JP5809864B2 - Biological light measurement device, waveform analysis method - Google Patents

Description

  The present invention relates to a living body light measuring device and the like that noninvasively measures the blood volume inside a living body with near infrared light.

  The biological light measurement device is a device that measures the blood circulation, hemodynamics, hemoglobin change, and the like inside the living body by utilizing the absorption of near infrared light by hemoglobin when passing through the living body. In the biological light measurement device, the blood volume inside the living body can be measured with almost no restraint on the subject and no harm to the subject. When the biological light measurement device is used, for example, a change in cerebral blood volume in the range of 2 to 3 cm from the surface of the head can be measured, and brain activation can be easily grasped. Such a brain function measurement method using a biological optical measurement device is called optical brain functional imaging, fNIRS (functional near-infrared spectroscopy), optical topography (registered trademark), or the like.

  As an application example of the biological optical measurement device in the medical field, for example, advanced medical treatment “assistant diagnosis of depressive symptoms using optical topography examination” can be cited (see Non-Patent Document 1). The outline of this test is as follows: (1) A test task in which a subject speaks as many Japanese words as possible starting with a specified initial letter while wearing a probe of a biological light measurement device (“language fluency” (2) While the subject is performing the examination task, the biological light measurement device measures the change in the brain activity state in the frontal lobe and temporal lobe of the subject as a waveform, (3) Furthermore, the biological light measuring device or the doctor analyzes the data and determines which pattern of mental disease the activation state of the brain (especially the cerebral cortex) for the examination subject matches. By determining, an accurate differential diagnosis is performed by assisting clinical diagnosis.

  Another example of the use of the biological optical measurement device in the medical field includes cerebral ischemia diagnosis (see Non-Patent Document 2). Non-Patent Document 2 describes that a change in oxygenated hemoglobin concentration associated with an oxygen inhalation task is measured and principal component analysis is performed on the measurement data. In Non-Patent Document 2, the left-right uniformity of the topography image (hereinafter, referred to as “weight value topography image”) created using the weight value (presence frequency) of the first principal component is measured before and after treatment. It is reported that it is different. When the report of Non-Patent Document 2 is viewed in detail, it can be seen that the treatment improvement region of the weight value topography image is a part, and that the hemoglobin change and the weight value topography image on the non-treatment side also change before and after the treatment. Thus, it can be expected that more clinical findings can be obtained by analyzing hemoglobin changes and topographic images in detail. It can also be expected that features for each brain functional region can be obtained by taking into account hemoglobin changes and weight value topography images after the second principal component.

Supervised by Masato Fukuda, edited by the Association of Applying Optical Topography Tests to Mental Health, “NIRS Waveform Clinical Interpretation-Advanced Medicine“ Optical Topography Test for Depression ”” published by Nakayama Shoten, April 2011 Akira Ebihara et al., “Diagnosis of cerebral ischemia by optical topography using oxygen inhalation”, Clinical Neurophysiology 37 (3): 77-84, 2009

  However, at present, it cannot be said that there are sufficient reports on examples of use of the biological optical measurement device in the medical field, and no reports have been analyzed in detail as described above. Therefore, oxygenated hemoglobin (oxy-hemoglobin), deoxy-hemoglobin, and total hemoglobin (total-hemoglobin), which is the sum of these, measured by a living body light measurement device (hereinafter collectively referred to as “total”) Is simply written as “hemoglobin”), and a method for confirming the waveform in detail and efficiently has not been devised sufficiently.

  The present invention has been made in view of the above-described problems, and an object of the present invention is to provide a biological light measurement device and the like that can confirm a hemoglobin waveform in detail and efficiently.

In order to achieve the above-described object, the first invention is provided with an irradiation probe for irradiating a subject with near-infrared light, and a detection probe for detecting the near-infrared light that has passed through the inside of the subject. And an apparatus main body that measures a hemoglobin waveform for each measurement channel that is a part between the irradiation probe and the detection probe and displays a measurement result. A biological light measurement device for classifying a plurality of clusters based on similarity of some or all of the measured hemoglobin waveforms, and generating a cluster waveform that is a waveform representative of the clusters means, together with the cluster waveform is information for identifying a display unit, the measurement result to display the correspondence cluster distribution information indicating a between the measuring channel and the cluster identification And a storage means for storing the cluster waveform and the cluster distributions for each information, the display means displays the cluster waveform and said cluster distributions for plurality of different identification information, and / or two different wherein a biological light measuring device which is characterized that you show the differences of the cluster distribution in the identification information.

According to a second aspect of the present invention, an irradiation probe that irradiates a subject with near-infrared light and a detection probe that detects the near-infrared light that has passed through the subject are arranged and attached to the subject. A living body optical measurement device configured to include a mounted portion and a device main body portion that measures a hemoglobin waveform for each measurement channel that is a portion between the irradiation probe and the detection probe and displays a measurement result. A waveform analysis method, a classification step for classifying into a plurality of clusters based on the similarity of some or all of the measured hemoglobin waveforms, and a generation step for generating a cluster waveform that is a waveform representative of the clusters; , together with the cluster waveform, and a display step of displaying the correspondence cluster distribution information indicating a between the measuring channel and the cluster, the information identifying the measurement result Comprising: a storage step of storing the cluster waveform and the cluster distributions for each that identification information, and said display step displays the cluster waveform and said cluster distributions for plurality of different identification information, and / or different Displays the difference of the cluster distribution in the two pieces of identification information
That it is a waveform analysis method characterized.

  According to the present invention, it is possible to provide a living body light measurement device or the like that can confirm a hemoglobin waveform in detail and efficiently. In particular, according to the biological light measurement device and the like according to the present invention, since the hemoglobin waveform is classified based on the similarity of the hemoglobin waveform, the characteristic component of the hemoglobin waveform is not separated as in the principal component analysis, etc. The feature in the brain function area can be clearly confirmed by the image showing the distribution of the waveform group.

The figure which shows an example of the hardware constitutions of a biological light measuring device The figure which shows an example of a mounting part The flowchart which shows the flow of processing of a living body light measuring device Figure showing a display example of the current cluster waveform and cluster distribution Figure showing a display example of a composite image The figure which shows the example of a display of the last cluster waveform and cluster distribution The figure which shows the example of a display of the difference of the hemoglobin waveform in several identification information The figure which shows the example of a display of the difference of the cluster distribution in several identification information The figure which shows the 1st example of a cluster feature-value display The figure which shows the 2nd example of a cluster feature-value display The figure which shows the example of a cluster distribution display about the number of several clusters

  The scope of application of the present invention is not limited by the purpose of diagnosis (differential diagnosis or the like) or the content of inspection (examination subject or the like). The present invention can be applied to any measurement data (hemoglobin waveform) acquired by a biological light measurement device. Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

  First, the hardware configuration of the biological light measurement device 1 will be described with reference to FIGS. 1 and 2. Note that the hardware configuration of the biological light measurement device 1 shown in FIGS. 1 and 2 is an example. The present invention is not limited by the hardware configuration of the biological light measurement device 1.

  As shown in FIG. 1, the biological light measurement device 1 includes a device main body 2 and a mounting portion 3. The apparatus main body 2 and the mounting portion 3 are connected by a transmission optical fiber 4 and a reception optical fiber 5. During the measurement by the biological light measurement device 1, the mounting unit 3 is mounted on the head of the subject 19.

  As shown in FIG. 2, the mounting portion 3 has passed through an irradiation probe 21 that irradiates the subject 19 with near-infrared light (light in the visible to infrared wavelength region) and the inside of the subject 19. A plurality of detection probes 22 for detecting near-infrared light are arranged. In order to detect information on the cerebral cortex from the top of the scalp, it is necessary that the distance between the irradiation position on the scalp and the detection position is about 30 mm apart. Therefore, the irradiation probes 21 and the detection probes 22 are alternately arranged in a square lattice pattern at intervals of 30 mm. Since the cerebral cortex can be measured with high sensitivity at each midpoint between the irradiation position and the detection position from the light propagation characteristics in the living body, the region between the adjacent irradiation probe 21 and detection probe 22 is set as the measurement channel 23.

  FIG. 2 schematically shows a state in which the subject 19 is wearing the mounting portion 3. The subject 19 wears the wearing part 3 so as to cover a part of the right head, a part of the upper front, and a part of the left head.

  In the enlarged schematic view of the mounting portion 3 shown in FIG. 2, the circle with “x” inside indicates the irradiation probe 21, and the circle with “●” inside indicates the detection probe 22. A small point between the irradiation probe 21 and the detection probe 22 indicates the measurement channel 23. In the example illustrated in FIG. 2, the mounting unit 3 includes 52 measurement channels 23 (CH1 to CH52). Note that the present invention is not limited to the number of measurement channels 23.

  Returning to the description of FIG. The apparatus main body 2 includes a measuring unit 6 and a computer 7. The measurement unit 6 and the computer 7 may be stored in one housing or may be stored in separate housings.

  The measuring unit 6 includes a laser diode 8, a detector 9, a lock-in amplifier 10, and an AD conversion 11. The laser diode 8 (light source unit) includes a plurality of optical modules that emit light having a plurality of wavelengths. The laser diode 8 includes an oscillating unit composed of a plurality of oscillators having different oscillating frequencies, and applies different modulations to the respective optical modules. Light of a plurality of wavelengths emitted from the laser diode 8 passes through the transmission optical fiber 4 and is irradiated on the head of the subject 19 from the irradiation probe 21.

  The detector 9 includes a photoelectric conversion element (a photodiode, a photomultiplier tube, or the like) that converts the light detected by the detection probe 22 and passing through the receiving optical fiber 5 into an electric signal corresponding to the amount of light. When a photodiode is used as the photoelectric conversion element, an avalanche photodiode capable of realizing highly sensitive optical measurement is preferable.

  The lock-in amplifier 10 receives an electrical signal from the detector 9 and selectively detects a modulation signal corresponding to the irradiation position and wavelength. Since the lock-in amplifier 10 is described in, for example, Japanese Patent No. 3599426, Japanese Patent No. 3839202, and the like, detailed description thereof is omitted.

  The AD converter 11 receives the analog signal from the lock-in amplifier 10 and converts it into a digital signal. The digital signal converted by the AD converter 11 is output to the computer 7.

  The computer 7 includes a digital board 12, a control unit 13, a storage unit 14, a display unit 15, an input unit 16, and the like. The digital board 12 is an interface for inputting a signal from the AD converter 11.

  The control unit 13 includes a CPU, a ROM, a RAM, and the like, and controls various devices of the biological light measurement device 1. In particular, the control unit 13 uses the detected light amounts of a plurality of wavelengths for each measurement channel 23 from the signal input via the digital board 12 (or the signal stored in the storage unit 14) to detect oxygen associated with brain activity. As a hemoglobin waveform (refer to the hemoglobin waveform 31x, etc. in FIG. 4), the temporal change of oxy-hemoglobin, deoxy-hemoglobin, and the total hemoglobin, which is the sum of these, Measure and display the measurement result on the display unit 15. Since this measurement method and imaging method are described in, for example, Japanese Patent No. 3599426, detailed description is omitted.

  The storage unit 14 is, for example, a built-in hard disk device, an external hard disk device, or various storage media (USB memory or the like), and stores signals input via the digital board 12 and processing results by the control unit 13. To do. The data stored in the storage unit 14 may be stored in a hospital (not shown) or a server on a network.

  The display unit 15 is a monitor such as a liquid crystal display, and displays a processing result by the control unit 13. The input unit 16 is an input device such as a mouse or a keyboard. The display unit 15 and the input unit 16 serve as a user interface that realizes an interactive process between the user (such as a doctor) and the computer 7. In addition, the display part 15 and the input part 16 may be integrated like a touchscreen type display etc., for example.

  Next, the details of the process of the biological light measurement apparatus 1 will be described with reference to FIGS. In addition, in order to obtain the processing results shown in FIGS. 4 to 11, the biological light measurement device 1 changes a part of the processing order shown in FIG. 3 or executes a part of the processing a plurality of times as necessary. Or omit some of the processing. In the following, in order to avoid confusion, the processing contents will be described in the order described in FIG.

  As shown in FIG. 3, the control unit 13 of the biological light measurement apparatus 1 measures the hemoglobin waveform of the subject 19 and stores it in the RAM, the storage unit 14 or the like (S1). As illustrated in FIG. 4, the control unit 13 assigns identification information 30 x that uniquely identifies the measurement result to the measurement result, and stores the measurement result for each measurement result. In FIG. 4, the measurement ID “1002” is assigned to the measurement result of the subject “S” and the measurement start time “July 1, 2011 13:10:10”.

  FIG. 4 shows the current measurement result, and FIG. 6 shows the previous measurement result. In order to distinguish these, “x” is added to the code of the element related to the current measurement result, and “y” is added to the code of the element related to the previous measurement result. Further, when these are collectively referred to without being distinguished, “x” and “y” are not added.

  As shown in FIG. 4, the hemoglobin waveform 31 x is measured for each measurement channel 23. The hemoglobin waveform 31x is displayed as a graph of the elapsed time (seconds) from the measurement start time on the horizontal axis and the measurement value (mM · mm) on the vertical axis. The measured value (mM · mm) is a hemoglobin concentration length that is a product of the hemoglobin concentration (mM) and the length (mm). The measured value is sampled every 0.1 seconds, for example. In the example shown in FIG. 4, the horizontal axis is in the range of 0 to 125 seconds, and the vertical axis is in the range of −0.2 to +0.4. Note that the vertical dotted lines at the elapsed time of 10 seconds and 70 seconds indicate the start time and end time of the feature amount calculation section 37 shown in FIGS.

  Returning to the description of FIG. The control unit 13 classifies into a plurality of clusters based on the similarity (particularly, the similarity of the entire waveform) of some or all of the measured hemoglobin waveforms 31, and calculates the cluster distribution 32 (S2). A cluster means a subset of data that is separated from a data set by some scale by a clustering technique, which is one of data analysis techniques. In the clustering method, the data included in the subset is classified so as to have a certain common characteristic (similarity of the entire waveform in the embodiment of the present invention). As for the number of clusters, the minimum value is two and the maximum value is the number of measurement channels 23. The cluster distribution 32 is information indicating the association between the measurement channel 23 and the cluster.

  Clustering methods are classified into hierarchical methods (such as the shortest distance method) and division optimization methods (such as k-average method). The hierarchical method is a method of generating a hierarchical structure by repeating sequential merging between clusters having the closest distance. Hierarchical methods are vulnerable to disturbances (noise, etc.), and there is a tendency that good results cannot be obtained with actual data. Therefore, in the embodiment of the present invention, the k-average method, which is a division optimization method that is robust against disturbance, is employed. The k-average method is a method of searching for a better division by dividing the input data group by setting the number of clusters to be k in advance, and repeatedly performing the division using this as an initial state. However, the present invention is not limited to the k-average method.

  First, the control unit 13 excludes, for example, the hemoglobin waveform 31 considered to be an artifact or the like from the entire measured hemoglobin waveform 31, and determines a classification target. The cause of the artifact includes the movement of the optical fiber, the change in the head position of the subject 19, the movement of the temporal muscle of the subject 19, and the like. The artifact determination is described in Non-Patent Document 1, for example.

  Next, the control unit 13 normalizes (z scores) the measurement values of the hemoglobin waveform 31 to be classified for each measurement channel 23. In the normalization process, for example, the average value and standard deviation of the measurement values of the entire measurement section (0 to 125 seconds) are used. When the sampling interval is 0.1 second, the number of measurement values in the entire measurement section (0 to 125 seconds) is 1250.

  Next, the control unit 13 uses vector data of normalized measurement values for each measurement channel 23 (in the above example, 1250-dimensional vector data) as an input data set, and uses a k-average method to calculate a plurality of data. Class. As the number of clusters, the number of iterations, and the initial value (representative data of each cluster in the initial state), for example, a default value stored in advance in the storage unit 14 or an input value input from the input unit 16 by the user can be used. good. Since the k-average method is a known clustering method, detailed description thereof is omitted.

  The cluster distribution 32x shown in FIG. 4 arranges a small region for each measurement channel 23 and shows the positional relationship between the measurement channels 23 two-dimensionally. Moreover, the cluster distribution 32x changes the form (shape, pattern, color, or combination thereof) of the small regions for each cluster. In FIG. 4, the first cluster is shown as a small circle with a “+” inside, and the second cluster is shown as a black small circle inside.

  In general, in order to accurately read the characteristics of the hemoglobin waveform 31 directly from the waveform, a considerable level of proficiency is required. Further, when the number of measurement channels 23 increases, even a highly skilled user takes a very long time. For example, the characteristic component of the hemoglobin waveform 31 can be separated by principal component analysis or the like, but it may hinder intuitive understanding.

  On the other hand, in the embodiment of the present invention, since the control unit 13 of the biological light measurement device 1 automatically calculates the cluster distribution 32, it takes time even when the user has a low level of proficiency or the number of measurement channels 23 is large. Measurement results can be read without Furthermore, in the cluster distribution 32x shown in FIG. 4, since the clustering result is superimposed on the image showing the positional relationship between the measurement channels 23, the features in the brain function region can be clearly confirmed. Further, since the cluster distribution 32 is a result of classification based on the similarity of the entire waveform, it does not hinder intuitive understanding.

  FIG. 5 shows another display example of the cluster distribution 32. FIG. 5A is a schematic diagram of a composite image in which the cluster distribution 32 is superimposed on the Broadman brain map. Broadman brain map is a common name for the anatomy and cytostructural division of the cerebral neocortex by Corbinian Broadman. In Broadman's original text, neurons in the cerebral cortex tissue are stained and visualized (in order to limit the patent drawing, it is converted to gray scale in FIG. 5), and a part of the tissue structure is uniform. Numbers from 1 to 52 are assigned as areas.

  FIG. 5A shows only the clustering results for some measurement channels 23 (CH28, CH29, CH38, CH39, CH49) in order to ensure readability. As can be seen from FIG. 5A, the boundary of the brain function area between 10 fields and 46 + 47 fields in the Broadman brain map is clearly expressed as the cluster distribution 32. In the case of the composite image shown in FIG. 5A, the clustering result can be confirmed in comparison with the anatomy / cytostructural division.

  FIG. 5B is a schematic diagram of a composite image in which the cluster distribution 32 is superimposed on an average brain image generated based on an unspecified number of brain images (such as MRI images). In FIG. 5B, the small area corresponding to the measurement channel 23 is deformed according to the shape of the average brain image, and the deformed small area is superimposed on the average brain image. The composite image in FIG. 5B can be rotated. That is, in FIG. 5B, the left head is the front, but when the user gives an instruction via the input unit 16, the right head can be the front. The composite image shown in FIG. 5B can be generated without the subject's 19 MRI image, and the features in the brain function area can be confirmed more intuitively.

  FIG. 5C is a schematic diagram of a composite image in which the cluster distribution 32 is superimposed on the MRI image of the subject 19. The composite image in FIG. 5C can also be rotated. With the composite image shown in FIG. 5C, the feature in the brain function area can be confirmed more accurately.

  Returning to the description of FIG. The control unit 13 generates a cluster waveform 33 for each cluster (S3). The cluster waveform 33 is a waveform representing a cluster, and is, for example, a graph of an average value at each sampling point of the hemoglobin waveform 31 included in each cluster. Further, the cluster waveform 33 may be, for example, a graph showing the median value, mode value, minimum value, maximum value, mode value, etc. at each sampling point of the hemoglobin waveform 31 included in each cluster. Further, the cluster waveform 33 is not limited to one, but may be a graph of a plurality of reference values (for example, minimum value, median value, maximum value, etc.).

  FIG. 4 shows a first cluster waveform 33ax and a second cluster waveform 33bx. As shown in FIG. 4, since there are as many cluster waveforms as the number of clusters, in order to distinguish these, “a” is used for the element code relating to the first cluster, and “b” is used for the element code relating to the second cluster. ". Further, when these are collectively referred to without being distinguished, “a” and “b” are not added.

  As shown in FIG. 4, in order to display all the hemoglobin waveforms 31 corresponding to the number of measurement channels 23, the display size of each waveform must be reduced, and the visibility is deteriorated. Further, it is more efficient to read the cluster waveforms 33 corresponding to the number of clusters than to read the hemoglobin waveforms 31 corresponding to the number of measurement channels 23.

  Returning to the description of FIG. The control unit 13 calculates a cluster feature 38 for each cluster (S4). The cluster feature amount 38 is a feature amount of the cluster waveform 33. Examples of the waveform feature amount include an integral value, an average value, a standard deviation, a centroid value, and an inclination value. The integral value is a value obtained by integrating the feature amount calculation section 37. The average value is an average value of measurement values in the feature amount calculation section 37. The standard deviation is an average value of measured values in the feature amount calculation section 37. The center-of-gravity value is the time during which the area of the measured value in the positive direction is halved throughout the entire period from the measurement start time to the measurement end time. The inclination value is an inclination of a predetermined time (for example, 5 seconds) from the start time of the feature amount calculation section 37. In the case where the feature amount calculation section 37 is an inspection task execution section, the integral value is the magnitude of the response in the inspection task, the centroid value is the reaction timing when viewed through the entire inspection, and the slope value is the reaction speed (initial activation) Is shown.

  Further, the control unit 13 stores the cluster distribution 32, the cluster waveform 33, and the cluster feature amount 38 in the storage unit 14 for each identification information 30 (S5). Then, the control unit 13 displays the cluster distribution 32, the cluster waveform 33, and the cluster feature amount 38 on the display unit 15 (S6). Examples of display modes include single display, comparison display, and difference display. In the single display, a single measurement result (a hemoglobin waveform 31 of a plurality of measurement channels 23) can be confirmed in detail and efficiently. In the comparison display and the difference display, it is possible to confirm in detail and efficiently a difference between a plurality of measurement results due to a temporal change due to treatment, a disease, or the like.

<Example of single display>
For example, the control unit 13 displays information shown in FIG. 4 as a single display. Further, the control unit 13 may display the cluster distribution 32x of FIG. 4 in place of the composite image shown in FIGS.

<First example of comparison display>
For example, the control unit 13 simultaneously displays the information shown in FIGS. 4 and 6 as a comparison display. The information shown in FIG. 4 is the current measurement result, and the information shown in FIG. 6 is the previous measurement result. In FIG. 6, a measurement ID “1001” is assigned to the measurement result of the subject “S” and the measurement start time “April 1, 2011 13:05:05”.

  As described above, the control unit 13 displays the cluster distribution 32 for a plurality of different pieces of identification information 30. Thereby, for example, when some treatment is given to the subject “S” within the period of both, the effect of the treatment can be easily confirmed.

<First example of comparison display + difference display>
Moreover, the control part 13 displays the information shown, for example in FIG. 7 as a comparison display and a difference display. FIG. 7 shows a hemoglobin waveform comparison display 34 and a hemoglobin waveform difference display 35. The hemoglobin waveform comparison display 34 is obtained by superimposing the hemoglobin waveform 31 x and the hemoglobin waveform 31 y for each measurement channel 23. The hemoglobin waveform difference display 35 is a waveform obtained by subtracting the measured value of the hemoglobin waveform 31y from the measured value of the hemoglobin waveform 31x at each sampling point. When the difference display 35 of the hemoglobin waveform is viewed, it can be clearly confirmed that the current measurement result is larger as the hemoglobin change.

<Second example of comparison display + difference display>
Moreover, the control part 13 displays the information shown, for example in FIG. 8 as a comparison display and a difference display. FIG. 8 shows a current cluster distribution 32x, a previous cluster distribution 32y, and a cluster distribution difference display 36.

  The cluster distribution difference display 36 is a difference between the current cluster distribution 32x and the previous cluster distribution 32y. In FIG. 8, the current cluster distribution 32x and the previous cluster distribution 32y are compared, and the measurement channel 23 having a difference in the clustering result is shown as a small circle with “■” inside. Looking at the difference display 36 of the cluster distribution, it can be seen that there is almost no difference between the current cluster distribution 32x and the previous cluster distribution 32y, and there is little change in the functional area between the current and previous times.

  In this way, the control unit 13 displays the cluster distribution 32 for a plurality of different pieces of identification information 30, or displays the difference between the cluster distributions 32 in two different pieces of identification information 30. Thereby, for example, when some treatment is given to the subject “S” within the period of both, the effect of the treatment can be easily confirmed.

<Third example of comparison display + difference display>
Moreover, the control part 13 displays the information shown, for example in FIG. 9 as a comparison display and a difference display. FIG. 9 shows a feature amount calculation section 37, a current first cluster waveform 33ax, a current second cluster waveform 33bx, a current first cluster feature value 38ax, a current second cluster feature value 38bx, and a previous first cluster. A waveform 33ay, a previous second cluster waveform 33by, a previous first cluster feature value 38ay, a previous second cluster feature value 38by, a first cluster feature value difference 39a, and a second cluster feature value difference 39b are shown. Yes.

  In the feature amount calculation section 37, the user can input an arbitrary value via the input unit 16. The current first cluster feature value 38ax is a feature value (integral value, average value, standard deviation, centroid value, inclination value) related to the first cluster of the current measurement result. The same applies to the current second cluster feature value 38bx, the previous first cluster feature value 38ay, and the previous second cluster feature value 38by. In addition, the specific numerical value shown in FIG. 9 is calculated from the value before normalization (z score formation).

  The first cluster feature value difference 39a is a value obtained by subtracting the previous first cluster feature value 38ay from the current first cluster feature value 38ax. The same applies to the difference 39b of the second cluster feature amount.

  In this way, the control unit 13 displays the cluster feature value 38 for a plurality of different pieces of identification information 30, or displays the difference between the cluster feature values 38 in two different pieces of identification information 30. Thereby, for example, when some kind of treatment is given to the subject “S” within both periods, the effect of the treatment can be quantitatively confirmed.

<Second example of comparison display>
Moreover, the control part 13 displays the information shown in FIG. 10, for example as a comparison display. FIG. 10 includes a feature amount calculation section 37, a current first cluster waveform 33ax, a current second cluster waveform 33bx, a current first cluster feature value 38ax, a current second cluster feature value 38bx, and a previous first cluster. A waveform 33ay, a previous second cluster waveform 33by, a previous first cluster feature value 38ay, a previous second cluster feature value 38by, and a cluster feature value scatter diagram 40 are shown.

  The cluster feature amount scatter diagram 40 is a scatter diagram in which the horizontal axis represents an integral value that is one of the cluster feature amounts 38 and the vertical axis represents the centroid value that is one of the cluster feature amounts 38. In the cluster feature amount scatter diagram 40 shown in FIG. 10, four points of the current first cluster, the current second cluster, the previous first cluster, and the previous second cluster are shown. Looking at the cluster feature amount scatter diagram 40, it is possible to intuitively understand the quantitative difference of the cluster feature amount 38.

<Modification>
In the above description, the cluster classification process (S2) for the same measurement result is performed once, but the present invention is not limited to this example. Hereinafter, as a modification, an example in which the cluster classification process (S2) for the same measurement result is performed a plurality of times will be described.

  FIG. 11 exemplarily shows a result of performing cluster classification processing three times with the number of clusters being 2, 3, and 4 for the same measurement result (measurement ID “1002”). The control unit 13 of the biological light measurement apparatus 1 performs cluster classification processing by the k-average method with the number of clusters set to 2, 3, and 4 for the same measurement result, and each cluster distribution 32, cluster waveform 33, the cluster feature 38 is calculated.

  In FIG. 11, three types of cluster distributions 32 are shown. In order to distinguish between them, “1” is added to the code of the element related to the cluster number of 2, “m” is added to the code of the element related to the cluster number of 3, and “n” is added to the code of the element related to the cluster number of 4. ing. Although not shown in FIG. 11, the control unit 13 may compare and display the cluster waveform 33 and the cluster feature 38 on the display unit 15.

  As described above, the control unit 13 performs the classification process a plurality of times with different numbers of clusters, and displays the cluster distribution 32, the cluster waveform 33, and the cluster feature amount 38 on the display unit 15 for the different numbers of clusters. This eliminates the need for the user to determine the number of clusters in advance before performing cluster classification processing. Then, the user can select the classification result of the optimum number of clusters and display detailed information about the selected classification result.

  Furthermore, the control unit 13 may automatically select the classification result of the optimal number of clusters after performing the classification process a plurality of times with different numbers of clusters. Two examples of this selection process will be described.

  For example, the control unit 13 recursively executes the k-average method from k = 2, compares the information criterion (Bayes information criterion, Akaike information criterion, etc.) before and after the cluster division, and the value is Process until no improvement. Then, the control unit 13 selects the classification result immediately before the classification result whose value is no longer improved as the classification result of the optimum number of clusters.

  Further, for example, the control unit 13 recursively executes the k average method from k = 2 to execute the cluster feature value 38, calculates the distance between the clusters before and after the cluster division, and is equal to or less than the threshold value. Processing is performed until the next cluster comes out. Then, the control unit 13 selects the classification result immediately before the classification result in which the space between the clusters that is equal to or less than the threshold appears as the classification result of the optimal number of clusters. Here, the distance between clusters is, for example, an integrated value, which is one of the cluster feature values 38, and one of the cluster feature values 38, as in the cluster feature value scatter diagram 40 shown in FIG. The Euclidean distance when the centroid value is the y coordinate.

  As described above, the control unit 13 selects the classification result of the optimum number of clusters based on the cluster feature value 38 from the classification results of a plurality of times. As a result, the user does not need to perform an operation of selecting the classification result of the optimal number of clusters. Then, the control unit 13 can display detailed information on the classification result of the optimal number of clusters on the display unit 15.

  The preferred embodiments of the biological light measurement device and the like according to the present invention have been described above with reference to the accompanying drawings, but the present invention is not limited to such examples. It will be apparent to those skilled in the art that various changes or modifications can be conceived within the scope of the technical idea disclosed in the present application, and these naturally belong to the technical scope of the present invention. Understood.

DESCRIPTION OF SYMBOLS 1 ......... Biometric optical measuring device 2 ......... Device main-body part 3 ......... Installation part 4 ......... Transmission optical fiber 5 ......... Reception optical fiber 6 ......... Measurement part 7 ......... Computer 8 ... ...... Laser diode 9 ...... Detector 10 ...... Lock-in amplifier 11 ...... AD conversion 12 ...... Digital board 13 ...... Control unit 14 ...... Storage unit 15 ...... Display unit 16 …… ... Input unit 19 ......... Subject 21 ......... Irradiation probe 22 ......... Detection probe 23 ......... Measurement channel 30 ......... Identification information 31 ......... Hemoglobin waveform 32 ......... Cluster distribution 33 ......... Cluster waveform 34 ......... Hemoglobin waveform comparison display 35 ......... Hemoglobin waveform difference display 36 ......... Cluster distribution difference display 37 ......... Feature amount calculation section 38 ......... Cluster feature amount 39 ......... Cluster The difference 40 ......... cluster feature quantities scatter diagram butterfly amount

Claims (6)

An irradiation probe that irradiates the subject with near-infrared light, and a detection probe that detects the near-infrared light that has passed through the inside of the subject are disposed, and a mounting portion that is mounted on the subject; A biological light measurement device comprising a device main body for measuring a hemoglobin waveform for each measurement channel that is a portion between the irradiation probe and the detection probe, and displaying a measurement result,
Classification means for classifying into a plurality of clusters based on the similarity of some or all of the measured hemoglobin waveforms;
Generating means for generating a cluster waveform which is a waveform representative of the cluster;
Along with the cluster waveform, a display means for displaying a cluster distribution which is information indicating an association between the measurement channel and the cluster;
Storage means for storing the cluster waveform and the cluster distribution for each identification information which is information for identifying a measurement result;
Equipped with a,
The display means displays the cluster waveform and the cluster distribution for a plurality of different pieces of identification information, and / or displays a difference between the cluster distributions in two different pieces of identification information.
Living body light measuring device comprising a call. The biological light measurement apparatus according to claim 1, wherein the classification unit classifies the plurality of clusters into a plurality of clusters based on similarity of the entire hemoglobin waveform. A calculation means for calculating a cluster feature value which is a feature value of the cluster waveform;
Further comprising
The storage means stores the cluster feature amount for each identification information,
The display means is different for a plurality of the identification information for displaying the cluster feature amount, and / or different in the two said identification information and displaying the difference between the cluster feature quantity claim 1 or claim Item 3. The biological light measurement device according to Item 2 . The classification means performs a plurality of classification processes with different numbers of clusters,
The generating means generates the cluster waveform for different numbers of clusters,
The calculation means calculates the cluster feature amount for a plurality of different clusters,
The display means is different for a plurality of number of clusters to display the cluster waveform and the cluster distribution, and / or, according to claim 3, wherein the displaying the clusters characteristic quantity for a plurality of different number of clusters Biological light measurement device. Selection means for selecting a classification result of the optimal number of clusters based on the cluster feature quantity from a plurality of classification results in the classification means,
The biological light measurement device according to claim 4 , further comprising: An irradiation probe that irradiates the subject with near-infrared light, and a detection probe that detects the near-infrared light that has passed through the inside of the subject are disposed, and a mounting portion that is mounted on the subject; A waveform analysis method executed by a biological light measurement device configured to measure a hemoglobin waveform for each measurement channel that is a portion between the irradiation probe and the detection probe and display a measurement result. ,
A classification step of classifying into a plurality of clusters based on the similarity of some or all of the measured hemoglobin waveforms;
Generating a cluster waveform that is a waveform representative of the cluster;
Along with the cluster waveform, a display step for displaying a cluster distribution which is information indicating correspondence between the measurement channel and the cluster;
A storage step of storing the cluster waveform and the cluster distribution for each identification information that is information for identifying a measurement result;
It includes,
The display step displays the cluster waveform and the cluster distribution for a plurality of different pieces of identification information, and / or displays a difference between the cluster distributions in two different pieces of identification information.
Waveform analysis wherein a call.