CN115530845A

CN115530845A - Method for detecting abnormal discharge in epilepsia electroencephalogram signal

Info

Publication number: CN115530845A
Application number: CN202211266426.2A
Authority: CN
Inventors: 王刚; 顾慧超
Original assignee: Changzhou Rishena Medical Equipment Co ltd
Current assignee: Changzhou Rishena Medical Equipment Co ltd
Priority date: 2022-10-17
Filing date: 2022-10-17
Publication date: 2022-12-30

Abstract

The invention relates to the technical field of epilepsia electroencephalogram signals, in particular to a method for detecting abnormal discharge in epilepsia electroencephalogram signals. The sharpness ratio algorithm flow is as follows: firstly, filtering the original electroencephalogram signals at 0.1-75Hz and trapping at 50Hz to remove direct-current components, high-frequency noise and power frequency noise. Counting all extreme points peak in the filtered signal, screening all found extreme points, if the amplitude variation degree between two continuous extreme points is larger than a certain threshold value, regarding the extreme point as an extreme point peak _ q meeting the condition, subtracting the amplitude of the point before 4ms and the amplitude of the point after 4ms from the amplitude of each extreme point peak _ q, taking an absolute value, calculating the average value of the two values, wherein the result is called sharpness, adding the sharpness of each peak _ q, and the ratio of the sharpness to the total number N of the peak _ q is the sharpness ratio sharpnessratio. The method has the advantages of small calculation amount and high real-time performance, can accurately detect epileptiform brain electrical activity, and provides help for timely rescue.

Description

Method for detecting abnormal discharge in epilepsia electroencephalogram signal

Technical Field

The invention relates to an epilepsy detection method, in particular to a method for detecting abnormal discharge in an epilepsy electroencephalogram signal.

Background

Epilepsy is a serious, recurrent, chronic neurological disorder, one of the most common neurological disorders in the world. Seizures are caused by an excessive synchronous firing of cerebral neurons and are characterized by disturbances or loss of consciousness, abnormal movement, mental or sensory disturbances, or disturbances of the autonomic nervous system. Due to the characteristics of epileptic seizures, the life of epileptic patients can be affected to a certain extent, such as incapability of driving and working, and even serious incapability of normal life. In addition, over time, seizures tend to become more frequent and more severe, possibly resulting in deterioration of other brain functions or physical damage.

Epilepsy can be broadly divided into two broad categories, focal and global, according to seizure symptoms, focal seizures meaning that only a portion of the cerebral hemisphere is affected, whereas in global seizures all regions of the brain are affected.

The traditional epilepsy treatment method mainly comprises drug treatment and surgical treatment.

Common epilepsy treatment drugs comprise sodium valproate, phenobarbital, gabapentin, phenytoin and the like, although the drugs are effective, the drugs usually have certain toxicity, side effects can be caused after long-term administration, and in addition, a large number of patients suffer from drug-resistant epilepsy and do not respond to the drugs.

Surgical methods include radical resection, such as hemisphectomy and corticotomy, and less aggressive resection of lesions, such as transection and stereoablation. The problem with the surgical approach is also apparent, first, that the surgical treatment is not necessarily completely successful and many epileptic patients are unable to get rid of seizures even after undergoing radical brain surgery. Second, these surgical procedures are often at high risk for complications and can result in long-term damage to the brain regions with important functions as well as to various cognitive and neurological functions.

Neurostimulation is an emerging method for treating epilepsy, and a certain electrical stimulation is directly given to neurons at the corresponding part or adjacent nerve tissues to achieve a therapeutic effect. Currently, the most clinically applied Neurostimulation techniques mainly include Vagus Nerve Stimulation (VNS), deep Brain Stimulation (DBS), and Reactive Neurostimulation (RNS).

VNS is an open-loop extracranial stimulator developed in 20 actual 80 s, which achieved the function of regulating the target area of the cerebral cortex by implanting a pulse generator in the left anterior chest, implanting electrodes in the neck, and performing electrical pulse stimulation on afferent fibers of the left vagus nerve of the neck.

DBS is an open-loop intracranial stimulator, accurate positioning is required to be realized before an operation, an electrode is implanted into a target area, and the excitation state of a nucleus where the electrode is located is changed by applying an electric stimulation signal with certain intensity to the electrode to realize nerve regulation. DBS regulates stimulation parameters, inhibits or activates neurons, releases corresponding nerve regulation factors, changes the intrinsic neurophysiological characteristics of an epileptic network, and improves the threshold value of epileptic seizure, thereby achieving the purpose of inhibiting epileptic seizure.

RNS is a closed-loop intracranial stimulator that continuously monitors neural activity at the focus of an epileptic seizure, and only responds to the stimulation when epileptiform activity is detected, firstly, the damage and side effects to the brain due to continuous stimulation can be reduced compared to open-loop neuromodulation methods that stimulate continuously or at a fixed time; second, the service life of the stimulator battery may be extended.

The RNS system judges the current state of a patient by collecting, processing and analyzing electroencephalogram signals and extracting useful characteristics through a corresponding algorithm, and outputs a control signal to an actuator according to the useful characteristics, adjusts stimulation parameters and gives corresponding stimulation. Therefore, for the RNS system, an algorithm for detecting epileptic activity (such as spike, spike-slow wave, etc.) in brain electrical signals is particularly important. The detection algorithm must meet several requirements, first, to be effective, it must be able to accurately and effectively detect epileptic activity in the signal. Secondly, to meet the real-time performance, the whole process needs to be completed within a short time from signal acquisition to signal processing and analysis to final instruction output so as to meet the requirement of real-time control. Finally, the feedback algorithm needs to consider the memory and calculation power of the single chip microcomputer, the used data memory cannot be excessive, and the calculation cannot be too complex.

There are three detection algorithms used in the RNS system of NeuroPace: power Change (Power Change), band pass filtering (Bandpass), and Area (Area). The power variation is also called Line Length (Line Length), and refers to the sum of distances between two adjacent sampling points in data of a period of time, which can reflect the power of a signal in a period of time. The disadvantage is that it is susceptible to interference, is not targeted, and any interference that causes fluctuations in the waveform is misidentified as epileptic activity. Band-pass filtering is used to detect activity in a specific frequency band, such as theta, alpha, beta, and gamma, and has the disadvantage that the detection frequency band needs to be manually set, and if the frequency band setting is not accurate, the detection effect is poor. The area is the area of a region enclosed by a signal waveform and an x axis, and in a discrete system, the area is the sum of absolute values of each sampling point, is too simple and is rarely used.

Disclosure of Invention

The invention aims to solve the defects and provides a method for detecting abnormal discharge in an epilepsia electroencephalogram signal.

In order to overcome the defects in the background art, the technical scheme adopted by the invention for solving the technical problems is as follows: the method for detecting abnormal discharge in the electroencephalogram signals of epilepsy is based on the fact that the electroencephalogram signals have different characteristics when epilepsy is in different stages, abnormal waves of spike waves and sharp waves can appear in the waveforms of the electroencephalogram signals due to the abnormal discharge of over-synchronization of cerebral neurons, and a sharpness ratio algorithm is provided aiming at the characteristics, and the algorithm flow is as follows:

s1, acquiring focus point electroencephalogram signals of an epileptic, dividing original signals according to one second, and analyzing the signals in each second;

s2, filtering the signal at 0.5-75Hz and trapping the signal at 50Hz, and removing direct-current components, high-frequency noise and power frequency noise;

s3, after the filtering in the step S2, down-sampling the signal, and if the sampling rate is low originally, omitting the step;

s4, counting all extreme points peak in the signals in the step S3, screening all extreme points, and if the amplitude variation degree between two continuous extreme points is larger than a certain threshold value thr1, the thr1 is not too large, so that a meaningful extreme point is prevented from being filtered, considering the extreme point as an extreme point peak _ q meeting the condition;

s5, subtracting the amplitude of the sampling point before 4ms and the amplitude of the sampling point after 4ms from the amplitude of each extreme point peak _ q in the step S4 respectively, taking an absolute value, then calculating the average value of the absolute value and the absolute value, calling the result as the sharpness of the extreme point, then adding the sharpness of each peak _ q, and taking the ratio of the sharpness of each peak _ q to the total number N of the peak _ q as the sharpness ratio, wherein the calculation formula is as follows:

s6, setting an individualized threshold thr2, comparing the sharpness ratio with the threshold, and if the sharpness ratio exceeds the threshold, indicating that abnormal discharge exists in one second, otherwise, indicating that no abnormal discharge exists.

The threshold thr2 in the step S6 depends on the level of the sharpness ratio of the patient in the stable state, and different patients have different thresholds which need to be set individually according to the actual electroencephalogram of the patient.

The invention has the beneficial effects that:

first, the sharpness ratio algorithm has the advantages that:

1. the typical characteristics of the epileptic abnormal wave are taken for description, so that the accuracy rate is high;

2. only one second of data is processed each time, and the down-sampling processing is carried out, and the calculation process is simple, so that the method occupies less memory, is faster in calculation, meets the real-time control requirement, and is suitable for a system with limited calculation power, namely RNS.

The second, sharpness ratio algorithm, has the significance of:

1. the epileptic seizure is the result of the continuous accumulation of abnormal discharge, and the epileptic seizure will occur when the discharge is accumulated to a certain degree, therefore, no matter in the interval of seizures, the prophase of seizures or the period of seizures, the epileptic seizure can be restrained to a certain degree by giving a certain intensity of stimulation as long as the abnormal discharge is detected;

2. the electroencephalogram signals are used for marking the epileptic seizure, the operation is usually performed by experienced neurologists, the work is time-consuming and labor-consuming, the epileptic seizure automatic detection algorithm can be further realized by means of the sharpness ratio algorithm, and the time is saved.

Drawings

The invention is further illustrated with reference to the following figures and examples.

FIG. 1 is a schematic structural diagram of typical abnormal discharge waveforms, which are spike, spike and spike-slow waves from left to right;

FIG. 2 is a schematic diagram of the structure of a filtered signal waveform;

FIG. 3 is a schematic diagram of the structure of a signal waveform after down-sampling;

FIG. 4 is a schematic diagram of the structure of all extreme points in the signal;

FIG. 5 is a schematic diagram of the structure of the qualified extreme points;

FIG. 6 is a schematic diagram of an enlarged view of the extreme points;

FIG. 7 is a structural diagram illustrating a brain waveform and its sharpness ratio without abnormal discharging phenomenon;

FIG. 8 is a structural diagram illustrating an electroencephalogram waveform with abnormal discharge phenomenon and its sharpness ratio;

fig. 9 is a schematic diagram of the structure of seizure detection.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

Epilepsy is a serious and recurrent nervous system disease, and is caused by abnormal discharges of brain neurons in an oversynchronous manner, wherein the abnormal discharges mainly represent sharp waves, spike waves, spine-slow waves and the like in an electroencephalogram, and the waveforms of the three typical abnormal waves are shown in fig. 1. During the interval of onset or in the early period of onset, the abnormal discharge lasts for several seconds to ten seconds, and the amplitude of the abnormal wave is dozens of microvolts to hundreds of microvolts. During the attack period, the abnormal discharge is violent, lasts for tens of seconds or even minutes, and the amplitude of the abnormal wave can reach thousands of microvolts.

A method for detecting abnormal discharge in electroencephalogram signals of epilepsy is based on the fact that when epilepsy is in different stages, electroencephalogram signals have different characteristics, abnormal waves of spike waves and sharp waves can appear in electroencephalogram signal waveforms due to abnormal discharge of over-synchronization of cerebral neurons, a sharpness ratio algorithm is provided for the characteristics, sharpness ratios describe the sharpness degree of the signal waveforms in unit time, and the higher the value of the sharpness ratio, the more the number of the sharp waves in the signals, the larger the sharpness and the higher the amplitude are. The epileptic activity of the brain is mainly manifested as abnormal discharge, which is mainly manifested as abnormal waves such as spike, and spine-slow wave in the electroencephalogram, and the more severe the abnormal discharge, the more spike and spike, the sharper the waveform, and the higher the amplitude. Therefore, the sharpness ratio can well describe the characteristics of an abnormal wave.

The sharpness ratio algorithm flow is as follows:

s1, acquiring focus point brain electrical signals of an epileptic, dividing original signals according to one second, and analyzing the signals in each second respectively;

s5, subtracting the amplitude of the sampling point before 4ms and the amplitude of the sampling point after 4ms from the amplitude of each extreme point peak _ q in the step S4 respectively, taking an absolute value, then calculating an average value of the absolute value and the absolute value, wherein the result is called sharpness of the extreme point, then adding the sharpness of each peak _ q, and the ratio of the sharpness of each peak _ q to the total number N of the peak _ q is sharpness ratio, wherein the calculation formula is as follows:

The threshold thr2 in step S6 depends on the level of sharpness ratio of the patient in a stable state, and different patients have different thresholds which need to be set individually according to the actual electroencephalogram of the patient.

Example (b):

firstly, dividing an original signal according to one second, analyzing by taking one second data as an example, wherein the one second data comprises one abnormal discharge, and the sampling rate is 1024Hz.

And step two, performing band-pass filtering on the signal at 0.5-75Hz, notching at 50Hz, and removing direct-current components, high-frequency noise and power frequency noise, wherein the waveform after filtering is shown in figure 2.

And step three, performing down-sampling on the signal, wherein the down-sampling aims to reduce unnecessary sampling points so as to improve the operation rate and improve the real-time property. After the filtering, according to the shannon sampling theorem, the sampling rate needs to be higher than twice of the highest frequency in the signal, if the sampling rate of the signal needs to be reduced to 256Hz, the signal needs to be filtered to below 128Hz in the second step, otherwise, high-frequency aliasing is caused. In addition, if the sampling rate is inherently low, this step is omitted. In this embodiment, the signal is reduced to 256Hz, and the waveform of the down-sampled signal is shown in FIG. 3.

Comparing fig. 2 and fig. 3, it can be seen that there is no change in the signal waveform before and after down-sampling, but the number of sampling points per second on the abscissa becomes one fourth of the original number, so as to facilitate the subsequent calculation.

And step four, counting all the extreme points peak in the signal, including a maximum value and a minimum value, as shown in fig. 4. Screening all extreme points, if the amplitude change degree L between two continuous extreme points is larger than a certain threshold value thr1, the extreme point is considered to be the extreme point peak _ q meeting the conditions, the purpose of screening is to eliminate non-pathological changes in signals, thr1 is not too large, and the filtering of meaningful extreme points is prevented. In this embodiment, the threshold thr1 is set to 20 μ V, and the extreme points after screening are shown in fig. 5.

Step five, subtracting the amplitude of the sampling point before 4ms and the amplitude of the sampling point after 4ms from the amplitude of each extreme point peak _ q meeting the step four respectively, taking an absolute value and then solving the average value of the two, wherein the result is called the sharpness of the extreme point, and the calculation formula is as follows:

the area between the abscissa 100-120 in fig. 5 is enlarged, as shown in fig. 6, taking one of the extreme points as an example, where:

V _{peak_q} ＝405.48μV，V _{peak_q-4ms} ，＝346.25μV，V _{peak_q+4ms} =381.41 μ V then the sharpness of this extreme point is:

adding the sharpness of each peak _ q, wherein the ratio of the sum of the sharpness of each peak _ q and the total number N of the peak _ q is the sharpness ratio, and the calculation formula is as follows:

the sharpness ratio of this one second signal was calculated to be 14.18.

Step six, setting a threshold thr2, and setting the threshold thr2 of the patient as 10, so that abnormal discharge exists in the one-second signal.

The effectiveness of the sharpness ratio is proved by taking the electroencephalogram data of another epileptic with abnormal discharge and without abnormal discharge for 10 seconds respectively, drawing the waveforms respectively, and calculating the sharpness ratio. Referring to fig. 7 and 8, the upper half represents signal waveforms, the abscissa represents time for 10 seconds, and the ordinate represents amplitude; the lower half is the sharpness ratio, with each point corresponding to each second of the calculation. In fig. 7, since there is no abnormal discharge, the signal is relatively gentle and the amplitude is relatively low, and the calculated value of the sharpness ratio is small. In fig. 8, the abnormal wave starts to appear in the signal from the third second, gradually increases, and becomes the most serious in the 5 th second, and then gradually decreases, and completely matches with the change trend described by the sharpness ratio waveform, when the abnormal wave appears in the signal, the value of the sharpness ratio increases, when the abnormal wave disappears in the signal, the value of the sharpness ratio decreases, therefore, the threshold thr2 of the patient may be set to 5, and when the sharpness ratio exceeds 5, it is determined that there is abnormal discharge.

In epileptic seizures, the abnormal electrical discharge is more intense and longer lasting than during the inter-seizure or pre-seizure phase. Taking the following one-time epileptic seizure electroencephalogram data of an epileptic patient as an example, the data duration is 383 seconds including one-time epileptic seizure, the sampling rate is 1024Hz, the waveform is drawn, and the sharpness ratio is calculated, as shown in fig. 9. Since a continuous and intense abnormal discharge occurs at the time of a seizure, the value of the sharpness ratio is higher and lasts for a longer time until it is reduced after the seizure is stopped. Between 197 and 276 seconds in fig. 9, i.e., within the dotted line in the figure, the value of the sharpness ratio continuously rises and is maintained at a high level for a long time, and thus it is judged that this interval is a seizure.

To further verify the effectiveness of the sharpness ratio algorithm to detect seizures, 8 epileptic patients (4 men and 4 women) were analyzed for approximately 319 hours of electroencephalogram data and seizures were detected therein, with the results shown in the following table. The results show that all 30 seizures were accurately detected, although the patient's seizure symptoms were not identical.

TABLE 1 results of seizure testing in eight patients

The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be considered to be within the technical scope of the present invention, and the technical solutions and the inventive concepts thereof according to the present invention should be equivalent or changed within the scope of the present invention.

Claims

1. A method for detecting abnormal discharge in an epileptic brain electrical signal is characterized in that the detection method is based on the fact that when epilepsy is in different stages, the brain electrical signal has different characteristics, abnormal waves of spike waves and sharp waves can appear in the brain electrical signal waveform due to abnormal discharge of oversynchronous neurons of a brain, and a sharpness ratio algorithm is provided aiming at the characteristics, and the algorithm flow is as follows:

s5, respectively using each extreme point peak _ q in accordance with the step S4Amplitude minus 4 _ms Amplitude of previous sample point and 4 _ms Then, the absolute value of the amplitude of the sampling point is taken and the average value of the absolute value and the absolute value is calculated, the result is called sharpness of the extreme point, then the sharpness of each peak _ q is added, the ratio of the sharpness ratio to the total number N of the peak _ q is the sharpness ratio, and the calculation formula is as follows:

2. The method for detecting abnormal discharge in electroencephalogram signals of epilepsy as claimed in claim 1, wherein: the threshold thr2 in the step S6 depends on the level of the sharpness ratio of the patient in the stable state, and different patients have different thresholds which need to be set individually according to the actual electroencephalogram of the patient.