Physiological signal processing method, application having the method, and processing system thereof

Abstract

The present invention relates to the field of biomedical signal processing, and in particular to a physiological signal processing method and an application and processing system thereof. The physiological signal processing method comprises: obtaining N×M filtered signals, that is, filtering a single channel signal with M cutoff frequencies f n,m Perform N×M filtering through at least one filter group; obtain an N×M-dimensional signal-to-noise ratio matrix SNR, that is, calculate the signal-to-noise ratio value SNR of each filtered signal separately n,m And form an N×M dimensional signal-to-noise ratio matrix SNR; where the filter bank is composed of N filters; and SNR n,m For a single-channel signal, it passes through the nth filter at the mth cutoff frequency f. n,m The signal-to-noise ratio value of the filtered signal is formed, n = 1, 2, ..., N, m = 1, 2, ..., M. The physiological signal processing method of the present invention is simple, ensures that the physiological signal is a valid signal, can correctly identify the characteristics when used in medical measurement, is convenient for medical evaluation and testing, has high accuracy and strong applicability.

Description

Technical Field

[0001] The present invention relates to the field of biomedical signal processing, and in particular to a physiological signal processing method, an application having the method, and a processing system thereof. Background Art

[0002] The recording and display of physiological signals are important components in biomedical measurements. Physiological signals such as electrocardiogram, heart sound, electroencephalogram, electromyogram, nerve potential or voltage are ultimately recorded or displayed in a certain way. According to the shape of the waveform and its changing law over time, doctors can analyze and judge diseases or physiological changes. Utilizing the feature that magnetic recording technology can record information for a long time, it is also possible to record sporadic diseases such as premature ventricular contractions, providing a basis for doctors to track the situations that have occurred to patients. Therefore, it is necessary to determine whether the features have been correctly identified, for example, the effective frequency of the signal, to detect the correctness of the signal collected by the acquisition device or to judge the acquisition quality; the determination of frequency components is used to determine the induced electroencephalogram components. A prerequisite for applying these technologies is at least an effective physiological signal. Summary of the Invention

[0003] The technical problem to be solved by the present invention is: to avoid errors in medical measurements, resulting in invalid acquisition of physiological signals and incorrect features of the acquired physiological signals. The present invention provides a physiological signal processing method, the processing method is simple, and when used in medical measurements, it can correctly identify features, facilitating medical evaluation and testing, with high accuracy and strong applicability.

[0004] The technical solution adopted by the present invention to solve its technical problems is: a physiological signal processing method, comprising:

[0005] Obtaining N×M filtered signals, that is, filtering a single-channel signal N×M times through at least one filter bank with M cut-off frequencies f n,m ;

[0006] Obtaining an N×M-dimensional signal-to-noise ratio matrix SNR, that is, calculating the signal-to-noise ratio value SNR of each filtered signal respectively n,m and forming an N×M-dimensional signal-to-noise ratio matrix SNR; where

[0007] the filter bank is composed of N filters; and

[0008] SNR n,m is the signal-to-noise ratio value of the filtered signal formed by the single-channel signal passing through the nth filter at the mth cut-off frequency f n,m , n = 1, 2,..., N, m = 1, 2,..., M.

[0009] Further, specifically,[[]]

[0010] The filter is any one of a low-pass filter, a high-pass filter, a band-pass filter, a band-stop filter, an IIR filter, or an FIR filter.

[0011] Furthermore, specifically,

[0012] The obtaining of the N×M dimensional signal-to-noise ratio matrix SNR includes:

[0013] Sliding windowing each filtered signal with a window width T and a sliding window step ΔT, and dividing it into K windows;

[0014] Obtaining the signal-to-noise ratio value of each window respectively through a signal-to-noise ratio calculation method, and then calculating the signal-to-noise ratio value SNR of each filtered signal n,m ;

[0015] Using the signal-to-noise ratio value SNR of each filtered signal n,m to form the matrix SNR; where

[0016] The signal-to-noise ratio calculation method includes any one of a time-domain method, a frequency-domain method, a combined method, or a variant method.

[0017] Furthermore, specifically,

[0018] The frequency-domain method includes:

[0019] Let s n,m (t) represent the sliding window signal of the filtered signal of the nth filter using the mth cut-off frequency f n,m in the kth window, k = 1, 2,..., K;

[0020] Calculate the power spectrum PS n,m (f) of s n,m (t);

[0021] Calculate the signal-to-noise ratio value of the kth window

[0022] and then calculate the signal-to-noise ratio value of each filtered signal

[0023] Furthermore, specifically,

[0024] The time-domain method includes:

[0025] Let s n,m (t) represent the sliding window signal of the filtered signal of the nth filter using the mth cut-off frequency f n,m in the kth window, k = 1, 2,..., K;

[0026] Calculate the maximum and minimum values of the absolute value of s n,m (t), and denote them as s max and s min;

[0027] Calculate the signal-to-noise ratio value SNR of the k-th window k = 20log10(s max / s min );

[0028] Furthermore, calculate the signal-to-noise ratio value of each filtered signal

[0029] Further, specifically,[[]]

[0030] The combination method includes:[[]]

[0031] Obtain the signal-to-noise ratio value SNR of each filtered signal through the frequency-domain method n,m , to obtain an N×M-dimensional signal-to-noise ratio matrix SNR f ;

[0032] Obtain the signal-to-noise ratio value SNR of each filtered signal through the time-domain method n,m , to obtain an N×M-dimensional signal-to-noise ratio matrix SNR t ; and

[0033] Calculate the linear combination of SNR f , SNR t , that is, the N×M-dimensional signal-to-noise ratio matrix SNR = α·SNR f +β·SNR t , where α and β are the weights of the signal-to-noise ratio obtained by two signal-to-noise ratio calculation methods.[[]]

[0034] Further, specifically,[[]]

[0035] The variant method includes:[[]]

[0036] Let s n,m (t) represent the sliding window signal of the filtered signal of the n-th filter using the m-th cut-off frequency f n,m at the k-th window, k = 1, 2,..., K;

[0037] Perform band-pass filtering on s n,m (t) with [f1, f2] and [f n,1 , f n,2 respectively, and the results are denoted as s n,m,1 (t) and s n,m,2 (t);

[0038] Calculate the signal-to-noise ratio value of the k-th window

[0039] SNR k = 10log10(∫|s n,m,1 (t)| 2 dt / ∫|s n,m,2 (t)|2 dt;

[0040] Furthermore, calculate the signal-to-noise ratio value SNR of each filtered signal n,m , that is, SNR n,m is the signal-to-noise ratio value SNR of K windows k average value.

[0041] Furthermore, specifically,

[0042] The sliding window method includes:

[0043] When △T < T, the sliding window is an overlapping sliding window to handle weak physiological signals with small energy changes;

[0044] When △T = T, the sliding window is a non-overlapping and non-spacing sliding window to handle physiological signals without obvious intermittent changes;

[0045] When △T > T, the sliding window is a non-overlapping and spaced sliding window to handle intermittent physiological signals.

[0046] Furthermore, specifically,

[0047] The physiological signal processing method further includes calculating SNR n,m for the cut-off frequency f n,m first-order difference diff(SNR n,m ), to obtain the first-order difference matrix of the signal-to-noise ratio matrix SNR of N×M dimensions.

[0048] Application of a physiological signal processing method as described above in determining the effective frequency range.

[0049] Furthermore, specifically, the filter bank includes a first filter bank and / or a second filter bank; where

[0050] The filters in the first filter bank are all low-pass filters with a gradually changing transition band to determine the upper limit of the effective frequency range of a certain physiological signal through the signal-to-noise ratio matrix SNR of N×M dimensions or its first-order difference;

[0051] and

[0052] The filters in the second filter bank are all high-pass filters with a gradually changing transition band to determine the lower limit of the effective frequency range of a certain physiological signal through the signal-to-noise ratio matrix SNR of N×M dimensions or its first-order difference.

[0053] Furthermore, specifically, the obtaining of N×M filtered signals includes:

[0054] Set the cut-off frequency f n,m are all within the frequency range [f1, f2] and are M specific cut-off frequencies f c,m ;

[0055] The single-channel signal is passed through a filter bank with M specific cut-off frequencies f c,m and traversed M times in the frequency range [f1, f2] in a step-by-step manner to obtain N×M filtered signals; and

[0056] The N×M-dimensional signal-to-noise ratio matrix SNR is as follows:

[0057] where

[0058] SNR n is the nth row of the signal-to-noise ratio matrix SNR, and the curve formed by it is the signal-to-noise ratio curve.

[0059] Furthermore, specifically, determining the upper limit of the effective frequency range of a certain physiological signal through the N×M-dimensional signal-to-noise ratio matrix SNR includes:

[0060] When the signal-to-noise ratio curves all show an upward trend and the wider the transition band of the filter, the higher the signal-to-noise ratio, it is determined that the specific cut-off frequency f c,m is in the effective frequency band of a certain physiological signal;

[0061] When the signal-to-noise ratio curve of the filter with the widest transition band drops significantly and becomes the lowest point, it is determined that the specific cut-off frequency f c,m has reached the boundary frequency point from signal to noise;

[0062] When the signal-to-noise ratio curves all show a downward trend and the wider the transition band of the filter, the lower the signal-to-noise ratio, it is determined that the specific cut-off frequency f c,m is in the noise frequency band;

[0063] Taking the specific cut-off frequency f c,m at the boundary frequency point from signal to noise as the upper limit of the effective frequency range of a certain physiological signal.

[0064] Furthermore, specifically, determining the lower limit of the effective frequency range of a certain physiological signal through the N×M-dimensional signal-to-noise ratio matrix SNR includes:

[0065] When the signal-to-noise ratio curves all show an upward trend and the wider the transition band of the filter, the higher the signal-to-noise ratio, it is determined that the specific cut-off frequency f c,m is in the effective frequency band of a certain physiological signal;

[0066] When the signal-to-noise ratio curve of the filter with the widest transition band drops significantly and becomes the lowest point, it is determined that the specific cut-off frequency f c,m has reached the boundary frequency point from signal to noise;

[0067] When the signal-to-noise ratio curves all show a downward trend and the wider the transition band of the filter, the lower the signal-to-noise ratio, it is determined that the specific cut-off frequency f c,m is in the noise frequency band;

[0068] Using a specific cut-off frequency f at the boundary frequency point of signal-to-noise as c,m the lower limit of the effective frequency range of a certain physiological signal.

[0069] Furthermore, specifically, obtaining the upper limit of the effective frequency range of a certain physiological signal through the first-order difference of the N×M-dimensional signal-to-noise ratio matrix SNR includes:

[0070] Setting the transition bands of N filters from wide to narrow, and letting diff(SNR n ) represent its first-order difference;

[0071] When diff(SNR n ) are all greater than 0 and SNR 1,m >SNR 2,m >…>SNR N,m , the corresponding specific cut-off frequency f c,m interval is the effective frequency range of a certain physiological signal, denoted as R1;

[0072] When diff(SNR n ) are partially less than 0 and SNR 1,m <SNR 2,m ≈…≈SNR N,m , the corresponding specific cut-off frequency f c,m interval is the physiological signal-noise boundary frequency band, denoted as R2;

[0073] When diff(SNR n ) are all less than 0 and SNR 1,m <SNR 2,m <…<SNR N,m , the corresponding specific cut-off frequency f c,m interval is the noise frequency range, denoted as R3;

[0074] Using the maximum or minimum or intermediate specific cut-off frequency f in R2 c,m as the upper limit of the frequency range of a certain physiological signal.

[0075] Furthermore, specifically, obtaining the lower limit of the effective frequency range of a certain physiological signal through the first-order difference of the N×M-dimensional signal-to-noise ratio matrix SNR includes:

[0076] Setting the transition bands of N filters from wide to narrow, and letting diff(SNR n ) represent its first-order difference;

[0077] When diff(SNR n ) are all greater than 0 and SNR 1,m >SNR 2,m >…>SNRN,m At this time, the corresponding specific cut-off frequency f c,m interval is the effective frequency range of a certain physiological signal, denoted as R1;

[0078] When diff(SNR n ) part is less than 0 and SNR 1,m <SNR 2,m ≈…≈SNR N,m At this time, the corresponding specific cut-off frequency f c,m interval is the physiological signal-noise boundary frequency band, denoted as R2;

[0079] When diff(SNR n ) are all less than 0 and SNR 1,m <SNR 2,m <…<SNR N,m At this time, the corresponding specific cut-off frequency f c,m interval is the noise frequency range, denoted as R3;

[0080] Use the maximum or minimum or intermediate specific cut-off frequency f c,m in R2 as the lower limit of the frequency range of a certain physiological signal.

[0081] Furthermore, specifically, when the physiological signal is a multi-channel signal, the method for determining the effective frequency range in the physiological signal further includes:

[0082] Respectively obtain the effective frequency ranges of Q groups of physiological signals Then the effective frequency range in the physiological signal is

[0083] An application of the physiological signal processing method as described above in determining the evoked component.

[0084] Furthermore, specifically, the number of the filter banks is at least two, namely the third filter bank and the fourth filter bank;

[0085] The third filter bank includes N band-pass filters, and the fourth filter bank includes M low-pass filters or M high-pass filters;

[0086] The obtaining of N×M filtered signals includes:

[0087] The third filter bank divides the single-channel signal into N sub-bands through band-pass filtering;

[0088] The fourth filter bank performs M times of swept low-pass filtering on the N sub-bands respectively with the cut-off frequency f n,m ; where

[0089] Set the specific cut-off frequency f c,mThe stimulation frequency is the cut-off frequency f n,m is n times the specific cut-off frequency f c,m , that is, the cut-off frequency f n,m = n·f c,m , n = 1, 2, …, N.

[0090] Furthermore, specifically, the N×M dimensional signal-to-noise ratio matrix SNR is:

[0091] and

[0092] Based on the N×M dimensional signal-to-noise ratio matrix SNR or its first-order cyclic difference to obtain the frequency component f of the evoked physiological signal V .

[0093] Furthermore, specifically, the obtaining of the frequency component f of the evoked physiological signal based on the N×M dimensional signal-to-noise ratio matrix SNR V includes:

[0094] Let SNR (m) be the m-th column of the N×M dimensional signal-to-noise ratio matrix SNR;

[0095] Calculate the corresponding comprehensive signal-to-noise ratio of SNR (m) , where the weighting value w is obtained n through the multiple frequency times n and the signal-to-noise ratio SNR n,m of each sub-band;

[0096] Obtain the frequency corresponding to the maximum comprehensive signal-to-noise ratio, that is, the frequency component of the evoked physiological signal

[0097]

[0098] Furthermore, specifically, the first-order cyclic difference is the first-order cyclic forward difference or the first-order cyclic backward difference;

[0099] Based on the first-order cyclic forward difference of the signal-to-noise ratio matrix SNR to obtain the frequency component f of the evoked physiological signal V includes:

[0100] Construct the first-order cyclic forward difference of the signal-to-noise ratio matrix SNR, that is

[0101]

[0102] Calculate the first-order cyclic forward difference comprehensive value of the signal-to-noise ratio corresponding to the stimulation frequency:

[0103]

[0104] Calculate the comprehensive signal-to-noise ratio SNR( m );

[0105] According to the SNR (m) and diff(SNR (m) ), determine the frequency components of the evoked physiological signal:

[0106] Threshold value.

[0107] Furthermore, specifically, when the physiological signal is a multi-channel signal, the method for determining the evoked component in the physiological signal further includes:

[0108] Obtain the frequency components of the evoked physiological signal for each of the Q groups of physiological data , and determine the frequency components of the evoked physiological signal where q = (1, 2,..., Q).

[0109] A physiological signal processing system, comprising:

[0110] A microprocessor, which adopts the physiological signal processing method as described above, or adopts the application in determining the effective frequency range as described above, or adopts the application in determining the evoked component as described above;

[0111] At least one filter bank;

[0112] Among them, the microprocessor includes a data processing module, which is adapted to calculate the N×M-dimensional signal-to-noise ratio matrix SNR or its first-order difference.

[0113] The beneficial effects of the present invention are as follows: The physiological signal processing method of the present invention has a simple processing method, ensures that the physiological signal is an effective signal, can correctly identify the characteristics during medical measurement, is convenient for medical evaluation and testing, and has high accuracy. And through further improvement, the upper or lower limit of the frequency can be accurately read, with high accuracy, and it is used to detect the correctness of the signal collected by the acquisition device or to determine the acquisition quality or to determine the electroencephalogram evoked component, and has strong applicability. BRIEF DESCRIPTION OF THE DRAWINGS

[0114] The present invention will be further described below with reference to the drawings and embodiments.

[0115] Figure 1 is the processing flow chart of Embodiment 1 of the present invention.

[0116] Figure 2 is the physiological signal spectrogram of Embodiment 2 of the present invention.

[0117] Figure 3 is the processing flow chart of Embodiment 2 of the present invention.

[0118] Figure 4 is the frequency response curve diagram of a specific embodiment of Embodiment 2 of the present invention.

[0119] Figure 5 It is the SNR curve graph of a specific embodiment of Embodiment 2 of the present invention.

[0120] Figure 6 It is the first-order difference curve graph of a specific embodiment of Embodiment 2 of the present invention.

[0121] Figure 7 It is the processing flow chart of Embodiment 3 of the present invention.

[0122] Figure 8 It is the visual stimulation interface graph of Embodiment 3 of the present invention. Detailed implementation manners

[0123] Now, the present invention will be further described in detail with reference to the accompanying drawings. These drawings are all simplified schematic diagrams, only illustrating the basic structure of the present invention in a schematic manner, so they only show the components related to the present invention.

[0124] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "transverse", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc. indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, so it cannot be understood as a limitation to the present invention. In addition, the features defined as "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present invention, unless otherwise stated, the meaning of "a plurality" is two or more.

[0125] In the description of the present invention, it should be noted that unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0126] Embodiment 1

[0127] As Figure 1 shown, it is Embodiment 1 of the present invention, a physiological signal processing method, including: obtaining N×M filtered signals, that is, taking a single-channel signal with M cut-off frequencies f n,mPerform filtering N×M times through at least one filter bank; obtain the N×M - dimensional signal - to - noise ratio matrix SNR, that is, calculate the signal - to - noise ratio value SNR of each filtered signal respectively n,m And form the N×M - dimensional signal - to - noise ratio matrix SNR; where the filter bank is composed of N filters; and SNR n,m Is the signal - to - noise ratio value of the filtered signal formed by the single - channel signal passing through the nth filter at the mth cut - off frequency f n,m Where n = 1, 2, …, N, m = 1, 2, …, M.

[0128] In the embodiment, the filter is any one of a low - pass filter, a high - pass filter, a band - pass filter, a band - stop filter, an IIR filter, or an FIR filter.

[0129] In the embodiment, obtaining the N×M - dimensional signal - to - noise ratio matrix SNR includes the following steps: perform sliding window on each filtered signal with a window width T and a sliding window step ΔT, and divide it into K windows; obtain the signal - to - noise ratio value of each window respectively through the signal - to - noise ratio calculation method, and then calculate the signal - to - noise ratio value SNR of each filtered signal n,m ; use the signal - to - noise ratio value SNR of each filtered signal n,m To form the matrix SNR; the signal - to - noise ratio calculation method includes any one of a time - domain method, a frequency - domain method, a combined method, or a variant method.

[0130] The frequency - domain method includes: let s n,m (t) represent the sliding - window signal of the filtered signal of the nth filter using the mth cut - off frequency f n,m In the kth window, k = 1, 2, …, K; calculate the power spectrum PS n,m (f) of s n,m (t), and the power spectrum PS n,m (f) is an index that describes the energy distribution of the signal at different frequencies based on the Fourier transform; calculate the signal - to - noise ratio value of the kth window:

[0131]

[0132] And then calculate the signal - to - noise ratio value of each filtered signal:

[0133]

[0134] The time - domain method includes:

[0135] Let s n,m (t) represent the sliding - window signal of the filtered signal of the nth filter using the mth cut - off frequency f n,m In the kth window, k = 1, 2, …, K; calculate the maximum and minimum values of the absolute value of s n,m (t), and denote them as s max And smin ; Calculate the signal-to-noise ratio (SNR) value of the k-th window:

[0136] SNR k = 20 log10(s max / s min );

[0137] Furthermore, calculate the SNR value of each filtered signal:

[0138]

[0139] The combination method includes:

[0140] Obtain the SNR value SNR of each filtered signal through the frequency-domain method n,m , and obtain an N×M-dimensional SNR matrix SNR f ; Obtain the SNR value SNR of each filtered signal through the time-domain method n,m , and obtain an N×M-dimensional SNR matrix SNR t ; And calculate the linear combination of SNR f , SNR t , that is, the N×M-dimensional SNR matrix SNR = α·SNR f + β·SNR t , where α and β are the weights of the SNR obtained by two SNR calculation methods, and α and β are empirical values obtained through testing convergence.

[0141] The variant method includes:

[0142] Let s n,m (t) represent the sliding window signal of the filtered signal of the n-th filter using the m-th cut-off frequency f n,m in the k-th window, k = 1, 2,..., K; perform band-pass filtering on s n,m (t) with [f1, f2] and [f n,1 , f n,2 , and denote the results as s n,m,1 (t) and s n,m,2 (t); Calculate the SNR value of the k-th window:

[0143] SNR k = 10 log10((|s n,m,1 (t)| 2 dt / ∫|s n,m,2 (t)| 2 dt));

[0144] Furthermore, calculate the SNR value SNR of each filtered signal n,m , that is, SNR n,m is the SNR value SNR of K windows kThe average value.

[0145] Furthermore, the sliding window method includes:

[0146] When △T < T, the sliding window is an overlapping sliding window to cope with weak physiological signals with small energy changes; when △T = T, the sliding window is a non-overlapping and non-gapped sliding window to cope with physiological signals without obvious intermittent changes; when △T > T, the sliding window is a non-overlapping and gapped sliding window to cope with intermittent physiological signals.

[0147] In the embodiment, the physiological signal processing method further includes calculating SNR n,m For the cut-off frequency f n,m The first-order difference diff(SNR n,m ), obtaining the first-order difference matrix of the N×M-dimensional signal-to-noise ratio matrix SNR.

[0148] A physiological signal processing method of the present invention has a simple processing method, ensures that the physiological signal is a valid signal, can correctly identify features during medical measurement and use, is convenient for medical evaluation and testing, has high accuracy, and strong applicability.

[0149] Embodiment 2

[0150] In the prior art, the definition of the frequency range of physiological signals such as electromyogram and electroencephalogram is usually obtained by authoritative scientific research institutions through specific experimental measurements, and most of them are broad empirical values (for example, the electromyogram frequency range is 20 - 1000 Hz, and the electroencephalogram frequency range is 1 - 100 Hz). However, using different acquisition means or for different application scenarios, the actual frequency range of the collected physiological signals is different. Specifically, needle electrode electromyogram has a wider frequency range than surface electromyogram; the frequency range of epileptic abnormal electroencephalogram is wider than that of normal electroencephalogram; hardware devices with good front-end amplification performance and larger bit widths can pick up physiological signals with a wider frequency band. The most commonly used method for determining the frequency range currently is to observe the frequency spectrum waveform of the physiological signal, as Figure 2 shown. It can be seen that due to the influence of strong DC components in its low-frequency band, it is difficult to accurately read the lower frequency limit; the high-frequency components of the physiological signal are weak and more likely to decay, resulting in no obvious boundary between the frequency upper limit and the noise band, and it is also difficult to determine.

[0151] As Figure 3As shown, this is the second embodiment of the present invention, an application in determining the effective frequency range. Using the physiological signal processing method described above, the difference from the first embodiment is that the filter bank includes a first filter bank and / or a second filter bank; wherein the filters in the first filter bank are all low-pass filters with a gradually changing transition band to determine the upper limit of the effective frequency range of a certain physiological signal through the N×M-dimensional signal-to-noise ratio matrix SNR or its first-order difference; and the filters in the second filter bank are all high-pass filters with a gradually changing transition band to determine the lower limit of the effective frequency range of a certain physiological signal through the N×M-dimensional signal-to-noise ratio matrix SNR or its first-order difference. Accurately reading the upper or lower limit of the frequency, with high accuracy, is used to detect the correctness of the signal collected by the acquisition device or to judge the acquisition quality, and also facilitates the user to select filtering parameters.

[0152] In the embodiment, obtaining N×M filtered signals includes the following steps: setting the cut-off frequency f n,m are all within the frequency range [f1, f2] and are M specific cut-off frequencies f c,m , the selection of the upper limit f1 and the lower limit f2 of the frequency range depends on the type of physiological signal to be processed (the theoretical frequency ranges of different physiological signals are different. Generally, the frequency range of electromyogram is 0 - 1000 Hz, and the frequency range of electroencephalogram is 0 - 100 Hz). If it is electromyogram, a larger frequency range is selected, and if it is electroencephalogram, the frequency range is correspondingly reduced, that is, (f2 - f1) of electromyogram > (f2 - f1) of electroencephalogram, and the specific cut-off frequency wherein, the larger M is, the smaller the step size of sweeping frequency within [f1, f2] at the specific cut-off frequency means; sweeping the single-channel signal M times within the frequency range [f1, f2] at M specific cut-off frequencies f c,m through the filter bank and obtaining N×M filtered signals; and the N×M-dimensional signal-to-noise ratio matrix SNR is:

[0153]

[0154] where SNR n is the nth row of the signal-to-noise ratio matrix SNR, and the curve formed by it is the signal-to-noise ratio curve.

[0155] It should be noted that for the M specific cut-off frequencies f c,m sweeping M times, when each filter in the filter bank sweeps, the corresponding specific cut-off frequency f in each filter c,m is the same.

[0156] In the embodiment, determining the upper limit of the effective frequency range of a certain physiological signal through the N×M-dimensional signal-to-noise ratio matrix SNR includes: when the signal-to-noise ratio curves all show an upward trend and the wider the transition band of the filter, the higher the signal-to-noise ratio, then it is determined that the specific cut-off frequency f c,mis within the effective frequency band of a certain physiological signal; when the signal-to-noise ratio curve of the filter with the widest transition band drops significantly and becomes the lowest point, it is determined that the specific cut-off frequency f c,m has reached the boundary frequency point from signal to noise; when the signal-to-noise ratio curves all show a downward trend and the wider the transition band of the filter, the lower the signal-to-noise ratio, it is determined that the specific cut-off frequency f c,m is within the noise frequency band; taking the specific cut-off frequency f c,m at the boundary frequency point from signal to noise as the upper limit of the effective frequency range of a certain physiological signal.

[0157] Determining the lower limit of the effective frequency range of a certain physiological signal through the N×M-dimensional signal-to-noise ratio matrix SNR includes: when the signal-to-noise ratio curves all show an upward trend and the wider the transition band of the filter, the higher the signal-to-noise ratio, it is determined that the specific cut-off frequency f c,m is within the effective frequency band of a certain physiological signal; when the signal-to-noise ratio curve of the filter with the widest transition band drops significantly and becomes the lowest point, it is determined that the specific cut-off frequency f c,m has reached the boundary frequency point from signal to noise; when the signal-to-noise ratio curves all show a downward trend and the wider the transition band of the filter, the lower the signal-to-noise ratio, it is determined that the specific cut-off frequency f c,m is within the noise frequency band; taking the specific cut-off frequency f c,m at the boundary frequency point from signal to noise as the lower limit of the effective frequency range of a certain physiological signal.

[0158] Preferably, obtaining the upper limit of the effective frequency range of a certain physiological signal through the first-order difference of the N×M-dimensional signal-to-noise ratio matrix SNR includes: setting the transition bands of N filters from wide to narrow, and letting diff(SNR n ) represent its first-order difference; when diff(SNR n ) are all greater than 0 and SNR 1,m >SNR 2,m >…>SNR N,m , the corresponding specific cut-off frequency f c,m interval is the effective frequency range of a certain physiological signal, denoted as R1; when diff(SNR n ) are partially less than 0 and SNR 1,m <SNR 2,m ≈…≈SNR N,m , the corresponding specific cut-off frequency f c,m interval is the physiological signal-noise boundary frequency band, denoted as R2; when diff(SNR n ) are all less than 0 and SNR 1,m <SNR 2,m <…<SNR N,m , the corresponding specific cut-off frequency f c,mThe interval is the noise frequency range, denoted as R3; the maximum or minimum or intermediate specific cut-off frequency f in R2 c,m is used as the upper limit of the frequency range of a certain physiological signal.

[0159] Obtaining the lower limit of the effective frequency range of a certain physiological signal through the first-order difference of the N×M-dimensional signal-to-noise ratio matrix SNR includes: setting the transition bands of N filters from wide to narrow, and letting diff(SNR n ) represent its first-order difference; when diff(SNR n ) are all greater than 0 and SNR 1,m >SNR 2,m >…>SNR N,m , the corresponding specific cut-off frequency f c,m interval is the effective frequency range of a certain physiological signal, denoted as R1; when diff(SNR n ) are partially less than 0 and SNR 1,m <SNR 2,m ≈…≈SNR N,m , the corresponding specific cut-off frequency f c,m interval is the physiological signal-noise boundary frequency band, denoted as R2; when diff(SNR n ) are all less than 0 and SNR 1,m <SNR 2,m <…<SNR N,m , the corresponding specific cut-off frequency f c,m interval is the noise frequency range, denoted as R3; the maximum or minimum or intermediate specific cut-off frequency f in R2 c,m is used as the lower limit of the frequency range of a certain physiological signal.

[0160] When determining the effective frequency range by taking the first-order difference of the N×M-dimensional signal-to-noise ratio matrix SNR, the processing speed is fast, and the effective frequency range can be obtained more intuitively with good accuracy.

[0161] It should be noted that in this embodiment, the larger N and M are, the smaller the sweep step is at the specific cut-off frequency f c,m , and the higher the accuracy is when calculating the upper or lower limit of the frequency.

[0162] Specifically, taking the electromyogram signal as an example, the filter bank consists of 3 filters. Set f1 = 40HZ, f2 = 220HZ, with a fixed step of 5HZ, and traverse 19 times. The specific cut-off frequency f c,m = 150HZ, and the frequency response curve is as shown in Figure 4 . Taking determining the upper limit of the frequency range of the electromyogram signal as an example to illustrate this embodiment, using the physiological signal processing method of Embodiment 1, 19 filtered signal-to-noise curves with changing cut-off frequencies and their first-order forward differences are obtained, as shown in Figure 5 andFigure 6 As shown, it can be determined that:

[0163] At the cut-off frequencies of 40 - 180 Hz: The signal-to-noise ratios after filtering of all 3 filters show an upward trend, or the first-order forward difference is greater than 0, and the wider the transition band, the higher the signal-to-noise ratio. Filter 1 > Filter 2 > Filter 3. A wider transition band introduces more signals, indicating that this cut-off frequency range of 40 - 180 Hz is within the effective frequency band of the EMG signal.

[0164] At the cut-off frequency of 185 Hz: The signal-to-noise ratio of Filter 1 with the widest transition band drops significantly, or the first-order forward difference is less than 0, making it the filter with the lowest signal-to-noise ratio. Filter 1 with the widest transition band starts to introduce noise, indicating that the cut-off frequency of 185 Hz is already at the boundary frequency point between signal and noise.

[0165] At the cut-off frequencies of 190 Hz - 220 Hz: The signal-to-noise ratios after filtering of all 3 filters show a downward trend, or the first-order forward difference is less than 0, and the wider the transition band, the lower the signal-to-noise ratio. Filter 1 < Filter 2 < Filter 3. A wider transition band introduces more noise, indicating that this cut-off frequency range is within the noise frequency band. It can be seen that 185 Hz can be used as the upper limit of the effective frequency of this EMG signal.

[0166] In some embodiments, when the physiological signal is a multi-channel signal, the method for determining the effective frequency range in the physiological signal further includes: respectively obtaining the effective frequency ranges of Q groups of physiological signals Then the effective frequency range in the physiological signal is where mean represents the mean operation.

[0167] It should be noted that when the physiological signal in the present invention is a multi-channel signal, spatial filtering such as montage processing, ICA (Independent Component Correlation Algorithm), or CCA (Canonical Correlation analysis) can be performed to reduce noise and dimensionality of the multi-channel signal, and then the physiological signal processing method of Embodiment 1 is used for calculation.

[0168] An application in determining the effective frequency range of the present invention adopts the physiological signal processing method of Embodiment 1, and through further improvement, accurately reads the upper or lower limit of the frequency, with high accuracy. It is used to detect the correctness of the signal collected by the acquisition device or judge the acquisition quality, and is also convenient for users to select filtering parameters.

[0169] Embodiment 3

[0170] Such as Figure 7As shown, an application in determining the inducing component, using the physiological signal processing method as above. The difference from the first embodiment is that the number of filter banks is at least two, namely the third filter bank and the fourth filter bank; the third filter bank includes N band-pass filters, and the fourth filter bank includes M low-pass filters or M high-pass filters; obtaining N×M filtered signals includes: the third filter bank divides the single-channel signal into N sub-bands through band-pass filtering; the fourth filter bank performs M times of swept low-pass filtering on the N sub-bands respectively; where, setting a specific cut-off frequency f n,m performs M times of swept low-pass filtering on each of the N sub-bands respectively; where, setting a specific cut-off frequency f c,m as the stimulation frequency, then the cut-off frequency f n,m is n times of the specific cut-off frequency f c,m , that is, the cut-off frequency f n,m =n·f c,m , n = 1, 2, …, N.

[0171] Specifically, the fourth filter bank performs M times of swept low-pass filtering on each sub-band with the cut-off frequency f n,m . Each time the cut-off frequency f n,m of the low-pass filter is different. The M times of swept filtering correspond to M cut-off frequencies, and each cut-off frequency is set in the low-pass filter. The fourth filter bank includes M low-pass filters, and the M low-pass filters mean that there are M cut-off frequencies set in the low-pass filter; similarly, the M high-pass filters mean that there are M cut-off frequencies set in the high-pass filter.

[0172] In the embodiment, the division of the N sub-bands can be in an overlapping manner or a non-overlapping manner.

[0173] In the embodiment, when the fourth filter bank performs M times of swept low-pass filtering on the N sub-bands respectively with the cut-off frequency f n,m , each time the filter in the fourth filter bank sweeps to a cut-off frequency f n,m , the signal-to-noise ratio will have

[0174] a step. The N×M-dimensional signal-to-noise ratio matrix SNR formed is:

[0175]

[0176] And obtaining the frequency component f V of the evoked physiological signal based on the N×M-dimensional signal-to-noise ratio matrix SNR or its first-order cyclic difference.

[0177] In the embodiment, obtaining the frequency component f V of the evoked physiological signal based on the N×M-dimensional signal-to-noise ratio matrix SNR includes: setting SNR (m) as the m-th column of the N×M-dimensional signal-to-noise ratio matrix SNR; calculating the comprehensive signal-to-noise ratio corresponding to SNR (m) :

[0178]

[0179] where the weighting value w n is obtained through the frequency multiplication times n and the signal-to-noise ratio SNR of each sub-band n,m ;

[0180] Obtain the frequency corresponding to the maximum comprehensive signal-to-noise ratio, that is, the frequency component f of the evoked physiological signal V :

[0181]

[0182] In the embodiment, the first-order cyclic difference is the first-order cyclic forward difference or the first-order cyclic backward difference;

[0183] Obtain the frequency component f of the evoked physiological signal based on the first-order cyclic forward difference of the signal-to-noise ratio matrix SNR V including: constructing the first-order cyclic forward difference of the signal-to-noise ratio matrix SNR, that is

[0184]

[0185] Or obtain the frequency component f of the evoked physiological signal based on the first-order cyclic backward difference of the signal-to-noise ratio matrix SNR V including: constructing the first-order cyclic backward difference of the signal-to-noise ratio matrix SNR, that is

[0186]

[0187] Calculate the comprehensive value of the first-order cyclic forward difference of the signal-to-noise ratio corresponding to the stimulation frequency:

[0188]

[0189] Calculate the comprehensive signal-to-noise ratio SNR corresponding to the stimulation frequency m ;

[0190] According to SNR (m) and diff(SNR (m) ), determine the frequency component of the evoked physiological signal:

[0191]

[0192] In the embodiment, when the physiological signal is a multi-channel signal, the method for determining the evoked component in the physiological signal further includes: obtaining the frequency components of the evoked physiological signal for each of the Q groups of physiological data , determine the frequency component of the evoked physiological signal:

[0193] where q = (1, 2,..., Q).

[0194] Among them, "mean" represents the mean operation.

[0195] It should be noted that in this embodiment, the physiological signal is an electroencephalogram (EEG) signal. As Figure 8 shown in the visual stimulation interface, there are a total of 5×8 bright blocks in the interface that flash at their respective calibrated different frequencies. The flashing frequency of the bright blocks traverses from 8 Hz to 15.8 Hz with a fixed step of 0.2 Hz. When in use, a person's eyes fixate on a certain bright block on the interface, and after receiving the visual stimulation of this bright block, a set of specific EEG signals will be generated in the occipital lobe area of the brain, generally the stimulation frequency and its harmonics; by performing M sweep low-pass filtrations separately for each sub-band to calculate the frequency components of the evoked physiological signal, the bright block that the human eyes keep fixating on in the interface can be located.

[0196] An application in determining evoked components of the present invention adopts the physiological signal processing method of Embodiment 1, and through further adjustment, is used to determine evoked EEG components, with high accuracy, providing an effective means for extracting the characteristic frequencies of evoked physiological signals.

[0197] Embodiment 4

[0198] A physiological signal processing system adopts the physiological signal processing method of Embodiment 1, or adopts the application in determining the effective frequency range of Embodiment 2, or adopts the application in determining evoked components of Embodiment 3. The physiological signal processing system includes:

[0199] A microprocessor;

[0200] At least one filter bank, connected to the microprocessor, and the physiological signal is transmitted to the microprocessor after passing through the filter bank;

[0201] Among them, the microprocessor includes: a data processing module, suitable for calculating the N×M-dimensional signal-to-noise ratio matrix SNR or its first-order difference.

[0202] Enlightened by the ideal embodiments of the present invention as described above, through the above description, relevant staff can completely make various changes and modifications without departing from the technical idea of this invention. The technical scope of this invention is not limited to the content in the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. A physiological signal processing method, characterized in that, Comprising: Obtain N×M filtered signals, that is, filter the single-channel signal at M cut-off frequencies f n,m Perform N×M times of filtering through at least one filter bank; Obtain the N×M dimensional signal-to-noise ratio matrix SNR, that is, calculate the signal-to-noise ratio value SNR of each filtered signal respectively n,m and form the N×M dimensional signal-to-noise ratio matrix SNR; where The filter bank is composed of N filters; and SNR n,m is the SNR value of the filtered signal formed by passing the single-channel signal through the nth filter at the mth cut-off frequency f n,m , where n = 1, 2, …, N and m = 1, 2, …, M; The obtaining of the N×M - dimensional signal - to - noise ratio matrix SNR includes: Sliding - windowing each filtered signal with a window width T and a sliding - window step ΔT, and dividing it into K windows; The signal-to-noise ratio values of each window are obtained through the signal-to-noise ratio calculation method, and then the signal-to-noise ratio value SNR of each filtered signal is calculated n,m ; Using the signal-to-noise ratio value SNR of each filtered signal n,m to form the matrix SNR; Among them, the way of the sliding - window includes: When ΔT < T, the sliding - window is an overlapping sliding - window to deal with weak physiological signals with small energy changes; When ΔT = T, the sliding - window is a non - overlapping and non - spaced sliding - window to deal with physiological signals without obvious intermittent changes; When ΔT > T, the sliding - window is a non - overlapping and spaced sliding - window to deal with intermittent physiological signals.

2. The physiological signal processing method according to claim 1, characterized in that The filter is any one of a low - pass filter, a high - pass filter, a band - pass filter, a band - stop filter, an IIR filter or an FIR filter.

3. The physiological signal processing method according to claim 1, characterized in that The signal - to - noise ratio calculation method includes any one of a time - domain method, a frequency - domain method, a combined method or a variant method.

4. The physiological signal processing method according to claim 3, characterized in that The frequency - domain method includes: Let s n,m (t) denote the sliding window signal of the filtered signal of the nth filter using the mth cut-off frequency f n,m in the kth window, where k = 1, 2, …, K; Calculate s n,m (t) power spectrum PS n,m (f); Calculate the signal-to-noise ratio value of the k-th window where f n，1 is the first cut-off frequency used by the n-th filter, and f n，2 is the second cut-off frequency used by the n-th filter; Furthermore, calculate the signal-to-noise ratio value of each filtered signal 5. The physiological signal processing method according to claim 3, characterized in that The time - domain method includes: Let s n,m (t) represent the sliding window signal of the filtered signal of the nth filter using the mth cut-off frequency f n,m in the kth window, where k = 1, 2, …, K; Calculate s n,m (t) the maximum and minimum values of the absolute value, denoted as s max and s min ; Calculate the signal-to-noise ratio value SNR of the k-th window k = 20 log10(s max / s min ); Furthermore, calculate the signal-to-noise ratio value of each filtered signal 6. The physiological signal processing method according to claim 3, characterized in that The combined method includes: Obtain the signal-to-noise ratio value SNR of each filtered signal through the frequency domain method n,m , and obtain the signal-to-noise ratio matrix SNR of N×M dimensions f ; Obtain the signal-to-noise ratio value SNR of each filtered signal through the time-domain method n,m , and obtain the signal-to-noise ratio matrix SNR of N×M dimensions t ; and Calculate SNR f and SNR t are linearly combined, that is, the N×M dimensional signal-to-noise ratio matrix SNR = α·SNR f +β·SNR t , where α and β are the weights of the signal-to-noise ratios obtained by two signal-to-noise ratio calculation methods.

7. The physiological signal processing method according to claim 3, characterized in that The variant method includes: Let s n,m (t) represent the sliding window signal of the filtered signal of the nth filter using the mth cut-off frequency f n,m in the kth window, where k = 1, 2, …, K; For s n,m (t), perform band-pass filtering on [f1, f2] and [f n,1 , f n,2 . Denote the result as s n,m,1 (t) and s n,m,2 (t), where f n，1 is the first cut-off frequency used by the nth filter, and f n，2 is the second cut-off frequency used by the nth filter; Calculating the signal - to - noise ratio value of the k - th window SNR k = 10 log10 (∫ |s n,m,1 (t)| 2 dt / ∫ |s n,m,2 (t)| 2 dt); Furthermore, calculate the signal-to-noise ratio value SNR of each filtered signal n,m , that is, SNR n,m is the average value of the signal-to-noise ratio values SNR k of K windows.

8. The physiological signal processing method according to claim 1, characterized in that The physiological signal processing method further includes calculating the SNR n,m for the cut-off frequency f n,m of the first-order difference diff(SNR n,m ), to obtain the first-order difference matrix of the N×M-dimensional signal-to-noise ratio matrix SNR.

9. Application of a physiological signal processing method according to any one of claims 1 - 8 in determining an effective frequency range; The filter bank includes a first filter bank and / or a second filter bank; where The filters in the first filter bank are all low - pass filters with a gradually changing transition band to determine the upper limit of the effective frequency range of a certain physiological signal through the N×M - dimensional signal - to - noise ratio matrix SNR or its first - order difference; And The filters in the second filter bank are all high - pass filters with a gradually changing transition band to determine the lower limit of the effective frequency range of a certain physiological signal through the N×M - dimensional signal - to - noise ratio matrix SNR or its first - order difference.

10. The application according to claim 9, characterized in that The obtaining of the N×M filtered signals includes: Set the cut-off frequency f n,m All are within the frequency range [f1, f2] and are M specific cut-off frequencies f c,m ; The single-channel signal is passed through a filter bank with M specific cut-off frequencies f c,m and traversed M times step by step within the frequency range [f1, f2], resulting in N×M filtered signals; and The N×M - dimensional signal - to - noise ratio matrix SNR is: wherein SNR n is the n-th row of the signal-to-noise ratio matrix SNR, and the curve formed by it is the signal-to-noise ratio curve.

11. The application according to claim 10, characterized in that Determining the upper limit of the effective frequency range of a certain physiological signal through the N×M - dimensional signal - to - noise ratio matrix SNR includes: When the signal-to-noise ratio curves all show an upward trend and the wider the transition band of the filter, the higher the signal-to-noise ratio, it is determined that the specific cut-off frequency f c,m is in the effective frequency band of a certain physiological signal; When the signal-to-noise ratio curve of the filter with the widest transition band drops significantly and becomes the lowest point, it is determined that the specific cut-off frequency f c,m has reached the boundary frequency point of signal-to-noise; When the SNR curves all show a downward trend and the wider the transition band of the filter, the lower the SNR, it is determined that the specific cut-off frequency f c,m is in the noise frequency band; Taking a specific cut-off frequency f at the boundary frequency point of signal-to-noise as c,m the upper limit of the effective frequency range of a certain physiological signal.

12. The application according to claim 10, characterized in that Determining the lower limit of the effective frequency range of a certain physiological signal through the N×M - dimensional signal - to - noise ratio matrix SNR includes: When the SNR curves all show an upward trend and the wider the transition band of the filter, the higher the SNR, it is determined that the specific cut-off frequency f c,m is in the effective frequency band of a certain physiological signal; When the SNR curve of the filter with the widest transition band drops significantly and becomes the lowest point, it is determined that the specific cut-off frequency f c,m has reached the boundary frequency point of signal to noise; When the SNR curves all show a downward trend and the wider the transition band of the filter, the lower the SNR, it is determined that the specific cut-off frequency f c,m is in the noise frequency band; Taking a specific cut-off frequency f at the boundary frequency point of signal-to-noise as c,m the lower limit of the effective frequency range of a certain physiological signal.

13. The application according to claim 10, characterized in that Obtaining the upper limit of the effective frequency range of a certain physiological signal through the first - order difference of the N×M - dimensional signal - to - noise ratio matrix SNR includes: Set the transition bands of N filters from wide to narrow, and let diff(SNR n ) represent its first-order difference; When diff(SNR n ) are all greater than 0 and SNR 1,m > SNR 2,m > … > SNR N,m At this time, the corresponding specific cut-off frequency f c,m interval is the effective frequency range of a certain physiological signal, denoted as R1; When the diff(SNR n ) part is less than 0 and SNR 1,m <SNR 2,m ≈…≈SNR N,m At this time, the corresponding specific cut-off frequency f c,m The interval is the physiological signal-noise boundary frequency band, denoted as R2; When diff(SNR n ) are all less than 0 and SNR 1,m <SNR 2,m <…<SNR N,m When, the corresponding specific cut-off frequency f c,m The interval is the noise frequency range, denoted as R3; Using the maximum or minimum or intermediate specific cut-off frequency f in R2 c,m as the upper limit of the frequency range of a certain physiological signal.

14. The application according to claim 10, characterized in that Obtaining the lower limit of the effective frequency range of a certain physiological signal through the first - order difference of the N×M - dimensional signal - to - noise ratio matrix SNR includes: Set the transition bands of N filters from wide to narrow, and let diff(SNR n ) represent its first-order difference; When diff(SNR n ) are all greater than 0 and SNR 1,m > SNR 2,m > … > SNR N,m , the corresponding specific cut-off frequency f c,m interval is the effective frequency range of a certain physiological signal, denoted as R1; When the diff(SNR n ) part is less than 0 and SNR 1,m < SNR 2,m ≈ … ≈ SNR N,m , the corresponding specific cut-off frequency f c,m interval is the physiological signal-noise boundary frequency band, denoted as R2; When diff(SNR n ) are all less than 0 and SNR 1,m < SNR 2,m < … < SNR N,m When, the corresponding specific cut-off frequency f c,m interval is the noise frequency range, denoted as R3; Using the maximum or minimum or intermediate specific cut-off frequency f in R2 c,m as the lower limit of the frequency range of a certain physiological signal.

15. The application according to claim 9, wherein: When the physiological signal is a multi-channel signal, the method for determining the effective frequency range in the physiological signal further includes: Obtain the effective frequency ranges of Q groups of physiological signals respectively Then the effective frequency range in the physiological signal is 16. An application of the physiological signal processing method according to any one of claims 1-8 in determining the evoked component; The number of the filter banks is at least two, namely a third filter bank and a fourth filter bank; The third filter bank includes N band-pass filters, and the fourth filter bank includes M low-pass filters or M high-pass filters; The obtaining of the N×M filtered signals includes: The third filter bank divides the single-channel signal into N sub-bands through band-pass filtering; The fourth filter bank performs M times of swept low-pass filtering on the N sub-bands respectively with a cut-off frequency f n,m ; wherein, Set a specific cut-off frequency f c,m as the stimulation frequency, then the cut-off frequency f n,m is n times the specific cut-off frequency f c,m , that is, the cut-off frequency f n,m = n·f c,m , n = 1, 2, …, N; The N×M-dimensional signal-to-noise ratio matrix SNR is: and Obtaining the frequency component f of the evoked physiological signal based on the N×M - dimensional signal - to - noise ratio matrix SNR or its first - order cyclic difference V .

17. The application according to claim 16, wherein: Obtaining the frequency component f of the evoked physiological signal based on the N×M - dimensional signal - to - noise ratio matrix SNR V including: Let SNR (m) be the m-th column of the N×M dimensional signal-to-noise ratio matrix SNR; Calculate SNR (m) The corresponding comprehensive signal-to-noise ratio where the weighting value w n is obtained through the multiplication factor n of each sub-band and the signal-to-noise ratio SNR n,m is obtained; Obtain the frequency corresponding to the maximum comprehensive signal-to-noise ratio, that is, the frequency component of the evoked physiological signal 18. The application according to claim 17, wherein: The first-order cyclic difference is a first-order cyclic forward difference or a first-order cyclic backward difference; Obtaining the frequency component f of the induced physiological signal based on the first-order cyclic forward difference of the signal-to-noise ratio matrix SNR V Including: Construct the first-order cyclic forward difference of the signal-to-noise ratio matrix SNR, that is Calculate the first-order cyclic forward difference comprehensive value of the signal-to-noise ratio corresponding to the stimulation frequency: Calculate the comprehensive signal-to-noise ratio SNR corresponding to the stimulation frequency (m) ; According to the SNR (m) and diff(SNR (m) ), determine the frequency components of the induced physiological signal:

19. The application according to claim 16, wherein: When the physiological signal is a multi-channel signal, the method for determining the evoked component in the physiological signal further includes: Obtain the frequency components of the induced physiological signals for each of the Q groups of physiological data Determine the frequency components of the induced physiological signals where q = 1, 2, …, Q.

20. A physiological signal processing system, characterized in that, Comprising: A microprocessor that adopts the physiological signal processing method according to claim 1, or adopts the application according to claim 9, or adopts the application according to claim 16; At least one filter bank; Wherein, the microprocessor includes a data processing module adapted to calculate the N×M-dimensional signal-to-noise ratio matrix SNR or its first-order difference.

Images