A method and system for monitoring physiological information after transcranial magnetic stimulation

Abstract

The present invention discloses a method and system for monitoring physiological information after transcranial magnetic stimulation. By obtaining a marked symmetry line and constructing a physiological convolution network, the change between the physiological category data corresponding to the latter time point and the physiological category data corresponding to the previous time point is found in the physiological matrix, and the relationship between the two physiological category data is determined. Thus, the physiological categories that affect the accuracy of the detection value are arranged, and each column in the arranged adjusted physiological matrix is ​​convoluted in sequence according to the time point. The relationship between the physiological categories is represented by the arrangement, and the data containing the relationship is input into the time convolution network for detection, so as to achieve a more accurate monitoring of the physiological information of transcranial magnetic stimulation, thereby monitoring the physiological state.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method and system for monitoring physiological information after transcranial magnetic stimulation. Background Art

[0002] At present, due to the non-invasive and objective characteristics of surface electromyography after transcranial magnetic stimulation, the key muscle groups of the paralyzed upper limb (elbow flexion, elbow extension, wrist flexion, wrist flexion, etc.) are collected by surface electromyography, and the time domain indicators: root mean square amplitude (RMS), integrated electromyographic value (IEMG) and frequency domain indicators: median frequency (MF), mean power frequency (MPF) are analyzed and extracted to quantitatively analyze the changes in muscle strength and muscle tension of the target muscle. The use of big data intelligent detection can detect users more conveniently and quickly, but the surface electromyography signals between the key muscle groups of the paralyzed upper limb (elbow flexion, elbow extension, wrist flexion, wrist flexion, etc.) are correlated, and the correlation affects the accuracy of big data detection of the user's physiological state. Summary of the invention

[0003] The purpose of the present invention is to provide a method and system for monitoring physiological information after transcranial magnetic stimulation, so as to solve the above-mentioned problems existing in the prior art.

[0004] In a first aspect, an embodiment of the present invention provides a method for monitoring physiological information after transcranial magnetic stimulation, comprising:

[0005] Obtain physiological information of multiple physiological categories monitored at multiple time points;

[0006] The physiological information is constructed into a two-dimensional matrix in the order of time points from small to large, to obtain a physiological matrix; the rows of the physiological matrix correspond to time points; the columns of the physiological matrix correspond to physiological categories;

[0007] Marking a symmetry line through a center point of the physiological matrix to obtain a marked symmetry line;

[0008] By using a physiological convolutional network, based on the physiological matrix and the marked symmetry line, the relationship between different categories of physiological information at adjacent time points is determined to obtain a relationship arrangement set;

[0009] Based on the relationship arrangement set, adjusting the columns in the physiological matrix to obtain an adjusted physiological matrix;

[0010] Based on the adjusted physiological matrix, convolution is performed sequentially according to time points to obtain a physiological state; the physiological state includes time domain indicators and frequency domain indicators.

[0011] Optionally, the physiological convolution network is used to determine the relationship between different categories of physiological information at adjacent time points based on the physiological matrix and the marked symmetry line to obtain a relationship arrangement set, including:

[0012] Mark the values ​​corresponding to the straight lines parallel to the marked symmetry lines in the physiological matrix to obtain a plurality of parallel symmetry lines; the parallel symmetry lines include the position corresponding to the jth page of the i-th row and the position corresponding to the j+1th page of the i+1-th row;

[0013] Obtaining a physiological convolution kernel; the physiological convolution kernel is a convolution kernel for convolving a physiological matrix;

[0014] Inserting rows and columns at the beginning of the physiological matrix and inserting rows and columns at the end of the physiological matrix to obtain a zero-filled physiological matrix; the values ​​in the added rows and columns of the zero-filled physiological matrix are 0;

[0015] Based on the zero-filled physiological matrix and the parallel symmetry lines, a physiological symmetry feature is obtained through the physiological convolution kernel; the physiological symmetry feature is a feature containing multiple physiological categories; multiple physiological symmetry features corresponding to multiple parallel symmetry lines;

[0016] Based on the multiple physiological symmetry features, a relationship arrangement set is obtained.

[0017] Optionally, obtaining physiological symmetry features based on the zero-filled physiological matrix and the parallel symmetry lines through the physiological convolution kernel includes:

[0018] Match the center point of the physiological convolution kernel with the position on the parallel symmetry line in the zero-filled physiological matrix, and perform convolution along the parallel symmetry line with a step size of 1 to obtain a first physiological symmetry feature map;

[0019] The number of rows of the first physiological symmetry feature map is equal to nm; n is the number of values ​​in the zero-filled physiological matrix through which the marking symmetry line passes; m is the number of columns of the physiological convolution kernel;

[0020] The first physiological symmetry feature map is input into a corresponding physiological symmetry network, and the relationship between multiple physiological categories is extracted to obtain a physiological symmetry feature.

[0021] Optionally, the structure of the physiological symmetric network includes:

[0022] The physiological symmetric network comprises n-m+2 layers;

[0023] Each layer of the physiological symmetric network includes a plurality of physiological feature convolution kernels; the plurality of physiological feature convolution kernels are convolved in the row direction of the corresponding first physiological symmetric feature map with a step size of 1;

[0024] The number of columns of the physiological feature convolution kernel is the same as the number of columns of the feature map; the number of rows of the physiological feature convolution kernel is 2.

[0025] Optionally, the plurality of parallel symmetry lines correspond to a plurality of physiological symmetry networks;

[0026] The number of convolution kernels in the last layer of multiple physiological symmetric networks is the same.

[0027] Optionally, obtaining a relationship arrangement set based on the multiple physiological symmetry features includes:

[0028] Inputting the physiological symmetry feature into a relationship detection network, extracting the output probability value, and obtaining a relationship detection value; obtaining multiple relationship detection values ​​corresponding to multiple physiological symmetry features;

[0029] Based on the multiple relationship detection values, a relationship arrangement set is obtained.

[0030] Optionally, obtaining a relationship arrangement set based on the multiple relationship detection values ​​includes:

[0031] Obtain a first relationship detection value and a second relationship detection value; the first relationship detection value and the second relationship detection value are two relationship detection values ​​among a plurality of relationship detection values;

[0032] The latest time point in the parallel symmetry lines corresponding to the first relationship detection value is the first time point; the latest time point in the parallel symmetry lines corresponding to the second relationship detection value is the second time point; the first time point and the second time point are adjacent; the second time point is later than the first time point;

[0033] Subtracting the absolute value of the first relationship detection value from the second relationship detection value as the relationship adjacent difference;

[0034] Marking the relationship adjacent differences and the physiological category corresponding to the second time point to obtain a physiological category importance value;

[0035] Arrange the physiological categories from large to small according to their importance values ​​to obtain a relationship arrangement set.

[0036] Optionally, adjusting the columns in the physiological matrix based on the relationship arrangement set to obtain the adjusted physiological matrix includes:

[0037] Extracting columns from the physiological matrix to obtain a plurality of physiological category columns;

[0038] The physiological category columns corresponding to the values ​​in the relationship arrangement set are arranged from top to bottom in sequence to obtain an adjusted physiological matrix.

[0039] Optionally, the adjusting the physiological matrix and sequentially convolving according to time points to obtain the physiological state includes:

[0040] According to the time points from early to late, the value corresponding to each row of the physiological matrix is ​​input into the temporal convolutional network to obtain the physiological state.

[0041] In a second aspect, an embodiment of the present invention provides a physiological information monitoring system after transcranial magnetic stimulation, comprising:

[0042] Acquisition module: obtains physiological information of multiple physiological categories monitored at multiple time points;

[0043] A two-dimensional matrix construction module: constructing a two-dimensional matrix of the physiological information in the order of time points from small to large to obtain a physiological matrix; the rows of the physiological matrix correspond to time points; the columns of the physiological matrix correspond to physiological categories;

[0044] Symmetry marking module: marking the symmetry line of the physiological matrix through the center point to obtain a marked symmetry line;

[0045] Arrangement module: through the physiological convolution network, based on the physiological matrix and the marked symmetry line, the relationship between different categories of physiological information at adjacent time points is determined to obtain a relationship arrangement set;

[0046] An adjustment module: adjusting the columns in the physiological matrix based on the relationship arrangement set to obtain an adjusted physiological matrix;

[0047] Time convolution module: based on the adjusted physiological matrix, convolutes sequentially according to time points to obtain the physiological state; the physiological state includes time domain indicators and frequency domain indicators.

[0048] Compared with the prior art, the embodiments of the present invention achieve the following beneficial effects:

[0049] An embodiment of the present invention also provides a method and system for monitoring physiological information after transcranial magnetic stimulation, the method comprising: obtaining physiological information of multiple physiological categories monitored at multiple time points; constructing a two-dimensional matrix of the physiological information in order from small to large time points to obtain a physiological matrix; the rows of the physiological matrix correspond to time points; the columns of the physiological matrix correspond to physiological categories; marking the physiological matrix with a symmetry line through a center point to obtain a marked symmetry line; through a physiological convolution network, based on the physiological matrix and the marked symmetry line, judging the relationship between different categories of physiological information at adjacent time points to obtain a relationship arrangement set; based on the relationship arrangement set, adjusting the columns in the physiological matrix to obtain an adjusted physiological matrix; based on the adjusted physiological matrix, convolving in sequence according to time points to obtain a physiological state; the physiological state includes time domain indicators and frequency domain indicators.

[0050] In the present invention, by obtaining the marked symmetry line and constructing a physiological convolution network, the change between the physiological category data corresponding to the next time point and the physiological category data corresponding to the previous time point is found in the physiological matrix, and the relationship between the two physiological category data is determined. Thus, the physiological categories that affect the accuracy of the detection value are arranged, and each column in the arranged adjusted physiological matrix is ​​convoluted in sequence according to the time point. The relationship between the physiological categories is represented by the arrangement, and the data containing the relationship is input into the time convolution network for detection, so as to achieve a more accurate monitoring of the physiological information of transcranial magnetic stimulation, thereby monitoring the technical effect of the physiological state. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 This is a flow chart of a method for monitoring physiological information after transcranial magnetic stimulation provided by an embodiment of the present invention.

[0052] Figure 2 It is a structural schematic diagram of marked symmetry lines and parallel symmetry lines in a method for monitoring physiological information after transcranial magnetic stimulation provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0053] The present invention will be described in detail below in conjunction with the accompanying drawings.

[0054] Example 1

[0055] like Figure 1 As shown, an embodiment of the present invention provides a method for monitoring physiological information after transcranial magnetic stimulation, the method comprising:

[0056] S101: Acquire physiological information of multiple physiological categories monitored at multiple time points.

[0057] Among them, the multiple physiological categories include surface electromyographic signals of the user's elbow flexion, elbow extension, wrist flexion and wrist extension; the physiological information is the value corresponding to the physiological category.

[0058] S102: constructing a two-dimensional matrix of the physiological information in the order of time points from small to large to obtain a physiological matrix; the rows of the physiological matrix correspond to time points; and the columns of the physiological matrix correspond to physiological categories.

[0059] In this embodiment, the number of time points is fixed to be the same as the number of physiological categories.

[0060] S103: Marking the symmetry line of the physiological matrix through the center point to obtain a marked symmetry line.

[0061] Wherein, the structure of the marking symmetry line is as follows Figure 2 shown.

[0062] S104: Using a physiological convolutional network, based on the physiological matrix and the marked symmetry line, the relationship between different categories of physiological information at adjacent time points is determined to obtain a relationship arrangement set.

[0063] S105: Based on the relationship arrangement set, adjust the columns in the physiological matrix to obtain an adjusted physiological matrix.

[0064] S106: Based on the adjusted physiological matrix, convolute sequentially according to time points to obtain a physiological state; the physiological state includes a time domain index and a frequency domain index.

[0065] Among them, the time domain indicators include root mean square amplitude (RMS) and integrated electromyographic value (IEMG), and the frequency domain indicators include median frequency (MF) and mean power frequency (MPF); according to the time domain indicators and frequency domain indicators, the changes in muscle strength and muscle tension of the target muscle are quantitatively analyzed to judge the user status.

[0066] Optionally, the physiological convolution network is used to determine the relationship between different categories of physiological information at adjacent time points based on the physiological matrix and the marked symmetry line to obtain a relationship arrangement set, including:

[0067] The values ​​corresponding to the straight lines parallel to the marked symmetry lines in the physiological matrix are marked to obtain multiple parallel symmetry lines; the parallel symmetry lines include the position corresponding to the i-th row and the j-th page, and the position corresponding to the i+1-th row and the j+1-th page.

[0068] Wherein, the structure of the parallel symmetry lines is as follows Figure 2 shown.

[0069] Wherein, i and j are natural numbers greater than 0.

[0070] In this embodiment, the parallel symmetry lines are as follows: Figure 1 shown.

[0071] Obtain a physiological convolution kernel; the physiological convolution kernel is a convolution kernel for convolving a physiological matrix.

[0072] Among them, the physiological convolution kernel is a two-dimensional convolution kernel; the number of rows of the physiological convolution kernel is equal to the number of columns which is less than the number of rows of the physiological matrix; the number of rows of the physiological convolution kernel is equal to the number of columns which is less than the number of columns of the physiological matrix; in this embodiment, the physiological convolution kernel is a 2*2 two-dimensional convolution kernel.

[0073] The beginning of the physiological matrix is ​​inserted into rows and columns, and the end of the physiological matrix is ​​inserted into rows and columns to obtain a zero-filled physiological matrix; the values ​​in the added rows and columns of the zero-filled physiological matrix are 0.

[0074] In this embodiment, one row and one column are added to the beginning of the physiological matrix, and one row and one column are added to the end of the physiological matrix.

[0075] Through the above method, the parallel symmetry lines correspond to 1 value in the corner, and the two-dimensional convolution kernel is 2*2, so zero padding is performed so that the two-dimensional convolution kernel can be multiplied with the two-dimensional matrix after zero padding.

[0076] Based on the zero-filled physiological matrix and the parallel symmetry lines, a physiological symmetry feature is obtained through the physiological convolution kernel; the physiological symmetry feature is a feature containing multiple physiological categories; and multiple physiological symmetry features are obtained after multiple parallel symmetry lines correspond to each other.

[0077] Based on the multiple physiological symmetry features, a relationship arrangement set is obtained.

[0078] Optionally, obtaining physiological symmetry features based on the zero-filled physiological matrix and the parallel symmetry lines through the physiological convolution kernel includes:

[0079] Match the center point of the physiological convolution kernel with the position on the parallel symmetry line in the zero-filled physiological matrix, and perform convolution along the parallel symmetry line with a step size of 1 to obtain a first physiological symmetry feature map;

[0080] The number of rows of the first physiological symmetry feature map is equal to nm; n is the number of values ​​in the zero-filled physiological matrix through which the marked symmetry line passes; and m is the number of columns of the physiological convolution kernel.

[0081] In this embodiment, n and m are natural numbers greater than 0; nm represents the length of the feature obtained after convolution along parallel symmetry lines.

[0082] The first physiological symmetry feature map is input into a corresponding physiological symmetry network, and the relationship between multiple physiological categories is extracted to obtain a physiological symmetry feature.

[0083] Optionally, the structure of the physiological symmetric network includes:

[0084] The physiological symmetric network comprises nm layers.

[0085] Among them, because each layer needs to fuse m adjacent features, and the last layer needs to contain features of all values ​​through which parallel symmetry lines pass, there are nm layers required.

[0086] Each layer of the physiological symmetric network includes multiple physiological feature convolution kernels; the multiple physiological feature convolution kernels are convolved in the row direction of the corresponding first physiological symmetric feature map with a step size of 1.

[0087] The number of columns of the physiological feature convolution kernel is the same as the number of columns of the feature map; the number of rows of the physiological feature convolution kernel is 2.

[0088] Optionally, the multiple parallel symmetry lines correspond to multiple physiological symmetry networks.

[0089] The number of convolution kernels in the last layer of multiple physiological symmetric networks is the same.

[0090] The number of convolution kernels determines the number of feature layers.

[0091] Optionally, obtaining a relationship arrangement set based on the multiple physiological symmetry features includes:

[0092] The physiological symmetry feature is input into the relationship detection network, and the output probability value is extracted to obtain the relationship detection value; multiple physiological symmetry features correspond to multiple relationship detection values.

[0093] The relationship detection network is a fully-connected neural network (FCNN). The relationship detection network is trained according to the labeled physiological state and the vectors containing all physiological categories at the same time point to obtain a trained relationship detection network.

[0094] The relationship detection value indicates the probability of obtaining the physiological state corresponding to the labeled data by inputting the physiological symmetry feature into the relationship detection network, and indicates the accuracy of the detection of the feature containing a part.

[0095] Based on the multiple relationship detection values, a relationship arrangement set is obtained.

[0096] Optionally, obtaining a relationship arrangement set based on the multiple relationship detection values ​​includes:

[0097] A first relationship detection value and a second relationship detection value are obtained; the first relationship detection value and the second relationship detection value are two relationship detection values ​​among a plurality of relationship detection values.

[0098] The latest time point in the parallel symmetry lines corresponding to the first relationship detection value is the first time point; the latest time point in the parallel symmetry lines corresponding to the second relationship detection value is the second time point; the first time point and the second time point are adjacent; the second time point is later than the first time point.

[0099] The absolute value of the second relationship detection value minus the first relationship detection value is taken as the relationship adjacent difference.

[0100] The relationship adjacent difference values ​​are marked corresponding to the physiological category corresponding to the second time point to obtain the physiological category importance value.

[0101] Among them, if the relationship detection value corresponding to the feature of the surface electromyographic signal that only includes elbow flexion is 0.8, and the relationship detection value corresponding to the feature of the surface electromyographic signal that includes elbow flexion is 0.85, 0.85-0.8=0.05, and the 0.05 is constructed into a key-value pair with body temperature.

[0102] Arrange the physiological categories from large to small according to their importance values ​​to obtain a relationship arrangement set.

[0103] Optionally, adjusting the columns in the physiological matrix based on the relationship arrangement set to obtain the adjusted physiological matrix includes:

[0104] Extracting columns from the physiological matrix to obtain a plurality of physiological category columns;

[0105] The physiological category columns corresponding to the values ​​in the relationship arrangement set are arranged from top to bottom in sequence to obtain an adjusted physiological matrix.

[0106] Optionally, the adjusting the physiological matrix and sequentially convolving according to time points to obtain the physiological state includes:

[0107] According to the time points from early to late, the value corresponding to each row of the physiological matrix is ​​input into the temporal convolutional network to obtain the physiological state.

[0108] Wherein, the rows of the physiological matrix correspond to time points.

[0109] Example 2

[0110] Based on the above-mentioned method for monitoring physiological information after transcranial magnetic stimulation, an embodiment of the present invention also provides a system for monitoring physiological information after transcranial magnetic stimulation, the system comprising an acquisition module, a two-dimensional matrix construction module, a symmetric marking module, an arrangement module, an adjustment module, a time convolution module and a warning module.

[0111] The acquisition module is used to obtain physiological information of multiple physiological categories monitored at multiple time points.

[0112] The two-dimensional matrix construction module is used to construct a two-dimensional matrix according to the order of time points from small to large for the physiological information to obtain a physiological matrix; the rows of the physiological matrix correspond to time points; and the columns of the physiological matrix correspond to physiological categories.

[0113] The symmetry marking module is used to mark the symmetry line of the physiological matrix passing through the center point to obtain a marked symmetry line.

[0114] The arrangement module is used to determine the relationship between different categories of physiological information at adjacent time points based on the physiological matrix and the marked symmetry line through a physiological convolutional network, and obtain a relationship arrangement set.

[0115] The adjustment module is used to adjust the columns in the physiological matrix based on the relationship arrangement set to obtain an adjusted physiological matrix.

[0116] The time convolution module is used to obtain the physiological state based on the adjusted physiological matrix and convolve in sequence according to time points; the physiological state includes time domain indicators and frequency domain indicators.

[0117] Furthermore, those skilled in the art will appreciate that although some embodiments herein include certain features included in other embodiments but not other features, the combination of features from different embodiments is meant to be within the scope of the present invention and to form different embodiments.

[0118] The various component embodiments of the present invention may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. It should be understood by those skilled in the art that a microprocessor or digital signal processor (DSP) may be used in practice to implement some or all of the functions of some or all of the components in the apparatus according to an embodiment of the present invention. The present invention may also be implemented as a device or device program (e.g., a computer program and a computer program product) for executing part or all of the methods described herein. Such a program implementing the present invention may be stored on a computer-readable medium, or may have the form of one or more signals. Such a signal may be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

Claims

1. A method for monitoring physiological information after transcranial magnetic stimulation, characterized in that: include: Obtain physiological information of multiple physiological categories monitored at multiple time points; Constructing a two-dimensional matrix of the physiological information in the order of time points from small to large to obtain a physiological matrix; The rows of the physiological matrix correspond to time points; The physiological categories corresponding to the columns of the physiological matrix; Marking a symmetry line through a center point of the physiological matrix to obtain a marked symmetry line; By using a physiological convolutional network, based on the physiological matrix and the marked symmetry line, the relationship between different categories of physiological information at adjacent time points is determined to obtain a relationship arrangement set; Based on the relationship arrangement set, adjusting the columns in the physiological matrix to obtain an adjusted physiological matrix; Based on the adjusted physiological matrix, sequentially convolute according to time points to obtain the physiological state; The physiological state includes time domain indicators and frequency domain indicators; The physiological convolution network is used to determine the relationship between different types of physiological information at adjacent time points based on the physiological matrix and the marked symmetry line, and obtain a relationship arrangement set, including: Mark the values ​​corresponding to the straight lines parallel to the marked symmetry lines in the physiological matrix to obtain a plurality of parallel symmetry lines; the parallel symmetry lines include the position corresponding to the jth page of the i-th row and the position corresponding to the j+1th page of the i+1-th row; Obtaining a physiological convolution kernel; the physiological convolution kernel is a convolution kernel for convolving a physiological matrix; Inserting rows and columns at the beginning of the physiological matrix and inserting rows and columns at the end of the physiological matrix to obtain a zero-filled physiological matrix; the values ​​in the added rows and columns of the zero-filled physiological matrix are 0; Based on the zero-filled physiological matrix and the parallel symmetry lines, a physiological symmetry feature is obtained through the physiological convolution kernel; the physiological symmetry feature is a feature containing multiple physiological categories; multiple parallel symmetry lines correspond to multiple physiological symmetry features; Based on the multiple physiological symmetry features, a relationship arrangement set is obtained.

2. The method for monitoring physiological information after transcranial magnetic stimulation according to claim 1, characterized in that: The method of obtaining physiological symmetry features based on the zero-filled physiological matrix and the parallel symmetry lines through the physiological convolution kernel includes: Match the center point of the physiological convolution kernel with the position on the parallel symmetry line in the zero-filled physiological matrix, and perform convolution along the parallel symmetry line with a step size of 1 to obtain a first physiological symmetry feature map; The number of rows of the first physiological symmetry feature map is equal to nm; n is the number of values ​​in the zero-filled physiological matrix through which the marking symmetry line passes; m is the number of columns of the physiological convolution kernel; The first physiological symmetry feature map is input into a corresponding physiological symmetry network, and the relationship between multiple physiological categories is extracted to obtain a physiological symmetry feature.

3. The method for monitoring physiological information after transcranial magnetic stimulation according to claim 2, characterized in that: The structure of the physiological symmetric network includes: The physiological symmetric network comprises n-m+2 layers; Each layer of the physiological symmetric network includes a plurality of physiological feature convolution kernels; the plurality of physiological feature convolution kernels are convolved in the row direction of the corresponding first physiological symmetric feature map with a step size of 1; The number of columns of the physiological feature convolution kernel is the same as the number of columns of the feature map; the number of rows of the physiological feature convolution kernel is 2.

4. The method for monitoring physiological information after transcranial magnetic stimulation according to claim 3, characterized in that: The plurality of parallel symmetry lines correspond to obtaining a plurality of physiological symmetry networks; The number of convolution kernels in the last layer of multiple physiological symmetric networks is the same.

5. The method for monitoring physiological information after transcranial magnetic stimulation according to claim 2, characterized in that: The step of obtaining a relationship arrangement set based on the multiple physiological symmetry features includes: Inputting the physiological symmetry feature into a relationship detection network, extracting the output probability value, and obtaining a relationship detection value; Multiple relationship detection values ​​are obtained corresponding to multiple physiological symmetry features; Based on the multiple relationship detection values, a relationship arrangement set is obtained.

6. The method for monitoring physiological information after transcranial magnetic stimulation according to claim 5, characterized in that: The step of obtaining a relationship arrangement set based on the multiple relationship detection values ​​includes: Obtain a first relationship detection value and a second relationship detection value; the first relationship detection value and the second relationship detection value are two relationship detection values ​​among a plurality of relationship detection values; The latest time point in the parallel symmetry lines corresponding to the first relationship detection value is the first time point; the latest time point in the parallel symmetry lines corresponding to the second relationship detection value is the second time point; the first time point and the second time point are adjacent; the second time point is later than the first time point; The absolute value of the second relationship detection value minus the first relationship detection value is taken as the relationship adjacent difference; Marking the relationship adjacent differences and the physiological category corresponding to the second time point to obtain a physiological category importance value; Arrange the physiological categories from large to small according to their importance values ​​to obtain a relationship arrangement set.

7. The method for monitoring physiological information after transcranial magnetic stimulation according to claim 1, characterized in that: The step of adjusting the columns in the physiological matrix based on the relationship arrangement set to obtain the adjusted physiological matrix includes: Extracting columns from the physiological matrix to obtain a plurality of physiological category columns; The physiological category columns corresponding to the values ​​in the relationship arrangement set are arranged from top to bottom in sequence to obtain an adjusted physiological matrix.

8. The method for monitoring physiological information after transcranial magnetic stimulation according to claim 1, characterized in that: The step of sequentially convolving the adjusted physiological matrix according to time points to obtain the physiological state includes: According to the time points from early to late, the value corresponding to each row of the physiological matrix is ​​input into the temporal convolutional network to obtain the physiological state.

9. A physiological information monitoring system after transcranial magnetic stimulation, characterized in that: include: Acquisition module: obtains physiological information of multiple physiological categories monitored at multiple time points; Two-dimensional matrix construction module: constructing a two-dimensional matrix of the physiological information in the order of time points from small to large to obtain a physiological matrix; The rows of the physiological matrix correspond to time points; The physiological categories corresponding to the columns of the physiological matrix; Symmetry marking module: marking the symmetry line of the physiological matrix through the center point to obtain a marked symmetry line; Arrangement module: through the physiological convolution network, based on the physiological matrix and the marked symmetry line, the relationship between different categories of physiological information at adjacent time points is determined to obtain a relationship arrangement set; An adjustment module: adjusting the columns in the physiological matrix based on the relationship arrangement set to obtain an adjusted physiological matrix; Time convolution module: based on the adjusted physiological matrix, convolute sequentially according to time points to obtain the physiological state; The physiological state includes time domain indicators and frequency domain indicators; The physiological convolution network is used to determine the relationship between different types of physiological information at adjacent time points based on the physiological matrix and the marked symmetry line, and obtain a relationship arrangement set, including: Mark the values ​​corresponding to the straight lines parallel to the marked symmetry lines in the physiological matrix to obtain a plurality of parallel symmetry lines; the parallel symmetry lines include the position corresponding to the jth page of the i-th row and the position corresponding to the j+1th page of the i+1-th row; Obtaining a physiological convolution kernel; the physiological convolution kernel is a convolution kernel for convolving a physiological matrix; Inserting rows and columns at the beginning of the physiological matrix and inserting rows and columns at the end of the physiological matrix to obtain a zero-filled physiological matrix; the values ​​in the added rows and columns of the zero-filled physiological matrix are 0; Based on the zero-filled physiological matrix and the parallel symmetry lines, a physiological symmetry feature is obtained through the physiological convolution kernel; the physiological symmetry feature is a feature containing multiple physiological categories; multiple parallel symmetry lines correspond to multiple physiological symmetry features; Based on the multiple physiological symmetry features, a relationship arrangement set is obtained.

Images