WO2018168909A1 - Parkinson's disease diagnosing device - Google Patents

Abstract

The objective of the present invention is to provide a Parkinson's disease diagnosing device capable of assessing Parkinson's disease from a pulse wave of a subject. A CPU 23 generates a group B comprising m elements and a group A comprising m+1 elements from pulse wave data (S106). The CPU 23 calculates a sample entropy on the assumption that group A is represented by an m-dimensional vector and group B is represented by an m-dimensional vector (step S107). The CPU 23 assesses, for a curve of m=2, whether the value at r=0.1 exceeds a threshold (0.3), and assesses that the subject has Parkinson's disease if said value exceeds the threshold (step S117). This makes it possible to assess psychiatric disorders on the basis of objective data.

Description

Parkinson's disease diagnostic device

The present invention relates to a Parkinson's disease diagnosis apparatus, and more particularly to an objective judgment process using a pulse wave.

It is difficult to clearly determine whether cerebral infarction is the cause of limb tremor that is characteristic of Parkinson's disease.

Patent Document 1 (Japanese Patent Laid-Open No. 2016-86787) discloses a diagnostic agent for autosomal dominant Parkinson's disease containing a reagent for mutation detection in CHCHD2.

Patent Document 2 (Japanese Patent Application Laid-Open No. 2016-75644) discloses a determination method for performing diagnosis by measuring two amino acids contained in human blood and measuring their ratio.

Patent Document 3 (Japanese Unexamined Patent Application Publication No. 2015-100525) discloses that a mental illness can be determined from the shape of a sample entropy curve.

However, the methods disclosed in Patent Documents 1 and 2 are complicated. Patent Document 3 discloses that it can be applied to mental illness, but does not mention whether it can be used for diagnosis of Parkinson's disease.

An object of the present invention is to provide a Parkinson's disease diagnosis apparatus that can easily and objectively determine Parkinson's disease.

1) A Parkinson's disease diagnosis apparatus according to the present invention includes:
A) Pulse wave storage means for storing pulse wave data that varies depending on the blood flow flowing through the measurement site,
B) First computing means having the following b1) to b12):
b1) Combination generation means for generating a plurality of combinations composed of m consecutive elements for N individual data from the pulse wave data,
b2) Assuming that each combination is an m-dimensional vector, the number of vectors existing within the distance d is calculated for each vector, and a statistical average value is calculated for the number of m-dimensional vectors. means,
C) Second computing means having the following c1) to c2):
c1) Combination generation means for generating a plurality of combinations composed of m + α elements for N pieces of individual data from the pulse wave data,
c2) Assuming that each combination is an m + α-dimensional vector, for each vector, the number of vectors existing within the distance d is calculated, and a statistical average value for the number of m + α-dimensional vectors is calculated m + α-dimensional average vector number calculation means,
C) m-dimensional average vector number randomness calculating means for obtaining microscopic randomness of the m-dimensional average vector number of the pulse wave based on the m-dimensional average vector number and the m + α-dimensional average vector number;
D) Parkinson's disease when the m = 2 m-dimensional average vector number randomness in the case of d = [variation in blood flow volume of the N individual data * 0.1] exceeds 0.3. Means for judging,
It has.

Thus, the objective judgment for diagnosing Parkinson's disease becomes possible by the microscopic randomness when the distance d = 0.1.

2) In the Parkinson's disease diagnosis apparatus according to the present invention, the degree of variation is a standard deviation. Therefore, an objective judgment for diagnosing Parkinson's disease can be made based on the microscopic randomness when the distance d = 0.1.

3) The Parkinson's disease diagnosis apparatus according to the present invention further includes a pulse wave measurement unit that measures pulse wave data that varies depending on the blood flow flowing through the measurement site. Therefore, there is no need to provide pulse wave data from the outside.

4) The Parkinson's disease determination degree computing device according to the present invention is:
A) Pulse wave storage means for storing pulse wave data that varies depending on the blood flow flowing through the measurement site,
B) First computing means having the following b1) to b12):
b1) Combination generation means for generating a plurality of combinations composed of m consecutive elements for N individual data from the pulse wave data,
b2) Assuming that each combination is an m-dimensional vector, the number of vectors existing within the distance d is calculated for each vector, and a statistical average value is calculated for the number of m-dimensional vectors. means,
C) Second computing means having the following c1) to c2):
c1) Combination generation means for generating a plurality of combinations composed of m + α elements for N pieces of individual data from the pulse wave data,
c2) Assuming that each combination is an m + α-dimensional vector, for each vector, the number of vectors existing within the distance d is calculated, and a statistical average value for the number of m + α-dimensional vectors is calculated m + α-dimensional average vector number calculation means,
C) m-dimensional average vector number randomness calculating means for obtaining microscopic randomness of the m-dimensional average vector number of the pulse wave based on the m-dimensional average vector number and the m + α-dimensional average vector number;
D) Output means for outputting m = 2 m-dimensional average vector number randomness as Parkinson's disease judgment degree when d = [degree of blood flow variation of the N individual data * 0.1]
It has.

Therefore, objective data for diagnosing Parkinson's disease can be obtained from the microscopic randomness when the distance d = 0.1.

5) In the Parkinson's disease determination degree computing device according to the present invention, the variation degree is a standard deviation. Therefore, an objective judgment for diagnosing Parkinson's disease can be made based on the microscopic randomness when the distance d = 0.1.

6) The Parkinson's disease determination degree calculation method according to the present invention is:
Store the pulse wave data that changes depending on the blood flow flowing to the measurement site,
For N pieces of individual data from the pulse wave data, a plurality of combinations composed of continuous m elements are generated, and each combination is an m-dimensional vector. Calculating a number, calculating a statistical average value for the number of m-dimensional vectors,
For N pieces of individual data from the pulse wave data, a plurality of combinations composed of continuous m + α elements are generated, and each combination is set as an m + α-dimensional vector. While calculating the number of existing vectors, calculating a statistical average value for the number of m + α-dimensional vectors,
Based on the number of m-dimensional average vectors and the number of m + α-dimensional average vectors, Parkinson's disease determination degree calculation method using a computer for obtaining the microscopic randomness of the m-dimensional average vector number of the pulse wave,
When d = [degree of variation in blood flow volume of the N individual data * 0.1], the m = 2 m-dimensional average vector number randomness is output as a Parkinson's disease determination degree.

Therefore, objective data for diagnosing Parkinson's disease can be obtained from the microscopic randomness when the distance d = 0.1.

7) The Parkinson's disease diagnosis apparatus according to the present invention is:
A pulse wave sensor that detects pulse wave data that varies depending on the blood flow flowing through the measurement site;
A processor that executes the following steps by means of program modules stored in a memory area;
Display to display the calculation results,
With
For N pieces of individual data from the pulse wave data, a plurality of combinations composed of continuous m elements are generated, and each combination is an m-dimensional vector. Calculating a number and calculating a statistical mean value for the m-dimensional vector number;
For N pieces of individual data from the pulse wave data, a plurality of combinations composed of continuous m + α elements are generated, and each combination is set as an m + α-dimensional vector. Calculating the number of existing vectors and calculating a statistical average value for the number of m + α-dimensional vectors;
Obtaining a microscopic randomness of the m-dimensional average vector number of the pulse wave based on the m-dimensional average vector number and the m + α-dimensional average vector number;
A step of displaying m = 2 m-dimensional average vector number randomness of m = 2 on the display as a Parkinson's disease determination degree in the case of d = [variation degree of blood flow of the N individual data * 0.1],
It has.

Therefore, objective data for diagnosing Parkinson's disease can be obtained from the microscopic randomness when the distance d = 0.1.

8) The Parkinson's disease diagnosis apparatus according to the present invention is:
A) Pulse wave storage means for storing pulse wave data that varies depending on the blood flow flowing through the measurement site,
B) First computing means having the following b1) to b12):
b1) Combination generation means for generating a plurality of combinations composed of m consecutive elements for N individual data from the pulse wave data,
b2) Assuming that each combination is an m-dimensional vector, the number of vectors existing within the distance d is calculated for each vector, and a statistical average value is calculated for the number of m-dimensional vectors. means,
C) Second computing means having the following c1) to c2):
c1) Combination generation means for generating a plurality of combinations composed of m + α elements for N pieces of individual data from the pulse wave data,
c2) Assuming that each combination is an m + α-dimensional vector, for each vector, the number of vectors existing within the distance d is calculated, and a statistical average value for the number of m + α-dimensional vectors is calculated m + α-dimensional average vector number calculation means,
C) m-dimensional average vector number randomness calculating means for obtaining microscopic randomness of the m-dimensional average vector number of the pulse wave based on the m-dimensional average vector number and the m + α-dimensional average vector number;
D) Representative value storage means for storing in advance representative values of m-dimensional average vector number randomness of healthy subjects,
E) an output means for outputting a deviation between the representative value and the m-dimensional average vector number randomness of the subject as a Parkinson's disease determination degree;
It has.

Therefore, the possibility of suffering from Parkinson's disease can be diagnosed by objective data.

In this specification, “representative value” includes an average value, a median value, and a mode value in a narrow sense. In addition, the “average value” in the narrow sense includes various arithmetic methods such as geometric average, harmonic average, generalized average, and weighted average. The “differential pulse wave data” includes data obtained by differentiating the pulse wave data at least once.

The features, other objects, uses, effects, etc. of the present invention will become clear by referring to the embodiments and drawings.

It is a functional block diagram of Parkinson's disease diagnostic device 1. It is a figure which shows the hardware constitutions of the Parkinson's disease diagnostic apparatus. It is an example of the waveform of pulse wave data. An overall flowchart is shown. It is a detailed flowchart of a sample entropy calculation process. A list of sample entropy when r and m are changed is shown. It is a display example of a representative of a healthy person and a sample entropy curve to be measured. FIG. 7A is a diagram showing sample entropy of a healthy person, and FIG. 7B is a diagram showing sample entropy of a Parkinson's disease patient. It is a figure which shows the one-dimensional arrangement | positioning analysis result of the sample entropy of r = 0.1 about the healthy subject and Parkinson's disease patient of FIG. 7A and FIG. 7B. 7A and 7B are quantile regression results of the minimum, 10%, 25%, 50%, 75%, 90%, and maximum values for the sample entropy of r = 0.1 for the healthy subjects and patients with Parkinson's disease. . It is a figure which shows the change of the sample entropy when not filtering and after filtering. It is a figure which shows other embodiment. FIG. 12A shows sample entropy of a healthy person, and FIG. 12B shows sample entropy of a Parkinson's disease patient. It is a partition analysis result. It is a partition analysis result. It is an example of the notation of a constellation graph. It is a flowchart of constellation graph notation. It is a figure for demonstrating the notation of a line graph.

1 ... Parkinson's disease diagnosis device 23 ... CPU
27 ... Memory

(First embodiment)
Functional Block Diagram FIG. 1 shows a functional block diagram of the Parkinson's disease diagnosis apparatus 1 according to the present invention. The Parkinson's disease diagnostic apparatus diagnostic apparatus 1 includes a fingertip pulse wave measuring unit 3, a filtering unit 4, a first calculating unit 7, a second calculating unit 9, an m-dimensional average vector number randomness calculating unit 10, and a determining unit 17. ing.

The finger plethysmogram measuring means 3 measures the finger plethysmogram and measures pulse wave data that changes depending on the blood flow flowing through the measurement site. The filtering unit 4 deletes a frequency component exceeding at least 10 Hz from the measured pulse wave data. The electroencephalogram storage means 5 stores pulse wave data from which frequency components exceeding 10 Hz have been removed.

The first calculation means 7 includes combination generation means 7a and m-dimensional average vector number calculation means 7c. The combination generation means 7a generates a plurality of combinations composed of m consecutive elements for N pieces of individual data from the pulse wave data. The m-dimensional vector distance calculation means 7c calculates the number of vectors existing in the distance d for each vector, with each combination as an m-dimensional vector, and calculates the statistical average value for the number of m-dimensional vectors. .

The second calculation means 9 includes a combination generation means 9a and an m + α-dimensional average vector number calculation means 9c. The combination generation unit 9a generates a plurality of combinations including m + α elements for N pieces of individual data from the pulse wave data. The m + α-dimensional vector distance calculation means 9c calculates the number of vectors existing in the distance d for each vector, with each combination as an m + α-dimensional vector, and statistics on the number of m + α-dimensional vectors. The average value is calculated.

Based on the m-dimensional average vector number randomness calculation means 10, the m-dimensional average vector number, and the m + α-dimensional average vector number, the microscopic randomness of the m-dimensional average vector number of the pulse wave is obtained. The determination means 17 notifies that it is a mental illness when exceeding the measured threshold value.

2. Hardware Configuration The hardware configuration of the psychiatric disorder diagnosis apparatus 1 will be described with reference to FIG. FIG. 1 shows an example of a hardware configuration in which the mental disease diagnosis apparatus 1 is configured using a CPU.

The mental illness diagnosis apparatus 1 includes a CPU 23, a memory 27, a hard disk 26, a monitor 30, an optical drive 25, an input device 28, a communication board 31, a pulse wave sensor 33, and a bus line 29. The CPU 23 controls each unit via the bus line 29 according to each program stored in the hard disk 26.

The hard disk 26 stores an operating system program 26o (hereinafter abbreviated as OS) and a main program 26p. The processing of the main program 26p will be described later.

The pulse wave sensor 33 is a general fingertip pulse wave measuring instrument. In the present embodiment, a finger plethysmograph is used that measures blood flow using infrared rays and obtains a finger plethysmogram from this blood flow. Specifically, the infrared light emitted from the light emitting element is reflected by the finger to be measured and received by the light receiving element. The intensity of the reflected light represents the blood flow rate. Therefore, the signal output from the light receiving element is a fingertip volume pulse wave. The signal from this light receiving element is output as digital data.

Data from the pulse wave sensor 33 is taken into the memory 27 and stored in the pulse wave storage unit 26m.

FIG. 3A shows an example of a fingertip pulse wave output from the pulse wave sensor 33. Although it is actually digital data, it is shown as a waveform in FIG. 3A.

The result storage unit 26k stores lost sample entropy obtained from the measured pulse wave as described later.

In this embodiment, Windows 10 (registered trademark or trademark) is adopted as the operating system program (OS) 26o, but the present invention is not limited to this.

Each program is read from the CD-ROM 25a storing the program via the optical drive 25 and installed in the hard disk 26. In addition to the CD-ROM, a program such as a flexible disk (FD) or an IC card may be installed on a hard disk from a computer-readable recording medium. Furthermore, it may be downloaded using a communication line.

In the present embodiment, the program stored in the CD-ROM is indirectly executed by the computer by installing the program from the CD-ROM to the hard disk 26. However, the present invention is not limited to this, and the program stored in the CD-ROM may be directly executed from the optical drive 25. Note that programs that can be executed by a computer are not only programs that can be directly executed by being installed as they are, but also programs that need to be converted into other forms (for example, those that have been compressed) In addition, those that can be executed in combination with other module parts are also included.

3. Flowchart FIG. 4 shows an overall flowchart of the diagnostic data generation process. The user brings the measurement target finger into contact with the pulse wave sensor 33 and causes the psychiatric disorder diagnosis apparatus 1 to execute measurement start processing. As a result, the CPU 23 starts reading and storing the finger plethysmogram (step S101 in FIG. 4). The result is stored in the memory 27. In the stored pulse wave data, the horizontal axis represents time, and the vertical axis represents blood flow.

CPU 23 performs a filtering process for removing frequency components exceeding 10 Hz (step S102). In this embodiment, the pulse wave data is subjected to spectrum analysis, and frequency components exceeding 30 Hz are set to zero, and then subjected to inverse Fourier transform. Thereby, pulse wave data from which frequency components exceeding 30 Hz are removed is obtained.

The CPU 23 performs N samplings from the pulse wave data (step S103). In the present embodiment, data for 5 seconds is adopted as a sampling rate of 200 times per second (N = 1000).

CPU 23 sets the number of elements to m = 2 (step S105).

The CPU 23 generates a set B composed of m elements and a set A composed of m + 1 elements while shifting N sampling values one by one (step S106). For example, N sampling values are k (1) for the first, k (2) for the second, k (3) for the third, k (4) for the fourth, k (5) for the fifth, ... nth is k (n), m = 2, and set B is composed of m elements, so B (1) = (k (1), k (2)), B (2) = (k (2), k (3)), B (3) = (k (3), k (4)) ..., B (Nm) = (k (Nm), k ( Nm)), Nm sets of data are generated. On the other hand, since the set A is composed of m + 1 elements, A (1) = (k (1), k (2), k (3)), A (2) = (k (2) , K (3), k (4)), A (3) = (k (3), k (4), k (5))..., A (Nm) = (k (Nm), k ( N-m + 1), k (N)) and Nm sets of data are formed. Here, as for the set B, the same number of sets as the set A are generated in order to match the number of sets of both as described later.

The CPU 23 calculates the sample entropy assuming that the set A is expressed by an m-dimensional vector and the set B is expressed by an m-dimensional vector (step S107). The sample entropy calculation method employed in this embodiment will be described with reference to FIG.

The CPU 23 sets the allowable value r as an initial value (step S20). In the present embodiment, the initial value of the allowable value r is set to 0.1.

The CPU 23 initializes the process number i (step S21). The CPU 23 calculates the difference between the i-th set and all other sets with each set of the set B as an m-dimensional vector, and the vector specified by the i-th set is within the distance σ · r. An average vector number B ^m _i (r) of existing vectors is obtained (step S23). Note that σ is a standard deviation of the sampled N sampling values. For the average number of vectors existing within the distance σ · r, a difference is calculated for each vector, and the number of vectors within which the difference is within the distance σ · r may be obtained.

Here, first, since i = 1, the CPU 23 obtains a distance between the vector B (1) and another vector B (2) = (k (2), k (3)). In the present embodiment, the maximum value of the difference between corresponding elements is adopted as the distance between vectors. For example, in the case of vector B (1) and vector B (2), the difference between k (2) of vector B (2) and k (1) of vector B (1) and k ( Of the differences between 3) and k (2) of vector B (1), the larger value is the distance. As for the vector distance, other vector distance calculation methods (for example, Euclidean distance, etc.) may be employed.

The CPU 23 sets the number of vectors in the distance σ · r to “1” if the calculated distance is smaller than the distance σ · r, and “0” otherwise. This is to determine how many vectors other than itself exist near the vector B (1). Similarly, the distance between the vector B (1) = (k (1), k (2)) and the vector B (3) = (k (3), k (4)) Repeat to find the distance between the vectors B (1) and B (Nm). The obtained number is totaled and the average is obtained. In this embodiment, since the distance to itself is not obtained, (Nm-1) distances are totaled and the average of them is obtained. Thereby, the average number of neighbors of the vector B ² ₁ (0.1) is obtained.

Similarly, the difference between the i-th set and all other sets is calculated with each set of set A as an m + 1-dimensional vector, and the distance σ · An average vector number A ^{m + 1} _i (r) of vectors existing in r is obtained (step S25).

The CPU 23 increments the process number i (step S27), and determines whether or not the process number i has been completed (step S29). In the present embodiment, as to whether or not the process has been completed, the number m of elements constituting the set is subtracted from the total number N and the value obtained by adding 1 is exceeded.

In this case, since the process has not been completed, the processes in steps S23 to S27 are repeated. As a result, the average number of neighbors of the vector B ² ₂ (0.1) and the vector A ² ₂ (0.1) is obtained. This gives the average number of neighbors for all vectors.

In step S29, when the CPU 23 obtains the average number of neighbors for all the sets for the sets A and B, the average vector number A ^{m + 1 for} all the sets.  (r) and average vector number B ^m  (r) is obtained (step S31).

In the present embodiment, the average number of vectors B ^m  For (r), the simple average, that is, the average number of vectors B ^m2 ₁ (0.1), B ^m2 ₂ (0.1), and B ^m2 ₃ (0.1) were summed up to find the average. The average calculation method may be used. The same applies to the average vector number A ^{m + 1} _i (r).

The CPU 23 calculates the average vector number A ^{m + 1 of} all sets.  (r) and average vector number B ^m  Sample entropy is obtained from (r) (step S33). In the present embodiment, (A ^{m + 1)}  (r) / B ^m  (r)) was obtained, and the value obtained by multiplying the natural logarithm by −1 was taken as the sample entropy.

SampEn (m, r, N) =-ln (A ^{m + 1}  (r) / B ^m  (r)) ... (Formula 1)
Here, m is the number of elements constituting the set, r is a distance threshold, and N is the number of samples.

In this way, the sample entropy SampEn (m2, r1) (hereinafter, SampEn is abbreviated as Sam) of the set A and the set B when the allowable value r = 0.1 and the number of elements m = 2 is obtained, and the result storage unit 26k (See FIG. 2).

The CPU 23 adds the increment r1 to the allowable value r (step S41 in FIG. 5), determines whether the allowable value r exceeds the set value (step S43), and if not, the value increased by r1. The sample entropy of set A and set B is obtained. Thereafter, similarly, the sample entropy Sam (m2, r2) in which the allowable value r is increased by r1 is obtained and stored in the result storage unit 26k (see FIG. 2). In this embodiment, r1 = 0.01, but the present invention is not limited to this.

Hereinafter, in the same manner, a sample entropy Sam (m2, rmax) of r = rmax is obtained.

In step S13, if the allowable value r exceeds the set value, the process is terminated. In the present embodiment, the set value is 0.7, but the present invention is not limited to this. In this way, sample entropy is obtained when the allowable value r is sequentially changed.

Thereby, sample entropy Sam (m2, r1) to Sam (m2, rmax) in the number of elements m = 2 shown in FIG. 6 is stored (see line 401).

The CPU 23 increments the number of elements m (step S109 in FIG. 4), determines whether the number of elements m exceeds the set value (step S111), and if not, performs the processing of steps S106 and S107.

Thus, sample entropy Sam (m3, r1) to Sam (m3, rmax) at the number of elements m = 3 is stored (see line 402 in FIG. 6).

Similarly, the sample entropy Sam (mmax, r1) to Sam (mmax, rmax) at the number of elements m = mmax is obtained from the sample entropy Sam (m4, r1) to Sam (m4, rmax) at the number of elements m = 4. Stored (see lines 403 and 409 in FIG. 6).

In this embodiment, the sample entropy up to m = 10 and r = 0.6 is obtained, but the present invention is not limited to this.

The CPU 23 reads the average of the sample entropy curves with different r in the case of m = 2 of the healthy person stored in advance and reads the representative curve, and monitors the representative curve and the measured sample entropy of the user together. (Step S115).

FIG. 6a shows a display example of a m = 2 representative value curve 50 and a measurement result curve 51 of a healthy person.

The CPU 23 determines whether the value at r = 0.1 exceeds the threshold for the m = 2 curve, and determines that it is a Parkinson's disease if it exceeds (step S117). In the present embodiment, the threshold value is set to 0.3.

Hereinafter, such a threshold will be described. The inventor hypothesized that there is some difference in the sample entropy curve between healthy subjects and Parkinson's disease patients, and compared various data. As a result, we found the following trends.

FIG. 7 shows the data distribution of 106 healthy subjects and 28 Parkinson's disease patients. FIG. 7A shows a change in sample entropy when r is changed in a healthy person and FIG. 7B is changed in a Parkinson's disease patient. When both are compared, the value in the case of m = 2 and r = 0.1 is different between healthy subjects and Parkinson's disease patients. Specifically, the sample entropy value of r = 0.1 is higher than that of healthy subjects (healthy subjects are about 0.2, Parkinson's disease patients 0.3 or more). FIG. 8 shows a one-way analysis of sample entropy values for m = 2 and r = 0.1. FIG. 9 shows the quantile regression analysis result.

Thus, it can be seen that the value of m = 2 and r = 0.1 is significant as a numerical value indicating the degree of Parkinson's disease. Therefore, objective determination of Parkinson's disease can be made based on the degree of Parkinson's disease.

In the present embodiment, the frequency component exceeding 30 Hz is removed from the measured pulse wave data, but the frequency component to be removed is not limited to this. The reason is that it is sufficient if noise can be removed. Further, any filtering method may be used. Furthermore, raw data without such filtering may be used. Specifically, although frequency components of 30 Hz or higher are removed, sample entropy can be obtained in the same manner even with pulse wave data that is not filtered. This will be described below.

FIG. 10A shows sample entropy values obtained based on pulse wave data not filtered. Thus, if there is no filtering, it has a jagged shape. The inventor considered the influence of power supply noise.

The change in the E sample entropy of m = 2 was examined by applying a 10-step low-pass filter of 10 Hz, 20 Hz,. FIG. 10B shows the results of filtering by 10 Hz, 20 Hz, and 30 Hz low-pass filters. Thus, in the case of a 30 Hz low-pass filter, the sample entropy of r = 0.1 was almost the same as the stepped data. On the other hand, when 10 Hz and 20 Hz low-pass filters were used, the sample entropy at r = 0.1 was lowered. Therefore, it is desirable to use a low-pass filter that removes 30 Hz or more for filtering. When a 40 Hz, 50 Hz,... 100 Hz low-pass filter was used, it was almost the same as 30 Hz. Therefore, the low-pass filter is not limited to 30 Hz.

4). Second Embodiment FIG. 11 shows a hardware configuration of a psychiatric disorder diagnosis apparatus 200 according to another embodiment. The psychiatric disorder diagnosis apparatus 200 includes a pulse wave sensor 203, a processor 204, a display 205, and a memory area 207 for storing data.

In this embodiment, the processor 204 is composed of a CPU, but it may be composed of a microprocessor or other processing unit capable of executing computer-executable instructions.

In this embodiment, the memory area 207 is composed of a flash memory, but RAM, ROM, EEPROM, or other memory, CD-ROM, DVD or other optical disk storage, magnetic cassette, magnetic tape, magnetic disk -It may consist of a computer storage medium, such as a storage or other magnetic storage device.

In the present embodiment, the display 205 is composed of an LCD, but may be composed of other displays such as an organic EL.

Each program module stored in the memory area 207 will be described. The memory area 207 stores a first calculation program module 220, a second calculation program module 230, an m-dimensional average vector number randomness calculation program module 210, a result display program module 251 and a filtering program module 253.

The first calculation program module 220 includes a combination determination program module 221 and an m-dimensional average vector number calculation program module 222. The second calculation program module 230 includes a combination determination program module 231 and an m + α-dimensional average vector number calculation program module 232.

The processor 204 uses the program modules stored in the memory area 207 to execute the same steps as in the first embodiment (see FIGS. 4 and 5).

The filtering program module 253 performs the calculation process in step S102 of FIG. The first arithmetic program module 220 and the second arithmetic program module 230 perform the arithmetic processing of step S103 and step S106. The third arithmetic program module 235 performs the processing from step S21 to step S33 in FIG.

The determination program module 251 performs the process of step S117.

The above embodiments can be grasped as the following devices.

A pulse wave sensor that detects pulse wave data that varies depending on the blood flow flowing through the measurement site;
A processor that executes the following steps by means of program modules stored in a memory area;
Display to display the calculation results,
With
B) For N pieces of individual data from the pulse wave data, a plurality of combinations composed of consecutive m elements are generated, and each combination is an m-dimensional vector, and each vector exists within a distance d. Calculating the number of vectors and calculating a statistical mean value for the number of m-dimensional vectors;
C) For N pieces of individual data from the pulse wave data, a plurality of combinations composed of consecutive m + α elements are generated, and each combination is set as an m + α-dimensional vector. Calculating the number of vectors present in the vector, and calculating a statistical average value for the number of m + α-dimensional vectors,
D) obtaining a microscopic randomness of the m-dimensional average vector number of the pulse wave based on the m-dimensional average vector number and the m + α-dimensional average vector number;
E) a step of displaying the m = 2 m-dimensional average vector number randomness of m = 2 in the case of d = [variation degree of blood flow of the N individual data * 0.1] on a display as a Parkinson's disease determination degree;
A Parkinson's disease diagnosis apparatus comprising:

In the present embodiment, the value detected by the pulse wave sensor 203 is A / D converted, and the digital data is given to the filtering program module 253 to remove 30 Hz or less. Prior to this, 30 Hz or less may be removed by a low-pass filter.

FIG. 12 shows a sample entropy value curve when r = 0.01 (1%) to 0.1 (10%) is added in FIG. As apparent from FIG. 12, the characteristic of the curve may not be stable in a region where r = 0.1 (10%) or less. This is considered to be due to the large influence of error and the measurement conditions as r = 0.01. However, the trend as a whole for such a region does not change from r = 0.1, so the determination may be made in this region.

Hereinafter, since the value of the sample entropy when r = 0.1 is an index for determining whether or not the patient is Parkinson's disease, it is hereinafter referred to as BPE (Border value of Perkinson Entropy).

13 and 14 show further measurement data. FIGS. 13 and 14 show the partition analysis results of a total of 157 people including 113 healthy subjects and 45 Parkinson's disease patients. From this partition analysis, it can be seen that if BPE is 0.3017 or less, 94.65% are Parkinson patients, and if BPE is less than 0.2189, 97.48% are healthy.

Fig. 15 shows a semicircular constellation graph. This is a notation example when the measurement result is displayed in 10 stages. The display angle is changed according to the sample entropy value.

It is also possible to display multiple measurement histories on such a constellation graph. The notation process in the case of displaying a plurality of measurement histories will be described with reference to FIG. Hereinafter, a case will be described in which the measurement result is measured in the measurement result storage unit together with the measurement date and time, and the latest seven time-series data are read out and displayed as a graph.

For the measured data, the CPU 23 extracts the latest seven time-series data stored in the measurement result storage unit (step S201), and defines seven concentric semicircles at equal intervals (step S203). . In the present embodiment, as shown in FIG. 17A, a reference arc C1 having a radius half that of the outermost arc C7 is defined, and the interval between the reference arc C1 and the arc C7 is divided into five equal parts. Two arcs C2 to C6 were defined, and a total of seven concentric circles were defined.

CPU 23 initializes process number k (step S205 in FIG. 16). The CPU 23 obtains an intersection pk between the kth straight line and the kth arc (step S207). In this case, since k = 1, an intersection point p1 between the straight line L1 and the arc C1 is obtained (see FIG. 17B).

The CPU 23 connects the point p (k-1) and the point pk with a line segment (step S209). In this case, since k = 1, the CPU 23 connects the points p0 and p1 with line segments. In the present embodiment, since the point p0 is the center of the arc, the line segment is the same as the straight line L1.

CPU 23 determines whether process number k is final (step S211). In this case, since it is not final, the process number k is incremented (step S213), and the intersection pk of the kth straight line and the kth arc is obtained (step S207). In this case, since k = 2, the intersection point p2 between the straight line L2 and the arc C2 is obtained.

The CPU 23 connects the point p (k-1) and the point pk with a line segment (step S209). In this case, since k = 2, the CPU 23 connects the points p1 and p2 with line segments. The CPU 23 determines whether or not the process number k is final (step S211). In this case, since it is not final, the process number k is incremented (step S213), and the processes of steps S207 to S213 are repeated.

As a result, as shown in FIG. 17C, a display history such as a line graph connecting points p0 to p7 is created. The CPU 23 displays the created display history screen data on the monitor 30 (see FIG. 2).

In this embodiment, a plurality of concentric circles are defined, and the intersection of the straight line defined by the time series data that is the basis of the history data and the corresponding concentric circles is obtained. Therefore, the end of the line segment can be positioned up to the top of the outermost arc a predetermined number of times without changing the angle of the straight line. In addition, since a straight line is defined from the center of the arc and the intersection with the corresponding arc is calculated, the change in the inclination of the straight line is expressed more greatly. Further, it can be displayed as history data without changing the slope of the straight line.

In the present embodiment, the threshold value “0.3” of the sample entropy value is set as the vertex value. As a result, if it is smaller than the above threshold value, it will tilt to the left as a whole, and if it is larger than the above threshold value, it will tilt to the right, so it is possible to intuitively determine whether you have Parkinson disease it can.

The angle β may be 90 degrees if the sample entropy value is 0.3, and 30 degrees when the sample entropy value increases by 0.1.

5). Other Embodiments In this embodiment, a sample entropy value of r = 0.1 and m = 2 is 0.3 or more with 5 units of data (N = 1000) as a unit at a sampling rate of 200 times per second. Was determined as a threshold. If measurement data for a longer time is used, the threshold value may be adjusted accordingly. For example, the sample entropy value decreases as the number of samples increases.

In the present embodiment, the number of samplings N = 1000, but the value of N is not limited to this, and N = 12000, and N = 36000 may be used. Even if the sampling rate is fine, the value of sample entropy does not change much. Because. This is because the sample entropy is a value indicating in which range a data group at a close position in space exists.

In the present embodiment, the diagnosis result that Parkinson's disease is suspected is notified, but a graph is displayed together with the representative value of a healthy person so that a doctor or the like can determine whether Parkinson's disease is present. .

Also, two thresholds may be provided to determine which measurement result belongs to. For example, if it is 0.3 or more and less than 0.4, it may be determined that “there is a possibility of Parkinson's disease”, and if it is 0.4 or more, it is determined that “the possibility of Parkinson's disease is high”.

In the present embodiment, the sample entropy value of r = 0.1 and m = 2 is determined to be 0.3 or more as a threshold value, but the representative value of the sample entropy curve of a healthy person is stored, and the deviation is the threshold value. When it exceeds, you may make it judge that it suffers from Parkinson's disease.

In the present embodiment, the case where the value of m is set to m and m + 1 has been described for the sets B and A, but the present invention is not limited to this, and may be m and m + 2, for example.

In the present embodiment, the sample entropy is obtained by the processing shown in FIGS. 4 and 5, but the present invention is not limited to this, and the sample entropy may be obtained by another calculation method.

In the present embodiment, the sample entropy is obtained, but any arithmetic expression representing a general redundancy, such as an entropy arithmetic expression of Kolomogorov or Shannon, can be employed.

In the present embodiment, the same number of sets as the set A are generated for the set B. However, the set B may be composed of N-m + 1 sets. This is because, if N is about 100 or more, the ratio of the sample entropy between the two is almost the same as the difference in error.

In this embodiment, in step S23 and step S25 in FIG. 5, the average is obtained by summing (Nm-1) distances, and the average is obtained. However, the present invention is not limited to this. The distances 0 may be added, and (Nm) distances may be summed up to obtain an average of them.

In the present embodiment, the distance d is r · σ (σ is a standard deviation of N samples), but is not limited thereto. In short, any numerical value indicating the variation of the pulse wave may be used.

In this embodiment, the pulse wave data is adopted, but it may be a velocity pulse wave or an acceleration pulse wave obtained by differentiating the pulse wave data. The velocity pulse wave can be obtained by differentiating the pulse wave data once, and the acceleration pulse wave can be obtained by differentiating the pulse wave data twice. Further, differential pulse wave data obtained by differentiating acceleration pulse waves one or more times may be used.

In addition, the terminal computer may be a mobile terminal instead of a personal computer. Alternatively, the terminal computer may not perform the calculation, but may transmit data necessary for the calculation to a network-connected computer and transmit the calculation result of the computer to the terminal computer.

In the above embodiment, in order to realize the function shown in FIG. 1, this is realized by software using a CPU. However, some or all of them may be realized by hardware such as a logic circuit.

In addition, you may make it make an operating system (OS) process a part of said program.

Although the present invention has been described as a preferred embodiment in the foregoing, it has been used for purposes of illustration and not limitation, and is intended to be used without departing from the scope and spirit of the present invention. Changes can be made within the scope of the claims.

Claims (8)

A) Pulse wave storage means for storing pulse wave data that varies depending on the blood flow flowing through the measurement site,
B) First computing means having the following b1) to b12):
b1) Combination generation means for generating a plurality of combinations composed of m consecutive elements for N individual data from the pulse wave data,
b2) Assuming that each combination is an m-dimensional vector, the number of vectors existing within the distance d is calculated for each vector, and a statistical average value is calculated for the number of m-dimensional vectors. means,
C) Second computing means having the following c1) to c2):
c1) Combination generation means for generating a plurality of combinations composed of m + α elements for N pieces of individual data from the pulse wave data,
c2) Assuming that each combination is an m + α-dimensional vector, for each vector, the number of vectors existing within the distance d is calculated, and a statistical average value for the number of m + α-dimensional vectors is calculated m + α-dimensional average vector number calculation means,
C) m-dimensional average vector number randomness calculating means for obtaining microscopic randomness of the m-dimensional average vector number of the pulse wave based on the m-dimensional average vector number and the m + α-dimensional average vector number;
D) Parkinson's disease when the m = 2 m-dimensional average vector number randomness in the case of d = [variation in blood flow volume of the N individual data * 0.1] exceeds 0.3. Means for judging,
A Parkinson's disease diagnosis apparatus comprising: The Parkinson's disease diagnostic device according to claim 1,
The degree of variation is a standard deviation;
Parkinson's disease diagnostic apparatus characterized by this. The Parkinson's disease diagnostic apparatus according to claim 1 or 2, further comprising:
Having a pulse wave measurement means for measuring pulse wave data that varies depending on the blood flow flowing through the measurement site;
Parkinson's disease diagnostic apparatus characterized by this. A) Pulse wave storage means for storing pulse wave data that varies depending on the blood flow flowing through the measurement site,
B) First computing means having the following b1) to b12):
b1) Combination generation means for generating a plurality of combinations composed of m consecutive elements for N individual data from the pulse wave data,
b2) Assuming that each combination is an m-dimensional vector, the number of vectors existing within the distance d is calculated for each vector, and a statistical average value is calculated for the number of m-dimensional vectors. means,
C) Second computing means having the following c1) to c2):
c1) Combination generation means for generating a plurality of combinations composed of m + α elements for N pieces of individual data from the pulse wave data,
c2) Assuming that each combination is an m + α-dimensional vector, for each vector, the number of vectors existing within the distance d is calculated, and a statistical average value for the number of m + α-dimensional vectors is calculated m + α-dimensional average vector number calculation means,
C) m-dimensional average vector number randomness calculating means for obtaining microscopic randomness of the m-dimensional average vector number of the pulse wave based on the m-dimensional average vector number and the m + α-dimensional average vector number;
D) Output means for outputting m = 2 m-dimensional average vector number randomness as Parkinson's disease judgment degree when d = [degree of blood flow variation of the N individual data * 0.1]
A Parkinson's disease determination degree computing device comprising: In the Parkinson's disease determination degree calculation device according to claim 4,
The degree of variation is a standard deviation;
Parkinson's disease determination degree arithmetic device characterized by the above. Store the pulse wave data that changes depending on the blood flow flowing to the measurement site,
For N pieces of individual data from the pulse wave data, a plurality of combinations composed of continuous m elements are generated, and each combination is an m-dimensional vector. Calculating a number, calculating a statistical average value for the number of m-dimensional vectors,
For N pieces of individual data from the pulse wave data, a plurality of combinations composed of continuous m + α elements are generated, and each combination is set as an m + α-dimensional vector. While calculating the number of existing vectors, calculating a statistical average value for the number of m + α-dimensional vectors,
Based on the number of m-dimensional average vectors and the number of m + α-dimensional average vectors, Parkinson's disease determination degree calculation method using a computer for obtaining the microscopic randomness of the m-dimensional average vector number of the pulse wave,
When the d = [degree of variation in blood flow volume of the N individual data * 0.1], the m = 2 m-dimensional average vector number randomness is output as a Parkinson's disease determination degree,
Parkinson's disease judgment degree calculation method characterized by the above. A pulse wave sensor that detects pulse wave data that varies depending on the blood flow flowing through the measurement site;
A processor that executes the following steps by means of program modules stored in a memory area;
Display to display the calculation results,
With
For N pieces of individual data from the pulse wave data, a plurality of combinations composed of continuous m elements are generated, and each combination is an m-dimensional vector. Calculating a number and calculating a statistical mean value for the m-dimensional vector number;
For N pieces of individual data from the pulse wave data, a plurality of combinations composed of continuous m + α elements are generated, and each combination is set as an m + α-dimensional vector. Calculating the number of existing vectors and calculating a statistical average value for the number of m + α-dimensional vectors;
Obtaining a microscopic randomness of the m-dimensional average vector number of the pulse wave based on the m-dimensional average vector number and the m + α-dimensional average vector number;
A step of displaying m = 2 m-dimensional average vector number randomness of m = 2 on the display as a Parkinson's disease determination degree in the case of d = [variation degree of blood flow of the N individual data * 0.1],
A Parkinson's disease diagnosis apparatus comprising: A) Pulse wave storage means for storing pulse wave data that varies depending on the blood flow flowing through the measurement site,
B) First computing means having the following b1) to b12):
b1) Combination generation means for generating a plurality of combinations composed of m consecutive elements for N individual data from the pulse wave data,
b2) Assuming that each combination is an m-dimensional vector, the number of vectors existing within the distance d is calculated for each vector, and a statistical average value is calculated for the number of m-dimensional vectors. means,
C) Second computing means having the following c1) to c2):
c1) Combination generation means for generating a plurality of combinations composed of m + α elements for N pieces of individual data from the pulse wave data,
c2) Assuming that each combination is an m + α-dimensional vector, for each vector, the number of vectors existing within the distance d is calculated, and a statistical average value for the number of m + α-dimensional vectors is calculated m + α-dimensional average vector number calculation means,
C) m-dimensional average vector number randomness calculating means for obtaining microscopic randomness of the m-dimensional average vector number of the pulse wave based on the m-dimensional average vector number and the m + α-dimensional average vector number;
D) Representative value storage means for storing in advance representative values of m-dimensional average vector number randomness of healthy subjects,
E) an output means for outputting a deviation between the representative value and the m-dimensional average vector number randomness of the subject as a Parkinson's disease determination degree;
A Parkinson's disease diagnosis apparatus comprising:

Images