JP2014124345A - Biosignal processor and biosignal processing method - Google Patents

Abstract

Provided are a biological signal processing apparatus and a biological signal processing method for accurately restoring an original biological signal when restoring the original biological signal from sampling data by spline interpolation.
A biological signal analysis system includes a detection unit that detects a biological signal from a living body, a compression unit 52 that band-compresses the biological signal by removing a high-frequency component of a predetermined value or more from the biological signal, and a compressed biological body. A sampling unit 53 that acquires sampling points from a signal at a predetermined sampling frequency and generates sampling data composed of a plurality of sampling points, and a plurality of interpolation points are added between the sampling points by spline interpolation. And up-sampling units (62, 63) for restoring biological signal data with higher temporal resolution.
[Selection] Figure 4

Description

  The present invention relates to a biological signal processing apparatus and a biological signal processing method.

  Conventionally, there is a device such as a Holter electrocardiograph that is worn on a patient and records a biological signal. Since such devices continue to record biological signals for a long time, the amount of data to be recorded becomes enormous.

  Therefore, in order to reduce the amount of data to be recorded, it is conceivable to keep the sampling frequency of the biological signal relatively low (decimation).

  However, if the sampling frequency is kept low, even if it is attempted to restore the original biological signal from the sampling data obtained by sampling, it cannot be accurately restored.

  Patent Document 1 discloses a technique for interpolating sampling data by a technique called spline interpolation in order to accurately restore the original biological signal from such sampling data.

JP 59-216282 A

  However, there is a problem that it is not always possible to accurately restore the sampled data simply by performing spline interpolation.

  The present invention has been made in view of the above circumstances, and when restoring the original biological signal from the sampling data by spline interpolation, the biological signal processing apparatus and the biological signal that restore the original biological signal with higher accuracy than before An object is to provide a processing method.

  In order to solve the above-described problems, a biological signal processing apparatus according to the present invention includes a detection unit that detects a biological signal from a living body, and removes a high-frequency component of a predetermined value or more from the biological signal and performs band compression on the biological signal A sampling unit that obtains sampling points from the compressed biological signal at a predetermined sampling frequency and generates sampling data composed of a plurality of sampling points, and a plurality of sampling points between the sampling points by spline interpolation. And an up-sampling unit that restores biological signal data with higher temporal resolution from the sampling data.

  According to the present invention, a highly accurate biological signal can be restored from sampling data.

It is a schematic block diagram of a biological signal analysis system. It is a block diagram which shows the hardware constitutions of a recording device. It is a block diagram which shows the hardware constitutions of an analyzer. It is a block diagram which shows the function structure of a biosignal analysis system. It is a flowchart which shows the procedure of a recording process. It is a figure which illustrates a biological signal (upper stage), a sampling interval (middle stage), and a sample point (lower stage). It is a figure for demonstrating the method of band compression. It is a figure which shows the schematic data structure of sampling data. It is a flowchart which shows the procedure of a restoration process. It is a figure for demonstrating the method of a spline interpolation process. It is a figure for demonstrating the method of an upsampling process. It is a figure which shows the schematic data structure of the reconfigure | reconstructed biological signal data. It is a figure which illustrates the decompress | restored biological signal. (A) It is a figure which shows the result of the evaluation experiment 1 (frequency 10Hz). (B) It is a figure which shows the result of the evaluation experiment 1 (frequency 15Hz). (C) It is a figure which shows the result of the evaluation experiment 1 (frequency 20Hz). (D) It is a figure which shows the result of the evaluation experiment 1 (frequency 25Hz). (A) It is a figure which shows the result of the evaluation experiment 2 (phase 0 degree). (B) It is a figure which shows the result of the evaluation experiment 2 (phase 30 degrees). (C) It is a figure which shows the result of the evaluation experiment 2 (phase 60 degrees). It is the figure where the result of evaluation experiment 2 was put together. It is a figure which shows the result of the evaluation experiment 3. (A) It is a figure which shows the result of the evaluation experiment 4 (attenuation rate of an amplitude). (B) It is a figure which shows the result of the evaluation experiment 4 (upper limit frequency of a non-attenuation zone). It is a figure which shows the result of the evaluation experiments 5 and 6 (waveform). It is a figure which shows the result of the evaluation experiment 5 (sampling position shift). It is a figure which shows the result of the evaluation experiment 6 (sampling position shift). It is a figure which shows the result of the evaluation experiment.

  Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In the description of the drawings, the same elements are denoted by the same reference numerals, and redundant description is omitted. In addition, the dimensional ratios in the drawings are exaggerated for convenience of explanation, and may be different from the actual ratios.

  FIG. 1 is a schematic configuration diagram of a biological signal analysis system 1 according to the present embodiment. FIG. 2 is a diagram illustrating a hardware configuration example of the recording apparatus 10 included in the biological signal analysis system 1. FIG. 3 is a diagram illustrating a hardware configuration example of the analysis device 20 included in the biological signal analysis system 1. FIG. 4 is a diagram illustrating a functional configuration example of the biological signal analysis system 1.

  Hereinafter, a schematic configuration of the biological signal analysis system 1 will be described with reference to FIGS.

(Whole biological signal analysis system 1)
As shown in FIG. 1, the biological signal analysis system 1 includes a recording device 10 and an analysis device 20.

  The recording device 10 detects and records a biological signal from a living body (subject) (not shown). For example, the recording apparatus 10 is a Holter electrocardiograph that detects an electrocardiogram signal (also referred to as “electrocardiogram data”) from a living body. Usually, the recording apparatus 10 is attached to a living body (subject) and continuously detects an electrocardiogram signal for a long time (for example, 24 hours).

  A biological signal such as an electrocardiogram signal is a continuous analog signal, and the recording apparatus 10 converts (A / D conversion) the detected biological signal into discrete digital data, and reduces some data in order to reduce the amount of data. Data is extracted and recorded at regular intervals. Hereinafter, a process of extracting a part of digital is called sampling (or downsampling), and digital data obtained by sampling is called sampling data. The sampling frequency Fsp that determines the sampling interval is, for example, 125 Hz, and is kept relatively low.

  On the other hand, the analysis device 20 restores the original biological signal from the sampling data recorded by the recording device 10. The analysis device 20 generates biological signal data having higher time resolution than the sampling data by adding a plurality of interpolation points to the sampling data. This is called restoration data. The analysis device 20 graphs the restored data and displays it on a display device such as a display as data obtained by restoring (reconstructing) the original biological signal. Thereby, the user can visually confirm the biological signal obtained from the living body (subject).

  The recording device 10 and the analysis device 20 as described above can transfer sampling data from the recording device 10 to the analysis device 20 by connecting to each other via wireless communication such as a computer network or a mobile phone. The sampling data may be recorded in the portable external memory 13, and the sampling data recorded in the recording device 10 can be moved to the analyzing device 20 by moving the external memory 13 to the analyzing device 20.

  The above computer network is a LAN (Local Area Network) in which computers and network devices are connected according to standards such as Ethernet (registered trademark), Token Ring, FDDI (Fiber Distributed Data Interface), etc., or LANs with dedicated lines. It consists of a connected WAN (Wide Area Network) or the like.

  Further, the recording device 10 and the analysis device 20 may be connected via BlueTooth (registered trademark) or infrared communication without using a computer network, or may be connected by USB communication via a USB cable.

  The portable external memory 13 may be a memory having any name, form, and structure as long as it is a portable memory such as a USB memory or an SD memory card.

(Recording device 10)
The recording device 10 includes an electrode 11 and a control device 12 as shown in FIGS. 1 and 2.

  The electrode 11 is attached to a living body and detects a biological signal. The number and arrangement of the electrodes can be appropriately changed according to the type of biological signal to be detected.

  The control device 12 is provided inside the casing as a main body (for example, a portion surrounded by a dotted line in FIG. 1), and controls the entire recording device 10. For example, as illustrated in FIG. 2, the control device 12 includes a connection interface (I / F) 31, an A / D converter 32, a band limiting filter 33, a CPU (Central Processing Unit) 34, a memory 35, A storage 36, an external memory slot 37, and a communication interface (I / F) 38 are provided, and these are connected to each other via a bus 39 for exchanging signals.

  The connection I / F 31 is an interface for ensuring electrical connection with the electrode 11, and inputs a biological signal detected at the electrode 11 to the recording apparatus 10.

  The biological signal input through the connection I / F 31 is amplified by the amplifier and supplied to the A / D converter 32.

  The A / D converter 32 converts the biological signal (analog signal) detected at the electrode 11 into digital data. Note that the sampling frequency Fa during A / D conversion is, for example, 1000 Hz. Further, when simply called “sampling frequency”, it means at least one of the above-described sampling frequency Fsp and A / D conversion sampling frequency Fa.

  The band limiting filter 33 is a filter that band-compresses the digital data of the biological signal converted by the A / D converter 32. For example, the band limiting filter 33 removes high frequency components from the biological signal.

  As a setting condition of the band limiting filter 33, for example, a frequency that becomes a boundary of a frequency band in which passage of a biological signal is blocked (attenuation rate is "1/1000" or less) (hereinafter, "stop band end frequency Fs"). Is referred to as 1/2 or less of the sampling frequency Fsp.

  For the band limiting filter 33 as described above, for example, “Kaiser Filter” is used.

  The CPU 34 is a control circuit composed of a processor or the like that executes control of the above-described units and various arithmetic processes according to a program. Each function of the recording apparatus 10 is exhibited by the CPU 34 executing a corresponding program. The

  The memory 35 is a high-speed accessible main storage device that temporarily stores programs and data as a work area. As the memory 35, for example, a DRAM (Dynamic Random Access Memory), an SDRAM (Synchronous Dynamic Access Memory), an SRAM (Static Random Access Memory), or the like is adopted.

  The storage 36 is an auxiliary storage device that stores various programs and various data. For the storage 36, for example, a flash memory, a solid state drive, a hard disk or the like is employed. Since the recording device 10 is assumed to be worn on a living body (subject) for a long time, the recording capacity of the storage 36 is set to a level at which 24-hour sampling data can be recorded at a sampling frequency Fsp of at least 125 Hz.

  The external memory slot 37 is an interface for reading / writing data (for example, sampling data) from / to the external memory 13 in a state where the portable external memory 13 is inserted. It should be noted that the recording capacity of the external memory 13 is set to be at least capable of recording 24-hour sampling data at a sampling frequency Fsp of 125 Hz, as with the storage 36.

  The communication I / F 38 is an interface for communicating with an external device (for example, the analysis device 20) via a computer network, and standards such as Ethernet (registered trademark), token ring, and FDDI are used. In addition, an interface for wireless communication using BlueTooth (registered trademark), infrared communication, USB communication, mobile phone, or the like may be provided.

  The recording apparatus 10 having the hardware configuration as described above has, as a functional configuration, a biological signal input unit 51, a band compression unit 52, a sampling unit 53, and a sampling data storage unit 54 as shown in the upper side of the drawing in FIG. And having.

  The biological signal input unit 51 is realized by, for example, the connection I / F 31 and the A / D converter 32 described above, and inputs the biological signal detected at the electrode 11 to the recording device 10 and converts it into digital data.

  The band compression unit 52 is realized by, for example, the band limiting filter 33 described above, and band-compresses the biological signal (digital data) input by the biological signal input unit 51.

  However, the band compression unit 52 is not limited to hardware such as the band limiting filter 33 and may be realized by software. For example, the band compression unit 52 may be realized by the CPU 34 reading a program installed in the storage 36 into the memory 35 and executing it.

  The sampling unit 53 samples the biological signal (digital data) subjected to band compression. For example, the sampling unit 53 down-samples the biological signal having the sampling frequency Fa of 1000 Hz that has been band-compressed by the band compressing unit 52 with the sampling frequency Fsp of 125 Hz, and generates sampling data composed of a plurality of sample points. Thereby, the data amount can be reduced (in this embodiment, the data amount is reduced to 1/8).

  Such a sampling unit 53 is realized by, for example, the CPU 34 reading a program installed in the storage 36 into the memory 35 and executing it.

  The sampling data storage unit 54 is realized by the storage 36 or the external memory 13 and stores sampling data down-sampled by the sampling unit 53.

(Analyzer 20)
The analysis device 20 is a terminal that acquires sampling data recorded in the recording device 10 and restores the original biological signal. For example, a desktop PC (personal computer) as shown in FIG. 1 may be used for the analysis device 20, or a portable terminal such as a tablet terminal, a smartphone, or a mobile phone may be used.

  As shown in FIGS. 1 and 3, the analysis device 20 includes a control device 21, a display device 22, and an input device 23.

  The control device 21 is provided inside the housing (for example, a portion surrounded by a dotted line in FIG. 1) and controls the entire analysis device 20. For example, as shown in FIG. 3, the control device 21 has a CPU 41, a memory 42, a storage 43, an external memory slot 44, and a communication interface (I / F) 45, which exchange signals. Are connected to each other via a bus 46.

  The CPU 41 is a control circuit configured by a multi-core processor or the like that executes control of the above-described units and various arithmetic processes according to a program. Each function of the analysis device 20 is executed by the CPU 41 executing a corresponding program. Demonstrated.

  The memory 42 is a high-speed accessible main storage device that temporarily stores programs and data as a work area. As the memory 42, for example, DRAM, SDRAM, SRAM, or the like is employed.

  The storage 43 is a large-capacity auxiliary storage device that stores various programs including an operating system and various data. As the storage 43, for example, a flash memory, a solid state drive, a hard disk, or the like is employed.

  The external memory slot 44 is an interface for reading / writing data (for example, sampling data) from / to the external memory 13 in a state where the portable external memory 13 is inserted.

  The communication I / F 45 is an interface for communicating with an external device (for example, the recording apparatus 10) via a computer network, and standards such as Ethernet (registered trademark), token ring, and FDDI are used. In addition, an interface for wireless communication using BlueTooth (registered trademark), infrared communication, USB communication, mobile phone, or the like may be provided.

  The display device 22 displays a screen for confirming the restoration data in the analysis device 20. As the display device 22, for example, a liquid crystal display, an organic EL (Electro-Luminescence), or the like as shown in FIG. 1 is employed.

  The input device 23 receives various inputs from the user. As the input device 23, a keyboard 23a, a mouse 23b, and the like as shown in FIG.

  As shown in FIG. 4, the analysis device 20 having the hardware configuration as described above includes an acquisition unit 61, an interpolation unit 62, an upsampling unit 63, and a restored data storage unit 64 as a functional configuration.

  The acquisition unit 61 acquires sampling data of a biological signal recorded in the recording device 10. The acquisition unit 61 communicates with the recording apparatus 10 via a computer network or the like, and acquires sampling data stored in the sampling data storage unit 54. As another acquisition method, the acquisition unit 61 may read sampling data from the external memory 13 inserted in the external memory slot 44, for example.

  The interpolation unit 62 locally selects a plurality of (for example, six) sample points from the sample points constituting the sampling data, and interpolates them with a spline interpolation function. Here, the spline interpolation function between two of the sample points is set as a spline interpolation function for the interpolation value. Then, the interpolation unit 62 repeats such interpolation, and interpolates between all the sample points included in the sampling data using a spline interpolation function. The spline interpolation function will be described later.

  The up-sampling unit 63 adds a plurality of interpolation points on each spline interpolation function obtained by the interpolation unit 62, and reconstructs restoration data having a higher time resolution (for example, 1000 Hz) than the sampling data.

  The restored data storage unit 64 stores the restored data reconstructed by the upsampling unit 63.

  The acquisition unit 61 as described above is realized by the CPU 41 reading the program installed in the storage 43 into the memory 42 and executing it, and controlling the external memory slot 44 or the communication I / F 45. The interpolation unit 62 and the upsampling unit 63 are realized by the CPU 41 reading out the program installed in the storage 43 to the memory 42 and executing it. The restored data storage unit 64 is realized by the storage 43 or the external memory 13, for example.

  Next, the operation of the recording apparatus 10 will be described.

<Recording process>
FIG. 5 is a flowchart showing a procedure of a process of recording a biological signal detected from a living body (subject) (hereinafter referred to as “recording process”). FIG. 6 is a diagram illustrating a biological signal detected from a living body (upper stage), a sampling interval (middle stage), and sampled sample points (lower stage). FIG. 7 is a diagram for explaining a band compression method. FIG. 8 is a diagram showing a schematic data structure of sampling data recorded in the recording apparatus 10.

  Hereinafter, the procedure of the recording process will be described with reference to FIGS.

  When the recording apparatus 10 receives a recording start instruction from the user in a state where the electrode 11 is attached to a living body (subject), the recording apparatus 10 starts the recording process shown in FIG.

  When the recording process is started, the recording device 10 functions as the biological signal input unit 51 and inputs the biological signal detected by the electrode 11 to the recording device 10 (step S101). The biosignal input here is, for example, an electrocardiogram signal having a waveform as shown in the upper part of FIG. In general, as shown in FIG. 6, basic waveforms such as P wave, Q wave, R wave, S wave, and T wave are observed in an electrocardiogram signal of a healthy person.

  Here, the biological signal input to the recording apparatus 10 is amplified by an amplifier.

  Next, the recording apparatus 10 converts the biological signal amplified by the amplifier into digital data (for example, the sampling frequency Fa is set to 1000 Hz) by the A / D converter 32, and then functions as the band compression unit 52. Then, band compression is performed (step S102). For the band compression in step S102, for example, a band limiting filter 33 as shown in FIG. 7, that is, “Kaiser Filter” is used. In the example shown in FIG. 7, the stop band edge frequency Fs of the band limiting filter 33 is set to ½ (for example, 62.5 Hz) of the sampling frequency Fsp. Note that “Fp” in the figure indicates a frequency that becomes a boundary of the passband, that is, a passband end frequency, and a section between the passband end frequency Fp and the stopband end frequency Fs is called an attenuation region.

  Then, the recording device 10 functions as the sampling unit 53, and down-samples the band-compressed biological signal (digital data) (step S103). For example, the downsampling in step S103 is performed at a predetermined sampling frequency Fsp. However, as described above, the sampling frequency Fsp is kept relatively low (for example, 125 Hz) in order to reduce the amount of data to be recorded.

  As a result of the downsampling in step S103, sample points with a sampling interval T (= 1 / Fsp) as shown in the middle and lower parts of FIG. 6 are obtained from the ECG signal subjected to band compression.

  Then, the recording apparatus 10 stores the sample point value obtained in step S103 as the sampling data 150 (step S104). The storage destination of the sampling data 150 here is the sampling data storage unit 54, that is, the storage 36, the external memory 13, and the like.

  In the sampling data 150, as shown in FIG. 8, a time 152 and a value 153 are associated with each sample point 151.

  The sample point 151 is information for identifying the sample point sampled in step S103. Time 152 is data indicating the time at which the sample point was sampled. The value 153 is data indicating the value (for example, potential) of the sampled sample point.

  Thereafter, the recording apparatus 10 ends the recording process.

  By executing the above recording process in the recording apparatus 10, it is possible to record sampling data on a biological signal detected from a living body (subject).

<Restore process>
Next, a process of restoring the original biological signal from the sampling data 150 recorded by the above-described recording process (hereinafter referred to as “restoration process”) will be described.

  FIG. 9 is a flowchart showing the procedure of the restoration process executed in the analysis apparatus 20. FIG. 10 is a diagram for explaining a method of spline interpolation processing performed during the restoration processing. FIG. 11 is a diagram for explaining a method of upsampling processing performed during the restoration processing. FIG. 12 is a diagram showing a schematic data structure of restoration data reconstructed during the restoration process.

  Hereinafter, the procedure of the restoration process will be described with reference to FIGS.

  For example, when receiving an instruction to start restoration from the user by operating the input device 23, the analysis device 20 starts the restoration processing illustrated in FIG.

  When the restoration process is started, the analysis device 20 functions as the acquisition unit 61 and acquires the sampling data 150 from the recording device 10 (step S201). For example, the analysis device 20 may acquire the sampling data 150 stored in the sampling data storage unit 54 of the recording device 10 via a computer network. Of course, the sampling data 150 recorded in the external memory 13 inserted in the external memory slot 44 may be read.

  Thereafter, the analysis device 20 functions as the interpolation unit 62, and interpolates between all the sample points included in the sampling data 150 using a spline interpolation function (step S202). In the present embodiment, the processing in step S202 is referred to as “local spline interpolation processing”.

  Hereinafter, the local spline interpolation processing will be described in detail. As shown in FIG. 10, in the local spline interpolation processing in step S <b> 202, first, the analysis apparatus 20 selects a plurality of sample points locally (continuously in time) from the sample points included in the sampling data 150. In the example shown in the figure, six sample points (black diamonds) are selected.

Next, the analysis apparatus 20 has the coordinates (x ₁ , y ₁ ), (x ₂ , y ₂ ) of each sample point in the range where the locally selected sample points exist (the vertical range shown in FIG. 10). ), ..., (x _i-1 , y _i-1 ), (x _i , y _i ), ..., (x _n-1 , y _n-1 ), (x _n , y _n ) Y = f (x). However, i is the sample point number shown in FIG. 10, and n represents the total number of locally selected sample points (“n = 6” in the example of FIG. 10).

  Then, the analysis apparatus 20 interpolates a plurality of locally selected sample points using a spline interpolation function. More specifically, the analysis device 20 performs spline interpolation on a plurality of locally selected sample points, and adds a plurality of interpolation points between two of the sample points (in the example shown in FIG. 10). Section C).

  Here, for the spline interpolation, for example, a cubic spline interpolation function such as the following formula (1) is used.

S _i (x) = a _i + b _i (x−x _i ) + c _i (x−x _i ) ² + d _i (x−x _i ) ³ ... Formula (1)
However, 1 ≦ i ≦ n−1 and x _i ≦ x ≦ x _{i + 1} are set. S _i (x) represents an i-th spline interpolation function used to add an interpolation point between the i-th and i + 1-th sample points, and a _i , b _i , c _i , d _i are Is a coefficient.

Then, the analysis device 20 obtains coefficients a _i , b _i , c _i , d _i that satisfy the following conditions 1 to 3, and determines a spline interpolation function.

(Condition 1)
S _i (x _i ) = f (x _i ) = y _i
S _i (x _{i + 1} ) = f (x _{i + 1} ) = y _{i + 1}
(Condition 2)
S _i-1 (x _i ) = S _i (x _i )
S ′ _i−1 (x _i ) = S ′ _i (x _i )
S ″ _i−1 (x _i ) = S ″ _i (x _i )
(Condition 3)
S ″ ₁ (x ₁ ) = S ″ _n−1 (x _n ) = 0 or
S ′ ₁ (x ₁ ) = f ′ (x ₀ ), S ′ _n−1 (x _n ) = f ′ (x _n )
The above condition 1 is a condition for the i-th and i + 1-th sample points to pass on the i-th spline interpolation function S _i (x). The condition 2 is a condition for smoothing the connection point between the ( _i−1 ) th spline interpolation function S _i−1 (x) and the i th spline interpolation function S _i (x). Further, the above condition 3 is a condition for allowing both ends of a range where the locally selected sample points exist (standing range shown in FIG. 10) to be free ends.

As described above, the analysis apparatus 20 obtains the coefficients a _i , b _i , c _i , and d _i satisfying the conditions 1 to 3 and obtains between the i-th and i + 1-th sample points (section C shown in FIG. 10). A spline interpolation function S _i (x) can be determined.

Then, as shown in FIG. 11, the analysis device 20 repeats the same spline interpolation process while shifting the sample points to be locally selected. That is, the analysis device 20 performs the spline interpolation function in the range shown in (1) of FIG. 11 (S ₁₁ , S ₁₂ , S ₁₃ , S ₁₄ , S ₁₅ ) to the section S ₁₃ (corresponding to the section C in FIG. 11). After determining S _i S ₁ (x), it corresponds to the section S ₂₃ (corresponding to the section D in FIG. 11) from the range (S ₂₁ , S ₂₂ , S ₂₃ , S ₂₄ , S ₂₅ ) shown in (2) of FIG. ) Spline interpolation function S _{i + 12} (x), the range shown in (3) of FIG. 11 (S ₃₁ , S ₃₂ , S ₃₃ , S ₃₄ , S ₃₅ ) to section S ₃₃ (corresponding to section E of FIG. 11) The spline interpolation function S _{i + 2 3} (x) is sequentially determined. The analysis device 20 connects the determined spline interpolation functions (S ₁₃ , S ₂₃ , S ₃₃ ) to create a continuous function (S ₁ , S ₂ , S ₃ ) as shown in (4) of FIG. . In this way, the analysis device 20 can interpolate between all the sample points included in the sampling data 150 using the spline interpolation function.

  When all the sample points are interpolated by the spline interpolation function in step S202, the analysis apparatus 20 functions as the upsampling unit 63 and adds a plurality of interpolation points on each spline interpolation function (step S203). At this time, the interpolation points are added at regular intervals as indicated by white circles in FIG. In this way, reconstructed data including the sample points originally included in the sample data 150 and the interpolation points added in step S203 is generated.

  Naturally, the time interval T ′ between the interpolation points is shorter (for example, 1/8 times) than the time interval T (= 1 / Fsp) between the sample points of the sampling data 150. Therefore, the restoration data is data having a higher time resolution (for example, the resampling frequency Fb is 1000 Hz) than the sampling data.

  Thereafter, the analysis device 20 continues to function as the upsampling unit 63 and stores it as the restored data 160 generated in step S203 (step S204). The storage destination of the restoration data 160 here is the restoration data storage unit 64, that is, the storage 43, the external memory 13, and the like.

  In the restored data 160, as shown in FIG. 12, a time 162 and a value 163 are associated with each sample point / interpolation point 161.

  The sample point / interpolation point 161 is information for identifying the sample point and the interpolation point. Time 162 is data indicating the time of the sample point and the interpolation point. The value 163 is data indicating the values (for example, potentials) of the sample points and the interpolation points.

  Such restoration data 160 is generated by adding the contents of the time 162 and the value 163 related to the interpolation point to the sampling data 150 (inside the dotted line in FIG. 12).

  Thereafter, the analysis device 20 ends the restoration process.

  The analysis device 20 can display the restored data 160 on the display device 42 based on an instruction from the user. For example, the analysis device 20 displays a graph in which the sampling points and the interpolation points included in the restoration data 160 are connected in time series as a restored biological signal.

  FIG. 13 is a diagram illustrating a restored biological signal.

  As shown in FIG. 13, the restored biological signal is displayed on the display device 42. The biological signal displayed here is created using the restoration data 160 having a higher time resolution than the sampling data 150 recorded in the recording apparatus 10. Therefore, the user can confirm the state of the living body (subject) based on the biological signal closer to the original biological signal.

  In the band compression in step S102, the stop band end frequency Fs of the band limiting filter 33 is set to ½ or less of the sampling frequency Fsp. This is due to the sampling theorem.

  Note that each processing unit in each flowchart described above is divided according to the main processing contents in order to facilitate understanding of the recording device 10 and the analysis device 20. The invention of the present application is not limited by the method of classification of the processing steps and the names thereof. The processing performed in the recording device 10 and the analysis device 20 can be divided into more processing steps. One processing step may execute more processes.

<Evaluation results>
Hereinafter, the evaluation result about said biological signal analysis system 1 is shown.

(Evaluation Experiment 1)
First, when a sine wave signal was input as a biological signal to the recording apparatus 10 described above, the waveform of the restored data by the analyzing apparatus 20 (hereinafter referred to as “restored waveform”) was evaluated.

  Specifically, assuming that the frequency of the sine wave signal input to the recording device 10 is 10 [Hz], the input waveform for one cycle of the sine wave signal and the restored waveform for one cycle by the analysis device 20 were compared. .

  FIG. 14A shows an input waveform (broken line) for one period of a 10 Hz sine wave signal input to the recording apparatus 10 and a waveform (dashed line) of sampling data obtained by the recording apparatus 10 when the sampling frequency Fsp is 125 Hz. (Hereinafter referred to as “sampling waveform”) and a restored waveform (solid line) for one cycle by the analysis apparatus 20. The analysis device 20 generates a restored waveform from the sampling waveform. As shown in FIG. 14A, the input waveform and the restored waveform almost overlap each other, and the maximum error in the amplitude was 0.071%. Therefore, it can be seen that the biological signal analysis system 1 of the present embodiment can appropriately restore the 10 Hz sine wave signal.

  A similar evaluation experiment was performed by changing the frequency of the sine wave signal input to the recording apparatus 10 (15 Hz, 20 Hz, 25 Hz). FIG. 14B shows the experimental results when the frequency of the sine wave signal input to the recording apparatus 10 is 15 Hz, and FIG. 14C shows the case where the frequency of the sine wave signal input to the recording apparatus 10 is 20 Hz. FIG. 14D shows the experimental results when the frequency of the sine wave signal input to the recording apparatus 10 is 25 Hz.

  Table 1 is data that summarizes the maximum error of the amplitude of 10 Hz, 15 Hz, 20 Hz, and 25 Hz. As shown in Table 1, it can be seen that the sine wave signal can be properly restored regardless of the frequency of the sine wave signal.

(Evaluation Experiment 2)
In the evaluation experiment 1 described above, the frequency of the sine wave signal input to the recording apparatus 10 was changed, and the input waveform of the input sine wave signal was compared with the restored waveform. On the other hand, in the evaluation experiment 2, the phase of the input sine wave signal was changed and the input waveform of the input sine wave signal was compared with the restored waveform.

  FIG. 15A shows an input waveform (broken line) for two periods of a sine wave signal having a phase of 0 ° input to the recording apparatus 10 and the recording apparatus 10 when the sampling frequency Fsp is 125 Hz with respect to this input waveform. FIG. 6 is a diagram showing a sampling waveform (chain line) obtained by, and a restored waveform (solid line) for two cycles by the analysis device 20. FIG. 15B shows an input waveform (broken line) for two periods of a sine wave signal having a phase of 30 ° input to the recording apparatus 10 and a restored waveform (solid line) for two periods by the analysis apparatus 20. FIG. FIG. 15C shows an input waveform (broken line) for two cycles of a sine wave signal having a phase of 60 ° input to the recording device 10 and a restored waveform (solid line) for two cycles by the analysis device 20. FIG.

  As shown in FIGS. 15A to 15C, the input waveform and the restored waveform almost overlap each other regardless of the phase of the input sine wave signal. Therefore, it can be seen that the biological signal analysis system 1 of the present embodiment can appropriately restore the original sine wave signal regardless of the phase of the sine wave signal. Note that the frequencies of the sine wave signals shown in FIGS. 15A to 15C are all 25 Hz as in FIG.

  FIG. 16 shows data summarizing the results of the evaluation experiment 3 when the input sine wave signal is 25 Hz.

  As shown in FIG. 16, when the frequency of the input sine wave signal is 25 Hz, the maximum error in amplitude does not change greatly and is low regardless of the phase of the sine wave signal (0 ° to 80 °).

(Evaluation Experiment 3)
In the evaluation experiments 1 and 2 described above, a single frequency sine wave signal was input to the recording apparatus 10. For example, in evaluation experiments 1 and 2, a sine wave signal having a single frequency of 3 seconds was input.

  On the other hand, in the evaluation experiment 3, a signal having a waveform formed by combining a plurality of sine wave signals having different frequencies and phases appropriately was input to the recording device 10 and the restored waveform by the analysis device 20 was evaluated.

  Specifically, a first sine wave signal having a frequency of 3 Hz and a phase of 0 °, a second sine wave signal having a frequency of 7 Hz and a phase of 20 °, and a third sine wave having a frequency of 13 Hz and a phase of 50 °. A signal having a waveform formed by combining a sine wave signal having a frequency of 19 Hz and a fourth sine wave signal having a phase of 110 ° and a fifth sine wave signal having a frequency of 29 Hz and a phase of 160 °. Is input to the recording apparatus 10. Then, the waveform (composite waveform) of the signal input to the recording device 10 was compared with the restored waveform by the analysis device 20.

  FIG. 17 shows a composite waveform (broken line) of a signal input to the recording apparatus 10, a sampling waveform (chain line) obtained by the recording apparatus 10 when the sampling frequency Fsp is 125 Hz, and an analysis apparatus 20 It is a figure which shows the decompression | restoration waveform (solid line) by. The analysis device 20 generates a restored waveform from the sampling waveform. As shown in FIG. 17, both waveforms are almost overlapped, and the maximum amplitude error was 0.191%. Therefore, it can be seen that the biological signal analysis system 1 of the present embodiment can appropriately restore even when a signal having a waveform formed by combining a plurality of sinusoidal signals having different frequencies and phases is input. Further, since the maximum error is small, it can be understood that frequency components other than the composite waveform are not generated in the restored waveform.

(Evaluation Experiment 4)
Next, when a sine wave signal was input to the recording apparatus 10 as a biological signal, the amplitude of the restored data by the analysis apparatus 20 was evaluated.

  Specifically, the frequency of the sine wave signal input to the recording device 10 is changed within a range of 1 [Hz] to Fsp / 2 [Hz] every 3 seconds, and one cycle of the restored data by the analysis device 20 The amplitude was measured and the minimum value was recorded.

  FIG. 18A is a diagram showing a restoration result when the sine wave signal input to the recording apparatus 10 is sampled at a sampling frequency Fsp of 125 Hz. FIG. 18A shows a graph (broken line) showing the amplitude (minimum value) of sampling data obtained by sampling at a sampling frequency Fsp of 125 Hz, and the amplitude (minimum) of restoration data from the sampling data by the analysis device 20. Value) is shown in a graph (solid line).

  As shown in FIG. 18A, the amplitude of the sampling data varies depending on the frequency of the input sine wave signal. On the other hand, it can be seen that the amplitude of the restored data does not vary depending on the frequency of the input sine wave signal, and an ideal result based on the design value of the band limiting filter 33 is obtained. In particular, in the restored data, the upper limit frequency of a band in which the amplitude is hardly attenuated (for example, a band in which the attenuation is less than −1 dB and is referred to as “non-attenuated band”) is high.

  In addition, although the same evaluation experiment was implemented by changing the sampling frequency Fsp (100 Hz, 200 Hz, 250 Hz, 500 Hz), ideal results were obtained in all cases (not shown).

  FIG. 18B is a graph in which the upper limit frequency of the non-attenuation band is plotted while changing the sampling frequency Fsp (100 Hz, 125 Hz, 200 Hz, 250 Hz, 500 Hz). A graph plotted with diamonds is a graph related to sampling data, and a graph plotted with squares is a graph related to restored data.

  As shown in FIG. 18B, regardless of the sampling frequency Fsp, the upper limit frequency of the non-attenuation band related to the restored data is about twice as high as the upper limit frequency of the non-attenuation band related to the sampling data. Therefore, it can be seen that the restored data is closer to the sine wave signal before sampling over a wider frequency band than the sampling data.

(Evaluation Experiment 5)
Next, when an electrocardiogram signal was input to the recording device 10, the vertex position of the R wave of the restoration data by the analysis device 20 was evaluated.

  Specifically, an electrocardiogram signal as shown in the upper part of FIG. 6 is input to the recording device 10, and the apex position of the R wave of the original electrocardiogram signal and the apex position of the R wave of the restored data by the analysis device 20 And compared.

  FIG. 19 shows an original electrocardiogram waveform (input waveform: rhombus plot), sampling data waveform (sampling waveform: triangle plot), and analysis apparatus 20 for the QRS wave of the electrocardiogram signal input to the recording apparatus 10. FIG. 5 is a diagram showing a waveform of restored data by (Restore waveform: circular plot).

  As shown in FIG. 19, the input waveform and the restored waveform almost overlap each other, and it can be seen that there is no deviation in the apex position (time direction deviation) of the R wave.

  Note that the same evaluation experiment was carried out by shifting the sampling relative position (triangular plot) with respect to the QRS wave.

  FIG. 20 shows the waveform of the original ECG signal (input waveform) and the waveform of the sampling data (sampling waveform) when the relative position (triangular plot) of sampling is shifted with respect to the QRS wave of the ECG signal, and restoration. It is a figure which shows the peak position of the R wave in the waveform (restoration waveform) of data.

  As shown in FIG. 20, even if the relative position of sampling with respect to the QRS wave of the electrocardiogram signal is shifted by 1 millisecond, a result is obtained in which the vertex position is not shifted between the input waveform and the restored waveform.

  It should be noted that not only the vertex position of the R wave but also the S wave and T wave of the electrocardiographic signal were obtained with no deviation in the vertex position (not shown).

(Evaluation Experiment 6)
Next, when an electrocardiogram signal was input to the recording device 10, the vertex value of the R wave of the restored data by the analysis device 20 was evaluated.

  Specifically, an electrocardiogram signal as shown in the upper part of FIG. 6 is input to the recording device 10, and the R wave apex value of the original electrocardiogram signal, the sampling data, and the R data restored by the analysis device 20 The peak value of the wave was compared.

  As shown in FIG. 19, the waveform of the original electrocardiogram signal and the waveform of the restored data are almost overlapped, and it can be seen that there is no shift in the peak value of the R wave (shift in the amplitude direction).

  Note that the same evaluation experiment was carried out by shifting the sampling relative position (triangular plot) with respect to the QRS wave.

  FIG. 21 shows the waveform of the original ECG signal (input waveform) and the waveform of the sampling data (sampling waveform) when the relative position of sampling (triangular plot) with respect to the QRS wave of the ECG signal is shifted. It is a figure which shows the waveform (restoration waveform) of data, and the vertex value of R wave in.

  As shown in FIG. 21, even when the sampling relative position of the electrocardiographic signal to the QRS wave is shifted by 1 millisecond, there is almost no shift in the peak value (mV) of the R wave between the input waveform and the restored waveform. Results were obtained.

  It should be noted that not only the peak value of the R wave but also the S wave and T wave of the electrocardiographic signal were obtained with almost no deviation in the peak value (not shown).

(Evaluation Experiment 7)
Next, when an electrocardiogram signal was input to the recording device 10, the R wave interval of the restored data by the analysis device 20 was evaluated.

  Specifically, an electrocardiogram signal as shown in the upper part of FIG. 6 is input to the recording device 10, and the R wave interval of the waveform (input waveform) of the original electrocardiogram signal and the restoration data by the analysis device 20 are displayed. The R wave interval of the waveform (restored waveform) was compared.

  FIG. 22 is a diagram illustrating the R wave interval (broken line) in the input waveform and the R wave interval (solid line) in the restored waveform in the recording apparatus 10.

  As shown in FIG. 22, it can be seen that the interval between the R waves in the input waveform and the interval between the R waves in the restored waveform are substantially overlapped and there is no deviation in the interval between the R waves.

<Theoretical background>
In each of the above-described evaluation experiments 1 to 4, experimental results when a sine wave signal is input to the recording apparatus 10 are shown. Although the frequency and phase values of the sine wave signal are limited, the present invention is not limited to the sine wave signal as in the above example, and the same result can be obtained no matter what signal is input to the recording apparatus 10. It is done. This is based on the following considerations.

  That is, any periodic function (including the biological signal of the above embodiment) can be expressed as a sum of a sine function and a cosine function by Fourier series expansion. If the frequency component included in the periodic function can be defined as finite, the developed sine function and cosine function are also limited to a finite frequency range. A periodic function that has been discretized by performing high-frequency cutoff that satisfies the sampling theorem has inherent amplitude characteristics, but it can be said that periodic information can be retained appropriately.

  From the above, a biological signal having a waveform that has been cut off at high frequencies by the band limiting filter 33 as in the above embodiment can be expressed by the sum of a sine function and a cosine function in a region below the cut off frequency.

  From each of the above evaluation experiments, it can be said that the sine function and cosine function having a predetermined frequency and having a finite frequency discretized can be restored by spline interpolation. At this time, a sine function and a cosine function having frequency components other than the predetermined frequency are hardly generated. For this reason, by applying this restored waveform to the coefficient calculation of Fourier series expansion, the Fourier series expansion can be appropriately expressed in the form of using the restored waveform. This is the basis for proper waveform reproduction using a restored waveform at a finite frequency.

<Modification>
Moreover, said embodiment intends to illustrate the summary of this invention, and does not limit this invention. Many alternatives, modifications, and variations will be apparent to those skilled in the art.

  For example, in the above embodiment, the recording device 10 and the analysis device 20 are described as separate devices, but this is not a limitation. For example, the recording apparatus 10 and the analysis apparatus 20 may be integrated. In addition, some processing that has been executed by the analysis apparatus 20 in the above-described embodiment may be executed by the recording apparatus 10. Moreover, you may make the analysis apparatus 20 perform some processes currently performed with the recording device 10 in the said embodiment.

  In the above embodiment, “Kaiser Filter” is used for the band compression in step S102, but the present invention is not limited to this.

In the above embodiment, a cubic spline interpolation function such as Equation (1) is used. However, the present invention is not limited to this, and a higher order spline interpolation function may be used. The coefficient _a i of the spline interpolating _function, b _i, c i, may change some conditions 1-3 was used in obtaining the _{d i.}

  In the above embodiment, interpolation points are added at regular intervals on each spline interpolation function in step S203. However, the present invention is not limited to this. May be.

  Moreover, although the said embodiment demonstrated the electrocardiogram signal as an example of a biomedical signal, it is not restricted to this. For example, it may be a myoelectric signal, an electroencephalogram, or a membrane potential of a nerve cell. Moreover, signals other than biological signals may be handled as the processing target.

  In the above embodiment, in the spline interpolation processing in step S <b> 202, the analysis apparatus 20 locally selects a plurality of sample points from the sample points included in the sampling data 150. The number of sample points to be selected at this time is preferably an even number, but is not limited thereto. Moreover, it may be more than six.

  In the above embodiment, the sampling frequency Fsp is 125 Hz, and the re-sampling frequency Fb of the biological signal data after the interpolation point is added is 1000 Hz. However, the present invention is not limited to this, and the re-sampling frequency Fb of the biological signal data may be 2000 Hz or the like.

  In the above embodiment, the biological signal input through the connection I / F 31 is amplified by the amplifier and supplied to the A / D converter 32. However, the present invention is not limited to this, and before supplying a biological signal to the A / D converter 32, noise (for example, a high frequency component) included in the biological signal by an analog filter (for example, a Butterworth filter) is removed. It may be removed.

  In the above embodiment, an example in which a digital filter is used as the band limiting filter 33 is described. However, the present invention is not limited to this, and an analog filter may be adopted as the band limiting filter 33. However, when an analog filter is employed, it is necessary to arrange a band limiting filter 33 between the connection I / F 31 and the A / D converter 32.

  The configuration of the biological signal analysis system 1 described above is not limited to the above-described configuration because the main configuration has been described in describing the features of the above-described embodiment. In addition, the configuration of the general biological signal analysis system 1 is not excluded.

  In addition, each functional configuration of the biological signal analysis system 1 described above is classified according to main processing contents in order to facilitate understanding of each functional configuration. The present invention is not limited by the way of classification and names of the constituent elements. Each functional configuration can be classified into more components according to the processing content. Moreover, it can also classify | categorize so that one component may perform more processes.

  The program for operating the biological signal analysis system 1 may be provided by a computer-readable recording medium such as a USB memory, a floppy (registered trademark) disk, a CD-ROM, or online via a network such as the Internet. May be provided in In this case, the program recorded on the computer-readable recording medium is usually transferred to and stored in the memory (35, 42), the storage (36, 43) or the like. Further, this program may be provided as, for example, a single application software, or may be incorporated in the software of each device as one function of the biological signal analysis system 1.

  The processing of each component described above can also be realized by a dedicated hardware circuit. In this case, it may be executed by one hardware or a plurality of hardware.

1 biological signal analysis system,
10 recording device,
11 electrodes,
12 Control device (recording device),
13 External memory,
20 analysis equipment,
21 Control device (analysis device),
22 display device,
23 input devices,
31 Connection I / F,
32 A / D converter,
33 Band-limiting filter,
34 CPU (recording device),
35 memory (recording device),
36 storage (recording device),
37 External memory slot (recording device),
38 Communication I / F (recording device),
41 CPU (analysis device),
42 Memory (analysis device),
43 Storage (analysis device),
44 External memory slot (analysis device),
45 Communication I / F (analysis device),
51 biological signal input unit,
52 band compression unit,
53 Sampling unit,
54 Sampling data storage unit,
61 Acquisition Department,
62 Interpolator,
63 Up-ampling section,
64 Restored data storage unit,
150 sampling data,
160 Restored data.

Claims (16)

A detection unit for detecting a biological signal from a living body;
A compression unit that removes high-frequency components of a predetermined value or more from the biological signal and band-compresses the biological signal;
A sampling unit that obtains sample points from a predetermined sampling frequency from the compressed biological signal, and generates sampling data composed of a plurality of sample points;
An up-sampling unit that adds a plurality of interpolation points between the sample points by spline interpolation and restores biological signal data with higher temporal resolution from the sampling data;
A biological signal processing apparatus.   The said compression part compresses the said biological signal with the high frequency cutoff filter which made the stop band edge frequency used as the boundary of the frequency band which blocks passage of a signal below 1/2 of the said sampling frequency. Biological signal processing apparatus.   The biological signal processing apparatus according to claim 2, wherein the high-frequency cutoff filter is a Kaiser Filter.   The biological signal processing apparatus according to claim 1, wherein the detection unit detects an electrocardiogram signal as the biological signal.   The upsampling unit performs cubic spline interpolation on a plurality of locally selected sample points, and adds a plurality of interpolation points between two of the sample points. 2. The biological signal processing apparatus according to item 1.   The living body according to claim 5, wherein the up-sampling unit performs cubic spline interpolation on an even number of locally selected sample points, and adds a plurality of interpolation points between two central points among the sample points. Signal processing device.   The upsampling unit performs cubic spline interpolation on six locally selected sample points, and adds a plurality of interpolation points between two central points among the sample points. Biological signal processing apparatus.   The biological signal processing apparatus according to claim 1, wherein the upsampling unit adds the plurality of interpolation points at equal intervals. A detection step of detecting a biological signal from the living body;
A compression step of removing a high-frequency component of a predetermined value or more from the biological signal and band-compressing the biological signal;
A sampling step for obtaining a sampling point at a predetermined sampling frequency from the compressed biological signal and generating sampling data composed of a plurality of sampling points;
A biological signal processing method for performing an up-sampling step of adding a plurality of interpolation points between the sample points by spline interpolation and restoring biological signal data with higher temporal resolution from the sampling data.   The said compression step compresses the said biological signal by the high frequency cutoff filter which made the stop band edge frequency used as the boundary of the frequency band which blocks passage of a signal below 1/2 of the said sampling frequency. Biological signal processing method.   The biological signal processing method according to claim 10, wherein the high-frequency cutoff filter is a Kaiser Filter.   The biological signal processing method according to claim 9, wherein the detecting step detects an electrocardiogram signal as the biological signal.   The upsampling step performs cubic spline interpolation on a plurality of locally selected sample points, and adds a plurality of interpolation points between two of the sample points. 2. The biological signal processing method according to item 1.   The living body according to claim 13, wherein the upsampling step performs cubic spline interpolation on an even number of locally selected sample points, and adds a plurality of interpolation points between two central points among the sample points. Signal processing method.   15. The up-sampling step performs cubic spline interpolation on six locally selected sample points, and adds a plurality of interpolation points between two central points among the sample points. Biological signal processing method.   The biological signal processing method according to claim 9, wherein the upsampling step adds the plurality of interpolation points at equal intervals.