Brain wave signal processing method and system

Abstract

The present invention provides a method and system for processing brain wave signals. The method comprises the following steps: performing spike wave detection on the brain wave signal to obtain a spike wave distribution waveform; performing transient frequency oscillation energy analysis on the brain wave signal to obtain multiple energy distribution waveforms; performing complexity analysis on the brain wave signal to obtain a complexity change waveform; obtaining a judgment result of a specified neural waveform based on the spike wave distribution waveform, the energy distribution waveform and the complexity change waveform, and outputting the judgment result.

Description

Technical Field

[0001] The present invention relates to a signal processing method, and in particular to a brain wave signal processing method and system. Background Art

[0002] The commonly used methods for measuring brain activity are: electrocorticography (ECoG) and electroencephalography (EEG). ECoG and EEG signals are commonly known as brain wave signals, and the principles of generating the two signals are almost the same. The difference between ECoG and EEG is that the measurement of ECoG signals is a semi-invasive method that opens the patient's skull and places electrodes flat on the cerebral cortex and under the dura mater, while the measurement of EEG signals is a non-invasive method that measures brain waves on the scalp.

[0003] Brain wave signals are most commonly used to diagnose epilepsy. During an epileptic seizure, according to the manifestation of epilepsy pathophysiology and the characteristics of brain wave signals, it can be roughly divided into the following stages: interictal, preictal, irregular phase, and bursting phase. Among them, the period between two epileptic seizures is called "inter-ictal", while "pre-ictal", "irregular phase" and "bursting phase" belong to the ictal phase of epileptic seizures. The irregular phase is of great significance for the diagnosis and evaluation of epilepsy, but existing algorithms are less able to effectively detect this feature. The reason lies in the manifestation of neurophysiology. The complex interaction between excitatory cells and inhibitory cells makes the brain wave signal show non-stationary characteristics. This brain wave signal shows irregular frequency and irregular amplitude changes, making it difficult for existing algorithm judgment models to accurately interpret. Summary of the invention

[0004] The present invention is directed to a brain wave signal processing method and system, which can accurately find a specified neural waveform in a brain wave signal.

[0005] According to an embodiment of the present invention, a brain wave signal processing method is to execute the following steps through a processor, including: obtaining a brain wave signal; performing spike detection on the brain wave signal to obtain a spike distribution waveform; performing instantaneous frequency oscillation energy analysis on the brain wave signal to obtain multiple energy distribution waveforms; performing complexity analysis on the brain wave signal to obtain a complexity change waveform; and obtaining a judgment result of a specified neural waveform based on the spike distribution waveform, the energy distribution waveform and the complexity change waveform, wherein the judgment result includes a time interval, the waveform of the time interval in the spike distribution waveform conforms to the spike distribution rule of the specified neural waveform, the waveform of the time interval of each energy distribution waveform conforms to the energy change rule of the specified neural waveform, and the waveform of the time interval in the complexity change waveform conforms to the complexity change rule of the specified neural waveform.

[0006] In the brain wave signal processing method according to an embodiment of the present invention, the above-mentioned step of performing spike wave detection on the brain wave signal to obtain a spike wave distribution waveform includes: dividing the brain wave signal into multiple sampling intervals using a window width; detecting the number of spike waves included in each sampling interval; and in a spike wave distribution waveform diagram, recording the number of spike waves included in each sampling interval along with a time series, thereby obtaining a spike wave distribution waveform.

[0007] In the brain wave signal processing method according to an embodiment of the present invention, after obtaining the spike wave distribution waveform, it also includes: judging whether the number of spike waves of multiple unit times in the spike wave distribution waveform conforms to the spike wave distribution rule. Among them, in two adjacent unit times in the time series, if the number of spike waves in the later unit time decreases compared with the number of spike waves in the earlier unit time, it is judged that the later unit time conforms to the spike wave distribution rule.

[0008] In the brain wave signal processing method according to an embodiment of the present invention, the above-mentioned step of performing instantaneous oscillation energy analysis on the brain wave signal to obtain an energy distribution waveform includes: performing empirical mode decomposition (EMD) on the brain wave signal to obtain multiple intrinsic mode signals; calculating the instantaneous frequency of each intrinsic mode signal in multiple sampling intervals, and then recording the instantaneous frequency corresponding to each sampling interval along the time series in the instantaneous frequency distribution diagram corresponding to each intrinsic mode signal, thereby obtaining multiple instantaneous frequency distribution waveforms corresponding to the multiple intrinsic mode signals; and performing energy calculation on the multiple instantaneous frequency distribution waveforms to obtain the multiple energy distribution waveforms.

[0009] In the brain wave signal processing method according to the embodiment of the present invention, after obtaining the energy distribution waveform, it also includes: judging whether the energy density of multiple unit times of each energy distribution waveform meets the energy change rule. Wherein, at the same time, in two adjacent unit times on the time series of each of these energy distribution waveforms, if the energy density of the later unit time decreases compared with the energy density of the earlier unit time, it is judged that the later unit time meets the energy change rule.

[0010] In the brain wave signal processing method according to an embodiment of the present invention, the above-mentioned step of performing complexity analysis on the brain wave signal to obtain a complexity variation waveform includes: dividing the brain wave signal into multiple sampling intervals using a window width; calculating the complexity of each sampling interval; in a complexity waveform diagram, recording the complexity corresponding to each sampling interval along with a time series, thereby obtaining a complexity waveform; dividing the complexity waveform into multiple calculation intervals using another window width; detecting the number of spikes included in each calculation interval; and in a complexity spike distribution diagram, recording the number of spikes included in each calculation interval along with a time series, thereby obtaining a complexity variation waveform.

[0011] In the brain wave signal processing method according to an embodiment of the present invention, after obtaining the complexity change waveform, it also includes: judging whether the number of spikes per unit time in the complexity change waveform meets the complexity change rule. The unit time with a spike number greater than a preset value is judged to meet the complexity change rule.

[0012] In the brain wave signal processing method according to the embodiment of the present invention, the above-mentioned designated neural waveform is an irregular period waveform used to determine epileptic seizures.

[0013] According to an embodiment of the present invention, a brain wave signal processing system includes: an output device, a storage device and a processor. The storage device includes a plurality of code segments. The processor is coupled to the output device and the storage device. The processor executes the code segments to: obtain brain wave signals; perform spike wave detection on the brain wave signals to obtain spike wave distribution waveforms; perform transient frequency oscillation energy analysis on the brain wave signals to obtain a plurality of energy distribution waveforms; perform complexity analysis on the brain wave signals to obtain complexity change waveforms; and obtain a judgment result of a specified neural waveform based on the spike wave distribution waveform, the energy distribution waveform and the complexity change waveform, wherein the judgment result includes a time interval, the waveform of the time interval in the spike wave distribution waveform conforms to the spike wave distribution rule of the specified neural waveform, the waveform of each time interval of the energy distribution waveform conforms to the energy change rule of the specified neural waveform, and the waveform of the time interval in the complexity change waveform conforms to the complexity change rule of the specified neural waveform.

[0014] Based on the above, the present invention combines spike wave detection, transient frequency oscillation energy analysis and complexity analysis to accurately find out irregular periods in brain wave signals. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 is a block diagram of a brain wave signal processing system according to an embodiment of the present invention;

[0016] Figure 2 is a block diagram of a brain wave signal processing module according to an embodiment of the present invention;

[0017] Figure 3 is a flow chart of a method for processing brain wave signals according to an embodiment of the present invention;

[0018] Figure 4 is a waveform diagram of a brain wave signal according to an embodiment of the present invention;

[0019] Figure 5 is a spike wave distribution waveform diagram according to an embodiment of the present invention;

[0020] Figure 6 is a waveform diagram of an intrinsic mode signal according to an embodiment of the present invention;

[0021] Figure 7 is a waveform diagram of instantaneous frequency distribution according to an embodiment of the present invention;

[0022] Figure 8 is an energy distribution waveform diagram according to an embodiment of the present invention;

[0023] Fig. 9 is a complexity waveform diagram according to an embodiment of the present invention;

[0024] Fig.10 is a complexity spike distribution diagram according to an embodiment of the present invention.

[0025] Description of Reference Numerals

[0026] 100: Brain wave signal processing system

[0027] 110: Processor

[0028] 120: Storage device

[0029] 130: Output device

[0030] 210: Spike detection module

[0031] 220: Energy Analysis Module

[0032] 230: Complexity Analyzer

[0033] 240: Identification module

[0034] S305-S325: Steps of the brain wave signal processing method

[0035] 510, 810, 1010: interval

[0036] x(t): brain wave signal

[0037] s(x(t)): spike distribution waveform

[0038] M1(x(t)), M2(x(t)), M3(x(t)): intrinsic mode signals

[0039] f(M1(x(t))), f(M2(x(t))), f(M3(x(t))): Instantaneous frequency distribution waveform

[0040] E(f(M1(x(t)))), E(f(M2(x(t)))), E(f(M3(x(t)))): Energy distribution waveform

[0041] En(x(t)): Complex waveform

[0042] S(En(x(t))): Complexity change waveform DETAILED DESCRIPTION

[0043] Reference will now be made in detail to exemplary embodiments of the present invention, examples of which are illustrated in the accompanying drawings. Whenever possible, the same reference numerals are used in the drawings and the description to refer to the same or like parts.

[0044] Figure 1 FIG. 1 is a block diagram of an electroencephalogram signal processing system according to an embodiment of the present invention. Figure 1 In the present invention, the brain wave signal processing system 100 is a device for analyzing brain wave signals. The brain wave signal processing system 100 can be a device with computing capabilities such as a desktop computer, a laptop computer, a smart phone, etc. The brain wave signal processing system 100 includes a processor 110, a storage device 120, and an output device 130.

[0045] The processor 110 is, for example, a central processing unit (CPU), a physical processing unit (PPU), a programmable microprocessor (Microprocessor), an embedded control chip, a digital signal processor (DSP), an application specific integrated circuit (ASIC) or other similar devices.

[0046] The storage device 120 is, for example, any type of fixed or removable random access memory (RAM), read-only memory (ROM), flash memory, hard disk or other similar device or a combination of these devices. The storage device 120 stores a plurality of code segments, which are executed by the processor 110 after being installed to implement the brain wave signal processing method described below.

[0047] The output device 130 is, for example, a display. When the processor 110 obtains the final judgment result, the judgment result can be transmitted to the output device 130. In addition, the processor 110 can also store the judgment result in the storage device 120, or transmit it to the server of the medical center via the network through a communication device (not shown).

[0048] Figure 2 is a block diagram of a brain wave signal processing module according to an embodiment of the present invention. Figure 2 The brain wave signal processing module shown is, for example, composed of multiple code segments and stored in the storage device 120. The brain wave signal processing module includes a spike wave detection module 210, an energy analysis module 220, a complexity analyzer 230 and an identification module 240. The spike wave detection module 210 is used to perform spike wave detection on the brain wave signal. The energy analysis module 220 is used to perform transient oscillation energy analysis on the brain wave signal. The complexity analyzer 230 is used to perform complexity analysis on the brain wave signal. The identification module 240 is used to obtain the time interval where the specified neural waveform is located based on the spike wave distribution waveform, the energy distribution waveform and the complexity change waveform. The above-mentioned specified neural waveform is an irregular phase waveform used to determine the epileptic seizure.

[0049] Figure 3 is a flow chart of a method for processing brain wave signals according to an embodiment of the present invention. Figure 3 In step S305, an electroencephalogram signal is obtained. For example, an electroencephalogram measuring instrument is used to obtain the electroencephalogram signal of the subject. Figure 4 shown. Figure 4 is a waveform diagram of an electroencephalogram signal according to an embodiment of the present invention. The electroencephalogram signal x(t) obtained at this time is a time domain signal. Next, the electroencephalogram signal processing system 100 receives the electroencephalogram signal x(t) from the electroencephalogram measuring instrument, and the processor 110 processes the electroencephalogram signal through the electroencephalogram signal processing module (such as Figure 2 As shown) is used to analyze the brain wave signal x(t).

[0050] For the time interval where the neural waveform is designated as an irregular phase, the brain wave signal processing module analyzes the pre-processed brain wave signal x(t) via the spike wave detection module 210, the energy analysis module 220, and the complexity analyzer 230. Here, the pre-processing is, for example, band-pass filtering, noise filtering, re-sampling, time-series segmentation, etc.

[0051] The spike detection module 210, the energy analysis module 220 and the complexity analyzer 230 respectively calculate the non-steady-state waveform characteristics (characteristics of the irregular period) with the default timing window. Afterwards, the spike detection module 210, the energy analysis module 220 and the complexity analyzer 230 transmit the obtained results to the recognition module 240, which interprets the specified neural waveform and finally outputs the results. The non-steady-state waveform characteristics include: increased amplitude density (i.e., fewer oscillations and larger amplitudes), frequency drift (band width, main frequency drift), larger waveform density intervals (relative to the initial onset (onset start)), and inconsistent intervals (larger randomness).

[0052] The following steps S310 , S315 , and S320 are used to respectively illustrate the spike detection module 210 , the energy analysis module 220 , and the complexity analyzer 230 . The steps may be executed simultaneously or separately, and the execution order of the steps is not limited herein.

[0053] In step S310, the spike wave detection module 210 performs spike wave detection on the brain wave signal to obtain a spike wave distribution waveform. Here, the spike wave detection module 210 divides the brain wave signal x(t) into multiple sampling intervals using a window width (first window width) and detects the number of spike waves in each sampling interval. Next, the spike wave detection module 210 records the number of spike waves included in each sampling interval in the spike wave distribution waveform diagram along with the time series, thereby obtaining the spike wave distribution waveform s(x(t)), as shown in FIG. Figure 5 shown. Figure 5 is a spike wave distribution waveform diagram according to an embodiment of the present invention. For example, assuming the window width is 5 seconds, take a signal from 0 to 5 seconds in the brain wave signal x(t), calculate how many spike waves there are in the signal within 5 seconds, and project the obtained spike wave number to Figure 5 The spike wave distribution waveform shown is at the 5th second. Similarly, the number of spike waves is recorded every 5 seconds. The time unit of the spike wave distribution waveform s(x(t)) is the window width (first window width).

[0054] In step S315, the energy analysis module 220 performs instantaneous frequency oscillation energy analysis on the brain wave signal to obtain a plurality of energy distribution waveforms. Figure 6 , Figure 7 , Figure 8 Let me give you an example. Figure 6 is a waveform diagram of an intrinsic mode signal according to an embodiment of the present invention. Figure 7 is a waveform diagram of instantaneous frequency distribution according to an embodiment of the present invention. Figure 8 is an energy distribution waveform diagram according to an embodiment of the present invention.

[0055] The energy analysis module 22 decomposes the brain wave signal x(t) to be analyzed into multiple intrinsic mode functions (in this embodiment, three intrinsic mode signals M1(x(t)), M2(x(t)), and M3(x(t)) are used as examples for explanation) by executing the empirical mode decomposition (EMD) method.

[0056] Next, the energy analysis module 22 calculates the instantaneous frequencies of the eigenmode signals M1(x(t)), M2(x(t)), and M3(x(t)) in a plurality of sampling intervals. Here, for the eigenmode signal M1(x(t)), the eigenmode signal M1(x(t)) is divided into a plurality of sampling intervals by a window width (a second window width), and then the instantaneous frequency of the signal in each sampling area is calculated by using, for example, Hilbert Transform, so as to obtain Figure 7 The instantaneous frequency distribution waveform f(M1(x(t))) is shown. Similarly, the eigenmode signal M2(x(t)) and the eigenmode signal M3(x(t)) are converted into the instantaneous frequency distribution waveform f(M2(x(t))) and the instantaneous frequency distribution waveform f(M3(x(t))) respectively. That is, in the instantaneous frequency distribution diagram corresponding to each eigenmode signal, the instantaneous frequency corresponding to each sampling interval is recorded along the time series, thereby obtaining the instantaneous frequency distribution waveform corresponding to each eigenmode signal.

[0057] Afterwards, the energy analysis module 22 performs energy calculations on the multiple instantaneous frequency distribution waveforms f(M1(x(t))), f(M2(x(t))), and f(M3(x(t))) to obtain multiple energy distribution waveforms E(f(M1(x(t)))), E(f(M2(x(t)))), and E(f(M3(x(t)))), such as Figure 8For example, the energy calculated by the energy analysis module 22 is spectral energy density or power spectral density.

[0058] In step S320, the complexity analyzer 230 performs complexity analysis on the brain wave signal to obtain a complexity variation waveform. Fig. 9 , Fig.10 Let me give you an example. Fig. 9 is a complexity waveform diagram according to an embodiment of the present invention. Fig.10 is a complexity spike distribution diagram according to an embodiment of the present invention.

[0059] The complexity analyzer 230 divides the brain wave signal x(t) into a plurality of sampling intervals using a window width (third window width) and calculates the complexity of each sampling interval. The third window width is, for example, 1 second. Here, the complexity analyzer 230 calculates the entropy of each sampling interval. In the complexity waveform diagram, the complexity analyzer 230 records the complexity (entropy) corresponding to each sampling interval along the time series, thereby obtaining the complexity waveform En(x(t)).

[0060] The entropy is, for example, Approximate Entropy (ApEn). Approximate entropy is a mathematical method for analyzing the complexity of information contained in a time series. By calculating the entropy value of a signal at different time-space scales, it provides a quantitative standard for testing the complexity of a signal. Approximate entropy is the calculation of the self-similarity difference of sampling intervals at two adjacent time scales. The higher the complexity, the higher the self-similarity ratio at a scale difference.

[0061] Next, the complexity analyzer 230 divides the complexity waveform En(x(t)) into a plurality of calculation intervals using another window width (fourth window width). The complexity analyzer 230 detects the number of spikes included in each calculation interval and calculates the number of spikes included in each calculation interval. Fig.10 In the complexity spike distribution diagram shown in FIG. 1 , the number of spikes included in each calculation interval is recorded along the time series, thereby obtaining the complexity change waveform S(En(x(t))). The fourth window width is, for example, 7 seconds. The complexity analyzer 230 takes a signal from 0 to 7 seconds in the complexity waveform En(x(t)), calculates how many spikes there are in the signal within 7 seconds, and projects the obtained number of spikes to Fig.10 The complexity spike wave distribution diagram shown is at the 7th second. Similarly, the number of spike waves is recorded every 7 seconds. The time unit of the complexity variation waveform S(En(x(t))) is the fourth window width.

[0062] In step S325, the identification module 240 obtains a judgment result of the specified neural waveform based on the spike distribution waveform s(x(t)), the energy distribution waveform E(f(M1(x(t)))), E(f(M2(x(t)))), E(f(M3(x(t)))) and the complexity change waveform S(En(x(t))). The judgment result includes the time segment in which the specified neural waveform is located in the brain wave signal. The waveform of the time interval in the spike distribution waveform s(x(t)) conforms to the spike distribution rule of the specified neural waveform, the waveforms of the time intervals of all energy distribution waveforms E(f(M1(x(t)))), E(f(M2(x(t)))), and E(f(M3(x(t)))) conform to the energy change rule of the specified neural waveform, and the waveform of the time interval in the complexity change waveform S(En(x(t))) conforms to the complexity change rule of the specified neural waveform.

[0063] Specifically, after obtaining the spike distribution waveform s(x(t)), the energy distribution waveform E(f(M1(x(t)))), E(f(M2(x(t)))), E(f(M3(x(t)))) and the complexity change waveform S(En(x(t))), the identification module 240 will determine whether the waveforms meet the corresponding rules one by one according to a pre-set unit time.

[0064] The identification module 240 determines whether the number of spikes per unit time in the spike distribution waveform s(x(t)) complies with the spike distribution rule. Among them, in two adjacent unit times in the time series, if the number of spikes per unit time in the later unit time decreases compared to the number of spikes per unit time in the earlier unit time, it is determined that the later unit time complies with the spike distribution rule. In this way, the spike detection module 210 determines whether the number of spikes decreases. For example, if Figure 5 The waveform in the illustrated interval 510 is determined to comply with the spike distribution rule.

[0065] For example, assuming that 10 seconds is used as the time unit for determination, the number of spikes (the total number of spikes) included in the first time unit of 0 to 10 seconds is 30, and the number of spikes (the total number of spikes) included in the second time unit of 10 to 20 seconds is 10, then the second time unit is determined to be in compliance with the spike distribution rule. Similarly, the identification module 240 will determine whether each unit time in the spike distribution waveform s(x(t)) complies with the spike distribution rule, and the identification module 240 sets the Boolean value corresponding to the unit time that complies with the spike distribution rule to "true", and sets the Boolean value corresponding to the unit time that does not comply with the spike distribution rule to "false".

[0066] In addition, the identification module 240 further determines whether the energy density of multiple unit times of the energy distribution waveforms E(f(M1(x(t)))), E(f(M2(x(t)))), and E(f(M3(x(t)))) conforms to the energy change rule. Among them, if the energy density of the later unit time decreases compared to the energy density of the earlier unit time in two adjacent unit times on the time series of each of these energy distribution waveforms, then it is determined that the later unit time conforms to the energy change rule. In other words, the energy density of the same unit time of the three energy distribution waveforms E(f(M1(x(t)))), E(f(M2(x(t)))), and E(f(M3(x(t)))) all decreases before it is determined that the unit time conforms to the energy change rule. For example, Figure 8 The waveform in the interval 810 shown is determined to comply with the energy change rule. Furthermore, the identification module 240 sets the Boolean value corresponding to the unit time that complies with the energy change rule to "true" and sets the Boolean value corresponding to the unit time that does not comply with the energy change rule to "false".

[0067] In addition, the identification module 240 further determines whether the number of spikes per unit time in the complexity change waveform S(En(x(t))) meets the complexity change rule. The unit time with a spike number greater than a preset value is determined to meet the complexity change rule. For example, Fig.10 The number of spikes in the waveform in the interval 1010 shown is greater than the preset value, so it is determined that the interval 1010 meets the complexity change rule. In addition, the identification module 240 sets the Boolean value corresponding to the unit time that meets the complexity change rule to "true" and sets the Boolean value corresponding to the unit time that does not meet the complexity change rule to "false".

[0068] Finally, the identification module 240 finds the unit time in which the corresponding Boolean value is "true" among the multiple unit times included in the spike distribution waveform s(x(t)), the energy distribution waveform E(f(M1(x(t)))), E(f(M2(x(t)))), E(f(M3(x(t)))) and the complexity change waveform S(En(x(t))), which represents the time interval where the specified neural waveform is located. For example, assuming that the Boolean value corresponding to the unit time of 100 to 110 seconds of the spike distribution waveform s(x(t)) is "true", the Boolean values ​​corresponding to the unit time of 100 to 110 seconds of the energy distribution waveforms E(f(M1(x(t)))), E(f(M2(x(t)))), and E(f(M3(x(t)))) are all "true", and the Boolean values ​​corresponding to the unit time of 100 to 110 seconds of the complexity change waveform S(En(x(t))) are all "true", it is determined that 100 to 110 seconds is the time interval where the specified neural waveform is located, that is, the irregular period.

[0069] In summary, the present invention combines spike wave detection, transient frequency oscillation energy analysis and complexity analysis, and utilizes the determination of spike wave distribution waveform to find out the characteristics before and after the attack, and from the determination of energy distribution waveform, the characteristics of irregular period and initial onset (Onset) can be distinguished, and from the determination of complexity change waveform, the characteristics of irregular period and explosion period can be distinguished. Based on this, the irregular period in the brain wave signal can be accurately found.

[0070] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for processing brain wave signals, characterized in that: The processor performs the following steps, including: Obtain brain wave signals; Performing spike wave detection on the brain wave signal to detect the number of spike waves corresponding to the first sampling intervals, thereby obtaining a spike wave distribution waveform; Performing transient frequency oscillation energy analysis on the brain wave signal to obtain a plurality of energy distribution waveforms; Performing complexity analysis on the brain wave signal to calculate multiple complexities of multiple second sampling intervals to obtain a complexity waveform, and detecting multiple numbers of spikes corresponding to multiple calculation intervals of the complexity waveform to obtain a complexity change waveform; and Based on the spike wave distribution waveform, the energy distribution waveform and the complexity change waveform, a time interval where a specified neural waveform is located is obtained, The number of spikes per unit time in the time interval in the spike distribution waveform is smaller than the number of spikes per unit time in the previous unit time. The energy density per unit time of the time interval of each of the multiple energy distribution waveforms is less than the energy density per unit time of the previous time interval. The number of spikes per unit time in the time interval in the complexity variation waveform is greater than a preset value, The step of performing the transient frequency oscillation energy analysis on the brain wave signal to obtain the energy distribution waveform comprises: Performing empirical mode decomposition on the brain wave signal to obtain a plurality of eigenmode signals; Calculating the instantaneous frequency of each of the eigenmode signals in a plurality of third sampling intervals, and then recording the instantaneous frequency corresponding to each of the third sampling intervals in a time series in an instantaneous frequency distribution diagram corresponding to each of the eigenmode signals, thereby obtaining a plurality of instantaneous frequency distribution waveforms corresponding to the eigenmode signals; and The spectrum energy density or the power spectrum density is calculated for the instantaneous frequency distribution waveform respectively to obtain the energy distribution waveform.

2. The brain wave signal processing method according to claim 1, characterized in that: The step of performing the spike wave detection on the brain wave signal to detect the number of the spike waves corresponding to the first sampling intervals to obtain the spike wave distribution waveform comprises: Dividing the brain wave signal into the plurality of first sampling intervals using a window width; Detecting the number of spikes included in each of the first sampling intervals; and In the spike wave distribution waveform diagram, the number of spike waves included in each of the first sampling intervals is recorded in time series, thereby obtaining the spike wave distribution waveform.

3. The brain wave signal processing method according to claim 1, characterized in that: After obtaining the spike wave distribution waveform, the method further includes: determining whether the number of spikes per unit time in the spike distribution waveform complies with the spike distribution rule, Among them, in two adjacent unit times in the time series, if the number of spikes in the later unit time decreases compared to the number of spikes in the earlier unit time, it is determined that the later unit time meets the spike distribution rule.

4. The brain wave signal processing method according to claim 1, characterized in that: After obtaining the energy distribution waveform, the method further includes: Determine whether the energy density of multiple unit times of the energy distribution waveform conforms to the energy change rule, Among them, at the same time, in the two adjacent unit times in the time series of the energy distribution waveform, if the energy density of the later unit time is lower than the energy density of the earlier unit time, it is determined that the later unit time meets the energy change rule.

5. The brain wave signal processing method according to claim 1, characterized in that: The step of performing the complexity analysis on the brain wave signal to obtain the complexity variation waveform comprises: Dividing the brain wave signal into the plurality of second sampling intervals by using a window width; In the complexity waveform diagram, the complexity corresponding to each of the second sampling intervals is recorded along a time series, thereby obtaining a complexity waveform; Dividing the complexity waveform into the plurality of calculation intervals using another window width; detecting the number of spikes included in each of the calculation intervals; and In the complexity spike distribution diagram, the number of spikes included in each calculation interval is recorded over time, thereby obtaining the complexity variation waveform.

6. The method for processing brain wave signals according to claim 1, characterized in that: After obtaining the complexity variation waveform, the method further includes: Determine whether the number of spikes per unit time in the complexity change waveform complies with the complexity change rule, The unit time in which the number of spikes is greater than the preset value is determined to be in compliance with the complexity change rule.

7. The brain wave signal processing method according to claim 1, characterized in that: The designated neural waveform is an irregular period waveform used to determine epileptic seizures.

8. A brain wave signal processing system, characterized in that: include: Output device; A storage device including a plurality of code segments; as well as A processor, coupled to the output device and the storage device, wherein the processor executes the code segment to: Obtain brain wave signals; Performing spike wave detection on the brain wave signal to detect the number of spike waves corresponding to the first sampling intervals, thereby obtaining a spike wave distribution waveform; Performing transient frequency oscillation energy analysis on the brain wave signal to obtain a plurality of energy distribution waveforms; Performing complexity analysis on the brain wave signal to calculate multiple complexities of multiple second sampling intervals to obtain a complexity waveform, and detecting multiple numbers of spikes corresponding to multiple calculation intervals of the complexity waveform to obtain a complexity change waveform; and Based on the spike wave distribution waveform, the energy distribution waveform and the complexity change waveform, a time interval where a specified neural waveform is located is obtained, The number of spikes per unit time in the time interval in the spike distribution waveform is smaller than the number of spikes per unit time in the previous unit time. The energy density per unit time of the time interval of each of the multiple energy distribution waveforms is less than the energy density per unit time of the previous time interval. The number of spikes per unit time in the time interval in the complexity variation waveform is greater than a preset value, The transient frequency oscillation energy analysis includes: Performing empirical mode decomposition on the brain wave signal to obtain a plurality of eigenmode signals; Calculating the instantaneous frequency of each of the eigenmode signals in a plurality of third sampling intervals, and then recording the instantaneous frequency corresponding to each of the third sampling intervals in a time series in an instantaneous frequency distribution diagram corresponding to each of the eigenmode signals, thereby obtaining a plurality of instantaneous frequency distribution waveforms corresponding to the eigenmode signals; and The spectrum energy density or the power spectrum density is calculated for the instantaneous frequency distribution waveform respectively to obtain the energy distribution waveform.

Images