (first biosignals) based on the reference potential. Note that the potentials measured by the measurement electrodes E0 to E14 may be body surface potentials based on the other measurement electrodes E0 to E14. By also measuring the body surface potentials based on the other measurement electrodes E0 to E14, it is possible to obtain first biosignals (biosurface potentials) for more channels even if the number of electrodes is the same.

[0046] Thus, the illustrated electrode device 10 has, as an example, 16 electrodes (m=16) and can measure the first biosignals of 15 (p=15) channels.

[0047] 1 and 3, the electrodes E0 to E15 are arranged in a two-dimensional array on a sheet 11. More specifically, the 16 electrodes E0 to E15 are arranged in a two-dimensional array of four columns and four rows. Note that the arrangement of the electrodes E0 to E15 is not limited to a two-dimensional array, and other arrangements are also possible.

[0048] The number m of the electrodes E0 to E15 included in the electrode device 10 is preferably at least 3 or more in order to obtain biosignals of multiple channels. The number m of the electrodes E0 to E15 included in the electrode device 10 is more preferably 9 to 25, and even more preferably 11 to 20.

[0049] The electrodes E0 to E15 included in the electrode device 10 are preferably integrated together to make the electrode device 10 compact and easy to handle. For example, it is preferable that all the electrodes E0 to E15 are integrated together so as to be located within a reference range X of a predetermined size.

[0050] Here, the reference range X is, for example, a square range large enough to position all of the electrodes E0 to E15 inside. The length of one side of the square constituting the reference range X is preferably 30 cm or less. If all of the electrodes E0 to E15 are integrated so as to fit within a square with one side of about 30 cm, it becomes easy to attach the electrodes E0 to E15 to a person's chest.

[0051] Moreover, the shorter the side length of the square constituting the reference range X, the more preferable, for example, 25 cm or less is more preferable, 20 cm or less is more preferable, and 15 cm or less is more preferable. The smaller the side length, the higher the degree of integration of the electrodes.

[0052] If the length of one side of the square constituting the reference range X is too small, it may be difficult to form the electrodes E0 to E15 or the wiring described below, so it is more preferable that the length is, for example, 5 cm or more, and even more preferable that the length is 10 cm or more.

[0053] In the embodiment, the first biological signal is acquired from a plurality of channels. That is, the number of channels p of the first biological signal is preferably an integer equal to or greater than 2. The larger the number of channels p of the first biological signal, the greater the amount of information required to reconstruct the second biological signal described below, and therefore it is preferable.

[0054] 4, n indicating the number of electrodes R, L, F, N, C1, C2, C3, C4, C5, and C6 is 10, and q indicating the number of channels of the second biosignal obtained by these electrodes is 12. That is, in the measurement for the standard 12-lead electrocardiogram, 10 electrodes (n=10) are used, and 12 (q=12) channels of the second biosignal are measured. In the measurement for the standard 12-lead electrocardiogram or other measurements of the second biosignal, n is preferably an integer of 3 or more, and q is preferably an integer of 2 or more.

[0055] The number of channels p obtained in the first arrangement is preferably equal to or greater than the number of channels q obtained in the second arrangement. When the number of channels p in the first arrangement is greater than the number of channels q in the second arrangement, many signals can be obtained in the first arrangement, so that the second biological signal can be reconstructed with high accuracy. For example, when the second arrangement is an electrode arrangement for a standard 12-lead electrocardiogram, the number of channels q may be 12 or less. When q=12, the number of channels p in the first arrangement using the electrode device 10 is preferably 10, 11, or 12 or more so as to be at least equal to q. In addition, the number of channels p in the first arrangement using the electrode device 10 is preferably 13 or more, which is larger than the number of channels q. It is more preferable that the number of channels p is sufficiently larger than the number of channels q, for example, 14 or more is preferable, and 15 or more is even more preferable.

[0056] The number of electrodes m in the first arrangement is preferably equal to or greater than the number of electrodes n in the second arrangement. When the number of electrodes m in the first arrangement is greater than the number of electrodes n in the second arrangement, it is advantageous because it is easier to obtain many signals in the first arrangement. For example, when the second arrangement is an electrode arrangement for a standard 12-lead electrocardiogram, the number of electrodes n=10. In this case, the number of electrodes m in the first arrangement using the electrode device 10 is preferably 9 or 10 or more so as to be at least equal to n. In addition, the number of electrodes m in the first arrangement using the electrode device 10 is preferably 11 or more, which is larger than the number of electrodes n. It is more preferable that the number of electrodes m is sufficiently larger than the number of electrodes n, for example, 12 or more is preferable, and 15 or more is even more preferable.

[0057] In addition, it is preferable that the electrodes are arranged at a higher density on the surface of the living body in the first arrangement than in the second arrangement.

[0058] Returning to Fig. 1, the first biosignals of multiple channels acquired from the electrode device 10 in the first arrangement are provided to the signal processing device 20. The signal processing device 20 generates second biosignals of multiple channels from the first biosignals of multiple channels using a generation model. Note that here, the second biosignals of a predetermined time series are generated from the first biosignals of a predetermined time series.

[0059] The signal processing device 20 includes a communication device 23 for receiving the first biosignal transmitted from the electrode device 10. The communication device 23 is, for example, a device for performing short-range wireless communication such as Bluetooth (registered trademark). The signal processing device 20 also includes a computer having a processor 21 and a storage device 22 connected to the processor. The communication device 23 may be provided inside or outside the computer.

[0060] The processor 21 is, for example, a CPU. The storage device 22 includes, for example, a primary storage device and a secondary storage device. The primary storage device is, for example, a RAM. The secondary storage device is, for example, a hard disk drive (HDD) or a solid state drive (SSD). The storage device 22 includes a computer program 22c executed by the processor 21.

[0061] The processor 21 reads and executes a computer program 22c stored in the storage device 22. The computer program 22c in the storage device 22 has a program code including instructions for causing a computer to operate as the signal processing device 20. The operation of the signal processing device 20 includes executing a process of acquiring p-channel first biosignals from m electrodes, inputting the acquired P-channel first biosignals to a generative model, and outputting a q-channel second biosignal from the generative model.

[0062] The storage device 22 has a storage area 22a for storing the acquired p-channel first biosignal, and a storage area 22b for storing the generated q-channel first biosignal.

[0063] Fig. 5 shows the procedure of the process executed by the signal processing device 20. In the process shown in Fig. 5, a second biosignal to be obtained by electrodes in a second arrangement different from the first arrangement is generated from a first biosignal acquired from the electrodes E0 to E15 in a first arrangement. That is, in the process shown in Fig. 5, a second biosignal required for diagnosis is reconstructed from the first biosignal that is actually measured.

[0064] First, in step S1, the signal processing device 20 acquires a first biosignal. For example, the signal processing device 20 receives a p-channel first biosignal from the electrode device 10 by wireless communication via the communication device 23. In step S2, the signal processing device 20 stores the received first biosignal in the storage device 22.

[0065] Fig. 6 shows time series data (signal waveforms) of 15 (p=15) channels of first biosignals (biosurface potentials) measured by 16 (m=16) electrodes E0 to E15. Fig. 6 shows waveforms of the first biosignals of a total of 15 channels from ch0 to ch14. Each of the channels from ch0 to ch14 is the potential at each of electrodes E0 to E14 measured with the potential of the reference electrode E15 as the reference potential.

[0066] In each signal waveform, the vertical axis indicates potential and the horizontal axis indicates time. Each signal waveform was measured simultaneously, and Fig. 6 shows the waveforms for the same period from 11.0 [s] to 13.0 [s]. For reference, the positions of P, R, T, Q, and S waves in the electrocardiogram are shown together with the signal waveform of ch0.

[0067] The 15-channel first biosignal shown in Fig. 6 indicates the potential at each position of the electrodes E0 to E14 densely arranged in the center of the human chest. In the following, as an example, an 8 (q=8) channel electrocardiogram (second biosignal) included in a standard 12-lead electrocardiogram is generated using these 15 (p=15) channel first biosignals. The 8-channel electrocardiogram generated below is, as an example, leads I, II, V1, V2, V3, V4, V5, and V6. Note that leads III, aV R Induction, aV L Induction, aV F A lead may also be generated.

[0068] In step S3 of FIG. 5, the signal processing device 20 uses the generation model 30 to generate a second biosignal of q channel for a period from the time series data of the first biosignal of p channel for the same period.

[0069] The generation model 30 is configured to output a second biosignal of a q channel when a first biosignal of a p channel is input. The generation model 30 is, for example, a trained model 30 that has been machine-learned to output a second biosignal of a q channel when a first biosignal of a p channel is input. An example of the machine learning is deep learning.

[0070] FIG. 7 shows an example of the generative model 30. FIG. 7(A) is an explanatory diagram of the machine learning phase of the generative model 30, and FIG. 7(B) is an explanatory diagram of the inference phase in which a second biosignal is reconstructed from a first biosignal. The generative model 30 includes, as an example, a convolutional neural network 111 and a fully connected layer 112, and is configured to obtain a feature amount 113 of the input first biosignal. The generative model 30 reconstructs the second biosignal from the feature amount 113 by the fully connected layer 114 and outputs it. Note that the input first biosignal is pre-processed in advance, such as noise removal, as necessary.

[0071] As shown in FIG. 7(A), in the learning phase, machine learning of a model 30, which is a neural network, is performed using a first biosignal for learning and a second biosignal for learning. The first biosignal for learning is, for example, a 15 (p=15) channel electrocardiogram acquired from 16 (m=16) electrodes E0 to E15 included in an electrode device 10 attached to the center of a person's chest (see FIG. 6). The second biosignal for learning is, for example, an 8 (q=8) channel electrocardiogram acquired from 10 (n=10) electrodes for measuring a standard 12-lead electrocardiogram. The 8-channel electrocardiogram includes leads I, II, V1, V2, V3, V4, V5, and V6.

[0072] In the learning phase, a first biosignal for training is paired with a second biosignal for training measured simultaneously for the same person during the same period. When the first biosignal for training is input to the model 30, machine learning of the model 30 is performed so that the second biosignal for training is reconstructed.

[0073] In the machine learning, pairs of the first biological signal for learning and the second biological signal for learning are used for a plurality of people. By using different learning signals for a plurality of people, the model 30 can generate the second biological signal with high accuracy regardless of individual differences.

[0074] In addition, the attachment position and attachment angle of the electrode device 10 on the human body may vary. However, by performing machine learning of the model 30 using a large number of learning signals including such variations, the second biosignal can be generated with high accuracy even if the attachment position and attachment angle are somewhat inappropriate. Therefore, attachment of the electrode device 10 of the embodiment may be somewhat less accurate, and is easier than attachment of electrodes for a standard 12-lead electrocardiogram. As a result, not only is it easier for medical staff to attach the electrode device 10, but the patient to whom the electrode device 10 is attached can also easily attach the electrode device 10 himself. Therefore, the electrode device 10 of the embodiment can be used to measure an electrocardiogram outside a medical facility such as at home. For this reason, the electrode device 10 of the embodiment can be used for remote medical care and the like.

[0075] As shown in FIG. 7(B), in the inference phase (corresponding to step S3 in FIG. 5), the first biosignal is input to the model 30. The model 30 outputs a second biosignal reconstructed from the first biosignal. The first biosignal input in the inference phase is, for example, a 15 (p=15) channel electrocardiogram acquired from 16 (m=16) electrodes E0 to E15 included in the electrode device 10 attached to the center of the person's chest (see FIG. 6). The second biosignal to be reconstructed is, for example, an 8 (q=8) channel electrocardiogram in a standard 12-lead electrocardiogram. The 8-channel electrocardiogram is lead I, lead II, lead V1, lead V2, lead V3, lead V4, lead V5, and lead V6.

[0076] FIG. 8 shows an 8-channel second biosignal (electrocardiogram) generated using a generative model 30 that is machine-learned using the first and second biosignals for learning from four people. To generate the second biosignal, time series data of 15 channels of the first biosignal for a predetermined period obtained from an electrode device 10 attached to the center of the subject's chest is input to the generative model 30. The predetermined period preferably has a duration of one heart beat or more. Note that the subjects are different from the four people whose learning signals were measured.

[0077] In each electrocardiogram in FIG. 8, the vertical axis indicates potential, and the horizontal axis indicates time. In FIG. 8, the “reconstructed” waveform shown by a solid line is a waveform reconstructed by the generative model 30. In FIG. 8, the “reference” waveform shown by a dotted line is a reference signal measured simultaneously with the first biosignal by attaching electrodes for a standard 12-lead electrocardiogram to the subject. The closer the “reconstructed” waveform is to the “reference”, the higher the accuracy of the reconstruction. As shown in FIG. 8, the “reconstructed” waveform is close to the “reference”, and the reconstruction is performed with high accuracy. Moreover, the “reconstructed” waveform shown in FIG. 8 is obtained by using the model 30 machine-learned using the learning signals of four people, and therefore it is expected that the accuracy of the reconstruction of the second biosignal can be further improved by using the model 30 machine-learned using a larger number of learning signals.

[0078] Returning to FIG. 5, in step S4, the signal processing device 20 stores the generated second biosignal in the storage device 22. The signal processing device 20 can also output the generated second biosignal to the outside (step S5). Outputting to the outside can be, for example, displayed on a display, printed by a printer, or transmitted to a network. The second biosignal output to the outside is used, for example, for diagnosis by a doctor. According to this embodiment, it is advantageous that an electrocardiogram included in a standard 12-lead electrocardiogram can be obtained without having the patient accurately wear electrodes for the standard 12-lead electrocardiogram.

[0079] It should be noted that the devices and systems according to the embodiments are not limited to use in measuring and generating electrocardiograms, but can also be used in measuring and generating other biological signals, such as skeletal muscle electromyograms or visceral smooth muscle electromyograms.

[0080] The present invention is not limited to the above-described embodiment, and various modifications are possible. [Explanation of symbols]

[0081] 1: Biosignal processing system 10: Electrode device 11: Sheet 11a: 1st page 11b: 2nd side 12: Wiring 15: Communication equipment 20: Signal processing device 21: Processor 22: Storage device 22a: Storage area 22b: Storage area 22c: Computer Programs 23: Communication equipment 30: Generative model 111: Convolutional Neural Networks 112 :Fully connected layer 113: Features 114 :Fully connected layer C1: Chest electrode C2: Chest electrode C3: Chest electrode C4: Chest electrode C5: Chest electrode C6: Chest electrode E0: Electrode E1: Electrode E2: Electrode E3: Electrode E4: Electrode E5: Electrode E6: Electrode E7: Electrode E8: Electrode E9: Electrode E10: Electrode E11: Electrode E12: Electrode E13: Electrode E14: Electrode E15: Electrode F: Limb electrode H: Heart L: Limb electrode N: Limb electrode R: Limb electrode X: Reference range m: Number of electrodes n: Number of electrodes p : Number of channels q : Number of channels

Claims

1. In the first configuration, m electrodes (where m is an integer greater than or equal to 3) are attached to the body, A signal processing device that inputs a first biological signal of p (where p is an integer of 2 or more) channels obtained from the m electrodes into a generative model and outputs a second biological signal of q (where q is an integer of 2 or more) channels from the generative model, Equipped with, The second biological signal of the q channel corresponds to the signal obtained from n (where n is an integer of 3 or more) electrodes attached to the body in the second configuration. The second configuration is an electrode configuration that differs from the first configuration in at least one of the number of electrodes and the mounting positions. The aforementioned m is greater than the aforementioned n, The generation model is configured to output a second biosignal consisting of fewer q channels when a first biosignal consisting of more p channels than q channels is input. Biosignal processing system.

2. The m electrodes attached to the living body in the first arrangement are mounted on a single base with the relative positions of the m electrodes fixed. The biosignal processing system according to claim 1.

3. The generative model is a trained model that has been machine-learned to output the second biosignal of the q channel when the first biosignal of the p channel is input. The biosignal processing system according to claim 1.

4. The second biological signal of the q channel is the signal of 12 channels in a standard 12-lead electrocardiogram or a signal of fewer than 12 channels included in a standard 12-lead electrocardiogram. The biosignal processing system according to claim 1.

5. The m electrodes on which the first biological signal of the p channel is acquired are all attached to the center of the chest of the living organism. The biosignal processing system according to claim 1.

6. A signal processing device that performs signal processing, The aforementioned signal processing is performed as follows: In the first configuration, a first biological signal is acquired from m electrodes (where m is an integer greater than or equal to 3) attached to the living body, with p channels (where p is an integer greater than or equal to 2). The first biological signal of the p channel is input to the generation model, and the second biological signal of the q channel (where q is an integer of 2 or more) is output from the generation model. This includes, The second biological signal of the q channel corresponds to the signal obtained from n (where n is an integer of 3 or more) electrodes attached to the body in the second configuration. The second configuration is an electrode configuration that differs from the first configuration in at least one of the number of electrodes and the mounting positions. The aforementioned m is greater than the aforementioned n, The generation model is configured to output a second biosignal consisting of fewer q channels than p channels when a first biosignal consisting of more p channels than q channels is input. Signal processing device.

7. A method for generating biological signals, In the first configuration, a first biological signal is acquired from m electrodes (where m is an integer greater than or equal to 3) attached to the living body, with p channels (where p is an integer greater than or equal to 2). The first biological signal of the p channel is input to the generation model, and the second biological signal of the q channel (where q is an integer of 2 or more) is output from the generation model. This includes, The second biological signal of the q channel corresponds to the signal obtained from n (where n is an integer of 3 or more) electrodes attached to the body in the second configuration. The second configuration is an electrode configuration that differs from the first configuration in at least one of the number of electrodes and the mounting positions. The aforementioned m is greater than the aforementioned n, The generation model is configured to output a second biosignal consisting of fewer q channels than p channels when a first biosignal consisting of more p channels than q channels is input. Methods for generating biological signals.

8. A computer program that causes a computer to perform signal processing, The aforementioned signal processing is performed as follows: In the first configuration, a first biological signal is acquired from m electrodes (where m is an integer greater than or equal to 3) attached to the living body, with p channels (where p is an integer greater than or equal to 2). The first biological signal of the p channel is input to the generation model, and the second biological signal of the q channel (where q is an integer of 2 or more) is output from the generation model. This includes, The second biological signal of the q channel corresponds to the signal obtained from n (where n is an integer of 3 or more) electrodes attached to the body in the second configuration. The second configuration is an electrode configuration that differs from the first configuration in at least one of the number of electrodes and the mounting positions. The aforementioned m is greater than the aforementioned n, The generation model is configured to output a second biosignal of fewer q channels when it receives a first biosignal of more p channels than q channels. Computer program.