Estimation method, estimation device, determination system, control program, and recording medium

Abstract

The present invention makes it possible to estimate the result of an amyloid PET test on the basis of a brain wave signal of a subject. An estimation method according to one embodiment of the present invention comprises: extracting, from a brain wave signal of a subject, a signal in a first frequency band and a signal in a second frequency band that is higher than the first frequency band; calculating a first matching rate indicating the degree of matching between the phase of the signal in the first frequency band and the phase of a signal obtained by applying a bandpass filter that passes a signal in the first frequency band on the envelope of the signal in the second frequency band; detecting the degree of accumulation of amyloid-β protein in the brain of the subject on the basis of the first matching rate; and estimating, on the basis of the degree of accumulation, whether the subject will be positive or negative in an amyloid PET test.

Description

Estimation method, estimation device, determination system, control program, and recording medium

[0001] The present invention relates to an estimation method, an estimation device, a determination system, a control program, and a recording medium for estimating the results of an amyloid PET test from an electroencephalogram signal of a subject.

[0002] Previous research has shown that amyloid beta protein accumulates in the brains of patients with Alzheimer's disease before the onset of the disease. Methods for examining the accumulation of amyloid beta protein in the brain include cerebrospinal fluid testing for Aβ42, amyloid PET testing, and blood testing.

[0003] Non-Patent Document 1 suggests that there is a relationship between mild cognitive impairment (MCI), diagnosed based on amyloid in cerebrospinal fluid in the brain, and a decrease in PAC (Phase Amplitude Coupling).

[0004] Musaeus, C.S. et al., “Electroencephalographic Cross-Frequency Coupling as a Sign of Disease Progression in Patients With Mild Cognitive Impairment: A Pilot Study”, Frontiers in Neuroscience, 14:790, 2020.

[0005] The above-mentioned cerebrospinal fluid test and amyloid PET test are highly invasive, and the amyloid PET test has problems such as high implementation costs and limited testing facilities. Also, the accuracy of determining the accumulation of amyloid β protein by blood test is unstable.

[0006] Since electroencephalography (EEG) is minimally invasive and relatively inexpensive, it would be of great clinical benefit if the results of amyloid PET tests could be predicted from EEG tests. However, the study described in Non-Patent Document 1 did not predict the results of amyloid PET tests.

[0007] One aspect of the present disclosure has been made in consideration of the above-mentioned problems, and aims to enable estimation of amyloid PET test results based on the electroencephalogram signal of a subject.

[0008] In order to solve the above problem, an estimation method according to one embodiment of the present invention extracts a signal of a first frequency band and a signal of a second frequency band that is a frequency band higher than the first frequency band from an EEG signal of a subject, calculates a first coincidence rate indicating the degree of coincidence between the phase of the signal of the first frequency band and the phase of a signal obtained by applying a band-pass filter that passes the signal of the first frequency band to the envelope of the signal of the second frequency band, detects the degree of accumulation of amyloid beta protein in the brain of the subject based on the first coincidence rate, and estimates whether the subject will test positive or negative in an amyloid PET test based on the degree.

[0009] In order to solve the above problem, an estimation device according to one embodiment of the present invention includes an extraction unit that extracts a signal of a first frequency band and a signal of a second frequency band that is a frequency band higher than the first frequency band from an EEG signal of a subject; a calculation unit that calculates a first coincidence rate that indicates the degree of coincidence between the phase of the signal of the first frequency band and the phase of a signal obtained by applying a band-pass filter that passes the signal of the first frequency band to the envelope of the signal of the second frequency band; a detection unit that detects the degree of accumulation of amyloid beta protein in the brain of the subject based on the first coincidence rate; and an estimation unit that estimates whether the subject will test positive or negative in an amyloid PET test based on the degree.

[0010] According to the present invention, it is possible to estimate the results of an amyloid PET test based on the electroencephalogram signal of a subject.

[0011] 1 is a block diagram illustrating an example of the configuration of an estimation system. It is an example of a diagram for explaining PAC processing. It is a flowchart illustrating an example of the flow of processing performed by an estimation device. It shows an example of the results of an amyloid PET test performed on a total of 110 men and women. It is an example of a diagram showing the positions of each electrode of an electroencephalograph attached to the head of a subject from an overhead perspective. It shows an example of the results of calculating PAC values ​​for each of subjects in the positive and negative groups of an amyloid PET test. It shows an example of t-values ​​and p-values ​​obtained from the data corresponding to FIG. 6. It is an example of a diagram illustrating, for each electrode position, PAC values ​​calculated by extracting gamma frequency band signals and theta frequency band signals. It is an example of a diagram illustrating, for each electrode position, t-values ​​and p-values ​​calculated by extracting gamma frequency band signals and theta frequency band signals. It is an example of a confusion matrix showing the estimation result of whether amyloid PET is positive or negative, and an example of an ROC curve corresponding to the estimation result.

[0012] [First Embodiment] Hereinafter, one embodiment of the present invention will be described in detail.

[0013] (Configuration of estimation system 100) The configuration of an estimation system 100 including an estimation device 1 that can estimate whether a subject will test positive or negative in an amyloid PET test based on the level of amyloid beta protein accumulation in the subject's brain will be described with reference to FIG. 1. FIG. 1 is a block diagram showing an example configuration of an estimation system 100 according to one aspect of the present invention. The estimation system 100 includes an estimation device 1 that acquires and analyzes electroencephalogram (EEG) signals from an electroencephalogram (EEG) 2. The electroencephalogram (EEG) 2 is a measuring instrument that is implanted in the subject's body and measures and acquires the subject's EEG signals. It is desirable that the measured EEG signals be those obtained when the subject is at rest.

[0014] 1 shows a configuration in which one estimation device 1 and a single electroencephalograph 2 are communicatively connected, but the estimation system 100 is not limited to this configuration. For example, the estimation system 100 may be configured to include two or more electroencephalographs 2. Note that in the estimation system 100, the functions of the estimation device 1 may be implemented by one or more computers.

[0015] (Configuration of Estimation Device 1) The configuration of the estimation device 1 will be described with reference to FIG. 1. As shown in FIG. 1, the estimation device 1 includes a control unit 10 and a storage unit 19. The control unit 10 controls each process in the estimation device 1. The control unit 10 can be realized by a logic circuit (hardware) formed on an integrated circuit (IC chip) or the like, or can be realized by software. When realized by software, the control unit 10 may be configured, for example, by a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), or a combination thereof. As shown in FIG. 1, the control unit 10 may include an acquisition unit 11, an extraction unit 12, a calculation unit 13, a detection unit 14, an estimation unit 15, and an output unit 16.

[0016] The acquisition unit 11 acquires the electroencephalogram signal of the subject from the electroencephalograph 2 .

[0017] The extraction unit 12 extracts, from the subject's EEG signal, a signal in a first frequency band and a signal in a second frequency band that is a frequency band higher than the first frequency band. In one aspect, the first frequency band may be a theta frequency band of 4 to 8 Hz, and the second frequency band may be a gamma frequency band of 70 to 100 Hz. This contributes to more accurate estimation, as described below.

[0018] The calculation unit 13 calculates a first coincidence rate indicating the degree of coincidence between the phase of the signal in the first frequency band and the phase of the signal obtained by applying a band-pass filter that passes the signal in the first frequency band to the envelope of the signal in the second frequency band. The calculation unit 13 may also calculate the phase of each signal by applying a Hilbert transform to the signal.

[0019] The detection unit 14 detects the degree of amyloid β-protein accumulation in the subject's brain based on the first match rate. The detection unit 14 may detect the degree of amyloid β-protein accumulation based only on the first match rate obtained from EEG signals at a specific location in the subject's brain. Furthermore, for example, the detection unit 14 may detect the degree of accumulation by referring to information previously stored in the memory unit 19, the information indicating a correspondence between the first match rate and the degree of amyloid β-protein accumulation in the brain. Here, the information indicating the correspondence may be configured to be specified for each attribute of the subject, such as gender or age. Furthermore, the first match rate may be configured to be specified for each location in the brain. The same applies to the second match rate, etc., described below.

[0020] The estimation unit 15 estimates whether the subject will test positive or negative in the amyloid PET test based on the degree of accumulation. For example, the estimation unit 15 may be configured to estimate that the subject will test positive in the amyloid PET test if the degree of accumulation is equal to or greater than a predetermined value. In one aspect, the processing of the extraction unit 12, the calculation unit 13, the detection unit 14, and the estimation unit 15 corresponds to PAC (Phase Amplitude Coupling) processing, which will be described later.

[0021] The output unit 16 outputs the estimation result by the estimation unit 15. Note that, when the first coincidence rate exceeds a predetermined threshold value, or when the estimation unit 15 estimates that the subject will test positive in the amyloid PET test, the output unit 16 may be configured to notify this to an external device or the like other than the estimation device 1.

[0022] The storage unit 19 is a storage device that at least temporarily stores software and various data used by the estimation device 1. As the storage unit 19, for example, a random access memory (RAM), a hard disk drive (HDD), a solid-state drive (SSD), a secure digital (SD) card, or an embedded multi-media controller (eMMC) can be applied.

[0023] (Configuration of EEG meter 2) The EEG meter 2 includes electrodes 21 and acquires the patient's electroencephalogram (EEG) as electroencephalogram signals via the electrodes 21. The EEG meter 2 preferably includes a plurality of electrodes 21 and acquires EEG signals at a plurality of positions in the brain. From another perspective, the acquisition unit 11 of the estimation device 1 preferably acquires EEG signals at a plurality of positions in the brain. Furthermore, the EEG meter 2 is preferably an implantable EEG meter that can be implanted in the patient's body. By using an implantable EEG meter, EEG signals can be acquired with high spatial resolution.

[0024] <PAC Processing> The PAC processing performed by the estimation device 1 will be described with reference to Fig. 2. Fig. 2 is an example diagram for explaining the PAC processing.

[0025] Graph 31 in Fig. 2 shows the electroencephalogram signal of the subject acquired by the acquisition unit 11. The horizontal axis of graph 31 and graphs 32 to 37 described below represents time (milliseconds). The information shown in graphs 32 to 38 in Fig. 2 is based on the electroencephalogram signal of the subject shown in graph 31. The output unit 16 may be configured to output the information exemplified in graphs 31 to 38 as soon as it is obtained.

[0026] Graph 32 shows a signal obtained by applying a high-frequency band-pass filter to the EEG signal, and graph 35 shows a signal obtained by applying a low-frequency band-pass filter to the EEG signal. As in this example, the extraction unit 12 applies different band-pass filters to the EEG signal shown in graph 31, and extracts the signals shown in graphs 32 and 35.

[0027] Graph 33 is an envelope obtained from the signal waveform of graph 32, and shows the amplitude of the signal in graph 32. Graph 36 shows the phase extracted by applying a Hilbert transform to the signal obtained by further applying a low-frequency band-pass filter to the envelope of graph 33.

[0028] Graph 34 shows the results of applying a fast Fourier transform to the subject's electroencephalogram in graph 31 to calculate the power spectrum density.

[0029] Graph 37 shows the phase extracted by applying the Hilbert transform to the signal of graph 35. In one aspect, the extraction unit 12 extracts the phases of graphs 36 and 37, but the calculation unit 13 may also extract the phases.

[0030] Formula 40 is a formula for calculating a PAC index value indicating the degree of agreement between the phase of graph 36 and the phase of graph 37 as a first agreement rate. x -U x ) indicates the phase difference between the phase L of graph 37 and the phase U of graph 36. The first coincidence rate calculated by the calculation unit 13 using calculation formula 40 indicates the degree of synchronization between the graphs, and is a value between 0 and 1. This first coincidence rate can also be said to be a value evaluating the coupling between the phase of the signal in the first frequency band corresponding to graph 37 and the amplitude of the signal in the second frequency band corresponding to graph 33. If there is a statistical relationship between the phase of graph 36 and the phase of graph 37, the average phase difference will not be 0, and the first coincidence rate will be a value greater than 0. Furthermore, graph 38 illustrates an example in which the average phase difference is 90 degrees and the value of the first coincidence rate is 0.367.

[0031] In relation to the calculation formula 40, the first coincidence rate is calculated by using a variable n indicating a time corresponding to a value, and calculating the phase of the waveform of the first frequency band at time n as φ ω (n), and the phase of the signal in the first frequency band obtained from the envelope of the signal in the second frequency band at time n is φ γω When (n), it can also be expressed as SI in the following equation (1).

[0032] In addition, the calculation unit 13 is not limited to a configuration in which the PAC index value is calculated as the first matching rate, but may also be configured to calculate and use a matching rate obtained using other frequency coupling methods as the first matching rate.

[0033] (Processing Performed by Estimation Apparatus 1) Next, the flow of processing performed by the estimation apparatus 1 will be described with reference to Fig. 3. Fig. 3 is a flowchart showing an example of the flow of processing performed by the estimation apparatus 1.

[0034] In step S101 (acquisition step), the acquisition unit 11 acquires an electroencephalogram signal of the subject from the electroencephalograph 2 .

[0035] In step S102 (first extraction step), the extraction unit 12 extracts a signal obtained by applying a high-frequency band pass filter and a signal obtained by applying a low-frequency band pass filter from the acquired electroencephalogram signal.

[0036] In step S103 (second extraction step), the extraction unit 12 extracts a phase for each of the band-pass filtered signals. The extraction unit 12 further applies a low-frequency band-pass filter to the envelope obtained from the signal waveform corresponding to graph 32 in FIG. 2 , which has been filtered through the high-frequency band-pass filter, and then applies a Hilbert transform to extract a phase corresponding to graph 36. The extraction unit 12 also applies a Hilbert transform to the signal corresponding to graph 35, which has been filtered through the low-frequency band-pass filter, to extract a phase corresponding to graph 37.

[0037] In step S104 (calculation step), the calculation unit 13 uses calculation formula 40 to calculate a first match rate for each of the extracted phases.

[0038] In step S105 (detection step), the detection unit 14 detects the degree of accumulation of amyloid β protein in the subject's brain based on the first coincidence rate.

[0039] In step S106 (estimation step), the estimation unit 15 estimates whether the subject will test positive or negative in the amyloid PET test, based on the degree of accumulation.

[0040] In step S107 (output step), the output unit 16 outputs the estimation result to an output device (not shown). The output device may be, for example, a display device having a display.

[0041] [Variant Example] Instead of the processing of steps S104 and S105 described above, a configuration may be adopted in which step S104' of calculating the substantial match rate and step S105' of detecting the degree of accumulation of amyloid beta protein based on the substantial match rate are respectively executed as follows.

[0042] (Step S104') In step S104', the calculation unit 13 first calculates a first coincidence rate (SI) for each of the extracted phases.

[0043] The calculation unit 13 also calculates a phase shuffle coincidence rate (SI'), which is a second coincidence rate, by dividing the first frequency band signal extracted from the EEG signal into multiple time segments and randomly shuffling the order of the segments to obtain a waveform, and then coupling the phase of the waveform with the amplitude of the second frequency band signal. For example, the calculation unit 13 divides the first frequency band signal extracted from the EEG signal sampled at a predetermined sample rate (e.g., 1000 Hz) into one-second data segments, and obtains a signal by shuffling the data before and after the one-second EEG signal data at a randomly selected time. The calculation unit 13 then calculates the phase shuffle coincidence rate (SI') by coupling the phase of the signal obtained by shuffling the data with the amplitude of the second frequency band signal. In other words, the calculation unit 13 calculates a phase shuffle matching rate, which is a second matching rate indicating the degree of matching between the phase of a signal obtained by temporally swapping the phase of the waveform of a first frequency band signal before and after an arbitrary time boundary, and the phase of a signal of the first frequency band obtained from the envelope of the signal of the second frequency band.

[0044] Next, the calculation unit 13 calculates the substantial match rate (SI'') by subtracting the phase shuffle match rate (SI') from the first match rate (SI).

[0045] (Step S105') In step S105', the detection unit 14 detects the degree of accumulation of amyloid β protein in the brain of the subject based on the substantial match rate.

[0046] According to the configuration of this modified example, depending on the physiological state of the subject, the degree of accumulation of amyloid β protein may be detected with higher accuracy than detection based on the first coincidence rate.

[0047] [Embodiment 2] The functions of the estimation device 1 (hereinafter referred to as the "device") can be realized by a program that causes a computer to function as the device, and a program that causes a computer to function as each control block of the device (particularly each unit included in the control unit 10).

[0048] In this case, the device includes a computer having at least one control device (e.g., a processor) and at least one storage device (e.g., a memory) as hardware for executing the program. The functions described in each of the above embodiments are realized by executing the program using the control device and storage device.

[0049] The program may be non-transitory and may be recorded on one or more computer-readable recording media. The recording media may or may not be included in the device. In the latter case, the program may be supplied to the device via any wired or wireless transmission medium.

[0050] Furthermore, some or all of the functions of the control blocks can be realized by logic circuits. For example, an integrated circuit in which a logic circuit that functions as each of the control blocks is formed is also included in the scope of the present invention. In addition, the functions of the control blocks can also be realized by, for example, a quantum computer.

[0051] Furthermore, each process described in each of the above embodiments may be executed by AI (Artificial Intelligence). In this case, the AI ​​may run on the control device or on another device (for example, an edge computer or a cloud server).

[0052] The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope of the claims. Embodiments obtained by appropriately combining the technical means disclosed in different embodiments are also included in the technical scope of the present invention.

[0053] [Summary] The estimation method according to aspect 1 of the present invention is a method of extracting, from an EEG signal of a subject, a signal of a first frequency band and a signal of a second frequency band that is a frequency band higher than the first frequency band, calculating a first coincidence rate indicating the degree of coincidence between the phase of the signal of the first frequency band and the phase of a signal obtained by applying a band-pass filter that passes the signal of the first frequency band to the envelope of the signal of the second frequency band, detecting the degree of accumulation of amyloid beta protein in the brain of the subject based on the first coincidence rate, and estimating whether the subject will test positive or negative in an amyloid PET test based on the degree.

[0054] An estimation method according to a second aspect of the present invention may be the method according to the first aspect, wherein the first frequency band is a theta frequency band and the second frequency band is a gamma frequency band.

[0055] An estimation method according to a third aspect of the present invention may be a method in the second aspect, in which the first frequency band is 4 to 8 Hz.

[0056] An estimation method according to a fourth aspect of the present invention may be a method in the second aspect, in which the second frequency band is 70 to 100 Hz.

[0057] An estimation method according to a fifth aspect of the present invention may be a method in any one of the first to fourth aspects described above, in which the first coincidence rate is calculated as SI in the following formula (1), where φω(n) is the phase of the waveform of the first frequency band at time n and φγω(n) is the phase of the signal of the first frequency band obtained from the envelope of the signal of the second frequency band at time n:

[0058] An estimation method according to a sixth aspect of the present invention may be a method of detecting the degree of accumulation based on a second coincidence rate obtained by subtracting the second coincidence rate from the first coincidence rate, the second coincidence rate being a second coincidence rate indicating the degree of coincidence between the phase of a signal obtained by temporally swapping the phase of a waveform before and after an arbitrary time point, and the phase of the signal in the first frequency band obtained from an envelope of the signal in the second frequency band, the method comprising: calculating a phase shuffle coincidence rate;

[0059] The estimation method according to aspect 7 of the present invention may be a method for notifying an external party that the first matching rate has exceeded a predetermined threshold value in any of aspects 1 to 6 above.

[0060] An estimation device according to aspect 8 of the present invention comprises an extraction unit that extracts a signal of a first frequency band and a signal of a second frequency band that is a higher frequency band than the first frequency band from an electroencephalogram signal of a subject; a calculation unit that calculates a first coincidence rate that indicates the degree of coincidence between the phase of the signal of the first frequency band and the phase of a signal obtained by applying a band-pass filter that passes the signal of the first frequency band to the envelope of the signal of the second frequency band; a detection unit that detects the degree of accumulation of amyloid beta protein in the brain of the subject based on the first coincidence rate; and an estimation unit that estimates whether the subject will test positive or negative in an amyloid PET test based on the degree.

[0061] A determination system according to aspect 9 of the present invention may be configured to include an electroencephalograph that is implanted in the subject's body and acquires the electroencephalogram signal, and the estimation device described in aspect 8 above.

[0062] A control program according to aspect 10 of the present invention is a control program for causing a computer to function as the estimation device described in aspect 8 above, and may be configured to cause a computer to function as the extraction unit, the calculation unit, the detection unit, and the estimation unit.

[0063] A recording medium according to an eleventh aspect of the present invention is a computer-readable recording medium on which the control program according to the tenth aspect is recorded.

[0064] The present invention will be further explained below with reference to examples shown in FIGS. 4 to 10, but the present invention is not limited to these.

[0065] Figure 4 shows an example of the results of an amyloid PET test performed on a total of 110 people, 42 men and 68 women. Figure 4 shows that of these 110 people, 16 men and 32 women, totaling 48 people, tested positive in the amyloid PET test, and 26 men and 36 women, totaling 62 people, tested negative. The "P-value" column in Figure 4 shows the P value obtained by testing the number of people in each group using Student's t-test.

[0066] At the time of this amyloid PET test, no significant difference was observed between positive and negative amyloid PET test results, regardless of age, gender, or results of the cognitive function test (MMSE: Mini Mental State Examination). Subjects with a positive amyloid PET test result were presumed to have Alzheimer's disease. On the other hand, subjects with a negative result were suspected of having a dementia other than Alzheimer's disease.

[0067] These 110 people were subjects, and electroencephalogram tests and calculation of PAC index values ​​were performed using a system similar to estimation system 100 as follows.

[0068] Figure 5 is an example of a diagram showing the positions of the electrodes of an electroencephalograph attached to the subject's head from an overhead perspective, with the top of the page corresponding to the front of the face.

[0069] Figure 6 shows an example of the results of calculating PAC values ​​for each of the subjects in the amyloid PET test positive and negative groups. In Figure 6 and Figure 7 described below, the horizontal axis represents four types of EEG signals in the first frequency band. The gamma frequency band of 70 to 100 Hz is commonly used as the second frequency band. The vertical axis corresponds to each position of the electrodes shown in Figure 5.

[0070] 6 shows the PAC values ​​of subjects in the amyloid PET test positive group using colors corresponding to the values ​​on gauge 511. For example, area 519 of graph 51 shows that the PAC value calculated from the gamma frequency band signal and the beta frequency band signal extracted from the EEG signal at position "T4" near the subject's right ear is approximately 0.06. Graph 52 shows the PAC values ​​of subjects in the amyloid PET test negative group using colors corresponding to the values ​​on gauge 521.

[0071] FIG. 7 shows an example of t-values ​​and p-values ​​obtained from data corresponding to FIG. 6 . Specifically, graph 53 in FIG. 7 shows t-values ​​obtained by performing a t-test on data corresponding to graphs 51 and 52 in FIG. 6 , using colors corresponding to each value in gauge 531. In graph 53, as shown in area 539, the t-values ​​corresponding to each PAC calculated from gamma frequency band signals and theta frequency band signals extracted from EEG signals at position "O2" near the back of the head of subjects in the amyloid PET test positive and negative groups were particularly high. This means that there is a particularly significant difference between the PAC values ​​calculated by extracting gamma frequency band signals and theta frequency band signals from the EEG signals at position "O2" of the subjects in the positive group and the negative group.

[0072] Graph 54 also shows the p-values ​​obtained from the t-values ​​shown in graph 53 using colors corresponding to the values ​​on gauge 541. In graph 54, the p-value shown in region 549 corresponding to region 539 was particularly low. This means that the PAC corresponding to region 539 fluctuates significantly depending on whether the patient is amyloid positive or negative.

[0073] In other words, it was shown that by referring to the PAC value calculated by extracting gamma frequency band signals and theta frequency band signals from the subject's EEG signal, particularly at position "O2," it is possible to estimate whether the subject will test positive or negative in an amyloid PET test.

[0074] In distribution chart 55 of Fig. 8, for subjects in the positive group in the amyloid PET test, PAC values ​​calculated by extracting gamma frequency band signals and theta frequency band signals, i.e., the PAC values ​​shown in the "Theta" column of graph 51, are illustrated for each electrode position shown in Fig. 5. In addition, in distribution chart 56, for subjects in the negative group, PAC values ​​calculated by extracting gamma frequency band signals and theta frequency band signals, i.e., the PAC values ​​shown in the "Theta" column of graph 52, are illustrated for each electrode position shown in Fig. 5.

[0075] In distribution diagram 57 of Fig. 9, t values ​​calculated by extracting gamma frequency band signals and theta frequency band signals for the positive group subjects and the negative group subjects, i.e., the t values ​​shown in the "Theta" column in graph 53, are shown for each electrode position shown in Fig. 5. In distribution diagram 58, p values ​​obtained from the t values ​​shown in distribution diagram 57, i.e., the p values ​​shown in the "Theta" column in graph 54, are shown for each electrode position shown in Fig. 5.

[0076] Table 59 in Figure 10 is an example of a confusion matrix showing the results of estimating whether amyloid PET was positive or negative, using the 110 subjects shown in Figure 4 as subjects, by referring to the PAC values ​​calculated by extracting gamma frequency band signals and theta frequency band signals. A fully trained support vector machine was used for the estimation. In using the support vector machine, the number of divisions (fold) was set to 10, the penalty parameter C, which represents the penalty given for misclassification of data, was set to 10, and the parameter γ used in the kernel function was set to 0.01.

[0077] In Table 59, the horizontal axis represents the estimated results based on the PAC value, and the vertical axis represents the results of the actual amyloid PET test. In Table 59, positives correspond to Label 1, and negatives correspond to Label 0. In Table 59, for a total of 110 samples, there were 28 true positives, 23 false positives, 20 false negatives, and 39 true negatives. The sensitivity, which can be calculated using the formula true positives / (true positives + false negatives) x 100, was 58%.

[0078] The solid line in graph 60 is an example of a receiver operating characteristic (ROC) curve corresponding to the estimation result. In graph 60, the false positive rate (FPR) on the horizontal axis represents the number of false positives / total number of negatives, and the true positive rate (TPR) on the vertical axis represents the number of true positives / total number of positives. In the results shown in graph 60, the area under the curve (AUC) value was 61%.

[0079] REFERENCE SIGNS LIST 1 Estimation device 2 Electroencephalograph 10 Control unit 11 Acquisition unit 12 Extraction unit 13 Calculation unit 14 Detection unit 15 Estimation unit 16 Output unit 19 Storage unit 100 Estimation system

Claims

1. A method for estimating whether the subject will test positive or negative in an amyloid PET test based on the degree of accumulation of amyloid beta protein in the brain, comprising: extracting a first frequency band signal and a second frequency band signal, the second frequency band being a higher frequency band than the first frequency band, calculating a first coincidence rate indicating the degree of coincidence between the phase of the first frequency band signal and the phase of a signal obtained by applying a bandpass filter that passes the first frequency band signal to the envelope of the second frequency band signal, detecting the degree of accumulation of amyloid beta protein in the brain of the subject based on the first coincidence rate, and estimating whether the subject will test positive or negative in an amyloid PET test based on the degree.

2. The estimation method according to claim 1, characterized in that the first frequency band is the theta frequency band and the second frequency band is the gamma frequency band.

3. The estimation method according to claim 2, characterized in that the first frequency band is 4 to 8 Hz.

4. The estimation method according to claim 2, characterized in that the second frequency band is 70 to 100 Hz.

5. The phase of the waveform of the first frequency band at time n is φ ω (n), and the phase of the signal of the first frequency band obtained from the envelope of the signal of the second frequency band at time n is φ γω 2. The method according to claim 1, wherein when n is (n), the first match rate is calculated as SI in the following formula (1):

6. The estimation method described in claim 5, further comprising: calculating a phase shuffle coincidence rate, which is a second coincidence rate indicating the degree of coincidence between the phase of a signal obtained by temporally swapping the phase of the waveform of the first frequency band signal before and after an arbitrary time boundary, and the phase of the first frequency band signal obtained from the envelope of the second frequency band signal; and detecting the degree of accumulation based on an actual coincidence rate obtained by subtracting the second coincidence rate from the first coincidence rate.

7. The estimation method according to claim 1, further comprising the step of notifying an external device that the first matching rate has exceeded a predetermined threshold value.

8. An estimation device comprising: an extraction unit that extracts a signal of a first frequency band and a signal of a second frequency band that is a higher frequency band than the first frequency band from the electroencephalogram signal of a subject; a calculation unit that calculates a first coincidence rate indicating the degree of coincidence between the phase of the signal of the first frequency band and the phase of a signal obtained by applying a bandpass filter that passes the signal of the first frequency band to the envelope of the signal of the second frequency band; a detection unit that detects the degree of accumulation of amyloid beta protein in the brain of the subject based on the first coincidence rate; and an estimation unit that estimates whether the subject will test positive or negative in an amyloid PET test based on the degree.

9. A determination system comprising: an electroencephalograph that is implanted in the subject's body and acquires the electroencephalogram signal; and the estimation device according to claim 8.

10. A control program for causing a computer to function as the estimation device according to claim 8, the control program causing a computer to function as the extraction unit, the calculation unit, the detection unit and the estimation unit.

11. A computer-readable recording medium having the control program according to claim 10 recorded thereon.