Brain wave measurement device, brain wave measurement system, and brain wave measurement program

Abstract

To provide a device for measuring an electroencephalogram by arranging a small number of electrodes and accurately detecting the occurrence of eyeball movement noise.SOLUTION: An electroencephalogram signal is measured from a biological signal including an electrooculogram (EOG) signal and an electroencephalogram signal obtained from four electrodes arranged above and below both eyes. This device 1 is provided with an EOG signal detection part 11 for detecting an eyeball movement component by using a correlation coefficient of a biological signal for each section of a prescribed length in the biological signal obtained from an electrode 2, a signal analysis part 13 for separating the eyeball movement component for a prescribed time biological signal including an eyeball movement positive section where the EOG signal is detected, an EOG signal removal part 15 for removing the eyeball movement component from the biological signal, and a brain wave signal generation part 17 for reconstituting the brain wave signal from the signal reconstituted by removing the eyeball movement component. The electroencephalogram signal generation unit 17 repeats signal analysis for separating the eyeball movement component from the reconstructed signal, and reconstructs the electroencephalogram signal by reliably removing the EOG signal noise included in the electroencephalogram signal when it is determined that there is no eyeball movement component.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an electroencephalogram (EEG) measuring device, an electroencephalogram measuring system, and an electroencephalogram measuring program. [Background technology]

[0002] In recent years, it has been suggested that epileptic discharges may be a physiological diagnostic marker for the early stages of Alzheimer's disease. Epileptic discharges occur more frequently during sleep, but in hospital EEG tests lasting 30 minutes to an hour, it is often not possible to record sleep EEG during the test, and the detection rate of epileptic discharges is low. EEG testing requires the attachment of approximately 20 electrodes (10-20 method) mainly around the head, which takes 1-2 hours, placing a heavy burden on medical staff, resulting in limited facilities that can perform the test.Furthermore, for patients, having many electrodes attached to the head can be a significant burden, as it disrupts sound sleep and restricts movement.

[0003] In conventional EEG measurement methods, it is time-consuming to remove hair when placing electrodes, and if the electrodes placed on the head come into contact with bedding, it is uncomfortable for the patient and interferes with sound sleep.It is expected that a new method for detecting epileptic waves in the temporal and frontal lobes will be realized by placing a small number of electrodes on the hairless face instead of the conventional electrodes placed on the head (hereinafter also referred to as head electrodes). Furthermore, removing artifacts (noise) is an important issue in EEG measurement. Among noise artifacts, electro-oculograms (EOG) generated by eye movement are one of the largest artifacts in EEG measurement. Compared to head electrodes, electrodes placed on the face (hereinafter referred to as facial electrodes) are closer to the eyes and are more susceptible to the influence of EOG signals. Therefore, EEG measurement using facial electrodes poses the problem of having to identify and remove EOG signals. In particular, inexperienced physicians have difficulty distinguishing between EOG and EEG signals, making it difficult to properly identify the periods in which EOG signals are occurring.

[0004] A conventional method involves attaching electrodes (10-20 method) around the head, removing noise from the EOG signal, and measuring the EEG signal. The EOG electrodes are used as a reference, and the EEG components of the EEG electrodes are estimated by linear regression. However, this method requires placing EOG electrodes separately from the electrodes (10-20 method). Furthermore, because the EOG electrodes themselves contain EEG components, using them as a reference also results in the loss of EEG signal information. There is also a method of creating a template by averaging eye movements in advance and subtracting the signal of this template, but this method has the problem that it can only remove EOG signals with a predetermined shape. There is also a method using blind source separation (a method of estimating the pre-mixed sound source signal from only the observed mixed signal), which does not depend much on the shape of the noise because it decomposes the EEG signal into signal sources, but there are problems such as the need to identify the location where the noise appears and the need to identify which of the separated signal sources originate from eye movement. Furthermore, there is a method for identifying EOG signals using a deep learning model, but this requires manual labeling of noise, there is little training data, and it can only handle specific electrode placements. Also, while it can handle simple blinks, it cannot handle the diverse eye movements that occur in real life.

[0005] On the other hand, there is a method that uses a statistical filter to detect noise components of eye movements from EEG signals (single time series data) from one electrode. However, this method captures sudden fluctuations in the time series, so there is a problem that epileptic discharges, which appear suddenly like noise components, and EEG signals during sleep are removed.

[0006] Furthermore, Patent Document 1 discloses a wearable biometric measurement device in the form of glasses or goggles that is worn by a user, and describes measuring the difference in potential between an electrode directly above the user's eyebrows that contacts a part directly above the eyebrows and a forehead electrode that contacts the user's forehead directly above the eyebrows as a frontalis myoelectric potential, and measuring not only the EEG signal but also the frontalis myoelectric potential to remove noise from the EEG signal due to the frontalis myoelectric potential generated in the frontalis muscle when the user looks up or raises their eyebrows (paragraph 0008 of the same patent document). Specifically, Patent Document 1 discloses a noise removal unit 41 shown in Figure 1 (in Figure 1, the inputs of 206 are listed as 102 and 108, but it is believed that they are actually 104 and 108), which removes noise from the left and right EEG signals using the left frontalis myoelectric potential, left upper and lower eye potential, right frontalis myoelectric potential, right upper and lower eye potential, and right and left eye potential (paragraph 0036 of the same patent document). However, in the device of Patent Document 1, the left EEG signal is measured (left EEG signal measuring unit 205) by measuring the potential difference between the left forehead electrode 103 and the left supra-auricular electrode 107 as an EEG signal for the left side of the user's head, and the right EEG signal is measured (right EEG signal measuring unit 206) by measuring the potential difference between the right forehead electrode 104 and the right supra-auricular electrode 108 as an EEG signal for the right side of the user's head (paragraph 0030 of the same document). Also, the left and right electrooculograms (left and right electrooculogram measuring unit 207) measure the potential difference between the left temple electrode 109 and the right temple electrode 110 as the electrooculograms in the left and right directions of the user (left and right electrooculograms). Furthermore, the right upper and lower electro-oculogram measuring unit 203 measures the potential difference between the electrode 102 directly above the right eyebrow and the electrode 105 directly below the right lower eyelid as the vertical electro-oculogram of the user's right eye (right upper and lower electro-oculogram), and the left upper and lower electro-oculogram measuring unit 204 measures the potential difference between the electrode 101 directly above the left eyebrow and the electrode 106 directly below the left lower eyelid as the vertical electro-oculogram of the user's left eye (left upper and lower electro-oculogram) (same paragraph 0029). As described above, the electrodes for the electrooculogram and the electrodes for the EEG signal are completely separated in the device of Patent Document 1. Therefore, the electrodes for the electrooculogram themselves contain EEG components, which poses a problem in that removing noise from the electrooculogram also erases the information in the EEG signal.

[0007] Patent Document 2 also discloses a method for removing noise components caused by eyeblink artifacts from recorded data of EEG signals. Specifically, first, EEGs are measured as time-series potential signals, and potential changes caused by eye movement are measured as an electrooculogram in parallel with the EEG measurement. The occurrence of a blink is detected from the potential changes in the measured electrooculogram. Next, a normalization parameter is calculated from the detected potential changes over time caused by the blink, and the data (blink artifact) in the time-series data of the EEG measurement is normalized using the normalization parameter. The normalized EEG data is then averaged, and the averaged EEG data is inversely normalized using the normalization parameter to obtain an estimated waveform of the eyeblink artifact, and the estimated waveform of the eyeblink artifact is removed from the time-series EEG data. In the method of Patent Document 2, an electroencephalograph connected with five electrodes is used to measure EEGs, and each electrode is attached to a location FP (F1) on the subject's scalp. Z , F Z , C Z , P Z , C3, and C4 (see Fig. 1 of Patent Document 2), and these electrodes detect electroencephalograms (EEGs) as potential changes, and potential changes caused by eye movements are detected as electrooculograms in parallel with the measurement of time series EEG data. In other words, there is a problem in that it is necessary to provide separate electrooculogram (EOG) electrodes to remove the component of eye movements from time series EEG data.

[0008] Furthermore, Patent Document 3 discloses a method of estimating EOG signals using only head electrodes without using EOG electrodes, by attaching electrodes (10-20 method) centered around the head. In the method of Patent Document 3, the time-series signal of the electroencephalogram is Fourier transformed into frequency components, and the presence or absence of artifacts of eyeblinks, lateral eye movements, electromyograms, and electrodes is determined. However, attaching electrodes (10-20 method) centered around the head still places a burden on the person being measured. [Prior art documents] [Patent documents]

[0009] [Patent Document 1] Japanese Patent Application Laid-Open No. 2017-185148 [Patent Document 2] Japanese Patent Application Publication No. 11-318843 [Patent Document 3] Japanese Patent Application Laid-Open No. 2004-350797 Summary of the Invention [Problem to be solved by the invention]

[0010] It is generally known that EOG signals generated by eye movements are detected with different polarities (opposite phase) between the upper and lower eyes, and in phase between the left and right eyes. However, because eye movements are complex and vary from patient to patient, it has been difficult for inexperienced doctors to identify EOG signals from EEG signals. Furthermore, EEG measurement using facial electrodes requires identifying and removing EOG signals, but there was a strong need for a system that would allow EEG measurement, including epileptic waves, to be performed by placing a small number of facial electrodes instead of head electrodes, thereby reducing the work of attaching the electrodes and reducing the burden on the person being measured. Furthermore, it is necessary to apply the same algorithm from wakefulness to sleep, when background EEG (brain wave activity that appears predominantly in the occipital lobe during rest, wakefulness, and closed eyes) varies greatly, without impairing epileptic discharges, which are sudden signals. It is also necessary to be able to detect everything from wakefulness to sleep, and the irregular eye movements that occur during wakefulness, without false positives of epileptic discharges, using a single algorithm.

[0011] In view of this situation, the present invention aims to provide an apparatus, system, and program that uses a small number of electrodes placed on the face to accurately detect the occurrence of eye movement noise and measure electroencephalogram signals including epileptic waves. [Means for solving the problem]

[0012] In order to solve the above problem, the electroencephalogram (EEG) measuring device of the present invention estimates eye movement from the correlation between four signals detected by four electrodes placed above and below both eyes, and measures EEG signals from biosignals including electrooculogram (EOG) signals and EEG signals obtained from the same four electrodes. In other words, the EEG measurement device of the present invention is a device that measures EEG signals from biosignals including EOG signals and EEG signals obtained from first to fourth electrodes placed above and below both eyes of the subject, and is equipped with: 1) an EOG signal detection unit that detects eye movement components using the correlation coefficient of the biosignal for each section of a predetermined length in the biosignal obtained from the electrodes; 2) a signal analysis unit that separates eye movement components from the biosignal for a predetermined time period including a positive eye movement section in which the EOG signal was detected; 3) an EOG signal removal unit that removes the eye movement components from the biosignal; and 4) an EEG signal generation unit that reconstructs an EEG signal from the signal from which the eye movement components have been removed. This device's configuration allows for the placement of a small number of facial electrodes instead of head electrodes, reducing the work of attaching electrodes and the burden on the subject, while enabling EEG measurements, including epileptic waves. By measuring EEG using a small number of facial electrodes, the electrodes can be used both for measuring electrooculography and EEG, and the EOG signal is removed using the correlation coefficient of each signal obtained from the small number of facial electrodes, reconstructing the EEG signal.

[0013] In the EOG signal removal unit of this device, it is preferable to repeat signal analysis to separate eye movement components from the signal from which the eye movement components have been removed, and if it is determined that there are no eye movement components, reconstruct the EEG signal. Considering that there will be multiple eye movement components to separate when the eyes move in multiple directions, such as up and down when following text while reading, the signal from which the eye movement components have been removed is again subjected to signal analysis to separate eye movement components, and if it is determined that there are no eye movement components, reconstruct the EEG signal, thereby reliably removing EOG signal noise contained in the EEG signal.

[0014] Here, the signal analysis used to separate the eye movement components is preferably principal component analysis. In this device, EEG measurements are performed using facial electrodes. Since the EOG signal component appears as a strong component in the biosignal, which includes the EOG signal and the EEG signal obtained from the facial electrodes, principal component analysis is used to decompose the biosignal into its principal components and separate the eye movement components. As signal analysis to separate the eye movement components, other signal separation methods such as independent component analysis (ICA), second-order blind identification (SOBI), and stationary subspace analysis can be used in addition to principal component analysis (PCA).

[0015] The EOG signal detection unit of this device also has a determination mode that detects eye movement components from time-series data of biosignals. There are three determination modes. The first determination mode detects eye movement components from the correlation (negative correlation) between the upper and lower electrodes of one eye and the correlation (positive correlation) between the upper electrodes of both eyes (this is called the first determination mode). In other words, the first determination mode determines that eye movement components have been detected when the correlation coefficient of the biosignals obtained from a pair of upper or lower electrodes of both eyes out of the first to fourth electrodes is equal to or greater than a predetermined positive value, and when the correlation coefficient of the biosignals obtained from a pair of upper or lower electrodes of the left or right eye is equal to or less than a predetermined negative value.

[0016] The second determination mode is based on the correlation (negative correlation) between the upper and lower electrodes of both eyes, the correlation (positive correlation) between the two upper electrodes of both eyes, and the correlation (positive correlation) between the two lower electrodes of both eyes (this is called the second determination mode). In other words, the second determination mode determines that an eye movement component has been detected when the correlation coefficient of the biosignals obtained from the two pairs of electrodes (upper and lower) of the first to fourth electrodes is equal to or greater than a predetermined positive value, and when the correlation coefficient of the biosignals obtained from the two pairs of electrodes (upper and lower) of the left and right eyes is equal to or less than a predetermined negative value.

[0017] Furthermore, the third determination mode is to detect the eye movement component from the correlation (negative correlation) between the upper electrode of the right eye and the lower electrode of the left eye (this is called the third determination mode). That is, the third determination mode determines that an eye movement component has been detected when the correlation coefficient of the biosignals obtained from a pair of electrodes above the left eye and below the right eye among the first to fourth electrodes is equal to or less than a predetermined negative value, and / or when the correlation coefficient of the biosignals obtained from a pair of electrodes above the right eye and below the left eye is equal to or less than a predetermined negative value.

[0018] The first determination mode is single (one pair of left and right, and one pair of top and bottom), while the second determination mode is double (two pairs of left and right, and two pairs of top and bottom), and the third determination mode is sash-style (one pair of top left and bottom right, and / or one pair of top right and bottom left). Because EOG signal generation conditions vary depending on the patient's eye movements and characteristics, this system automatically selects the evaluation mode based on the EOG signal generation conditions, or allows doctors and other professionals to select the optimal evaluation mode. By superimposing the EOG signal generation areas detected in each evaluation mode on the detected EEG signal data, it can assist doctors and other professionals with specialized knowledge in identifying the correct EOG signal. Doctors and other professionals can select the evaluation mode they deem optimal based on the EOG signal detection results displayed on a display or other device to identify the EOG signal. This makes it possible to avoid misidentifying high-amplitude, low-frequency components (slow-wave components) that appear during sleep.

[0019] The present device may further include an EOG signal removal method selection unit that automatically selects the above-mentioned determination mode. That is, the present device may further include an EOG signal removal method selection unit that automatically selects, as the optimum determination mode, the determination mode having the highest absolute value of the correlation coefficient calculated in the first and second determination modes. Alternatively, the present device may further include an EOG signal removal method selection unit that automatically selects, as the optimum determination mode, the determination mode having the highest absolute value of the correlation coefficient calculated in the first to third determination modes. By providing an EOG signal removal method selection unit, it becomes possible to avoid misidentifying high-amplitude low-frequency components (slow wave components) that appear during sleep.

[0020] The present device may also be provided with an EOG signal removal method selection unit, allowing the user to manually select the above-mentioned determination mode. That is, the present device may further include an EOG signal removal method selection unit that allows the user to select one determination mode from the first to third determination modes. The present device is configured so that the user can select one determination mode from a plurality of determination modes according to the detection result of the EOG signal. Here, the user is assumed to be a person different from the subject (such as a doctor with specialized knowledge), but the subject himself or herself may make the selection. Preferably, the device further includes a display unit that displays the positive eye movement interval (the interval in which the EOG signal is detected in the acquired time-series data of the biosignal) detected by the EOG signal detection unit by superimposing it on the biosignal. The display unit can accept a user's instruction to correct the displayed positive eye movement interval and redisplay it reflecting the instruction. The display unit is a display connected to the device and can be displayed on a monitor screen, tablet terminal screen, smartphone screen, etc. The EOG signal removal method selection unit displays the positive eye movement section superimposed on the biosignal on the display unit for each of the first to third judgment modes, or displays the first to third judgment modes so that they can be identified using colored frames or the like, and allows the user to select one of the first to third judgment modes using a mouse or the like.

[0021] In the EOG signal detection unit of this device, it is preferable that the biosignal include signals obtained from the fifth and sixth electrodes, which are placed in contact with the left and right temples of the subject, in addition to the first to fourth electrodes. By including the signals obtained from the fifth and sixth electrodes in the biosignal, the accuracy of the electroencephalogram measurement can be further improved. Furthermore, in addition to the first to fourth electrodes and the fifth and sixth electrodes, it is preferable to include in the biosignal a signal obtained from a seventh electrode placed in contact with the center of the subject's head. By adding the seventh electrode, which is a head electrode, in addition to the facial electrodes, the accuracy of EEG measurement can be further improved. Note that reference electrodes (placed on both earlobes) and a ground electrode are necessary for measuring biosignals, and these will be explained later.

[0022] Next, an electroencephalogram measuring system including this device will be described. The electroencephalogram measurement system of the present invention is a system comprising first to fourth electrodes, a biosignal input unit to which biosignals obtained from these electrodes are input, a jig for placing these electrodes above and below both eyes of the subject, and the electroencephalogram measurement device of the present invention described above. Here, the biosignal input unit is a part that measures the potential from the electrodes and performs A / D conversion, such as the electroencephalogram signal input unit of an electroencephalograph. The jig places electrodes above and below both eyes of the subject, and can maintain the same placement regardless of eye or facial movements. Examples of such devices include adhesive ones that allow the electrodes to be fixed to the skin, eye mask-like ones that allow electrodes to be placed above and below both eyes, and eyeglass or goggle-like ones with electrodes placed on a frame that can be worn on the face.

[0023] Next, the electroencephalogram measuring program of the present invention will be described. The EEG measurement program of the present invention is a program for measuring EEG signals from biosignals including electrooculogram (EOG) signals and EEG signals obtained from first to fourth electrodes placed above and below both eyes of the subject, and is a program for causing a computer to execute the following steps. 1) EOG signal detection step: An eye movement component is detected for each section of a predetermined length in a biosignal obtained from an electrode, using a correlation coefficient of the biosignal. 2) Signal analysis step: Separation of eye movement components is performed for a biological signal of a predetermined time period including a positive eye movement section in which an EOG signal is detected. 3) EOG signal removal step: removes eye movement components from the biosignal. 4) EEG signal generation step: EEG signals are reconstructed from signals from which eye movement components have been removed.

[0024] In the EOG signal removal step of the program of the present invention, it is preferable to repeat signal analysis to separate the eye movement components from the signal from which the eye movement components have been removed, and if it is determined that there are no eye movement components, to reconstruct the EEG signal. When the eyes move in multiple directions, there are multiple eye movement components that need to be separated. Signal analysis is then repeated to separate the eye movement components from the signal from which the eye movement components have been removed. If it is determined that there are no eye movement components, the EEG signal is reconstructed, thereby reliably removing the EOG signal noise contained in the EEG signal. [Effects of the Invention]

[0025] According to the present invention, it is possible to accurately detect the occurrence of eye movement noise using a small number of electrodes placed on the face, and to measure electroencephalograms including epileptic waves. Furthermore, according to the present invention, the number of electrodes is significantly reduced, which reduces the work of attaching electrodes and reduces the burden on the person being measured, thereby enabling electroencephalogram measurement. Furthermore, according to the present invention, by using electrodes placed above and below the eye movement, it is possible to detect irregular eye movements from wakefulness to sleep and waking activity using a single algorithm, without falsely detecting epileptic discharges. [Brief explanation of the drawings]

[0026] [Figure 1] Functional block diagram of the electroencephalogram measuring device of Example 1 [Figure 2] An explanatory diagram of the correlation coefficient of signals from four electrodes (first to fourth electrodes) [Figure 3] 1. An explanatory diagram of first to third determination modes [Figure 4] An illustration of the difference between detecting whether or not blinks are included in signal time series data [Figure 5] EOG signal detection unit processing diagram [Figure 6] An explanatory diagram of the processing of the signal analysis unit, EOG signal removal unit, and EEG signal generation unit. [Figure 7] An explanatory diagram of the processing of the signal analysis unit [Figure 8] An example of a monitor display screen [Figure 9] Epileptic discharge signal waveform diagram [Figure 10] Schematic diagram of the arrangement of the first to sixth electrodes [Figure 11] Schematic diagram of the arrangement of the first to seventh electrodes [Figure 12] Schematic diagram of actual electrode arrangement [Figure 13] An example of time-series data of biosignals obtained from each electrode [Figure 14] Functional block diagram of an electroencephalogram measuring device according to a second embodiment [Figure 15] Functional block diagram of an electroencephalogram measurement system according to a third embodiment [Figure 16] Flow diagram of the electroencephalogram measurement program of Example 4 DETAILED DESCRIPTION OF THE INVENTION

[0027] An example of an embodiment of the present invention will be described in detail below with reference to the drawings. Note that the scope of the present invention is not limited to the following examples and illustrated examples, and many modifications and variations are possible. [Example]

[0028] Fig. 1 shows a functional block diagram of one embodiment of the electroencephalogram (EEG) measuring device of the present invention. In the electroencephalogram (EEG) measuring device 1 shown in Fig. 1, four biosignals detected by four electrodes 2 placed above and below both eyes of a subject are input to the device 1 by a biosignal input unit 3, eye movement is estimated from the correlation between the four signals, and an EEG signal is reconstructed from the biosignals including the EOG signal and the EEG signal obtained from the same four electrodes, and output to a signal output unit 4. The EEG measurement device 1 is composed of an EOG signal detection unit 11 that detects eye movement components using a correlation coefficient for each predetermined length section in the time series data of the biosignal obtained from the electrode 2, a signal analysis unit 13 that separates the eye movement components from the biosignal for a predetermined time period including the eye movement positive section in which the EOG signal was detected, an EOG signal removal unit 15 that removes the eye movement components from the biosignal, and an EEG signal generation unit 17 that reconstructs an EEG signal from the signal from which the eye movement components have been removed. The EOG signal removal unit 15 repeatedly performs signal analysis processing in the signal analysis unit 13, which separates the eye movement components, on the signal from which the eye movement components have been removed, and if it is determined that there are no eye movement components, the processing is transferred to the EEG signal generation unit 17, which reconstructs the EEG signal.

[0029] The electrodes 2 can be electrodes that are normally used for measuring electroencephalograms or electrooculography. The size of the electrodes 2 is, for example, approximately 10 mm x 20 mm. Adhesive stickers or eye masks can be used as jigs (not shown) for placing the electrodes 2 above and below both eyes of the person being measured. The biosignal input unit 3 measures the potential from the electrodes 2, performs A / D conversion, and converts the analog signal into a digital signal, as in a commercially available electroencephalograph, and inputs the digital signal. The electroencephalogram measuring device 1 inputs the signal obtained from the biosignal input unit 3, reconstructs the electroencephalogram signal, and outputs it to the signal output unit 4.

[0030] First, we will explain the EOG signal detection unit 11 in the electroencephalogram measurement device 1. When eye movement occurs, an EOG signal is detected from four electrodes 2 placed above and below both eyes of the person being measured. The EOG signal is generated when there is a standing potential in the eyeball, with a positive potential on the cornea side and a negative potential on the retina side, and when there is eye movement, the potential closest to the electrode is picked up, resulting in a potential difference. The EOG signal detection unit 11 detects the EOG signal using a correlation coefficient for each section of a predetermined length (e.g., 1 second) in the time series data (several minutes to several hours) of the biosignal obtained from the electrode 2 via the biosignal input unit 3.

[0031] As shown in Figure 2, within the same 1-second cutout section of four electrodes (first to fourth electrodes) placed above and below both eyes of the subject, there is a high positive correlation coefficient between the EOG signal section (1-second cutout section) from the first electrode (Fp1) and the EOG signal section from the second electrode (Fp2).Similarly, there is a high positive correlation coefficient between the EOG signal section from the third electrode (PG1) and the EOG signal section from the fourth electrode (PG2). On the other hand, the negative correlation coefficient is high between the EOG signal section from the first electrode (Fp1) and the EOG signal section from the third electrode (PG1).Similarly, the negative correlation coefficient is high between the EOG signal section from the second electrode (Fp2) and the EOG signal section from the fourth electrode (PG2). Therefore, for example, if the correlation coefficient between the EOG signal section from the first electrode (Fp1) and the EOG signal section from the second electrode (Fp2) is greater than or equal to a positive threshold, and the correlation coefficient between the EOG signal section from the first electrode (Fp1) and the EOG signal section from the third electrode (PG1) is less than or equal to a negative threshold, it can be determined that eye movement has occurred, i.e., that an eye movement component has been detected.

[0032] As shown in Fig. 3, the electroencephalogram measuring device of the present invention has three modes for determining that an eye movement component has been detected. The image in Fig. 3(1) shows the first determination mode, in which an eye movement component is determined to have been detected (single determination) when the correlation coefficient of the signals obtained from a pair of electrodes on the upper left and right sides of both eyes among the first to fourth electrodes is equal to or greater than a positive threshold and the correlation coefficient of the signals obtained from a pair of electrodes on the upper and lower sides of the right eye is equal to or less than a negative threshold. Note that, although not shown, the first determination mode may also determine that an eye movement component has been detected when the correlation coefficient of the signals obtained from a pair of electrodes on the lower left and right sides of both eyes is equal to or greater than a positive threshold and the correlation coefficient of the signals obtained from a pair of electrodes on the upper and lower sides of the left eye is equal to or less than a negative threshold.

[0033] 3(2) shows the second determination mode, in which an eye movement component is determined to have been detected (double determination) when the correlation coefficients of the signals obtained from the two pairs of electrodes, the upper and lower ones of both eyes, among the first to fourth electrodes, are both above a positive threshold, and when the correlation coefficients of the signals obtained from the two pairs of electrodes, the upper and lower ones of the left and right eyes, are both below a negative threshold. While the first determination mode makes a single determination using one pair of electrodes, one left and one upper and one lower, the second determination mode makes a double determination using two pairs of electrodes, one left and one upper and two lower, and the determination result of the detection of the eye movement component may differ.

[0034] Furthermore, the image in FIG. 3(3) shows a third determination mode, in which an eye movement component is determined to have been detected (sash determination) when the correlation coefficient of the signals obtained from a pair of electrodes above the left eye and below the right eye among the first to fourth electrodes is equal to or less than a negative threshold. Although not shown, the third determination mode may also determine that an eye movement component has been detected when the correlation coefficient of the signals obtained from a pair of electrodes above the right eye and below the left eye is equal to or less than a negative threshold. Compared to the first and second determination modes, the third determination mode presumably favors the detection of eye movement components when the distance between the electrodes is longer, since this lowers the correlation of the electroencephalograms.

[0035] Fig. 4 visualizes the difference in detection of whether or not a blink is included (processing target or non-processing target) for each predetermined time (1 second) interval in the time-series data of the signal from the electrodes, depending on the first to third determination modes. Fig. 4 shows that in the first determination mode, one blink interval was detected in the time-series data 6 seconds after the start of measurement, in the second determination mode, two blink intervals were detected in the time-series data 3 seconds and 4 seconds after the start of measurement, and in the third determination mode, four blink intervals were detected in the time-series data 2 to 4 seconds and 6 seconds after the start of measurement.

[0036] The processing of the EOG signal detection unit 11 will be described with reference to FIG. 5. The EOG signal detection unit 11 divides the entire period of the biosignal into 1-second intervals (S01). When detecting eye movement components using the first determination mode, the EOG signal detection unit 11 calculates the correlation coefficient between the first electrode (Fp1) and the second electrode (Fp2) and the correlation coefficient between the first electrode (Fp1) and the third electrode (PG1) for each interval (S02). The absolute value of the calculated correlation coefficient is compared with a predetermined threshold (S03). If the absolute value of the correlation coefficient is equal to or greater than the threshold, the interval is determined to be an eye movement positive interval (S04). If the absolute value of the correlation coefficient is less than the threshold, the interval is determined to be an eye movement negative interval (S05). Then, the processing from S02 is repeated for the next 1-second interval, and whether or not the interval is an eye movement positive interval is determined over the entire period of the time series data. 5 shows an example in which the determination is made in the first determination mode, but similarly, the determination can also be made in another determination mode (the second determination mode or the third determination mode). In this way, the occurrence period of eye movement is identified using the first to third determination modes set by the combination of four electrodes (the first to fourth electrodes) arranged above and below both eyes.

[0037] The processing of the signal analysis section 13, the EOG signal removal section 15, and the electroencephalogram signal generation section 17 in the electroencephalogram measurement device 1 will be described with reference to FIG. The signal analysis unit 13 separates eye movement components from the biosignal for a predetermined time (e.g., 5 seconds) including the positive eye movement interval (each second) detected by the EOG signal detection unit 11. For example, principal component analysis can be used to separate the eye movement components. This is because, among the signals obtained from the four electrodes placed above and below both eyes, the signal with the highest signal intensity can be inferred to be an EOG signal. As shown in FIG. 6, the signal analysis unit 13 performs principal component analysis on the predetermined time including the positive eye movement interval (S11), decomposes the signal data for the predetermined time (5 seconds) of the biosignal into principal components (S12), and selects the principal component with the highest correlation coefficient with the original signal data for the predetermined time as the EOG signal (S13). Because a certain sample length is required to ensure the accuracy of principal component analysis, we performed principal component analysis on the biosignal for a predetermined time (e.g., 5 seconds) including the positive eye movement interval (each second).

[0038] In order to remove eye movement components from the biosignal, the EOG signal removal unit 15 estimates the principal components as EOG signals, removes the selected principal components, and reconstructs the data (S14). Then, a threshold is used to determine whether the reconstructed data indicates eye movement (S15). If the threshold is met, the reconstructed data is subjected to the processing of the signal analysis unit 13 and the processing of the EOG signal removal unit 15, i.e., the processing of S11 to S15. This repetition is performed because eye movements in multiple directions (e.g., two directions) within a predetermined time (5 seconds) may be separated into separate principal components. If the result of the determination (S15) is below the threshold, the EEG signal generation unit 17 reconstructs (generates) an EEG signal from the signal from which the eye movement components have been removed.

[0039] FIG. 7 is an explanatory diagram of the processing performed by the signal analysis unit 13. The section in which the EOG signal is detected is a section of time-series data of the biosignal for a predetermined time (5 seconds) including a positive eye movement section (1 second section). FIG. 7 shows an example of a section cut out from the biosignal from the third electrode (PG1) for a predetermined time (5 seconds) including a positive eye movement section. This is then subjected to principal component analysis (PCA) to detect features. The normalized correlation coefficient between each feature and the original signal is evaluated. For example, for the signal from the third electrode (PG1), the correlation coefficient with principal component 1 (PC1) is calculated as 0.7, the correlation coefficient with principal component 2 (PC2) is calculated as 0.2, etc., and the one with the highest normalized correlation coefficient is selected. The selected PC1 can be estimated to be the EOG signal. The EOG signal removal unit 15 removes PC1 and reconstructs the signal.

[0040] Figure 8 shows an example of a screen output from the electroencephalogram measuring device 1 to the signal output unit 4 and displayed on a monitor or the like. In the display example in Figure 8, the biosignal obtained from the electrode 2 and the interval area of ​​the EOG signal detected in each determination mode are shown with two types of dashed lines and one type of solid line, and the positive eye movement interval detected in each of the three determination modes is displayed. A doctor or other person can select the determination mode depending on the detection results. Furthermore, if an erroneous detection occurs, the doctor or other person can point to the correct display position and the display can be re-displayed.

[0041] Figure 9 shows signal waveforms of epileptic discharges as an example of actual EEG measurements. Figure 9(1) shows the signal waveform before signals estimated to be EOG signals are excluded, and Figure 9(2) shows the signal waveform after signals estimated to be EOG signals are excluded by the EEG measurement device 1. In both cases, epileptic discharges are observed in the waveform regions enclosed by lines. Comparing Figures 9(1) and 9(2) reveals that the EEG measurement device 1 is able to accurately detect the occurrence of eye movements and measure epileptic discharges.

[0042] In this example, four biosignals detected by four electrodes (electrodes 1 to 4) placed above and below the subject's eyes are used. However, as shown in FIG. 10, in addition to the first to fourth electrodes (Fp1, Fp2, PG1, PG2), signals obtained from two electrodes (electrodes 5 and 6: T1 and T2) placed in contact with the subject's left and right temples may also be included in the biosignals. Placing additional electrodes near the left and right temples, which reduces the electrode attachment work and places less strain on the patient, can improve the accuracy of the EEG signal. For example, the two electrodes (T1 and T2) can be sealed electrodes placed at the hairline. 11, in addition to the first to sixth electrodes, a signal obtained from a seventh electrode (Fz) placed in contact with the center of the subject's head may be included in the biosignal. By reducing the number of head electrodes, which place a heavy burden on the patient, to one, the patient's burden can be reduced and the accuracy of detecting epileptic waves can be improved. The shape and size of the electrodes are not limited to those shown in the figure, and as shown in Figure 12, in actual measurements, it is necessary to attach reference electrodes (A1, A2) to the left and right earlobes and an earth electrode (Z) to the forehead.

[0043] Figure 13 shows an example of time-series data (waveforms) of biosignals obtained from each electrode (electrodes 1 to 7). The electrode arrangement is as shown in Figure 12. For the first, third, and fifth electrodes located on the left side of the face, the potential of the left earlobe is used as the reference; for the second, fourth, and sixth electrodes located on the right side of the face, the potential of the right earlobe is used as the reference; and for the seventh electrode (Fz) located near the center of the head, the potential of the left and right earlobes is used as the reference. Data for each electrode is obtained by recording the potential difference between the reference electrode and each electrode. Using the earlobe as the reference electrode allows for screening of generalized brain activity. Horizontal connection is a method for measuring the potential difference between two points, allowing you to capture the phase reversal of the signal. [Example]

[0044] FIG. 14 shows a functional block diagram of another embodiment of the electroencephalogram measuring device of the present invention. The electroencephalogram measuring device 10 shown in FIG. 14 further includes an EOG signal removal method selection unit 12 and a display unit 19 in addition to an EOG signal detection unit 11, a signal analysis unit 13, an EOG signal removal unit 15, and an electroencephalogram signal generation unit 17. The EOG signal removal method selection unit 12 can automatically select one of the first to third determination modes described in the first embodiment, or the user (e.g., a doctor) can manually select one of the three determination modes according to the situation. In the case of automatic selection, the determination mode with the highest absolute value of the correlation coefficient calculated in each of the first to third determination modes is determined as the optimal determination mode. In the case of manual selection, the positive eye movement intervals detected in each of the three determination modes are displayed on a screen such as that shown in FIG. 8, and the user is prompted to select one of the first to third determination modes. As already described above, in FIG. 8, the biosignals obtained from the electrodes and the positive eye movement intervals detected in each determination mode are indicated by two types of dashed lines and one type of solid line, allowing the user to distinguish the positive eye movement intervals detected in the first to third determination modes. The display unit 19 displays the screen shown in FIG. 8 to prompt the user to select one of the first to third determination modes. The determination mode can be selected, for example, by operating a button on the screen. It is also possible to display the detection results in different colors for each of the first to third determination modes. In addition, for each determination mode, it is also possible to display a signal from which the EOG signal has been removed. Furthermore, if the position is misaligned in the detection result, the user can correct the display by indicating the correct position of the EOG signal on the screen using the display unit 19. In this way, the occurrence interval of eye movement can be identified using each determination mode, and the determination mode can be selected automatically or manually depending on the situation. [Example]

[0045] (Electroencephalogram measurement system) FIG. 15 shows a functional block diagram of the electroencephalogram measuring system of the present invention. The electroencephalogram measurement system of the present invention comprises an electroencephalogram measurement device including at least first to fourth electrodes 2, a biosignal input unit 3 to which biosignals obtained from these electrodes are input, a jig 5 for placing these electrodes 2 above and below both eyes of the subject, and an electroencephalogram signal detection unit 11, a signal analysis unit 13, an EOG signal removal unit 15, and an electroencephalogram signal generation unit 17. In the electroencephalogram measurement device, the biosignal input unit 3 measures potentials from the electrodes 2 and performs A / D conversion. The jig 5 is not particularly limited as long as it can maintain the electrodes in the same positions despite eye or facial movements. However, preferred are adhesive stickers that can securely attach electrodes to the skin, eye masks that can be attached to both ears and allow electrodes to be placed above and below both eyes, or eyeglasses or goggles that have electrodes arranged on a frame and can be worn on the face. [Example]

[0046] (Electroencephalogram measurement program) FIG. 16 shows a flow diagram of the electroencephalogram measurement program of the present invention. The EEG measurement program is a program that measures EEG signals from biological signals including EOG signals and EEG signals obtained from four electrodes (first to fourth electrodes) placed above and below both eyes of the subject, and is a program that causes a computer to execute an EOG signal detection step, a signal analysis step, an EOG signal removal step, and an EEG signal generation step.

[0047] In the EOG signal detection step, the entire section of the biosignal obtained from the electrodes is divided into sections of a predetermined length (e.g., every 1 second) (S101), and a correlation coefficient is calculated for each section using the first to third determination modes described in the above embodiment (S102). The absolute value of the correlation coefficient is compared with a threshold (S103). If the correlation coefficient is equal to or greater than the threshold (≧threshold), the section is determined to be a positive eye movement section (S104). If the correlation coefficient is less than the threshold (<threshold), the section is determined to be a negative eye movement section (S105). If the determination has been made for the entire section, the EOG signal detection process ends (S106), and the process proceeds to the next signal analysis step.

[0048] In the signal analysis step, principal component analysis is performed for a predetermined time period including positive eye movement (S111), the signal data is decomposed into principal components (S112), and the component with the highest correlation coefficient with the original signal data for the predetermined time period is selected (S113).

[0049] Then, in the EOG signal removal step, the selected principal component is removed and the data is reconstructed (S121), and the reconstructed data is judged using a threshold to determine whether or not it indicates eye movement (S122). If the reconstructed data is judged to be equal to or greater than the threshold (≧threshold), the signal analysis steps (S111 to S113) and the EOG signal removal steps (S121, S122) are repeated for the reconstructed data. If the reconstructed data is judged to be less than the threshold (<threshold), the process proceeds to the EEG signal generation step, where an EEG signal is reconstructed based on a signal obtained by removing the eye movement component from the biosignal obtained from the electrodes (S131). [Industrial Applicability]

[0050] The present invention is useful for measuring electroencephalographic signals. [Explanation of symbols]

[0051] 1,10 Electroencephalogram (EEG) measuring device 2 electrodes 3. Biological signal input section 4 Signal output section 5 Jig 11 EOG signal detection unit 12 EOG signal removal method selection section 13 Signal analysis section 15 EOG signal removal section 17 EEG signal generation unit 19 Display section 100 EEG measurement system

Claims

1. An apparatus for measuring an electrooculogram (EOG) signal and an electroencephalogram signal from biosignals including an electroencephalogram signal obtained from first to fourth electrodes arranged above and below both eyes of a subject, an EOG signal detection unit that detects an eye movement component for each section of a predetermined length in the biosignal obtained from the electrodes using a correlation coefficient of the biosignal; a signal analysis unit that separates eye movement components from the biological signal for a predetermined time period including a positive eye movement section in which an EOG signal is detected; an EOG signal removal unit that removes eye movement components from the biological signal; an electroencephalogram signal generator that reconstructs an electroencephalogram signal from the signal from which the eye movement component has been removed; An electroencephalogram measuring device comprising:

2. In the EOG signal removal unit, The electroencephalogram measuring device of claim 1, characterized in that it repeatedly performs signal analysis to separate eye movement components from a signal from which eye movement components have been removed, and if it is determined that there are no eye movement components, it reconstructs the electroencephalogram signal.

3. 2. The electroencephalogram measuring device according to claim 1, wherein the signal analysis for separating eye movement components is principal component analysis.

4. The electroencephalogram measuring device of any one of claims 1 to 3, characterized in that the EOG signal detection unit has a first determination mode in which it determines that an eye movement component has been detected when the correlation coefficient of the biological signals obtained from a pair of electrodes on the upper or lower sides of both eyes, among the first to fourth electrodes, is greater than or equal to a predetermined positive value, and when the correlation coefficient of the biological signals obtained from a pair of electrodes on the upper or lower sides of the left eye or right eye is less than or equal to a predetermined negative value.

5. The electroencephalogram measuring device of claim 4, characterized in that the EOG signal detection unit has a second determination mode in which it determines that an eye movement component has been detected when the correlation coefficient of the biological signals obtained from two pairs of electrodes, one on the upper and one on the lower side of each eye, among the first to fourth electrodes, is equal to or greater than a predetermined positive value, and when the correlation coefficient of the biological signals obtained from two pairs of electrodes, one on the upper and one on the right eye, is equal to or less than a predetermined negative value.

6. The electroencephalogram measuring device of claim 5, characterized in that the EOG signal detection unit has a third determination mode in which it determines that an eye movement component has been detected when the correlation coefficient of the biosignal obtained from a pair of electrodes above the left eye and below the right eye among the first to fourth electrodes is equal to or less than a predetermined negative value, and / or when the correlation coefficient of the biosignal obtained from a pair of electrodes above the right eye and below the left eye is equal to or less than a predetermined negative value.

7. The electroencephalogram measuring device of claim 5, further comprising an EOG signal removal method selection unit that automatically selects the judgment mode having a higher absolute value of the correlation coefficient calculated in the first and second judgment modes as the optimal judgment mode.

8. The electroencephalogram measuring device of claim 6, further comprising an EOG signal removal method selection unit that automatically selects the judgment mode having the highest absolute value of the correlation coefficient calculated in the first to third judgment modes as the optimal judgment mode.

9. 7. The electroencephalogram measuring device according to claim 6, further comprising an EOG signal removal method selection unit that allows a user to select one determination mode from the first to third determination modes.

10. An electroencephalogram (EEG) measuring device as described in any one of claims 1 to 3, further comprising a display unit that displays the positive eye movement interval detected by the EOG signal detection unit by superimposing it on the biological signal.

11. The electroencephalogram measuring device of claim 10, characterized in that the display unit accepts a user's instruction to correct the section in the displayed detection result of the positive eye movement section, and re-displays the section reflecting the correction instruction.

12. An electroencephalogram measuring device as described in any one of claims 1 to 3, characterized in that in the EOG signal detection unit, in addition to the first to fourth electrodes, signals obtained from fifth and sixth electrodes contacted to the left and right temples of the person being measured are included in the biosignal.

13. The electroencephalogram measuring device of claim 12, characterized in that in the EOG signal detection unit, in addition to the first to sixth electrodes, a signal obtained from a seventh electrode contacted near the center of the scalp of the person being measured is included in the biological signal.

14. First to fourth electrodes; a biosignal input unit to which a biosignal obtained from the electrode is input; A jig for placing the electrodes above and below both eyes of the subject; The electroencephalogram measuring device according to any one of claims 1 to 3, An electroencephalogram measurement system comprising:

15. A program for measuring an electrooculogram (EOG) signal and an electroencephalogram signal from biosignals including an electroencephalogram signal obtained from first to fourth electrodes arranged above and below both eyes of a subject, an EOG signal detection step of detecting an eye movement component for each section of a predetermined length in the biosignal obtained from the electrodes using a correlation coefficient of the biosignal; a signal analysis step of separating eye movement components from the biological signal for a predetermined period of time including a positive eye movement section in which an EOG signal is detected; an EOG signal removal step of removing an eye movement component from the biological signal; an electroencephalogram signal generating step of reconstructing an electroencephalogram signal from the signal from which the eye movement component has been removed; A brain wave measurement program that allows a computer to execute the above.

16. In the EOG signal removal step, The electroencephalogram measurement program of claim 15, characterized in that it repeats signal analysis to separate eye movement components from a signal from which eye movement components have been removed, and if it is determined that there are no eye movement components, it reconstructs the electroencephalogram signal.

Images