Brain wave analysis device, brain wave analysis program, exercise assistance system, and exercise assistance method

Abstract

The invention relates to a brain wave analysis device, a brain wave analysis program, an exercise assistance system, and an exercise assistance method. A brain wave analysis device (16) is provided with: a signal acquisition unit (40) for acquiring the timing of brain wave signals of a subject (12) to be analyzed; and a calculation unit (58) for obtaining a natural frequency relating to the motion intention of the analyst (12) on the basis of the time-series correlation frequency characteristics of the brain wave signal acquired when the analyst (12) is resting.

Description

Technical Field

[0001] The present invention relates to an electroencephalogram analysis device, an electroencephalogram analysis program, a movement assistance system, and a movement assistance method. Background Art

[0002] Conventionally, in various fields such as medical care, health, nursing, and sports, there have been technologies that can detect the actions or states of a subject to be analyzed through various sensors, and use the obtained detection data to analyze the movement state, etc. of the subject to be analyzed. For example, there is a technology called a brain-computer interface that can analyze electroencephalogram signals, then identify the characteristic frequencies representing the movement intention or brain state of the subject to be analyzed from them, and control mechanical equipment based on the signal intensity of the characteristic frequencies included in the electroencephalogram signals (refer to Non-Patent Document 1, etc.).

[0003] Figure 12 is a schematic diagram of a conventional method for measuring the characteristic frequency of a subject to be analyzed. The horizontal axis in the figure represents time (unit: s), and the vertical axis represents frequency (unit: Hz). Herein, "time" represents the time elapsed from the resting state to the movement imagination state, and t = 0 corresponds to the moment of state transition. In addition, the darkness of the color in the figure represents the magnitude of the signal intensity at the corresponding time and corresponding frequency, and the lighter the color, the greater the signal intensity.

[0004] As can be seen from this figure, shortly after the transition to the movement imagination state (t > 0), a band-shaped region with a relatively low signal intensity appears on the figure, and this region extends in the time axis direction. The characteristic frequency corresponding to this band-shaped region is called the individual SMR-ERD frequency (Individual Sensorimotor Rhythm Event-Related Desynchronization Frequency, hereinafter referred to as "ISF"), and it is known that it has a high correlation with the movement intention.

[0005] Prior Art Documents

[0006] Non-Patent Documents

[0007] [Non-Patent Document 1] "Neurophysiological predictor of SMR-based BCI performance", B. Blankertz et al., NeuroImage, Volume 51, p1303 - 1309, 2010 Summary of the Invention

[0008] Technical Problem to be Solved by the Invention

[0009] However, when using the above-mentioned existing measurement methods, in order to improve the identification accuracy of the natural frequency, during the process of transitioning from the resting state to the motor imagery state, it is necessary to increase the number and duration of "trials". Therefore, not only does the identification time of the natural frequency become longer, but the mental and physical burdens on the subject being analyzed also increase.

[0010] Specifically, assuming that the resting state lasts for 5 seconds, the motor imagery state lasts for 5 seconds, and the number of trials is 20 times, then the measurement time of the natural frequency is approximately 3 minutes. Especially for patients with paralyzed parts, the time required to move the relevant parts is longer. For example, the measurement time may be as long as 15 minutes. As the measurement time extends, the mental and physical burdens on the subject being analyzed also increase.

[0011] In view of the above problems, the present invention is proposed. The object of the present invention is to provide an electroencephalogram analysis device and an electroencephalogram analysis program, which can significantly shorten the time required for the identification process when analyzing the electroencephalogram signal of the subject being analyzed and identifying the natural frequency related to the motor intention or brain state. In addition, a movement assistance system and a movement assistance method are provided.

[0012] Technical means for solving the problem

[0013] The electroencephalogram analysis device in the first aspect of the present invention includes the following parts: an acquisition unit for acquiring the time series of the electroencephalogram signal of the subject being analyzed; an operation unit for obtaining the natural frequency related to the motor intention or brain state of the subject being analyzed based on the relevant frequency characteristics of the time series of the electroencephalogram signal acquired by the acquisition unit when the subject being analyzed is at rest.

[0014] The electroencephalogram analysis device in the second aspect of the present invention further includes the following parts: a calculation unit for calculating the sample value of the peak frequency in the frequency characteristics; an estimation unit for obtaining the estimated value of the peak frequency based on the population of the sample values calculated by the calculation unit. Among them, the operation unit converts the estimated value obtained by the estimation unit into the natural frequency according to a pre-defined conversion rule.

[0015] The electroencephalogram analysis device in the third aspect of the present invention obtains the estimated value through the estimation unit and converts the estimated value into the natural frequency through the operation unit, so that the natural frequency can be obtained without acquiring the electroencephalogram signal when the subject being analyzed has a motor intention.

[0016] The electroencephalogram analysis device in the fourth aspect of the present invention obtains the estimated value through the estimation unit and converts the estimated value into the natural frequency through the operation unit, so that the natural frequency can be obtained when the subject being analyzed moves without using the paralyzed part.

[0017] In the electroencephalogram analysis device according to the fifth aspect of the present invention, starting from the moment when the electroencephalogram signal measurement begins, acquisition is performed by the acquisition unit and calculation is performed by the calculation unit every unit time, so as to accumulate the overall sample values. Based on the overall sample values, the estimation unit uses the sequential Bayesian method to obtain the estimation value for each unit time.

[0018] The electroencephalogram analysis device according to the sixth aspect of the present invention further includes a determination unit, which is used to judge whether an end condition is satisfied each time the estimation unit performs an estimation. When the determination unit judges that the end condition is satisfied, the estimation unit ends the estimation of the peak frequency.

[0019] In the electroencephalogram analysis device according to the seventh aspect of the present invention, the peak frequency is the alpha frequency within the alpha band, the natural frequency is the individual SMR-ERD frequency, and the conversion rule is expressed in the form of an identity function or a linear function, where the identity function or the linear function takes the estimation value as the independent variable.

[0020] In the electroencephalogram analysis device according to the eighth aspect of the present invention, the conversion rule is defined according to the specific situation of the person to be analyzed.

[0021] In the electroencephalogram analysis device according to the ninth aspect of the present invention, the acquisition unit acquires the first time series of the electroencephalogram signal measured when the person to be analyzed is at rest or the second time series of the electroencephalogram signal measured when the person to be analyzed performs motor imagery. The operation unit only uses the first time series to perform a first operation to obtain the natural frequency.

[0022] In the electroencephalogram analysis device according to the tenth aspect of the present invention, the operation unit switches between the first operation and the second operation, where the second operation uses both the first time series and the second time series to obtain the natural frequency.

[0023] The electroencephalogram analysis device according to the eleventh aspect of the present invention further includes a prompting unit, which is used to send a prompting message to the person to be analyzed to prompt the person to be analyzed to maintain a resting state.

[0024] The electroencephalogram analysis device according to the twelfth aspect of the present invention further includes an auxiliary control unit, which controls a motion assistance device based on the natural frequency obtained by the operation unit, so as to assist the motion of the person to be analyzed.

[0025] In the electroencephalogram analysis program according to the thirteenth aspect of the present invention, one or more computers are caused to execute an acquisition step for acquiring the time series of the electroencephalogram signal of the person to be analyzed, and an operation step for obtaining the natural frequency related to the motion intention or brain state of the person to be measured based on the relevant frequency characteristics of the time series of the electroencephalogram signal acquired when the person to be analyzed is at rest.

[0026] The motion assistance system in the fourteenth aspect of the present invention includes the following devices: the electroencephalogram analysis device in the above twelfth aspect; an electroencephalograph for measuring the electroencephalogram of the person to be analyzed and sending the obtained electroencephalogram signal to the electroencephalogram analysis device; a motion assistance device that is controlled by the electroencephalogram analysis device to perform related operations, thereby assisting the person to be analyzed in exercising.

[0027] The motion assistance method in the fifteenth aspect of the present invention uses a system having the following devices: an electroencephalograph for measuring the electroencephalogram of the person to be analyzed and outputting an electroencephalogram signal; an electroencephalogram analysis device for analyzing the electroencephalogram signal sent from the electroencephalograph; a motion assistance device that is controlled by the electroencephalogram analysis device to perform related operations, thereby assisting the person to be analyzed in exercising. Among them, the method includes the following steps: an acquisition step, in which the electroencephalogram analysis device uses the electroencephalograph to measure the electroencephalogram of the person to be analyzed at rest and acquires the time sequence of the electroencephalogram signal; an operation step, in which the electroencephalogram analysis device obtains the natural frequency related to the motion intention or brain state of the person to be analyzed according to the relevant frequency characteristics of the acquired time sequence of the electroencephalogram signal; a calibration step, in which the obtained natural frequency is set as the calibration parameter of the motion assistance device and calibration is performed; an assistance step, in which the calibrated motion assistance device is controlled to operate, thereby assisting the person to be analyzed in exercising.

[0028] Advantages of the Invention

[0029] According to the present invention, when analyzing the electroencephalogram signal of the person to be analyzed and identifying the natural frequency related to the motion intention, the time required for identification can be significantly shortened. Description of the Drawings

[0030] Figure 1 is the overall structure diagram of the BMI system integrating the electroencephalogram analysis device in an embodiment of the present invention.

[0031] Figure 2 is Figure 1 the functional module diagram of the shown processor and memory.

[0032] Figure 3 is using Figure 1 the flowchart of the motion assistance method of the shown BMI system.

[0033] Figure 4 is Figure 1 the flowchart of the analysis operation of the shown electroencephalogram analysis device.

[0034] Figure 5 is the graph of the change of the electroencephalogram signal over time in the resting state.

[0035] Figure 6 It is a schematic diagram of the calculation method of the IAF sample value.

[0036] Figure 7 It is a schematic diagram of the convergence of the IAF estimated value obtained by the sequential Bayesian method.

[0037] Figure 8 It is a schematic diagram of the time shortening effect generated by the abort handling at convergence.

[0038] Figure 9 It is a schematic diagram of the correlation between IAF and ISF.

[0039] Figure 10 It is a probability density distribution diagram of the deviation between IAF and ISF.

[0040] Figure 11 It is a schematic diagram of the effect of the electroencephalogram analysis method in this embodiment.

[0041] Figure 12 It is a schematic diagram of the existing method for measuring the natural frequency of the subject to be analyzed. Detailed implementation mode

[0042] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. For the sake of easy understanding, the same constituent elements and steps in each figure will be denoted by the same reference numerals as much as possible, and repeated descriptions will be omitted.

[0043] [Structure of BMI system 10]

[0044] [Overall structure]

[0045] Figure 1 It is an overall structure diagram of a brain-computer interface system (hereinafter referred to as "BMI system 10") integrating the electroencephalogram analysis device 16 in an embodiment of the present invention. The BMI system 10 can analyze the electroencephalogram emitted by the subject 12 and assist the subject 12 in movement according to the analysis result. Specifically, the BMI system 10 includes an electroencephalograph 14, an electroencephalogram analysis device 16, and a movement assistance device 18.

[0046] For example, the electroencephalograph 14 can be a head-mounted device that can measure the electroencephalogram emitted by the head 12h of the subject 12. The electroencephalograph 14 detects an electrical signal through electrodes (not shown in the figure) and outputs it to the electroencephalogram analysis device 16.

[0047] The electroencephalogram analysis device 16 is a computer that can analyze the brain states of the person to be analyzed 12, such as movement intention, fatigue, and cognition, based on the electroencephalogram signals measured by the electroencephalograph 14. Specifically, the electroencephalogram analysis device 16 includes an operation unit 22, a prompting unit 23, a sensor controller 24, a processor 26, and a memory 28.

[0048] The operation unit 22 enables users (including the person to be analyzed 12 and medical workers) to perform various operations. The operation unit 22 can be an input device including operation buttons, microphones, etc., or an output device including a display panel, speakers, etc.

[0049] The prompting unit 23 is an output device that, according to the instructions of the processor 26, sends a prompting message (hereinafter referred to as "request message") to the person to be analyzed 12 to prompt the person to be analyzed to maintain a resting state. For example, the prompting unit 23 is composed of components such as a display panel, a lamp, and speakers. The prompting methods of the request message include the following: text guidance or voice guidance, lighting up the lamp, outputting various sounds, etc. In addition, the prompting of the request message is not limited to being performed by the prompting unit 23 of the electroencephalogram analysis device 16, and can also be prompted by someone other than the person to be analyzed 12 (such as the operator of the electroencephalogram analysis device 16).

[0050] The sensor controller 24 is a control circuit for performing various controls on the electroencephalograph 14. For example, the sensor controller 24 can perform various signal processes including sampling process, low-pass filtering process, A / D conversion process, etc. including sensor synchronization process. Therefore, the sensor controller 24 can obtain an electrical signal representing the electroencephalogram of the person to be analyzed 12 (i.e., electroencephalogram signal) at a certain sampling interval, and send the electroencephalogram signal to the processor 26. The specific sampling interval can be a value between several tens of milliseconds and several hundreds of milliseconds.

[0051] The processor 26 is used to control each component of the electroencephalogram analysis device 16 as a whole. The processor 26 can be a general-purpose processor including a CPU (Central Processing Unit) or an MPU (Micro-Processing Unit), or a dedicated processor including an FPGA (Field Programmable Gate Array) or a GPU (Graphics Processing Unit).

[0052] The memory 28 is a non-volatile storage medium including a ROM (Read Only Memory) or a RAM (Random access memory), and is used to store the programs and data required for the processor 26 to control each component.

[0053] The motion assistance device 18 is a device for assisting or helping the subject 12 in moving the target part (in this figure example, it is the arm 12a). In addition to the arm 12a, the target part can also be the hand, foot, finger, knee, elbow, etc., as well as various body parts that can perform stretching / bending movements. The motion assistance device 18 can be a "wearable robot" that assists the subject 12 in performing stretching / bending movements of the body part by driving an actuator, or an "illusion induction device" that gives an illusion stimulus through vision or touch to assist the patient in performing stretching / bending movements of the body part.

[0054] <Functional module>

[0055] Figure 2 is Figure 1 The functional module diagram of the shown processor 26 and memory 28. The processor 26 works as a signal acquisition unit 40 (equivalent to the "acquisition unit"), a frequency identification unit 42, and an assistance control unit 44 by reading and executing an electroencephalogram analysis program from the memory 28.

[0056] The signal acquisition unit 40 acquires the time series of the electroencephalogram signal of the subject 12 through the sensor controller 24 ( Figure 1 ). In this way, the electroencephalogram signals within a unit time are acquired one by one. The length of the unit time can be equal to the sampling interval or an integer multiple of the sampling interval. The time series of the electroencephalogram signal includes the time series measured when the subject 12 is at rest (hereinafter referred to as the "first time series") and the time series measured when the subject 12 performs motor imagery (hereinafter referred to as the "second time series").

[0057] The frequency identification unit 42 analyzes the electroencephalogram signal acquired by the signal acquisition unit 40 to identify the characteristic frequency related to the motion intention of the subject 12. The characteristic frequency can be the individual ERD frequency (Event-related desynchronization Frequency) or the individual SMR-ERD frequency (i.e., ISF), etc.

[0058] Among them, the "SMR-ERD frequency" refers to the frequency within the 8 - 13 Hz alpha band of the scalp electroencephalogram measured near the motor cortex, where the event-related desynchronization (ERD) of the movement-related response is more significant. The SMR-ERD frequency varies between individuals and fluctuates within the range of 8 - 13 Hz. Therefore, in order to clearly indicate that it is an individual-specific frequency, it is usually called the individual SMR-ERD frequency (i.e., ISF).

[0059] The frequency identification unit 42 specifically includes a preprocessing unit 50, a calculation unit 52, an estimation unit 54, a determination unit 56, and an operation unit 58.

[0060] The preprocessing unit 50 preprocesses the time series of the electroencephalogram signals acquired by the signal acquisition unit 40, which is a necessary step for calculating the Individual Alpha Frequency (hereinafter referred to as "IAF"). This preprocessing includes: [1] "filtering processing" including moving average; [2] "frequency conversion processing" including FFT (Fast Fourier Transform); [3] "trend removal processing" for removing 1 / f noise from the frequency characteristics (or power spectrum), etc.

[0061] Among them, the "alpha frequency" refers to the frequency corresponding to the peak signal intensity in the 8 - 13Hz alpha frequency band in the scalp electroencephalogram that reflects the collective activities of brain nerve cells. The alpha frequency varies among individuals and fluctuates within the range of 8 - 13Hz. Therefore, to clearly indicate that it is a frequency unique to an individual, it is usually called the individual alpha frequency (i.e., IAF).

[0062] The calculation unit 52 calculates the IAF of the subject 12 based on the frequency characteristics of the electroencephalogram signals obtained by the preprocessing unit 50, thereby obtaining the sample value for each time series (hereinafter referred to as "IAF sample value"). Specifically, the calculation unit 52 detects the maximum peak within a specific frequency band (here the alpha frequency band) in the frequency characteristics, and calculates the frequency corresponding to this maximum peak as the IAF sample value. In addition, in addition to the alpha frequency band (8 - 13Hz), at least one of the delta frequency band (1 - 3Hz), theta frequency band (3 - 7Hz), beta frequency band (14 - 30Hz), and gamma frequency band (above 30Hz) can also be selected as the specific frequency band, and which one to select can be determined according to the analysis object.

[0063] The estimation unit 54 uses the totality of the IAF sample values calculated by the calculation unit 52 to estimate the IAF of the subject 12, thereby obtaining the estimated value for each totality (hereinafter referred to as "IAF estimated value"). As an estimation method, various statistical methods such as the Bayesian method or the sequential Bayesian method (or Kalman filter) can be used. For example, the estimation unit 54 can use the sequential Bayesian method to obtain the estimated value for each unit time based on the totality accumulated since the start time of the electroencephalogram signal measurement.

[0064] The determination unit 56 determines whether the IAF estimated value satisfies the end condition each time the estimation unit 54 performs an estimation. If the end condition is satisfied, it instructs the estimation unit 54 to end the estimation process. As an example, the end conditions can include the following: [Condition 1] The IAF estimated value has converged (for example, the change amount does not exceed a certain threshold); [Condition 2] The number of estimations by the estimation unit 54 exceeds a certain threshold; [Condition 3] A certain amount of time has passed since the start time of the electroencephalogram signal measurement; [Condition 4] A combination of the above Conditions 1 - 3, etc.

[0065] When the determination unit 56 determines that the end condition is satisfied, the operation unit 58 converts the most recently obtained IAF estimated value into an inherent frequency (here, ISF) related to the motion intention of the subject 12 according to a pre-defined conversion rule, thereby obtaining a conversion value for each parsing operation (hereinafter referred to as "ISF conversion value"). This conversion rule can be a general rule applicable to the subject 12, or different rules formulated according to the specific situation of the subject 12. In addition, when the conversion rule is a function with IAF as the independent variable, the function shape can be: [1] linear functions such as identity functions and linear functions; [2] non-linear functions such as polynomial functions with an exponent greater than or equal to 2 and exponential functions.

[0066] It should be noted that the operation unit 58 only uses the above first time sequence to perform the arithmetic processing for solving ISF (hereinafter referred to as "first operation"), but at the same time, it can also use both the first time sequence and the second time sequence to perform the arithmetic processing for solving ISF (hereinafter referred to as "second operation"). In this case, the operation unit 58 can switch between the first operation and the second operation as needed. Regarding the switching between these two operations, for example, it can be manually switched through the input operation of the operation unit 22( Figure 1 ), or automatically switched according to the analysis result of the electroencephalogram signal acquired by the signal acquisition unit 40.

[0067] The auxiliary control unit 44 controls the motion assist device 18 according to the ISF conversion value obtained by the operation unit 58 (hereinafter referred to as "auxiliary control"). The auxiliary control unit 44 includes the following parts: a setting unit 60 for setting an ISF set value suitable for the subject 12; a determination unit 62 for determining the control amount of the motion assist device 18 according to the frequency characteristics of the electroencephalogram signal and the ISF set value set by the setting unit 60.

[0068] On the other hand, the subject information 70, frequency information 72, and conversion information 74 are stored in the memory 28, and these information correspond to each other.

[0069] The subject information 70 includes various information related to the subject 12. For example, it includes the identification information of the subject 12, personal information, the diagnosis result of the subject 12, the degree of recovery, the type and usage record of the motion assist device 18, etc. The frequency information 72 includes various information related to the peak frequency or inherent frequency. For example, it includes IAF sample values, IAF estimated values, and ISF conversion values, etc. The conversion information 74 includes various information that can determine the conversion rule. For example, it includes the type of function shape, coefficients, degrees, LUT (look-up table), etc.

[0070] [Operation of the BMI system 10]

[0071] The specific construction of the BMI system 10 in this embodiment is as described above. Next, with reference to Figure 3 and Figure 4 the flowchart of Figures 5 to 11 , the operation of the BMI system 10 will be described in detail, especially the movement assistance operation of the electroencephalogram analysis device 16.

[0072] <Explanation of the movement assistance method>

[0073] Figure 3 is a flowchart of the movement assistance method of the BMI system 10 shown in Figure 1 . In step SP100, the "wearing" process is executed to wear the electroencephalogram device 14 on the head of the subject 12. In step SP102, the "start" process is executed to monitor the electroencephalogram emitted by the subject 12. In step SP104, the "confirmation" process is executed to check whether the subject 12 is in a resting state. If the subject 12 is not in a resting state (step SP104: NO), the process remains at step SP104 until the subject 12 is in a resting state. If the subject 12 is in a resting state (step SP104: YES), the process proceeds to the next step, which is step SP106.

[0074] In step SP106, the "calibration" process is executed to calibrate the movement assistance device 18. Specifically, after step SP102 is completed, the electroencephalogram analysis device 16 analyzes the electroencephalogram signals sent one by one from the electroencephalogram device 14 to obtain the inherent frequency (i.e., ISF) of the subject 12, and sets this ISF value as the calibration parameter of the movement assistance device 18.

[0075] In step SP108, the "assistance" process is executed to perform neurorehabilitation treatment on the subject 12. Specifically, the electroencephalogram analysis device 16 controls the movement assistance device 18 calibrated in step SP106 to operate. By operating the movement assistance device 18, assistance is provided to the movement of the subject 12.

[0076] <Analysis operation of the electroencephalogram analysis device 16>

[0077] Figure 4 is Figure 1Flowchart of the analysis operation of the brain wave analysis device 16 shown. For example, before performing rehabilitation treatment on the arm 12a of the person to be analyzed 12, the brain wave analysis device 16 executes this flowchart to identify the inherent frequency (i.e., ISF) of the person to be analyzed 12. This flowchart is not only executed once at the start of the rehabilitation treatment, but can also be executed one or more times during the process as needed. In addition, this operation can be directly confirmed through the analysis of the brain wave program, or indirectly confirmed by comparing the ideal signal waveform analog-input to the electroencephalograph 14 with the output result of the brain wave analysis device 16.

[0078] In Figure 4 step SP10, the signal acquisition unit 40 acquires the time series of the brain wave signals per unit time through the electroencephalograph 14 and the sensor controller 24.

[0079] Figure 5 is a graph of the change of the brain wave signal over time in the resting state. In the graph, the horizontal axis represents time (unit: s), and the vertical axis represents the brain potential (unit: mV). It can be seen from this graph that the brain wave signal fluctuates slightly above and below the reference value, presenting a complex waveform.

[0080] In Figure 4 step SP12, the frequency identification unit 42 (specifically, the preprocessing unit 50) preprocesses the time series of the brain wave signals acquired in step SP10. Through this step, the frequency characteristics of the brain wave signals can be obtained.

[0081] In step SP14, the frequency identification unit 42 (specifically, the calculation unit 52) calculates the IAF sample value based on the frequency characteristics obtained in the preprocessing of step SP12.

[0082] Figure 6 is a schematic diagram of the calculation method of the IAF sample value. In the graph, the horizontal axis represents frequency (unit: Hz), and the vertical axis represents signal intensity (unit: dimensionless). It can be seen from this graph that this frequency characteristic has a maximum peak near 11 [Hz]. That is to say, the calculation result of the IAF sample value is 11 [Hz].

[0083] In Figure 4 step SP16, the processor 26 detects whether the estimation timing of the IAF (hereinafter referred to as "estimation timing") has arrived. If the estimation timing has not arrived (step SP16: NO), the processor 26 will return to step SP10 and sequentially repeat steps SP10 - SP16 until the estimation timing arrives. If the estimation timing has arrived (step SP16: YES), the processor 26 will enter the next step, that is, step SP18.

[0084] In step SP18, the frequency identification unit 42 (specifically, the estimation unit 54) estimates the IAF by applying the sequential Bayesian method cumulatively calculated one by one in step SP14. For example, according to the Bayesian method, when calculating the posterior probability p(IAF = A|Data), the prior probability p(IAF = A) is used for calculation according to the following formula (1). It should be noted that in the sequential Bayesian method, when calculating the current posterior probability, the previous posterior probability needs to be regarded as the current prior probability.

[0085]

Formula 1

[0086]

[0087] In step SP20, the frequency identification unit 42 (specifically, the determination unit 56) determines whether the IAF estimated value obtained in step SP18 satisfies the end condition. If the end condition is not satisfied (step SP20: NO), the processor 26 returns to step SP10 and sequentially repeats steps SP10 - SP20 until the end condition is satisfied.

[0088] Figure 7 It is a schematic diagram of the convergence of the IAF estimated value obtained by the sequential Bayesian method. In the figure, the horizontal axis represents time (unit: s), and the vertical axis represents the estimation error (unit: Hz). Among them, "time" is the time elapsed from the time point (t = 0) when the sequential Bayesian method starts estimation. And "estimation error" is the difference (i.e., deviation) between the estimated value of the IAF and the actual value.

[0089] The solid line in the figure represents the average value of the data of 180 healthy adults. In addition, the lower dotted line is the boundary line of "average value - 2σ" (σ: standard deviation), and the upper dotted line is the boundary line of "average value + 2σ". It can be seen from this figure that although the estimation error is large at the beginning, as time goes by, the estimation error will decrease exponentially and finally converge to zero or a value close to zero.

[0090] For example, assuming that the change of the IAF estimated value follows a Gaussian distribution, then by setting the estimation time to 26 s, the estimation error of 95.7% (i.e., the range of 4σ) of the people among healthy adults can be controlled within 1 Hz. Compared with Figure 12 the existing measurement methods (several minutes to more than ten minutes) shown, the required time is greatly shortened.

[0091] Figure 8 is a schematic diagram of the time shortening effect caused by the abort processing at convergence. The horizontal axis of the histogram represents the estimation error of an individual (refer to Figure 6)The estimated time required to control within 1 Hz, i.e., the convergence time (unit: s). According to this histogram, the frequency is the highest when the convergence time is within 4 s, and as the convergence time increases, the frequency gradually decreases.

[0092] As can be seen from this figure, the median of the convergence time is equivalent to "11 s", and the 2σ boundary line of the convergence time is equivalent to "30 s". That is to say, when the estimation error of 95.7% of healthy adults is controlled within 1 Hz, the abort process can be implemented at convergence (complying with the above [Condition 1]), thereby further shortening the estimated time by approximately 2 / 3 (30 s - 11 s = 19 s).

[0093] On the other hand, returning to Figure 4 step SP20 in

[0094] In Figure 4 step SP22 of

[0095]

Formula 2

[0096] ISF = g(IAF)…(2)

[0097] Figure 9 is a schematic diagram of the correlation between IAF and ISF. This figure is based on the counting results of IAF and ISF data pairs of 187 healthy adults, and is represented by the shade of color in the grid. The lighter the color (closer to white), the larger the counting value. As can be seen from this figure, there is a certain positive correlation between ISF and IAF (r = 0.62).

[0098] Figure 10 is the probability density distribution diagram of the deviation between IAF and ISF. Among them, the "deviation" is defined as Δ = IAF - ISF (unit: Hz). As can be seen from this figure, this distribution has the highest probability at the deviation Δ = 0 Hz.

[0099] That is to say, the correlation shown in Figure 9 and Figure 10 can be referred to, and ISF can be calculated according to the following formula (3). In this case, formula (2) becomes g(x) = x, indicating that the so-called identity transformation holds.

[0100]

Formula 3

[0101] ISF = IAF…(3)

[0102] In Figure 4 in step SP22, the auxiliary control unit 44 (specifically, the setting unit 60) sets the ISF conversion value obtained in step SP20. Through this step, the auxiliary control unit 44 can control the motion assistance device 18 to meet the needs of the subject 12 to be analyzed.

[0103] <Effect of the above method>

[0104] Figure 11 is a schematic diagram showing the effect of the electroencephalogram analysis method in this embodiment. The horizontal axis in the figure represents the time elapsed since the start of electroencephalogram measurement (unit: s). The upper bar graph shows the specific situation of the ISF recognition time in the "comparative example" (refer to Figure 12 ). The lower bar graph shows the specific situation of the ISF recognition time in the "example" (refer to Figure 3 and Figure 4 ).

[0105] In the "existing example", if the time to reach the resting state is T1, the time to maintain the motor imagery state is T2, and the number of measurement executions is 20 times, then the measurement time of the natural frequency is 20Tc (where Tc = T1 + T2). In particular, for the subject 12 with a paralyzed part, the time required to move that part is relatively long. For example, the measurement time may be as long as 200 s to 1000 s. As the measurement time lengthens, the mental and physical burdens on the subject 12 will also increase.

[0106] On the other hand, in the "example", the subject 12 reaches the resting state and starts the estimation of ISF. Assuming that the time required for this series of processes until the estimation ends is T3, as shown in Figure 7 and Figure 8 , T3 is approximately 30 s. That is to say, without obtaining the electroencephalogram signal when the subject 12 has a motor intention, or without the subject 12 needing to move the paralyzed part, the ISF can be recognized in a short time. This significantly reduces the mental and physical burdens on the subject 12.

[0107] [Summary of the embodiment]

[0108] As described above, the motion assistance system (here, the BMI system 10) in this embodiment includes the following devices: an electroencephalogram analysis device 16; an electroencephalograph 14 for measuring the electroencephalogram of the subject 12 and sending the obtained electroencephalogram signal to the electroencephalogram analysis device 16; a motion assistance device 18 that is controlled by the electroencephalogram analysis device 16 to perform related operations, thereby assisting the subject 12 to perform movements.

[0109] The electroencephalogram analysis device 16 includes the following parts: an acquisition unit (here, the signal acquisition unit 40) for acquiring the time series of the electroencephalogram signals of the subject 12; an arithmetic unit 58 for obtaining an inherent frequency related to the movement intention or brain state of the subject 12 based on the relevant frequency characteristics of the acquired time series of the electroencephalogram signals.

[0110] In addition, according to the electroencephalogram analysis method and electroencephalogram analysis program in the present embodiment, one or more computers (or the processor 26) execute an acquisition step (SP10) for acquiring the time series of the electroencephalogram signals of the subject 12, and an arithmetic step (SP22) for obtaining an inherent frequency related to the movement intention of the subject 12 based on the relevant frequency characteristics of the time series of the electroencephalogram signals acquired when the subject 12 is at rest.

[0111] In addition, according to the movement assistance method and movement assistance program in the present embodiment, one or more computers (or the processor 26) in addition to executing the above acquisition step (SP10) and arithmetic step (SP22), also execute the following steps: a calibration step (SP24), in which the obtained inherent frequency is set as the calibration parameter of the movement assistance device 18; an assistance step (SP108), in which the calibrated movement assistance device 18 is controlled to operate so as to assist the subject 12 in moving.

[0112] As described above, by obtaining an inherent frequency related to the movement intention or brain state of the subject 12 based on the relevant frequency characteristics of the time series of the electroencephalogram signals acquired when the subject 12 is at rest, it is possible to use the acquired electroencephalogram signals to identify the inherent frequency when the subject 12 maintains a resting state, that is, without requiring additional movement. So that when analyzing the electroencephalogram signals of the subject 12 and identifying the inherent frequency related to the movement intention or brain state, the time required for identification can be significantly shortened.

[0113] In addition, the electroencephalogram analysis device 16 further includes the following parts: a calculation unit 52 for calculating the sample value of the peak frequency in the frequency characteristics; an estimation unit 54 for obtaining an estimated value of the peak frequency based on the overall of the calculated sample values. Among them, the arithmetic unit 58 can convert the estimated value obtained by the estimation unit 54 into an inherent frequency according to a predefined conversion rule.

[0114] That is to say, the electroencephalogram analysis device 16 obtains an estimated value of the peak frequency through the estimation unit 54 and converts the estimated value into an inherent frequency through the arithmetic unit 58, so that it is possible to obtain the inherent frequency without acquiring the electroencephalogram signals when the subject 12 has a movement intention, or without requiring the subject 12 to move the paralyzed part. This significantly reduces the mental and physical burdens on the subject 12.

[0115] In addition, starting from the moment when the electroencephalogram signal measurement begins, acquisition is performed by the signal acquisition unit 40 and calculation is performed by the calculation unit 52 for each unit time, so as to accumulate the overall sample values. The estimation unit 54 can use the sequential Bayesian method to obtain the estimation value for each unit time based on the overall sample values.

[0116] In addition, if the electroencephalogram analysis device 16 further includes a determination unit 56 for determining whether an end condition is satisfied each time the estimation unit 54 performs an estimation, when the determination unit 56 determines that the end condition is satisfied, the estimation unit 54 can end the estimation of the peak frequency. By aborting the estimation process when the end condition is satisfied, the time required to identify the natural frequency can be further shortened.

[0117] In addition, when the peak frequency is the alpha frequency (i.e., IAF) within the alpha band and the natural frequency is the individual SMR-ERD frequency (i.e., ISF), the conversion rule can be expressed in the form of an identity function or a linear function, where the identity function or the linear function takes the IAF estimated value as the independent variable. Using a more concise conversion rule is beneficial to shortening the time required for the operation process.

[0118] In addition, the conversion rule can be defined according to the specific situation of the subject 12 to be analyzed. Thus, a conversion more in line with the characteristics of the subject 12 can be performed.

[0119] In addition, when the signal acquisition unit 40 acquires the first time series of the electroencephalogram signal measured when the subject 12 is at rest or the second time series of the electroencephalogram signal measured when the subject 12 performs motor imagery, the operation unit 58 can perform a first operation using only the first time series to obtain the natural frequency. In addition, the operation unit 58 can also switch between the first operation and the second operation, where the second operation uses both the first time series and the second time series to obtain the natural frequency.

[0120] In addition, the electroencephalogram analysis device 16 can further include a prompting unit for sending a prompting message to the subject 12 to prompt the subject 12 to maintain a resting state. In addition, the electroencephalogram analysis device 16 can further include an auxiliary control unit 44, which controls the movement assistance device 18 based on the natural frequency obtained by the operation unit 58, thereby assisting the movement of the subject 12. In this way, the subject 12 can be analyzed or assisted in moving.

[0121] [Variant Example]

[0122] It should be noted that the present invention is not limited to the above-described embodiments. Of course, modifications and adjustments can be freely made without departing from the gist of the present invention. Alternatively, within the scope where there is no technical contradiction, the respective components can be arbitrarily combined. Or, within the scope where there is no technical contradiction, the execution order of the steps in the flowchart can be modified.

[0123] In the above-described embodiments, the related situation where the electroencephalogram analysis device 16 analyzes the electroencephalogram signal and controls the movement assistance device 18 is illustrated, but the system configuration is not limited thereto. For example, the analysis unit and the control unit can be separately provided, and the required data can be transmitted between the two through wired communication or wireless communication. In this case, the analysis process can also be executed by a server device in the cloud or a server device deployed locally. In particular, when the server device is a server device in the cloud, the server device can be a computer cluster that constructs a distributed system.

[0124] Explanation of reference numerals

[0125] 10... BMI system (movement assistance system), 12... subject to be analyzed, 14... electroencephalograph, 16... electroencephalogram analysis device (computer), 18... movement assistance device, 26... processor, 28... memory, 40... signal acquisition unit (acquisition unit), 42... frequency identification unit, 44... auxiliary control unit, 50... preprocessing unit, 52... calculation unit, 54... estimation unit, 56... determination unit, 58... operation unit.

Claims

1. An electroencephalogram analysis device, comprising: An acquisition unit configured to acquire the time series of the electroencephalogram signal of the subject to be analyzed; An operation unit configured to obtain an inherent frequency related to the movement intention or brain state of the subject to be analyzed based on the relevant frequency characteristics of the time series of the electroencephalogram signal acquired by the acquisition unit when the subject to be analyzed is at rest.

2. The electroencephalogram analysis device according to claim 1, further comprising: A calculation unit configured to calculate a sample value of the peak frequency in the frequency characteristics; An estimation unit configured to obtain an estimated value of the peak frequency based on the population of the sample values calculated by the calculation unit, wherein the operation unit converts the estimated value obtained by the estimation unit into the inherent frequency according to a pre-defined conversion rule.

3. The electroencephalogram analysis device according to claim 2, wherein the estimated value is obtained by the estimation unit, and the estimated value is converted into the inherent frequency by the operation unit, so that the inherent frequency can be obtained without acquiring the electroencephalogram signal when the subject to be analyzed has a movement intention.

4. The electroencephalogram analysis device according to claim 2, wherein the estimated value is obtained by the estimation unit, and the estimated value is converted into the inherent frequency by the operation unit, so that the inherent frequency can be obtained when the subject to be analyzed moves without using the paralyzed part.

5. The electroencephalogram analysis device according to claim 2, wherein starting from the moment when the electroencephalogram signal measurement begins, the acquisition unit acquires and the calculation unit calculates every unit time, so as to accumulate the population of the sample values, and the estimation unit obtains the estimated value of each unit time by using the sequential Bayesian method based on the population of the sample values.

6. The electroencephalogram analysis device according to claim 4, wherein it further comprises a determination unit configured to judge whether an end condition is satisfied each time the estimation unit makes an estimation, and when the determination unit judges that the end condition is satisfied, the estimation unit ends the estimation of the peak frequency.

7. The electroencephalogram analysis device according to claim 2, wherein the peak frequency is the alpha frequency in the alpha band, the inherent frequency is the individual SMR-ERD frequency, the conversion rule is represented in the form of an identity function or a linear function, and the identity function or the linear function takes the estimated value as the independent variable.

8. The electroencephalogram analysis device according to claim 2, wherein the conversion rule is defined according to the specific situation of the subject to be analyzed.

9. The electroencephalogram analysis device according to claim 1, wherein the acquisition unit acquires the first time series of the electroencephalogram signal measured when the subject to be analyzed is at rest or the second time series of the electroencephalogram signal measured when the subject to be analyzed imagines movement, and the operation unit only uses the first time series for the first operation to obtain the inherent frequency.

10. The electroencephalogram analysis device according to claim 9, wherein The arithmetic unit switches between the first arithmetic operation and the second arithmetic operation, and the second arithmetic operation uses both the first timing and the second timing to obtain the natural frequency.

11. The electroencephalogram analysis device according to claim 1, wherein a prompting unit is further included, which is configured to send a prompting message to the subject to prompt the subject to maintain a resting state.

12. The electroencephalogram analysis device according to claim 1, wherein an auxiliary control unit is further included, which controls a movement assistance device based on the natural frequency obtained by the arithmetic unit, so as to assist the movement of the subject.

13. An electroencephalogram analysis program, wherein one or more computers are caused to execute an acquisition step and an arithmetic step. In the acquisition step, the timing of the electroencephalogram signal of the subject is acquired. In the arithmetic step, based on the relevant frequency characteristics of the timing of the electroencephalogram signal acquired when the subject is at rest, the natural frequency related to the movement intention or state of the subject under test is obtained.

14. A movement assistance system, comprising: the electroencephalogram analysis device according to claim 12; an electroencephalograph for measuring the electroencephalogram of the subject and sending the obtained electroencephalogram signal to the electroencephalogram analysis device; a movement assistance device, which performs relevant operations under the control of the electroencephalogram analysis device, so as to assist the subject in moving.

15. A movement assistance method, using a system having the following devices: an electroencephalograph for measuring the electroencephalogram of the subject and outputting an electroencephalogram signal; an electroencephalogram analysis device for analyzing the electroencephalogram signal sent from the electroencephalograph; a movement assistance device, which performs relevant operations under the control of the electroencephalogram analysis device, so as to assist the subject in moving, and the movement assistance method includes the following steps: an acquisition step, in which the electroencephalogram analysis device uses the electroencephalograph to measure the electroencephalogram of the subject at rest and acquires the timing of the electroencephalogram signal; an arithmetic step, in which the electroencephalogram analysis device obtains the natural frequency related to the movement intention or brain state of the subject according to the relevant frequency characteristics of the acquired timing of the electroencephalogram signal; a calibration step, in which the obtained natural frequency is set as the calibration parameter of the movement assistance device and calibration is performed; an assistance step, in which the calibrated movement assistance device is controlled to operate, so as to assist the subject in moving.

Images