Method, device and program for setting analysis conditions for brain waves, magnetoencephalogram or near-infrared spectroscopy, method, device and program for estimating activity region or intracerebral network in brain, and method and device for applying stimulus to estimated activity region or subject

Abstract

Provided is an estimation method for estimating an activity region or an intracerebral network in the brain of a subject, on the basis of brain waves, a magnetoencephalogram, or near-infrared spectroscopy. The estimation method includes: a second acquisition step S11 for acquiring the brain waves, a magnetoencephalogram, or near-infrared spectroscopy data of the subject; and a second estimation step S12 for analyzing the brain waves, the magnetoencephalogram, or the near-infrared spectroscopy data acquired in the second acquisition step S11 and thereby estimating an activity region or an intracerebral network in the brain of the subject.

Description

Method, device and program for setting analysis conditions for electroencephalography, magnetoencephalography or near-infrared spectroscopy, method, device and program for estimating brain active areas or brain networks, and method and device for applying stimulation to estimated active areas or subjects

[0001] The present invention relates to a technique for estimating brain active regions or brain networks.

[0002] The following non-patent literature 1 and 2 report that when a stimulus is applied to an aphasic patient, the active areas in the brain are measured using fMRI images, and aphasia is treated by applying electrical stimulation to the active areas.

[0003] Fridriksson et al,,”Transcranial Direct Current Stimulation vs Sham Stimulation to Treat Aphasia After Stroke”, JAMA Neurol. 2018;75(12):1470-1476. doi:10.1001 / jamaneurol.2018.2287, Internet〈URL: https: / / jamanetwork.com / journals / jamaneurology / fullarticle / 2696529〉Baker et al., “Using Transcranial Direct-Current Stimulation to Treat Stroke Patients With Aphasia”, Stroke. 2010;41:1229-1236, Internet〈URL: https: / / www.ahajournals.org / doi / 10.1161 / STROKEAHA.109.576785〉Illman et al., “Comparing MEG and EEG in detecting the ~20-Hz rhythm modulation to tactile and proprioceptive stimulation,” Neuroimage 215, 116804, Internet〈URL: https: / / doi.org / 10.1016 / J.NEUROIMAGE.2020.116804〉Legget et al., “MEG and EEG demonstrate similar test-retest reliability of the 40 Hz auditory steady-state response,” Int. J. Psychophysiol. 114, 16, Internet〈URL: https: / / doi.org / 10.1016 / J.IJPSYCHO.2017.01.013〉

[0004] However, MRI devices for capturing fMRI images are expensive, and only a limited number of facilities have them installed.

[0005] The present invention has been made in consideration of the above problems, and aims to provide a method for inexpensively and simply estimating active regions or brain networks in the brain.

[0006] In order to solve the above problems, the present invention includes the following aspects. Item 1. An estimation method for estimating active regions or brain networks in a subject's brain based on electroencephalograms, magnetoencephalograms, or near-infrared spectroscopy, comprising: a second acquisition step of acquiring electroencephalogram, magnetoencephalogram, or near-infrared spectroscopy data of the subject; and a second estimation step of analyzing the electroencephalogram, magnetoencephalogram, or near-infrared spectroscopy data acquired in the second acquisition step to estimate active regions or brain networks in the subject's brain. Item 2. The estimation method according to Item 1, wherein the second acquisition step acquires the electroencephalogram, magnetoencephalogram, or near-infrared spectroscopy data when a stimulus is applied to the subject. Item 3. The estimation method according to Item 2, wherein the second estimation step estimates active regions in the brain when the stimulus is applied to the subject. Item 4. Item 4. A stimulation method for providing electrical stimulation, magnetic stimulation, ultrasonic stimulation, optical stimulation, visual stimulation, auditory stimulation, tactile stimulation, or mechanical stimulation to the subject in accordance with the active area or brain network estimated by the estimation method according to any one of items 1 to 3. Item 5. A stimulation method according to item 4, for providing electrical stimulation to the estimated active area. 4. A condition setting method for setting analysis conditions for electroencephalogram (EEG), magnetoencephalogram (MEG), or near-infrared spectroscopy data for estimating an active area or brain network in the brain by the estimation method according to any one of items 1 to 3, the condition setting method comprising: an image acquisition step for acquiring brain images of a subject taken by an MRI device or a PET device; an identification step for identifying an active area or brain network in the brain of the subject based on the brain images; a first acquisition step for acquiring electroencephalogram (EEG), magnetoencephalogram (MEG), or near-infrared spectroscopy data of the subject; a first estimation step for analyzing the electroencephalogram (EEG), magnetoencephalogram (MEG), or near-infrared spectroscopy data acquired in the first acquisition step using a plurality of analysis conditions to estimate an active area or brain network in the brain of the subject for each of the analysis conditions; a comparison step for comparing, for each of the analysis conditions, the active area estimated in the first estimation step with the active area or brain network identified in the identification step; and a condition selection step for selecting one analysis condition from the plurality of analysis conditions based on a comparison result in the comparison step.Item 7. The condition setting method according to Item 6, wherein the first acquisition step acquires the electroencephalogram, magnetoencephalogram, or near-infrared spectroscopy data when a stimulus is applied to the subject. Item 8. The estimation method according to Item 7, wherein the first estimation step estimates an active area in the brain when the stimulus is applied to the subject. Item 9. The condition setting method according to Item 8, wherein the first estimation step performs a signal source estimation analysis on the electroencephalogram during real stimulation and the electroencephalogram during control stimulation, extracts a first principal component or average of a time-series signal, a frequency spectrum signal, or a functional connectivity index using the time-series signal or the frequency spectrum signal for each of pre-divided brain areas, and estimates the active area by taking the difference between the first principal component or average during real stimulation and the first principal component or average during control stimulation, and the analysis conditions include at least one of: whether to perform absolute value processing of the difference; whether to perform smoothing processing on the first principal component or average; and the time of the analysis area. Item 10. Item 11. A condition setting method according to any one of Items 6 to 10, wherein in the first estimation step, a first principal component or an average of a time series signal, a frequency spectrum signal, or a functional connectivity index using the time series signal or the frequency spectrum signal is extracted for signals from each electrode attached to the subject for the EEG during actual stimulation and the EEG during control stimulation, and the active area is estimated by taking the difference between the first principal component or average during actual stimulation and the first principal component or average during control stimulation, and the analysis conditions include at least one of: whether to perform absolute value processing of the difference, whether to perform smoothing processing on the first principal component or average, and the time of the analysis area. Item 12. A condition setting method according to any one of Items 6 to 10, wherein the brain image is an fMRI image taken by the MRI device. An estimation device that estimates an active area or an intracerebral network in a subject's brain based on electroencephalograms, magnetoencephalograms, or near-infrared spectroscopy, comprising: a second acquisition unit that acquires electroencephalogram, magnetoencephalogram, or near-infrared spectroscopy data of the subject; and a second estimation unit that analyzes the electroencephalogram, magnetoencephalogram, or near-infrared spectroscopy data acquired by the second acquisition unit to estimate the active area or the intracerebral network in the subject's brain.Item 13. A stimulation device that provides electrical stimulation, magnetic stimulation, ultrasonic stimulation, optical stimulation, visual stimulation, auditory stimulation, tactile stimulation, or mechanical stimulation to the subject in accordance with the active region or brain network estimated by the estimation device according to Item 12. Item 14. A condition setting device that sets analysis conditions for the electroencephalogram, magnetoencephalogram, or near-infrared spectroscopy data used by the estimation device according to Item 12 to estimate active areas or brain networks in the brain, the condition setting device comprising: an image acquisition unit that acquires brain images of a subject taken by an MRI device or a PET device; an identification unit that identifies active areas or brain networks in the subject's brain based on the brain images; a first acquisition unit that acquires the electroencephalogram, magnetoencephalogram, or near-infrared spectroscopy data of the subject; a first estimation unit that analyzes the electroencephalogram, magnetoencephalogram, or near-infrared spectroscopy data acquired by the first acquisition unit using a plurality of analysis conditions to estimate active areas or brain networks in the subject's brain for each of the analysis conditions; a comparison unit that compares, for each of the analysis conditions, the active areas or brain networks estimated by the first estimation unit with the active areas or brain networks identified by the identification unit; and a condition selection unit that selects one analysis condition from the plurality of analysis conditions based on a comparison result by the comparison unit. An estimation program for estimating an active area or brain network in a subject's brain based on electroencephalograms, magnetoencephalograms, or near-infrared spectroscopy, the estimation program causing a computer to execute the following steps: a second acquisition step of acquiring electroencephalogram, magnetoencephalogram, or near-infrared spectroscopy data of the subject; and a second estimation step of analyzing the electroencephalogram, magnetoencephalogram, or near-infrared spectroscopy data acquired by the second acquisition step to estimate an active area or brain network in the subject's brain.Item 16. A condition setting program for setting analysis conditions for electroencephalogram, magnetoencephalogram, or near-infrared spectroscopy data for estimating active areas or brain networks in the brain by the estimation method according to Item 1, the condition setting program causing a computer to execute the following steps: an image acquisition step for acquiring brain images of a subject taken by an MRI device or a PET device; an identification step for identifying active areas or brain networks in the subject's brain based on the brain images; a first acquisition step for acquiring electroencephalogram, magnetoencephalogram, or near-infrared spectroscopy data of the subject; a first estimation step for analyzing the electroencephalogram, magnetoencephalogram, or near-infrared spectroscopy data acquired in the first acquisition step using a plurality of analysis conditions to estimate active areas or brain networks in the subject's brain for each of the analysis conditions; a comparison step for comparing, for each of the analysis conditions, the active areas or brain networks estimated in the first estimation step with the active areas or brain networks identified in the identification step; and a condition selection step for selecting one analysis condition from the plurality of analysis conditions based on the comparison result in the comparison step. Item 18. A computer-readable recording medium having recorded thereon the estimation program according to Item 15. Item 19. A computer-readable recording medium having recorded thereon the condition setting program according to Item 16.

[0007] According to the present invention, active regions or brain networks can be estimated inexpensively and simply.

[0008] 1 is a block diagram showing the configuration of an analysis and stimulation system; FIG. 2 is a flowchart showing the processing steps of a condition setting method; FIG. 3 is a flowchart showing the processing steps of an estimation method and a stimulation application method; FIG. 4 is a diagram for explaining the content of a language task; FIG. 5 is a graph showing the average reaction speed after stimulation is applied under each condition in a language task; FIG. 6 is a diagram showing FC that is significantly correlated with the average error angle in Example 3; FIG. 7 is a diagram showing FC that is significantly correlated with the detection angle in Example 3.

[0009] Hereinafter, an embodiment of the present invention will be described with reference to the accompanying drawings. Note that the present invention is not limited to the following embodiment, and various modifications are possible without departing from the spirit of the present invention.

[0010] (Overall Configuration) Fig. 1 is a block diagram showing the configuration of an analysis and stimulation system 1 according to one embodiment of the present invention. The analysis and stimulation system 1 is a system for analyzing active areas in the brain of a subject and applying electrical stimulation to the active areas, and includes an MRI device 2, an electroencephalogram (EEG) measuring device 3, an analysis device 4, and a stimulation device 5. In this embodiment, the subject is an aphasia patient, but the present invention is not limited to aphasia patients. The subject may also be a healthy individual.

[0011] The analysis device 4 has the functions of both the condition setting device and the estimation device described in the claims, and can be configured as a general-purpose computer. The analysis device 4 includes, as its hardware configuration, a processor such as a CPU or GPU (not shown), a main memory device such as a DRAM or SRAM (not shown), and an auxiliary memory device 40 such as an HDD or SSD. The auxiliary memory device 40 stores a condition setting program P1, an estimation program P2, etc.

[0012] The auxiliary storage device 40 may be external to the analysis device 4. The analysis device 4 may also be configured with multiple computers.

[0013] The analysis device 4 has, as functional blocks, an image acquisition unit 41, an identification unit 42, a first acquisition unit 43, a first estimation unit 44, a comparison unit 45, a condition selection unit 46, a second acquisition unit 47, and a second estimation unit 48. The image acquisition unit 41, the identification unit 42, the first acquisition unit 43, the first estimation unit 44, the comparison unit 45, and the condition selection unit 46 are realized by the processor of the analysis device 4 reading a condition setting program P1 into a main storage device and executing it. The second acquisition unit 47 and the second estimation unit 48 are realized by the processor of the analysis device 4 reading an estimation program P2 into a main storage device and executing it. The condition setting program P1 and the estimation program P2 may be downloaded to the analysis device 4 via a communication network such as the Internet, or the condition setting program P1 and the estimation program P2 may be recorded on a computer-readable non-transitory recording medium such as a CD-ROM or an SD card and installed on the analysis device 4 via the storage medium.

[0014] (Condition Setting) The function of the analysis device 4 as a condition setting device will be described with reference to FIG. 2. FIG. 2 is a flowchart showing the processing steps of a condition setting method according to this embodiment. This condition setting method is a method for setting analysis conditions for EEG, MEG, or NIR spectroscopy data in order to estimate active brain regions or brain networks using EEG, MEG, or NIR spectroscopy. In this embodiment, this method sets analysis conditions for EEG in order to estimate brain regions to which electrical stimulation should be applied for the treatment of aphasia patients. This condition setting method is also implemented in clinical trials using one or more subjects. The subjects may be aphasia patients or healthy individuals.

[0015] In step S1 (image acquisition step), the image acquisition unit 41 of the analysis device 4 acquires a brain image (fMRI image) of the subject captured by the MRI device 2. In this embodiment, an MRI image of the subject is captured in the MRI device 2 while providing the subject with a visual stimulus (for example, a language task in which the subject is shown an image of an object and asked to name the object). The captured MRI image is transmitted from the MRI device 2 to the analysis device 4 and acquired by the image acquisition unit 41.

[0016] In step S1, brain image data of the subject may be acquired from a database, data held by a third party, etc. For example, if brain image data of the subject measured by a third party is available as open data, the data may be acquired by downloading it.

[0017] In step S2 (identification step), the identification unit 42 identifies the active areas in the brain when the subject is stimulated based on the brain image. This allows the active areas in the brain to be accurately identified. The method for identifying the areas in the brain is not particularly limited, and they may be identified by coordinates or by areas subdivided based on an anatomical atlas such as the Human Connectome Project (HCP, Glasser et al.).

[0018] In step S3 (first acquisition step), the first acquisition unit 43 acquires electroencephalograms obtained when the stimulus is applied to the subject. In this embodiment, the electroencephalogram of the subject is measured by the electroencephalogram measuring device 3 while applying the same type of stimulus as that applied in step S1 to the subject. The measured electroencephalogram data is transmitted from the electroencephalogram measuring device 3 to the analysis device 4 and acquired by the first acquisition unit 43.

[0019] In step S3, similar to step S1, the subject's electroencephalogram data may be obtained from a database or data held by a third party.

[0020] In step S4 (first estimation step), the first estimation unit 44 analyzes the EEG acquired in step S3 under multiple analysis conditions and estimates the brain activation area when the subject is stimulated for each analysis condition. The analysis method is not particularly limited, but signal source estimation analysis can be used. For example, signal source estimation analysis is performed on the EEG during real stimulation and the EEG during control stimulation. For each pre-divided brain region (each region in the HCP atlas), a first principal component or average of a time series signal, a frequency spectrum signal, or a functional connectivity index using the time series signal or the frequency spectrum signal is extracted, and the activation area is estimated by calculating the difference between the first principal component during real stimulation and the first principal component during control stimulation. Multiple analysis conditions can be set by combining factors such as: whether or not to perform absolute value processing of the difference; whether or not to perform smoothing on the first principal component or average; and the duration of the analysis area. Examples of analysis conditions are shown in Table 1.

[0021]

[0022] The first estimation unit 44 analyzes the electroencephalogram for each of these analysis conditions 1 to n and estimates the active areas in the brain. Note that the number of analysis conditions, n, is not particularly limited. Also, it is not necessary to include all combinations of elements in the analysis conditions. Table 2 shows an example of analysis conditions and estimated areas.

[0023]

[0024] Note that steps S3 and S4 may be performed before step S1 or S2.

[0025] In step S5 (comparison step), the comparison unit 45 compares the active area estimated in step S4 with the active area identified in step S2 for each of the analysis conditions.

[0026] If there are multiple subjects, the above steps S1 to S5 are repeated while changing subjects (step S7) until all subjects have been analyzed (YES in step S6). Note that steps S1 to S5 may be performed simultaneously for multiple subjects.

[0027] In step S8 (condition selection step), the condition selection unit 46 selects one analysis condition from the plurality of analysis conditions based on the comparison result in step S5. If there is one subject, the condition is selected from analysis conditions 1 to n under which the estimated region matches or most closely resembles the region identified in step S2. If there are multiple subjects, the condition is extracted for each subject from analysis conditions 1 to n under which the estimated region matches or most closely resembles the region identified in step S2, and the most extracted analysis condition is selected.

[0028] As described above, the condition setting method according to this embodiment sets the analysis conditions selected in step S8 as analysis conditions for estimating active regions in the brain of a given or unknown subject based on electroencephalograms. While area estimation using electroencephalogram analysis is simpler and less expensive than area identification using fMRI images, it is difficult to improve accuracy unless the analysis conditions are appropriately set. In contrast, in this embodiment, the accuracy of the analysis conditions is verified based on the results of a comparison between the area estimated by electroencephalogram analysis and the area identified by fMRI images. This allows for the setting of analysis conditions that enable highly accurate area estimation. Therefore, by using the set analysis conditions, active regions in the brain can be estimated inexpensively and easily.

[0029] (Estimation and Stimulation) Figure 3 is a flowchart showing the processing steps of the estimation method and stimulation method according to this embodiment. The estimation method is a method of estimating active areas in the subject's brain based on electroencephalograms, and includes steps S11 and S12. The stimulation method is a method of applying electrical stimulation to the active areas estimated by the estimation method, and includes step S13. The subject may be the same as the subject described above, but this embodiment assumes an aphasia patient who has never had their active areas measured in their brain.

[0030] In step S11 (second acquisition step), the second acquisition unit 47 of the analysis device 4 acquires electroencephalograms (EEGs) obtained when the stimulus is applied to the subject (aphasic patient). Specifically, the subject's EEGs are measured by the EEG measuring device 3 while the subject is given the same type of stimulus (language task) as that given to the test subject in the above-described condition setting method. The measured EEG data is transmitted from the EEG measuring device 3 to the analysis device 4 and acquired by the second acquisition unit 47.

[0031] In step S11, similar to steps S1 and S3, the subject's electroencephalogram data may be acquired from a database or the like.

[0032] In step S12 (second estimation step), the second estimation unit 48 analyzes the EEG acquired in step S11 to estimate the activated brain region when the stimulus is applied to the subject. In the present embodiment, the second estimation unit 48 analyzes the EEG acquired in step S11 using the analysis conditions set by the above-described condition setting method to estimate the activated brain region when the stimulus is applied to the subject. The analysis conditions set by the above-described condition setting method are analysis conditions that enable highly accurate region estimation. Therefore, the estimation method according to the present embodiment can accurately estimate the activated brain region in response to a language task in a subject who is an aphasic patient through EEG analysis.

[0033] In step S13, the stimulation device 5 applies electrical stimulation to the active area estimated by the estimation method according to this embodiment via the stimulation electrodes 50. This is expected to have an effect of improving the language ability of the subject.

[0034] (Additional Notes) The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope of the claims. Forms obtained by appropriately combining the technical means disclosed in the embodiments are also included in the technical scope of the present invention.

[0035] In the above embodiment, fMRI images taken by the MRI device 2 are used to accurately identify active regions in the subject's brain, but a PET device may be used instead of the MRI device 2.

[0036] The electroencephalogram analysis method can also be modified as appropriate. In the above embodiment, a signal source estimation analysis is performed to extract the first principal component or average of a time-series signal, a frequency spectrum signal, or a functional connectivity index using the time-series signal or the frequency spectrum signal. However, instead of performing a signal source estimation analysis, a first principal component or average of a time-series signal, etc., of a signal from each electrode attached to a subject may be extracted. Furthermore, in addition to the above-mentioned time-series signals, frequency power values, time-series fluctuations of frequency power values, correlations between brain region (or electrode) time-series signals, time-series fluctuations of correlations between brain region (or electrode) time-series signals, inter-frequency coupling values ​​in brain region (or electrode) frequency signals, time-series fluctuations of inter-frequency coupling in brain region (or electrode) frequency signals, and functional connectivity between brain regions (or electrodes) may be calculated, and the first principal component or average of these may be extracted.

[0037] In the above embodiment, in steps S3 and S11, the EEG of the subject and the object is acquired using an EEG measuring device 3. However, a magnetoencephalogram (MEG) or near-infrared spectroscopy device may be used instead of the EEG measuring device 3. In this case, in step S12, the MEG or near-infrared spectroscopy data is analyzed to estimate active areas in the brain. EEGs are obtained by measuring the electrical current generated by the firing of neurons from the scalp, while MEGs are obtained by measuring the magnetic field generated by this electrical current. In fact, it is known that the waveforms of electrical current and magnetic field at each frequency exhibit similar behavior (e.g., Non-Patent Documents 3 and 4). Therefore, it is possible to use a magnetoencephalogram (MEG) instead of an EEG measuring device.

[0038] In the above embodiment, in step S12, the electroencephalogram acquired in step S11 is analyzed under the analysis conditions set by the condition setting method according to this embodiment, but the analysis conditions are not limited to this. For example, the analysis conditions may be set by repeatedly estimating an active area in the brain and electrically stimulating the estimated active area while changing the analysis conditions.

[0039] In the above embodiment, the estimation method focused on brain activity regions, but brain networks may also be estimated. In this case, there is no need to apply stimulation to the subject or the subject in steps S1, S3, and S11. The brain network is represented, for example, by the degree of synchronization of time-series signals of multiple regions in the HCP atlas (e.g., correlation of time-series signals or phase differences between various frequency bands). First, the brain network is identified by calculating the degree of synchronization of time-series signals between pixels using time-series signals for each pixel or anatomical region of the fMRI image. Next, the brain network is estimated by calculating the degree of inter-electrode synchronization using time-series signals for each electrode or signal source of the EEG while changing the analysis conditions. For each analysis condition, the analysis conditions can be set by comparing the brain network estimated from the EEG with the brain network identified from the fMRI image.

[0040] Examples of the present invention will be described below, but the present invention is not limited to the following examples.

[0041] Example 1 In Example 1, an experiment was conducted on 13 subjects to verify the validity of the condition setting method and estimation method according to the present invention. The experiment was conducted over two days, with active brain regions identified using fMRI on the first day and active brain regions estimated using electroencephalograms on the second day. Analysis conditions suitable for estimation using electroencephalograms were then selected, and the estimation accuracy was verified.

[0042] (Identification by fMRI) On the first day, identification of active regions in the brain by fMRI was carried out according to steps S1 and S2 shown in FIG.

[0043] In step S1, MRI images were taken while each subject was given a language task in the MRI device 2. Specifically, as shown in Figure 4(a), 11 real stimulus blocks (R blocks) and 10 control (contrast stimulus) blocks (C blocks) were alternately repeated during the MRI measurement. In the R blocks, images of objects such as animals were presented as shown in Figure 4(b), and in the C blocks, unclear images were presented as shown in Figure 4(c). In each block, seven images and a switching cue were presented in sequence, with each image presented for 2.5 seconds. Each time an image appeared, the subject was asked to respond as quickly as possible to the name of the object if they recognized it, or to "I don't know" if they did not.

[0044] In step S2, the analysis device 4 was used to identify, based on the MRI images, the areas of activity in the brain that were activated when the subject was stimulated. Specifically, the following processes were performed: - Preprocessing (correction of shifts in functional images due to body movement during imaging, conversion of the subject's brain to standard brain coordinates, and spatial smoothing) - Statistical processing (estimation of brain activity for each voxel (pixel) using a general linear model) - Segmentation of the brain based on the HCP atlas - Calculation of the average brain activity across all blocks for each of the R and C blocks - Calculation of the difference between the average brain activity in the R block and the average brain activity in the C block - Calculation of the absolute value of the difference These processes identified the areas of activity in the brain that were required for recalling the meaning of an object and its name.

[0045] Furthermore, the magnitude of the absolute value for each region was ranked, and the region with the largest absolute value was identified for the frontal and temporal-parietal regions.

[0046] (Estimation by Electroencephalogram) On the second day, estimation of the active areas in the brain by electroencephalogram was performed according to steps S3 and S4 shown in FIG.

[0047] In step S3, the subject's electroencephalogram (EEG) was measured using the electroencephalogram (EEG) measuring device 3 while the subject was given the same type of stimulus (language task) as in step S1. However, because fMRI and EEG signal changes at different speeds in response to stimuli, the timing of image presentation, etc., was changed from step S1. Specifically, during EEG measurement, R and C blocks were presented randomly, with R and C blocks appearing 40 times each. In the R block, an image of an object such as an animal was presented for 1 second, and in the C block, an unclear image was presented for 1 second. The interval between each block was 3 to 6 seconds.

[0048] A 64-channel electroencephalogram measuring device was used as the electroencephalogram measuring device 3. The conditions for measuring the electroencephalogram were as follows: Electrode placement: International 10-20 system Bandpass filter: 0.5 Hz to 60 Hz

[0049] In step S4, the EEG was analyzed under multiple analysis conditions using the analysis device 4, and the active areas in the brain were estimated for each of the analysis conditions. The following processes were performed in the analysis: - Noise removal (eyeball components, etc.) using independent component analysis - Averaging the waveforms of all trials for each of the R block and C block - Signal source estimation to estimate signals on the cerebral cortex - Extraction of the first principal component for each area in the HCP atlas

[0050] Twelve analysis conditions were prepared by combining the following elements: - Whether or not to process the absolute value of the difference between the first principal component in the R block and the first principal component in the C block - Whether or not to smooth the waveform of the first principal component (Yes: 1-15 Hz, No: 1-40 Hz) - Time of the analysis area (0-0.8 s, 0-1.0 s, or 0-1.5 s)

[0051] The most active area was estimated for each analysis condition. Based on the above, identification by fMRI and estimation by EEG were performed.

[0052] (Setting of analysis conditions) Next, the active areas identified by fMRI were compared with the active areas estimated by EEG, and one analysis condition was selected from 12. In this example, the analysis conditions shown in Table 3 were selected for each of the two regions, the frontal and temporal-parietal.

[0053]

[0054] (Verification of Estimation Accuracy) The EEG acquired in step S3 above was analyzed using the analysis conditions shown in Table 3 to estimate active regions in the brain. Specifically, for the frontal region, active regions were estimated using the following steps 1 to 7: 1. Band-pass filter processing of EEG 2. Noise removal (eyeball components, etc.) using independent component analysis 3. Averaging waveforms of all trials for each of R block and C block 4. Performing signal source estimation 5. Extracting the first principal component for each region in the HCP atlas 6. Performing smoothing processing on the first principal component 7. Extracting the absolute value of the difference between the first principal component (average brain activity) in the R block and the first principal component (average brain activity) in the C block for a range of 1.5 seconds from the time the image was presented, and using this as the amount of activity 8. Identifying the area with the greatest activity in the target area of ​​the frontal brain. In this way, in "4." and "5.", in order to identify the area suitable for (electrical) stimulation, the high-dimensionality (region subdivision) achieved by signal source estimation is deliberately reduced using principal component analysis to summarize the characteristic signals for each specific region.

[0055] Additionally, for the temporal to parietal regions, active areas were estimated using the following steps 1 to 8. 1. Bandpass filter processing of EEG 2. Noise removal using independent component analysis (eyeball components, etc.) 3. Averaging the waveforms of all trials for each of the R block and C block 4. Performing signal source estimation 5. Extracting the first principal component for each region in the HCP atlas 6. Performing smoothing processing on the first principal component 7. Extracting the absolute value of the difference between the average brain activity in the R block and the average brain activity in the C block for a range of 0.8 seconds from the time the image was presented, and using this as the amount of activity 8. Identifying the region with the greatest activity in the target temporal to parietal regions Note that in step 6, a lowpass filter of order 4 and 15 Hz was applied to the first principal component.

[0056] Then, the estimation accuracy was verified by comparing the estimated region (hereinafter referred to as the EEG estimated region) with the region identified by fMRI (hereinafter referred to as the fMRI identified region).

[0057] The validation results for the frontal region are shown in Table 4.

[0058]

[0059] The verification results for the temporal to parietal regions are shown in Table 5.

[0060]

[0061] Regarding accuracy, "very close" means that the EEG estimated region and the fMRI specific region are adjacent in the HCP-based region division, and "close" means that the EEG estimated region and the fMRI specific region are located across one region in the HCP-based region division. Also, "within a single electrode area" means that the EEG estimated region and the fMRI specific region are within a 5 cm x 5 cm area, which is the size of the stimulating electrode.

[0062] As described above, we were able to estimate the active areas in the brain with high accuracy for 11 out of 13 subjects (85%) in both the frontal and temporal-parietal regions.

[0063] [Example 2] In Example 2, the effect of providing electrical stimulation on improving language function was examined. Specifically, the stimulation effect was compared under the following conditions for one subject out of the 13 subjects who participated in the experiment in Example 1. (Condition 1) An electrode was placed in the area of ​​maximum brain activity (hereinafter referred to as L-max) estimated in Example 1, and electrical stimulation was applied at 2 mA for 15 minutes. (Condition 2) An electrode was placed 2.5 cm behind L-max, and electrical stimulation was applied at 2 mA for 15 minutes. (Condition 3) An electrode was placed at L-max, and sham stimulation was applied for 30 seconds.

[0064] The stimulus presentation for each condition was conducted at least two days apart. After the stimulus presentation for each condition, the subjects were given a naming task. Specifically, subjects were presented with an image of an object and asked to name it as quickly as possible, and their reaction time was measured.

[0065] Figure 5 is a graph showing the average reaction speed after stimulation under each condition in the language task. Condition 1 (L-max) and Condition 2 (L-max + 2.5 cm) resulted in faster reaction speeds than Condition 3 (sham stimulation), confirming that Condition 1 was particularly effective.

[0066] Example 3 In Example 3, we investigated whether brain networks can be estimated based on electroencephalograms (EEG). Specifically, to verify whether sensory integration function, which declines with age, can be detected from electroencephalograms (EEG), we examined whether there was a correlation between movement accuracy when visual rotation interference was performed during an arm-reaching task, and the connectivity between brain regions at rest analyzed from EEG (functional connectivity, rsEEG-FC), and performance on a physical motor function test. Visual rotation interference during this reaching task has been shown to increase the angle at which the arm's reach point deviates from the target point (mean error angle) in elderly people, and to increase the angle at which visual rotation interference can be detected (detection angle) (Yoshimura et al., Brain Sciences, 2020). Therefore, we investigated the correlation between these two indicators and indicators from a physical motor function test, which is routinely conducted as part of a fall risk survey, in elderly people aged 75 years or older. Results indicated a correlation between both indicators. Therefore, we next examined whether the rsEEG-FC, which was significantly correlated with these two indices (mean error angle and detection angle), included areas related to sensory integration. Previous fMRI studies have shown that the motor cortex is associated with these indices (Yoshimura et al., Brain Sciences, 2020; Inagaki et al., Preprints, 2024). Therefore, we examined the presence or absence of rsEEG-FC centered on the motor cortex, and found connectivity with areas related to visual processing. These results suggest the validity of functional connectivity analysis using EEG, i.e., the estimation of brain networks based on EEG.

[0067] (Method) rsEEG measurements were taken from 64 electrodes for 5 minutes on 10 elderly people aged 75 years or older (77.8 ± 2.48 years). Participants were instructed to look at a crosshair displayed on a PC screen without thinking about anything in particular.

[0068] As preprocessing for the EEG signals, a notch filter with a cutoff frequency of 50 Hz and a bandpass filter with a passband of 1-40 Hz were applied to remove power supply noise and other noise not related to brain activity. Next, independent component analysis was performed, and independent components with high noise content were visually judged and removed. After referencing using the average of all electrodes, downsampling was performed at 512 Hz to shorten calculation time and reduce data volume.

[0069] Next, to estimate signals for each anatomical brain region, we performed cortical-level signal source estimation using minimum norm estimation on the preprocessed rsEEG data. First, we created a forward model using the source space and boundary element model of a commonly used standard brain template. Next, we created a noise covariance matrix from this forward model and the rsEEG data, and then created an inverse operator from the rsEEG data, forward model, and noise covariance matrix. Finally, we performed cortical-level signal source estimation using the MNE method from the rsEEG data and inverse operator. Finally, we performed principal component analysis within each of the 360 ​​brain regions of interest from the Human Connectome Project (HCP), and extracted the first principal component as the time-series signal for each region.

[0070] To calculate rsEEG-FC, a correlation matrix was created from the time-series signals of each region. Previous studies using fMRI images confirmed significant FC in the left primary motor cortex. Therefore, the left primary motor cortex (L_4) of the HCP was used as the seed region, and the other regions were used as target regions to calculate the correlation values ​​for each pair.

[0071] Group-level analysis was performed using a general linear model (GLM). Pairwise correlation values ​​were used as the dependent variable, and age, gender, and motor indices (mean error angle and detection angle) calculated from the analysis of the motor task data were used as explanatory variables. FC between ROIs with high correlations with motor indices was estimated. Results were corrected for multiple comparisons using the Bonferroni method.

[0072] (Results) For ROI pairs between target regions, with the left primary motor cortex (L_4) as the seed region, we estimated FC between ROIs that correlated significantly with the mean error angle or the detection angle. Three significant connections were identified for the mean error angle, and five for the detection angle.

[0073] The FCs that showed a significant correlation with the mean error angle were: first, the connection with L_V4t as the target area (t = -4.42, p = 0.004), second, the connection with R_47s as the target area (t = -4.25, p = 0.008), and third, the connection with L_IP0 as the target area (t = -3.89, p = 0.036) (see Figure 6). In all cases, the higher the motor accuracy of the participants, i.e., the smaller the mean error angle, the stronger the FC (a negative connection).

[0074] The FCs that were significantly correlated with the detection angle were: 1) the connection with the R_IP2 target region (t = -4.41, p = 0.004); 2) the connection with the R_p47r target region (t = -4.15, p = 0.012); 3) the connection with the L_TPOJ1 target region (t = -4.11, p = 0.014); 4) the connection with the R_PGs target region (t = -4.07, p = 0.017); and 5) the connection with the L_LO1 target region (t = -3.83, p = 0.046) (see Figure 7). In all cases, participants who were more aware of smaller visual rotation angles, i.e., with smaller detection angles, showed stronger negative FCs.

[0075] The present invention is particularly suitable for treating patients with aphasia, but is not limited thereto. For example, the present invention can also be applied to a method for estimating the active area in the brain of a stroke patient with paralysis when stimulation is applied to the paralyzed area and then electrically stimulating the active area. Furthermore, the present invention can also be applied to a method for estimating the active area in the brain of a patient with impaired cognitive function due to brain trauma when cognitive stimulation is applied and then electrically stimulating the active area.

[0076] Furthermore, the present invention can be applied not only to the treatment of diseases but also to improving the performance of healthy individuals. For example, the present invention can be applied to a method of estimating the area in the brain associated with a specific body part where the accuracy of movement needs to be improved, and then electrically stimulating that active area.

[0077] Although the present invention is preferably applied to the estimated brain active area by electrical stimulation, the present invention is not limited to this. For example, the present invention can also be applied to cases where magnetic stimulation, ultrasonic stimulation, optical stimulation, visual stimulation, auditory stimulation, tactile stimulation, mechanical stimulation, etc. are applied.

[0078] REFERENCE SIGNS LIST 1 Analysis and stimulation system 2 MRI device 3 Electroencephalogram measuring device 4 Analysis device 40 Auxiliary storage device 41 Image acquisition unit 42 Identification unit 43 First acquisition unit 44 First estimation unit 45 Comparison unit 46 Condition selection unit 47 Second acquisition unit 48 Second estimation unit 5 Stimulation device 50 Stimulating electrode P1 Condition setting program P2 Estimation program

Claims

1. A method for estimating active areas or brain networks in a subject's brain based on electroencephalograms, magnetoencephalograms or near-infrared spectroscopy, comprising: a second acquisition step of acquiring electroencephalogram, magnetoencephalogram or near-infrared spectroscopy data of the subject; and a second estimation step of analyzing the electroencephalogram, magnetoencephalogram or near-infrared spectroscopy data acquired in the second acquisition step to estimate active areas or brain networks in the subject's brain.

2. The estimation method according to claim 1, wherein the second acquisition step acquires the electroencephalogram, magnetoencephalogram or near-infrared spectroscopy data when a stimulus is applied to the subject.

3. The estimation method according to claim 2, wherein the second estimation step estimates the active area in the brain when the stimulus is applied to the subject.

4. A stimulation method for providing electrical, magnetic, ultrasonic, optical, visual, auditory, tactile or mechanical stimulation to the subject in accordance with the active area or brain network estimated by the estimation method described in claim 1.

5. The stimulation method according to claim 4, further comprising the step of: providing electrical stimulation to the estimated active area.

6. A condition setting method for setting analysis conditions for the electroencephalogram, magnetoencephalogram or near-infrared spectroscopy data for estimating active areas or brain networks in the brain by the estimation method of claim 1, comprising: an image acquisition step for acquiring a brain image of a subject taken by an MRI device or a PET device; an identification step for identifying active areas or brain networks in the subject's brain based on the brain images; a first acquisition step for acquiring electroencephalogram, magnetoencephalogram or near-infrared spectroscopy data of the subject; a first estimation step for analyzing the electroencephalogram, magnetoencephalogram or near-infrared spectroscopy data acquired in the first acquisition step under a plurality of analysis conditions to estimate active areas or brain networks in the subject's brain for each of the analysis conditions; a comparison step for comparing, for each of the analysis conditions, the active areas estimated in the first estimation step with the active areas or brain networks identified in the identification step; and a condition selection step for selecting one analysis condition from the plurality of analysis conditions based on the comparison result in the comparison step.

7. A condition setting method according to claim 6, wherein in the first acquisition step, the electroencephalogram, magnetoencephalogram or near-infrared spectroscopy data is acquired when a stimulus is applied to the subject.

8. The estimation method according to claim 7, wherein the first estimation step estimates an active area in the brain when the stimulus is applied to the subject.

9. A condition setting method as described in claim 8, wherein in the first estimation step, a signal source estimation analysis is performed on the EEG during real stimulation and the EEG during control stimulation, a first principal component or average of a time series signal, a frequency spectrum signal, or a functional connectivity index using the time series signal or the frequency spectrum signal is extracted for each of pre-divided brain regions, and the active region is estimated by taking the difference between the first principal component or average during real stimulation and the first principal component or average during control stimulation, and the analysis conditions include at least any of: whether or not to perform absolute value processing of the difference, whether or not to perform smoothing processing on the first principal component or average, and the time of the analysis area.

10. A condition setting method as described in claim 8, wherein in the first estimation step, a first principal component or average of a time series signal, a frequency spectrum signal, or a functional connectivity index using the time series signal or the frequency spectrum signal is extracted for the signals of each electrode attached to the subject for the EEG during real stimulation and the EEG during control stimulation, and the active area is estimated by taking the difference between the first principal component or average during real stimulation and the first principal component or average during control stimulation, and the analysis conditions include at least any of: whether or not to perform absolute value processing of the difference, whether or not to perform smoothing processing on the first principal component or average, and the time of the analysis area.

11. A condition setting method according to any one of claims 6 to 10, wherein the brain image is an fMRI image taken by the MRI device.

12. An estimation device that estimates active areas or brain networks in a subject's brain based on electroencephalograms, magnetoencephalograms, or near-infrared spectroscopy, the estimation device having: a second acquisition unit that acquires electroencephalogram, magnetoencephalogram, or near-infrared spectroscopy data of the subject; and a second estimation unit that analyzes the electroencephalogram, magnetoencephalogram, or near-infrared spectroscopy data acquired by the second acquisition unit to estimate active areas or brain networks in the subject's brain.

13. A stimulation device that provides electrical, magnetic, ultrasonic, optical, visual, auditory, tactile or mechanical stimulation to the subject in accordance with the active area or brain network estimated by the estimation device described in claim 12.

14. A condition setting device that sets analysis conditions for the electroencephalogram, magnetoencephalogram or near-infrared spectroscopy data to estimate active areas or brain networks in the brain by the estimation device described in claim 12, comprising: an image acquisition unit that acquires brain images of a subject taken by an MRI device or a PET device; an identification unit that identifies active areas or brain networks in the subject's brain based on the brain images; a first acquisition unit that acquires electroencephalogram, magnetoencephalogram or near-infrared spectroscopy data of the subject; a first estimation unit that analyzes the electroencephalogram, magnetoencephalogram or near-infrared spectroscopy data acquired by the first acquisition unit under a plurality of analysis conditions and estimates active areas or brain networks in the subject's brain for each of the analysis conditions; a comparison unit that compares, for each of the analysis conditions, the active areas or brain networks estimated by the first estimation unit with the active areas or brain networks identified by the identification unit; and a condition selection unit that selects one analysis condition from the plurality of analysis conditions based on a comparison result by the comparison unit.

15. An estimation program for estimating active areas or brain networks in a subject's brain based on electroencephalograms, magnetoencephalograms or near-infrared spectroscopy, comprising: a second acquisition step for acquiring electroencephalogram, magnetoencephalogram or near-infrared spectroscopy data of the subject; and a second estimation step for analyzing the electroencephalogram, magnetoencephalogram or near-infrared spectroscopy data acquired by the second acquisition step to estimate active areas or brain networks in the subject's brain.

16. A condition setting program for setting analysis conditions for the electroencephalogram, magnetoencephalogram or near-infrared spectroscopy data for estimating active areas or brain networks in the brain by the estimation method of claim 1, comprising: an image acquisition step for acquiring brain images of a subject taken by an MRI device or a PET device; an identification step for identifying active areas or brain networks in the subject's brain based on the brain images; a first acquisition step for acquiring electroencephalogram, magnetoencephalogram or near-infrared spectroscopy data of the subject; a first estimation step for analyzing the electroencephalogram, magnetoencephalogram or near-infrared spectroscopy data acquired in the first acquisition step under a plurality of analysis conditions to estimate active areas or brain networks in the subject's brain for each of the analysis conditions; a comparison step for comparing, for each of the analysis conditions, the active areas or brain networks estimated in the first estimation step with the active areas or brain networks identified in the identification step; and a condition selection step for selecting one analysis condition from the plurality of analysis conditions based on a comparison result in the comparison step.

17. A computer-readable recording medium having the estimation program according to claim 15 recorded thereon.

18. A computer-readable recording medium having the condition setting program according to claim 16 recorded thereon.

Images