CN119326423A - A method for real-time determination of signal quality in brain impedance measurement - Google Patents

Abstract

The invention belongs to the field of signal quality analysis, in particular to quality analysis of human noninvasive brain impedance measurement signals, and particularly discloses a method for judging signal quality in real time in brain impedance measurement, which is characterized in that brain wave signals formed by head excitation are respectively subjected to signal stability, signal intensity and noise interference degree analysis, so that the signal quality is judged according to analysis results, the multidimensional judgment of the brain wave signal quality is realized, then, during the brain wave signal acquisition process, the voltage electrode is used for recording brain wave signals, and meanwhile, the impedance monitoring equipment and the displacement sensor are used for acquiring the contact impedance and the contact movement information between the current electrode and the scalp, so that the measurement parameter range of the contact impedance is expanded, on the basis, the acquired contact impedance and the acquired contact movement data are analyzed, more comprehensive data support can be provided for the identification of brain wave signal quality abnormality, and the accuracy and the reliability of abnormality classification are further improved.

Description

Method for real-time judging signal quality in brain impedance measurement

Technical Field

The invention belongs to the field of signal quality analysis, in particular to quality analysis of a human noninvasive brain impedance measurement signal, and particularly discloses a method for judging signal quality in real time in brain impedance measurement.

Background

Brain wave measurement is a technology for recording brain electrical signals, and has wide application requirements in various fields, from medical diagnosis to scientific research to man-machine interaction and education, and provides valuable information sources to help us to better understand brain functions and the performances thereof under disease states.

In the electroencephalogram acquisition process, a current electrode is usually adopted to be in contact with the scalp, a high-frequency alternating current excitation signal is applied, and a voltage electrode is responsible for detecting and recording an output signal generated by the high-frequency alternating current excitation signal so as to form brain waves. However, due to non-ideal factors such as electromagnetic interference and skin relaxation in the environment, the quality of brain wave signals may be affected, and in particular, the signal quality problem caused by skin relaxation is particularly obvious. Therefore, a careful evaluation of the quality of the brain wave signal is necessary, and such an evaluation should be performed on the basis of a measurement of the contact impedance formed by the electrode and the skin to ensure the reliability and accuracy of the signal.

The traditional Chinese patent of the invention with the publication number of CN109730675A provides an electroencephalogram signal quality detection system and an electroencephalogram signal quality detection method, and an electroencephalogram signal is accessed by utilizing a signal processing module to output an electroencephalogram signal quality parameter, wherein the electroencephalogram signal quality parameter comprises a signal quality ratio and a lead line impedance, so that whether the problem of acquisition environmental influence exists in the process of accessing the electroencephalogram signal to the signal processing module can be detected in real time through the signal quality ratio and the lead line impedance. However, the signal quality ratio defined in this scheme is mainly used for evaluating the stability of brain wave signals in the process of transmitting the brain wave signals to the low noise amplifier through the lead wire, however, there is a limitation in evaluating the signal quality by the signal quality ratio alone because it cannot comprehensively reflect the overall quality condition of the brain wave signals.

In addition, the lead wire impedance defined in the scheme only reflects impedance change caused by contact pressure between the electrode and the scalp, but relative movement between the electrode and the scalp also has a significant influence on signal quality in an actual measurement process, so that measurement of the contact impedance by the scheme appears to be too immobilized, so that judgment on brain wave signal quality according to the contact impedance measurement lacks comprehensive data support, and when brain wave signals are abnormal and the lead wire impedance is normal, signal abnormality is easily and erroneously attributed to other factors, thereby causing erroneous judgment on signal abnormality.

Disclosure of Invention

In order to overcome the defects, the invention discloses a method for judging the signal quality in real time in brain impedance measurement, and the corresponding technical problems can be effectively solved.

The object of the invention is achieved by a method for determining signal quality in real time in brain impedance measurement comprising the steps of placing a current electrode and a voltage electrode at specified positions of the head while providing an impedance monitoring device and a displacement sensor at the respective specified positions.

An excitation signal is applied to the head through the current electrode by using a high-frequency alternating current signal, a voltage signal is acquired by the voltage electrode to form an brain wave signal, and the contact impedance and the contact movement of the current electrode and the head are acquired by the impedance monitoring equipment and the displacement sensor to form a contact impedance time sequence set and a contact movement time sequence set.

The formed brain wave signals are subjected to frequency spectrum transformation after being subjected to filtering treatment, so that the period components and the period length are identified by analyzing the peak values in the spectrogram, and the acquisition period is selected according to the period length.

The method comprises the steps of intercepting measurement signals in an acquisition period from continuous brain wave signals and dividing the measurement signals into a plurality of period windows.

The signal stability is analyzed by performing standard deviation calculation on the signals in each period window and comparing the standard deviation of the signals of adjacent period windows.

And extracting the signal amplitude range of each period window to analyze the signal intensity.

And carrying out noise detection on the signals of each period window to analyze noise interference degree.

And importing the signal stability, the signal strength and the noise interference degree in the acquisition period into a signal quality judgment model to obtain a signal quality score of the acquisition period, comparing the signal quality score with a preset threshold, and carrying out abnormal prompt when the signal quality score is lower than the preset threshold.

When an abnormal prompt occurs, intercepting a contact impedance time sequence set and a contact movement time sequence set in an acquisition period where the abnormal prompt time is positioned, and carrying out abnormal pointing identification.

By combining all the technical schemes, the brain wave signal quality judging method has the advantages that 1, brain wave signals formed by head excitation are respectively analyzed for signal stability, signal intensity and noise interference, so that the signal quality is judged according to analysis results, multidimensional judgment of the brain wave signal quality is realized, various key attributes of the signal quality can be covered on the whole, and comprehensiveness of signal quality judgment are ensured by providing different visual angles through each dimension.

2. According to the invention, during the brain wave signal acquisition process, the voltage electrode is utilized to record the brain wave signal, and meanwhile, the impedance monitoring equipment and the displacement sensor are utilized to acquire the contact impedance and the contact movement information between the current electrode and the scalp, so that the measurement parameter range of the contact impedance is expanded. On the basis, by analyzing the acquired contact impedance and contact movement data, more comprehensive data support can be provided for identifying the brain wave signal quality abnormality, and the accuracy and reliability of abnormality classification are further improved.

Drawings

The invention will be further described with reference to the accompanying drawings, in which embodiments do not constitute any limitation of the invention, and other drawings can be obtained by one of ordinary skill in the art without inventive effort from the following drawings.

FIG. 1 is a diagram of the steps of the method of the present invention.

Fig. 2 is a schematic diagram of the whole process of acquisition and use of brain wave signals, contact impedance and contact movement displacement in the invention.

FIG. 3 is a flow chart of the anomaly orientation recognition implementation in the present invention.

Detailed Description

The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

Referring to fig. 1, the present invention proposes a method for determining signal quality in real time in brain impedance measurement, comprising the steps of placing a current electrode and a voltage electrode at specified positions of a head while providing an impedance monitoring device and a displacement sensor at the corresponding specified positions.

In the operation of the above scheme, the current electrode is placed at a designated position of the head for applying a high frequency alternating current signal. The voltage electrodes are placed adjacent or symmetrical to the current electrodes for collecting the voltage signals, and the electrode layout is typically chosen to obtain the maximum signal-to-noise ratio.

An impedance monitoring device and displacement sensor are placed under or in close proximity to the current electrode for monitoring in real time the contact impedance and contact movement displacement between the electrode and the scalp.

An excitation signal is applied to the head through the current electrode by using a high-frequency alternating current signal, a voltage signal is collected by the voltage electrode to form an brain wave signal, and the contact impedance and the contact movement displacement of the current electrode and the head are collected by the impedance monitoring equipment and the displacement sensor to form a contact impedance time sequence set and a contact movement time sequence set.

It is to be understood that contact resistance refers to the resistance encountered when current flows between the electrode and the skin surface, and in general, contact resistance is related to the contact state of the electrode with the skin, such as contact pressure, and proper contact pressure can improve the contact of the electrode with the skin and reduce the contact resistance. Too much pressure may lead to skin damage and too little pressure may lead to poor contact, contact resistance is also associated with skin conditions, such as skin relaxation changes, skin relaxation changes (e.g. muscle contraction, perspiration, etc.) affecting contact resistance, high contact resistance leading to more noise being mixed in the signal, since high resistance increases the resistance in the signal transmission path, resulting in more random noise being introduced into the signal which may lead to reduced signal quality, affecting signal reliability.

Again, it should be appreciated that contact movement displacement refers to movement displacement between the electrode and the designated location of the head, typically in direct relation to the relative movement between the electrode and the skin. In particular, such movement may be a result of movement of the head, movement of the electrode itself, or a combination of both. Poor contact may result when the contact position between the electrode and the skin is changed, thereby increasing contact resistance and affecting signal quality.

The formed brain wave signals are subjected to frequency spectrum transformation after being subjected to filtering treatment, so that the period components and the period length are identified by analyzing the peak values in the spectrogram, and the acquisition period is selected according to the period length.

It is to be understood that after the brain wave signal is formed, the brain wave signal needs to be filtered firstly to remove unnecessary frequency components, so that useful signals are reserved, the filtered signals are clearer, visual analysis is easier to perform, and clear data support is provided for signal quality judgment.

It should be further understood that performing the spectral transformation on the filtered brain wave signal is to transform the signal in the time domain into the frequency domain spectrum using fourier transformation, so that the energy distribution of the signal on different frequencies can be obtained.

In the operation which can be realized by the scheme, the peak value detection algorithm is used for identifying the peak value in the spectrogram, wherein the maximum peak value corresponds to the period component of the signal, because the period component can generate significant energy concentration in the frequency spectrum, namely, the peak value in the frequency spectrum, the period component existing in the signal can be determined by analyzing the peak position in the spectrogram, the frequency corresponding to the peak value is extracted from the spectrogram to be the period frequency, and the period length is obtained by taking the reciprocal of the period frequency.

It is noted that the resulting brain wave signals are periodic due to the fact that the neuronal activity in the brain exhibits a regular synchronous discharge pattern.

In the operation that above-mentioned scheme further can be realized, select the collection period according to cycle length and select the collection period can be chosen by the duration of a plurality of cycle length, for example 5 cycle length, select the cycle length of a certain amount can ensure that the data of gathering is representative, can cover a complete periodic pattern promptly, through gathering the data of a plurality of cycles simultaneously, can reduce the bias that causes because of the abnormality in the single cycle for the data is more comprehensive, stable, reduces the influence of accidental factor.

The brain wave signals in the acquisition period are intercepted from the continuous brain wave signals and divided into a plurality of period windows, wherein each period window corresponds to a signal with a period length, and in the example that the acquisition period is 5 periods, the number of the period windows obtained by division is 5.

According to the invention, the formed brain wave signals are subjected to periodic component identification, so that the signals are subjected to periodic window division based on the period length, the signals in each periodic window have higher internal normalization, so that the data are more comparable among different periods, the subsequent signal quality judgment is facilitated, meanwhile, the abnormal signals are easier to identify due to the periodic window division, the abnormal signals possibly break the periodic mode, and in addition, the signals are divided into a plurality of periodic windows for processing, so that the calculation burden of processing a large amount of data at one time can be reduced, and the processing efficiency is improved.

The signal stability is analyzed by calculating the standard deviation of the signals in each period window and comparing the standard deviation of the signals of the adjacent period windows, and the specific analysis is that the standard deviation of the signals of the later period window in the adjacent period window and the standard deviation of the signals of the preceding period window are subjected to the absolute value difference and divided by the period length to obtain the change rate of the standard deviation of the signals of the adjacent period windows.

It should be noted that the standard deviation of the signal reflects the fluctuation of the signal in each window.

Counting the number of adjacent period windows which can be divided in an acquisition period, wherein the acquisition period is an example value of 5 period lengths, 4 adjacent period windows are obtained through division, namely 1 and 2,2 and 3,3 and 4,4 and 5, the signal standard deviation change rate of each adjacent period window is taken as the maximum value, the minimum value is the maximum value, the result of dividing the maximum value by the maximum signal standard deviation change rate is taken as the reduction number, the value 1 is taken as the reduced number, and the signal stability is obtained through the reduction operation.

The invention realizes the signal fluctuation analysis of the single periodic window by carrying out standard deviation calculation on the signal of each periodic window in the acquisition period, and carries out signal standard deviation change rate analysis by means of the signal standard deviation of the adjacent periodic window, thus obtaining the signal stability, ensuring the whole coverage of the signal change in the whole acquisition period, and capturing the fine fluctuation of the signal in each periodic window, so that the signal stability analysis is more scientific and reliable.

The signal amplitude range of each period window is extracted to analyze the signal intensity, and the specific analysis is that the signal amplitude range of each period window in the acquisition period is formed into a signal amplitude change curve in a two-dimensional coordinate system constructed by taking the period window as a horizontal axis and taking the signal amplitude range as a vertical axis.

It should be added that, in order to ensure that the drawing of the signal amplitude variation curve is reliable and truly reflects the actual situation, it is recommended to select an acquisition period comprising a greater period length, so as to provide sufficient sample data for the signal amplitude range within each period window, thereby enhancing the representativeness and reliability of the curve.

It should be noted that the signal amplitude range refers to the difference between the maximum value and the minimum value in the signal in each period window, and in general, the larger the signal amplitude range, the larger the amplitude fluctuation of the signal, and thus the larger the energy of the signal and the larger the signal strength in the same duration.

Marking inflection points in the signal amplitude change curve, extracting horizontal segments from the signal amplitude change curve based on the marked inflection points, and further calculating the average signal amplitude range of the horizontal segments as a main signal amplitude range.

It should be noted that, although there is a certain periodicity in the formed brain wave signal, there is a certain difference in the signal amplitude ranges of adjacent periods due to the inherent randomness and noise influence of the signal itself, which results in that the formed signal amplitude variation curve is not a straight line with perfect level but presents a fluctuating state, and the signal amplitude variation curve can be divided into different segments according to the position of the inflection point by inflection point identification (the inflection point is usually the position of the local maximum or minimum in the data curve), wherein the horizontal segment refers to the portion where the slope of the curve is close to zero. Horizontal segment refers to a state where the signal amplitude range remains relatively stable for a longer period of time. This stability generally means that the signal is in a more optimal state of operation, i.e. the signal is less disturbed, and the horizontal segment represents a typical behavior of the signal in normal operation, which can be used as a basis for evaluating the signal strength.

The signal strength is calculated by comparing the subject signal amplitude range with the reference signal amplitude range.

It should be understood that, since the signal amplitude range is dimensional, the main signal amplitude range needs to be compared with the reference signal amplitude range to eliminate the dimension when the signal intensity analysis is performed, and the main signal amplitude range and the reference signal amplitude range may be divided to obtain the signal intensity, for example.

Specifically, the reference signal amplitude range can be selected according to the historical brain wave measurement signals, and specifically, the historical brain wave measurement signals formed under the excitation of the same high-frequency alternating current signals can be used as the reference signal amplitude range by taking the average value of the signal amplitude ranges of the historical brain wave measurement signals.

And carrying out noise detection on the signals of each period window to analyze noise interference degree.

The improved operation of the scheme is implemented by extracting signal envelopes of all periodic windows, further performing coincidence comparison on the signal envelopes of adjacent periodic windows, marking non-coincidence areas from the signal envelopes, and obtaining the envelope length and the time interval corresponding to the non-coincidence areas.

It should be added that the envelope of a signal is a trend line of the amplitude of the signal over time, in particular the upper and lower limit curves of the signal, which are usually used to represent the trajectories of the maxima and minima of the signal.

It is further added that by comparing the envelopes of adjacent periodic windows, it is possible to find areas that should normally coincide but that differ significantly in some time periods. These non-coincident regions are often caused by noise, which can cause abnormal changes in signal amplitude. The envelope length and the time interval corresponding to the non-coincident region are obtained, and the influence range of noise can be further quantized.

And respectively calculating the difference between the envelope length and the time interval corresponding to the non-coincident region formed by all adjacent periodic windows to obtain the difference between the non-coincident envelope length and the non-coincident time interval, and further adding the differences to obtain the envelope coincidence difference of noise detection.

In the optional operation of the scheme, the non-coincident envelope length difference is calculated by substituting the envelope length corresponding to each non-coincident region into the calculation processCalculating to obtain a non-overlapping envelope length difference DD _long, wherein l _i represents the envelope length corresponding to the i-th non-overlapping region, i represents the number of the non-overlapping region, i=1, 2. From this equation, the larger the envelope length difference corresponding to each non-overlapping region is, the larger the non-overlapping envelope length difference is.

In a further optional operation of the above solution, the non-overlapping time interval difference is calculated by extracting a start time and an end time from time intervals corresponding to the non-overlapping regions, respectively, and obtaining the duration.

And comparing the starting time corresponding to each non-coincident region to obtain the starting interval duration.

And comparing the starting time corresponding to each non-coincident region to obtain the ending interval duration.

Substituting the duration of each non-coincident region into the calculated time period by combining the starting interval duration and the ending interval durationObtaining a non-coincident time interval difference DD _time, wherein Deltat _start、△t_end respectively represents a start interval duration and an end interval duration, DN _i represents the duration of an i-th non-coincident region, and T represents the period length. The longer the starting interval duration, the longer the ending interval duration and the longer the duration difference of the non-overlapping areas are, the larger the difference of the non-overlapping time intervals is.

It should be explained that the non-overlapping area formed by the envelopes of the adjacent periodic windows is caused by noise, wherein the defined non-overlapping envelope length difference represents the difference of the signal amplitude variation range caused by noise, the non-overlapping time interval difference represents the difference of the signal amplitude variation time span caused by noise, and when the two differences are larger, the influence range and the time span of the noise are unstable, and the influence degree of the noise is larger, so that the influence degree of the noise can be quantified by the non-overlapping envelope length difference and the non-overlapping time interval difference.

And carrying out frequency spectrum transformation on the signals in each period window, calculating the signal to noise ratio of each period window from the obtained spectrogram, and taking the minimum signal to noise ratio as the signal to noise ratio of noise detection.

It should be appreciated that the signal-to-noise ratio is a measure of the ratio of signal energy to noise energy in a signal.

It is to be appreciated that the minimum signal-to-noise ratio generally represents the worst signal quality condition over all periodic windows. In signal processing we are concerned about the behaviour of the signal under the most adverse conditions, as this is the limit of the performance of the signal processing system.

Further analyzing noise interference degree by substituting envelope coincidence difference and signal-to-noise ratio of noise detection into analysis typeAnd obtaining noise interference NL, wherein ED and SR respectively represent envelope coincidence difference and signal-to-noise ratio of noise detection, and a and b respectively represent specific gravity factors corresponding to the envelope coincidence difference and the signal-to-noise ratio. Wherein the larger the envelope coincidence difference of the noise detection is, the smaller the signal to noise ratio is, and the larger the noise interference degree is.

It should be noted that, when noise detection is performed, the time domain angle (envelope coincidence difference) and the frequency domain angle (signal to noise ratio) of the signal are combined to perform noise detection, so that more comprehensive information can be provided, and the accuracy, the robustness and the adaptability of noise detection can be improved. The method can better understand the characteristics of the signals and improve the noise detection effect.

It should be further noted that, since the signal-to-noise ratio may directly reflect the ratio of useful information to noise in a signal, the signal-to-noise ratio is an important indicator for evaluating the noise interference level in many cases. Thus, in the calculation of the set noise interference level, the specific gravity factor corresponding to the signal to noise ratio can be set to be slightly larger, and the specific gravity factors corresponding to the envelope overlap difference and the signal to noise ratio are respectively 0.4 and 0.6.

And importing the signal stability, the signal strength and the noise interference degree in the acquisition period into a signal quality judgment model to obtain a signal quality score of the acquisition period, comparing the signal quality score with a preset threshold, and carrying out abnormal prompt when the signal quality score is lower than the preset threshold.

In the above embodiment of the operation, the signal quality evaluation model may set the effective value of each parameter of the signal stability, the signal strength and the noise interference according to the signal characteristics (such as the frequency) of the excitation signal, because some frequencies of the excitation signal are more susceptible to noise, and then the signal stability, the signal strength and the noise interference in the acquisition period are compared with the corresponding effective values to obtain the signal quality score.

When abnormal prompt occurs, intercepting a contact impedance time sequence set and a contact movement time sequence set in an acquisition period of the abnormal prompt time to conduct abnormal pointing recognition, wherein the specific recognition process is that the contact impedance time sequence set in the acquisition period of the abnormal prompt time is compared with normal contact impedance, and a time proportion value which does not accord with the normal contact impedance is counted.

In the above, the normal contact impedance value depends on the specific application scenario, and it is generally desirable for the brain wave signal to have a contact impedance lower than 5 kiloohms (kΩ).

The method comprises the steps of carrying out mutation analysis on a contact impedance time sequence set and a contact movement time sequence set in an acquisition period where an abnormal prompt time is located to obtain mutation indexes corresponding to contact impedance and contact movement, wherein the mutation analysis comprises the steps of respectively carrying out average value calculation and standard deviation calculation on the contact impedance time sequence set and the contact movement time sequence set, and taking percentage calculation on the contact impedance and the mutation indexes corresponding to contact movement after dividing the standard deviation by the average value, wherein the higher the mutation index is, the larger the fluctuation of data is indicated.

And comparing the variation indexes corresponding to the contact impedance and the contact movement with the configured safety variation indexes respectively, and simultaneously comparing the time proportion value which does not accord with the normal contact impedance with the configured safety proportion value, thereby outputting parameters higher than the safety variation indexes or higher than the safety proportion value as abnormal directions.

The above safety mutation index and the safety ratio value are initially set, wherein the purpose is to assist in abnormality direction recognition.

The identification of specific abnormal orientations in the above is shown in fig. 3.

Further, the abnormal pointing identification further comprises abnormal time positioning, and the method specifically comprises the following steps of collecting abnormal pointing time sequences in a two-dimensional coordinate system taking time as a horizontal axis and marking a plurality of points in the two-dimensional coordinate system taking abnormal pointing as a vertical axis in an acquisition period where abnormal prompting time is located, and drawing an abnormal pointing change curve.

In the example of the above operation, the abnormality-oriented timing set is the contact impedance timing set when the abnormality is oriented to contact impedance fluctuation, and the abnormality-oriented timing set is the contact movement timing set when the abnormality is oriented to contact movement fluctuation.

The whole process of acquisition and use of the midbrain wave signals, the contact impedance and the contact movement displacement is shown in fig. 2.

And respectively acquiring slopes of points on the abnormal directional change curve, and carrying out mutation identification on the slopes of the adjacent points, wherein the mutation identification can be carried out by comparing the absolute value differences of the slopes of the adjacent points with the absolute value differences of the critical slopes, and when the absolute value difference of the slope of a certain adjacent point is higher than the absolute value difference of the critical slope, the fact that the slopes of the adjacent points have the mutation is identified, and then the time period of the adjacent point on the transverse axis is taken as abnormal time.

The invention can carry out targeted processing of abnormal pointing by positioning abnormal time.

The foregoing is merely illustrative of the structures of this invention and various modifications, additions and substitutions for those skilled in the art of describing particular embodiments without departing from the structures of the invention or exceeding the scope of the invention as defined by the claims.

Claims (10)

1. A method for real-time determination of signal quality in brain impedance measurement, characterized in that it comprises the following steps: placing current electrodes and voltage electrodes at designated locations on the head, and setting impedance monitoring equipment and displacement sensors at corresponding designated locations; A high-frequency alternating current signal is used to apply an excitation signal to the head through the current electrode, and the voltage electrode collects the voltage signal to form an electroencephalogram signal, and an impedance monitoring device and a displacement sensor collect the contact impedance and contact movement displacement between the current electrode and the head to form a contact impedance time series set and a contact movement time series set; The formed brain wave signal is filtered and then transformed into a spectrum, thereby identifying the periodic component and period length by analyzing the peak value in the spectrum diagram, and then selecting the acquisition time period according to the period length; The brain wave signals within the acquisition period are intercepted from the continuous brain wave signals and divided into a number of period windows; The signal stability is analyzed by calculating the standard deviation of the signal in each cycle window and comparing the standard deviation of the signals in adjacent cycle windows. Extract the signal amplitude range of each period window to analyze the signal strength; Perform noise detection on the signal of each period window to analyze the noise interference degree; The signal stability, signal strength and noise interference degree during the acquisition period are introduced into the signal quality evaluation model to obtain the signal quality score of the acquisition period, and compared with the predetermined threshold. When the signal quality score is lower than the predetermined threshold, an abnormal prompt is issued; When an abnormal prompt appears, the contact impedance time series set and the contact movement time series set within the acquisition period of the abnormal prompt time are intercepted to identify the abnormal direction. 2. A method for real-time determination of signal quality in brain impedance measurement according to claim 1, characterized in that: the analysis of signal stability is as follows: The standard deviation of the signal in the next cycle window in the adjacent cycle window is obtained by taking the absolute value of the difference between the standard deviation of the signal in the previous cycle window and the period length to obtain the rate of change of the signal standard deviation in the adjacent cycle window; The number of adjacent period windows that can be divided into the statistical collection period is then calculated, and the maximum value of the signal standard deviation change rate of each adjacent period window is taken. The subtrahend is taken after dividing the maximum value by the maximum signal standard deviation change rate, and the value 1 is taken as the minuend, thereby constructing a subtraction operation to obtain the signal stability. 3. A method for real-time determination of signal quality in brain impedance measurement according to claim 1, characterized in that: the analysis of signal strength refers to the following process: The signal amplitude range of each period window in the acquisition period is used to form a signal amplitude variation curve in a two-dimensional coordinate system with the period window as the horizontal axis and the signal amplitude range as the vertical axis; An inflection point is marked in the signal amplitude change curve, and a horizontal segment is extracted from the signal amplitude change curve based on the marked inflection point, and then an average signal amplitude range of the horizontal segment is calculated as the main signal amplitude range; The signal strength is calculated by comparing the subject signal amplitude range with the reference signal amplitude range. 4. A method for real-time determination of signal quality in brain impedance measurement according to claim 1, characterized in that: the noise detection is implemented as follows: Extract the signal envelope of each period window, and then compare the signal envelopes of adjacent period windows, mark the non-overlapping areas, and obtain the envelope length and time interval corresponding to the non-overlapping areas; The envelope lengths and time intervals corresponding to the non-overlapping areas formed by all adjacent periodic windows are respectively calculated to obtain the non-overlapping envelope length difference and the non-overlapping time interval difference, and then the two differences are added to obtain the envelope overlap difference of noise detection; The signal in each period window is transformed into a spectrum, and the signal-to-noise ratio of each period window is calculated from the obtained spectrum diagram, and the minimum signal-to-noise ratio is taken as the signal-to-noise ratio of noise detection. 5. A method for real-time determination of signal quality in brain impedance measurement according to claim 4, characterized in that: the difference in length of the non-coincident envelope refers to the following calculation process: Substitute the envelope lengths corresponding to each non-overlapping area into the calculation formula The non-overlapping envelope length difference DD _long is calculated, where l _i represents the envelope length corresponding to the i-th non-overlapping area, i represents the non-overlapping area number, i=1, 2, ..., n, and n represents the number of non-overlapping areas. 6. A method for real-time determination of signal quality in brain impedance measurement as claimed in claim 4, characterized in that: the difference between the non-overlapping time intervals is calculated as follows: Extract the start time and end time from the time interval corresponding to each non-overlapping area, and obtain the duration; Compare the start times corresponding to the non-overlapping areas to obtain the start interval duration; Compare the start time corresponding to each non-overlapping area to obtain the end interval duration; Substitute the duration of each non-overlapping area into the calculation formula, combined with the start interval duration and the end interval duration. The non-overlapping time interval difference DD _time is obtained, where △t _start and △t _end represent the start interval duration and the end interval duration respectively, DN _i represents the duration of the i-th non-overlapping area, and T represents the cycle length. 7. A method for real-time determination of signal quality in brain impedance measurement according to claim 4, characterized in that: the analysis of noise interference degree is as follows: Substitute the envelope overlap difference and signal-to-noise ratio of noise detection into the analytical formula The noise interference degree NL is obtained, where ED and SR represent the envelope overlap difference and signal-to-noise ratio of noise detection respectively, and a and b represent the weighting factors corresponding to the envelope overlap difference and signal-to-noise ratio respectively. 8. A method for real-time determination of signal quality in brain impedance measurement according to claim 1, characterized in that: the abnormal direction identification is as follows: The contact impedance time series set and the contact movement time series set in the acquisition period of the abnormal prompt time are subjected to variation analysis to obtain the variation index corresponding to the contact impedance and the contact movement; The contact impedance time series set in the acquisition period of the abnormal prompt time is compared with the normal contact impedance, and the time proportion value that does not meet the normal contact impedance is counted; The variation indexes corresponding to the contact impedance and contact movement are compared with the configured safety variation indexes respectively. At the same time, the time proportion value that does not meet the normal contact impedance is compared with the configured safety proportion value. The parameters higher than the safety variation index or higher than the safety proportion value are output as abnormal indications. 9. A method for real-time determination of signal quality in brain impedance measurement according to claim 8, characterized in that: the variation analysis is implemented as follows: The mean value and standard deviation of the contact impedance time series set and the contact movement time series set are calculated respectively, and the variation index corresponding to the contact impedance and contact movement is calculated using the mean value and standard deviation. 10. A method for real-time determination of signal quality in brain impedance measurement according to claim 1, characterized in that: the abnormal direction identification also includes abnormal time location, and the specific operations are as follows: During the collection period of the abnormal prompt time, the abnormal direction time series set is marked with several points in a two-dimensional coordinate system with time as the horizontal axis and the abnormal direction as the vertical axis to draw an abnormal direction change curve; The slope of each point on the abnormal pointing change curve is obtained respectively, and the slopes of adjacent points are identified for mutation. If a mutation is found in the slope of an adjacent point, the time period of the adjacent point on the horizontal axis is taken as the abnormal time.